G. I. Davies - A Critical and Exegetical Commentary On Exodus 1 - 18 Volume 2 - in Two Volumes - Commentary On Exodus 11 - 18-Bloomsbury Publishing PLC (2020)

The
INTERNATIONAL CRITICAL
COMMENTARY
on the Holy Scriptures of the Old and
New Testaments
GENERAL EDITORS
G. I. DAVIES, F.B.A.
Emeritus Professor of Old Testament Studies in the University of Cambridge
Fellow of Fitzwilliam College
AND
C. M. TUCKETT
Emeritus Professor of New Testament in the University of Oxford
Fellow of Pembroke College
FORMERLY UNDER THE EDITORSHIP OF
J. A. EMERTON, F.B.A., C. E. B. CRANFIELD, F.B.A.

and G. N. STANTON
General Editors of the New Series
S. R. DRIVER
A. PLUMMER
C. A. BRIGGS
Founding Editors
ii
A CRITICAL AND EXEGETICAL
COMMENTARY
ON
EXODUS 1–18
BY
G. I. DAVIES, F.B.A.
Emeritus Professor of Old Testament Studies in the University of Cambridge
Fellow of Fitzwilliam College
IN TWO VOLUMES
VOLUME 2
Commentary on Exodus 11–18

T&T CLARK
Bloomsbury Publishing Plc
50 Bedford Square, London, WC1B 3DP, UK
1385 Broadway, New York, NY 10018, USA
BLOOMSBURY, T&T CLARK and the T&T Clark logo are trademarks of
Bloomsbury Publishing Plc
First published in Great Britain 2020
Copyright © G. I. Davies, 2020
G. I. Davies has asserted his right under the Copyright, Designs and Patents Act,
1988, to be identified as the Author of this work.
All rights reserved. No part of this publication may be reproduced or transmitted

in any form or by any means, electronic or mechanical, including photocopying,
recording, or any information storage or retrieval system, without prior permission in
writing from the publishers.
Bloomsbury Publishing Plc does not have any control over, or responsibility for, any
third-party websites referred to or in this book. All internet addresses given in this
book were correct at the time of going to press. The author and publisher regret any
inconvenience caused if addresses have changed or sites have ceased to exist, but
can accept no responsibility for any such changes.
The NewJerusalemU, GraecaU and TranslitLSU fonts used to print this work
are available from Linguist’s Software, Inc., PO Box 580, Edmonds,
WA 98020-0580 USA.
Tel (425) 775-1130. www.linguistsoftware.com
A catalogue record for this book is available from the British Library.
A catalog record for this book is available from the Library of Congress.
ISBN: HB: 978-0-5676-8871-2

ePDF: 978-0-5676-8872-9
Series: International Critical Commentary
Typeset by Forthcoming Publications (www.forthpub.com)
To find out more about our authors and books visit www.bloomsbury.com
and sign up for our newsletters.
CONTENTS OF VOLUME 2
Introductory Note vii
PASSOVER AND DEPARTURE

(11.1–13.22)
Chapter 11.1-10
The Announcement of the Final Plague 1
Chapter 12.1-20
Instructions from Yahweh about the Passover
and Unleavened Bread 27
Chapter 12.21-27
Moses Instructs the Israelites about the Passover 75
Chapter 12.28-42, 50-51

The Final Plague and the Israelites’ Release 101
Chapter 12.43-49
Further Instructions from Yahweh about the Passover Meal 138
Chapter 13.1-16
Laws about the Consecration of the Firstborn
and the Festival of Unleavened Bread 157
Chapter 13.17-22
Aspects of Israel’s Departure 195
THE DELIVERANCE AT THE SEA

(14.1–15.21)
Chapter 14.1-31
The Crossing of the Sea and the Destruction
of the Egyptians 216
vi CONTENTS
Chapter 15.1-21
Two Songs Celebrating the Deliverance of the Israelites
and the Destruction of an Egyptian Force 283
THE JOURNEY TO THE MOUNTAIN OF GOD

(15.22–18.27)
Chapter 15.22-27
Sweet Water and Laws at Marah and Elim 398
Chapter 16.1-36
Manna and Quails in the Wilderness of Sin 423
Chapter 17.1-7
Water from the Rock at Rephidim 487
Chapter 17.8-16
Victory over Amalek 513
Chapter 18.1-12
The Coming of Jethro
and Further Celebration of the Exodus 546
Chapter 18.13-27
Jethro’s Advice about a Judicial System
for Israel and His Departure 578
I N T R O D U C TO RY N OT E
This volume represents the second part of the commentary on

Exodus 1–18, for which the main introductory matter (including
abbreviations and bibliography) can be found at the beginning of
volume 1.
G.I.D.
viii
PASS O V E R A N D D E PA RTURE
( 11 . 1 – 1 3 . 2 2 )
C ha p t e r 11. 1- 10
Th e An n ou nc e m e nt of t he F inal P l ag ue
Moses’ confrontations with Pharaoh have reached a decisive point

and 10.28-29 at first sight suggest that there is nothing more to
be said. Accordingly in 11.1-3 Moses receives from Yahweh an
assurance that after one more plague (the nature of which is not
yet stated) Pharaoh will let Israel go, and the (plural) ‘you’ in
v. 1 and the sequel in v. 2 make clear that this is a message for
Israel (as later in 12.1 and 14.1-2), not Pharaoh. But in vv. 4-8a
Moses still has one last word for Pharaoh before he finally leaves
him in 11.8b, and it provides him with a solemn warning of the
form which the final plague will take. Although introduced as a
message from Yahweh in v. 4, its wording is quite different from
v. 1. Some aspects of it recall earlier warnings that Moses has been
instructed to give to Pharaoh, but in other ways it takes a new
form which suits its climactic position in the ongoing dialogue
(see the Explanatory Note on vv. 4-6), and no indication is given
of Pharaoh’s response before Moses departs. Instead the chapter
ends (vv. 9-10) with a summary of the earlier plague-narrative,
in part another divine word to Moses and in part the narrator’s
comment, which reaffirms Pharaoh’s refusal to heed Yahweh’s
demand and explains it in terms both of its deeper purpose (v. 9)
and its theological cause (v. 10).
The Masoretic section-divisions mark breaks at the beginning and
end of the chapter, but also those noted above before verses 4 and 9
(likewise SP, except that it begins v. 4 with additional material; see
Text and Versions). The break before v. 9 is well attested at Qumran
(4QpalExl, 4QpalExm, 4QExc) and 4QpalExl also has the division at
the end of the chapter: 4QpalExm apparently did not. The presence
2 EXODUS 1–18
of a division before v. 4 in 4QpalExl and 2QExa has been inferred

indirectly, but there is no clear evidence. Nor is there any evidence
to determine whether there was one at the beginning of the chapter.
Both the brief instruction given to Moses for Israel (vv. 1-3) and
the summary and conclusion to the earlier plagues (vv. 9-10: in this
case together with the long passage about the Passover ritual in
12.1-28) interrupt the main narrative sequence about Moses’ final
confrontation with Pharaoh, which resumes with the coming of the
final plague in 12.29ff. This has since the late nineteenth century
provided a basis for the analysis of ch. 11 into underlying sources
and editorial additions.
Knobel pioneered the long-standing division of the passage between

three main authors, but while he assigned vv. 4-8 (Kriegsbuch = J) and 9-10
(Grundschrift = P) to different sources, he saw vv. 1-3 in a remarkably modern
way as the work of the Jehovistic redactor (i.e. RJE in later terminology:
pp. 89-91, cf. Num.-Jos., pp. 532, 548). From Wellhausen (Composition,
p. 68) and Dillmann onwards vv. 1-3 were attributed to the second early
source, i.e. E, but with little positive justification except for the assumption
in them that the Israelites lived among the Egyptians and not separately (cf.
Carpenter/Harford-Battersby, p. 95: this view is ascribed to E on the basis
of 10.21-23 on p. 88, but without any real justification in the text there).
Conversely some unease was expressed about the ascription of vv. 9-10 to the
P source, mainly because the future interpretation then given to v. 9 seemed
out of place in the narrative here (so already Wellhausen, ibid.) and some
commentators attributed them, or part of them, to a late redactor.1 Minor
redactional additions were also posited in vv. 1, 5 and 7: only Gressmann
went further with his suggestions for redistribution and rearrangement of the
verses and speculation about ‘missing’ features of the text (Mose, pp. 97-98
n. 1; Anfänge, pp. 41, 44). Such drastic measures found no acceptance, at
least for the time being. In fact a simpler analysis was proposed by Rudolph,
who assigned vv. 1-3 to J as well as vv. 4-8, but then had to suggest that
these sections were originally in the opposite order, to preserve what he took
to be the close connection between v. 4 and 10.29 (Elohist, pp. 21-23, 275).
Rudolph’s elimination of an E element in the passage, though weakly justified,
was taken up by Noth, but by attributing certain verses to a supplementer of
J rather than by a rearrangement of the text. Initially Noth simply attributed
vv. 1-3 (and v. 4aα) to JS (ÜGP, p. 32), but later he held that it was vv. 7-8
1
McNeile oddly ascribed them to RJE (pp. xvii, 61). Smend (p. 133) and
Eissfeldt (p. 270*) suspected the presence of the ‘third early source’ (i.e. J1, L)
here, because of some parallels to 4.21.
11.1-10 3
which were secondary (and possibly the whole of vv. 4-8), because a shift to
a further conversation with Pharaoh was ‘in no way possible for Moses in J
after 10.28f.’ (pp. 71-72, ET, pp. 92-93): now it was the words intended for
Israel in vv. 1-3 which were the original contribution of J.2
The older source-critical consensus continued to be maintained, however,
by Fohrer, Hyatt and Childs, with Fohrer even separating off vv. 2-3 for his N
source (= J1, L) because he believed that 12.35-36, which are clearly associ-
ated with them, had such an origin and the generosity of the Egyptians does
not fit the plague-narratives of the other sources (p. 82). But Steingrimmson
marked a foretaste of the redaction-critical analyses that were to become
widespread. Even the oldest nucleus of the passage (vv. 4, 5a, 6) was of late
origin according to him (p. 218), and the remainder was introduced in a series
of stages (p. 163). Kohata’s analysis was similar, but for her vv. 4-6 were from
J (following 10.1-19) and only (possibly) v. 10a from P (pp. 122-26). The
precise origin of the remainder was not defined. Houtman shows himself to
be well aware of such suggestions, but as usual confines himself to exploring
how ‘the [final] narrator’ constructed the text with the needs of ‘the reader’
in mind rather than Moses’ location or the likely chronology (pp. 129-30; cf.
Childs, p. 161).
For L. Schmidt (Beobachtungen, pp. 50-57) v. 8b followed 10.28-29 (J)
immediately, and so nothing else in vv. 1-8 can be from J. But vv. 1 and 4-8a
are from RJE and vv. 9-10 are from P: Schmidt saw, as P. Heinisch had done
over fifty years earlier, that the Hebrew imperfect in v. 9a could have an
iterative sense and so readily be taken to summarise previous events (p. 56).
Only vv. 2-3 came from a later (and undefinable) stage of redaction. Levin’s
analysis (pp. 338-39) is in some ways similar to this, although his Yahwist
is later and his nucleus (the whole of v. 8) belongs not to it but to the major
stage of amplification of J which introduced the plague-narrative in what
for him was its original form: so 11.8 is the sequel not to 10.28-29 but to
9.13-16. Verses 4-6 were added subsequently (like 12.29-30), and v. 7 later
still; likewise apparently vv. 1-3, since they overlap with the original J source
in 3.21-22 and 12.35-36 (pp. 334, 338). Verses 9-10 are apparently from P
(p. 336).
A simpler and in some ways more traditional ‘supplementary’ analysis
was offered by Blum. Verses 4-8 belong to the pre-Deuteronomistic plague-
narrative (p. 13), which was disturbed by the insertion of vv. 1-3 as part
of the ‘unselbständigen’ Kd layer which also includes 3.1–4.18; 5.22–6.1;
12.21-27; 13.3-16 and parts of ch. 14 (pp. 35-36). Verses 9-10 are from P
(to be precise Kp), added as a supplement to the existing narrative and so
2
Kohata and others have also seen an address to Israel as the primary text in
the passage, but in vv. 4-6(7) because the addressee in v. 4a is not defined (Jahwist,
p. 123 and n. 187).
4 EXODUS 1–18
composed specifically for their present position (p. 254).3 Blum’s analysis
was closely followed by F. Ahuis (Exodus 11,1 – 13,16, pp. 102-104), except
that he also attributed vv. 9-10 to the Deuteronomist(!), like several additions
which he detected in 12.1-20. Van Seters shared Blum’s view of vv. 9-10, but
was content to ascribe the whole of vv. 1-8 to J, though with 10.28-29 trans-
posed to the middle of v. 8 (pp. 108, 121 n. 30: like Heinisch): in his view the
‘despoiling’ motif as an aspect of Yahweh’s ‘victory’ over Egypt fitted well
into the overall narrative plot.
In the most recent scholarship the attribution of vv. 9-10 to P (mostly as
an independent source) has been generally accepted (only Propp [p. 313] and
for v. 9 Gertz [pp. 181-82: EndR] demur), but for vv. 1-8 a variety of views
are held. Schmidt (also Graupner, pp. 66-67) retains the old bipartite division,
emphasising the intrusiveness of vv. 1-3 but also their connection to passages
from J (6.1) or supplementary to it (in his view 3.20-22) and the absence of any
distinctively Deuteronomistic features. So he assigns vv. 4-8 to J and vv. 1-3
to JS.4 Propp thinks that v. 1 is the original introduction to the main narrative
in vv. 4-8, which he attributes to E, while vv. 2-3 are from J (pp. 313-15: cf.
Dozeman, p. 255). K. Schmid (in scattered references to particular verses)
seems to view vv. 1-8 as belonging to the pre-Priestly independent Exodus
narrative, but he does not discuss any sub-division within them. Gertz, by
contrast, traces a very detailed process of development. In vv. 4-8 the original
core (part of a supplement to the pre-Priestly Exodus narrative) comprises
only vv. 4aβb, 5a, 6a, 7aα, 8b, with the remainder being added by EndR or
later (pp. 180-82). Verses 1-3 are also from EndR, because of their relation-
ship to other passages (10.24-26 as well as 3.18-22 and 5.22–6.1) already
attributed to this layer: in Gertz’s larger understanding of the plague-narrative
they were needed to underline the divine authority for what Moses says in
vv. 4-8* after these verses had been separated from their original authorisation
in 9.13-16 by the insertion of the previous three plagues.
Most analyses of this passage have taken their starting-point at

least from the view that vv. 1-3, 4-8 and 9-10 stem from different
authors. Occasionally it has been suggested that v. 1, or even the
whole of vv. 1-3, was designed to lead into (or in Rudolph’s case
out of) vv. 4-8 and less often that vv. 4-6 provide further words
addressed to Israel, continuing from vv. 1-3. But the contents and
3
In view of the fact that they are actually out of place at this stage in the
narrative, these verses are the Achilles’ heel (or one of them) for a supplementary
understanding of P.
4
This is therefore a simpler (and more traditional) view of the passage than
that set out earlier by Schmidt’s pupil Kohata. But both agree that E cannot be
present in vv. 1-3 (cf. Schmidt, p. 462, for his reasons).
11.1-10 5
grammatical references (cf. ‘you’ in vv. 1, 7 and 8) in the two

sections make it clear that they take the narrative in two different
directions and, unlike vv. 4-8, vv. 1-3 have no point of contact with
the preceding context either. Despite contrary opinion (voiced most
vigorously by Gressmann but also taken up by L. Schmidt), there is
no difficulty in seeing v. 4 (with or without its introductory formula)
as the continuation of 10.29: Moses is not said to leave Pharaoh at
the end of ch. 10 and it is very natural that he should only do so in
10.8 after giving him a warning of the final plague which will by
its very severity instil into him and his courtiers the ultimate lesson
about Israel’s special place in Yahweh’s concerns.
If, as seems necessary, a distinction in origin between vv. 1-3(a),
which are clearly linked in some way to 3.20-22 and 12.35-36, and
vv. 4-8 is to be made, there are four main possibilities: (a) vv. 1-3
are a redactional addition to the old narrative in vv. 4-8 (and 10.24-
29*); (b) they come from a parallel old account of the Exodus story;
(c) they are part of a Deuteronomistic layer in the composition of
Exodus (‘Kd’); and (d) they belong to the final redaction of the
Pentateuch (EndR). The strongest objection to (b) is its presumption
that Israelites lived in among the Egyptians and not separately in
the land of Goshen. This is not supported by 10.23: in fact 3.20-22;
11.1-3 and 12.35-36 are probably the only passages which involve
such a presumption. Against (c) it has been pointed out (by W.H.
Schmidt) that there are no Deuteronomistic features in these verses,
whatever may (or may not) be said about such features in other
elements of the ‘layer’ identified by Blum. Similarly there is no real
justification for (d) (or indeed for the view that the verses are of
yet later origin), since there is no sign here of any contact with the
Priestly narrative.5 Gertz’s arguments for an EndR origin are based
on attributions of other passages which there is no good reason to
follow. Blenkinsopp’s observations about the very similar motif in
the account of the first return from Babylon (Ezra 1.4, 6 etc.) are
more plausible (Pentateuch, p. 155) but in view of other probably
imaginary elements in that account it could itself have been
modelled on the Exodus passages. It is therefore most likely that
5
A connection has often been made with the materials needed to make the
Tabernacle (25.1-7; 35.4-9, 21-29), but beyond the phrase ‘all sorts of gold
objects’ (kōl kelê zāhāb) in 35.22 there is no close correspondence of wording or
any allusion to an Egyptian origin for what was offered there.
6 EXODUS 1–18
at least for vv. 1-3a (a) is correct and we should think of an origin
with a redactor who knew the combined old accounts of the Exodus
story (i.e. RJE), into which this and the related passages (3.21-22 and
12.35-36) were inserted. With these verses he provided an explicit
account of Moses’ reception of a word from Yahweh about the final
plague and the authorisation for the Israelites’ ‘plundering’ of the
Egyptians. It is probable (see the Explanatory Note) that v. 3b was
designed more as an introduction to vv. 4-8 than as the conclusion
of vv. 1-3a and so belonged to the main account here.
Verses 4-8 make coherent sense as they stand and there is no more
reason to eliminate motifs from vv. 6-7 than there was when they
occurred in earlier episodes of the plague-narrative. Some differ-
ences from the narrative in 12.29-32 have been seen as reasons for
regarding v. 5aβ and v. 8a as secondary, but the inconsistency is
just as problematic then (if not more so) as it is in the present text.
The different ways of describing the antithesis to Pharaoh’s son are
not strictly contradictory and can perhaps be put down to delib-
erate variation, which is more likely in an original author than a
redactor. The fact that in 12.31-32 it is Pharaoh himself and not his
‘servants’ who plead with Moses to leave can be put down to a very
plausible development in the narrative plot. In 10.28 Pharaoh has
ruled out any further contact with Moses, so it is natural that Moses
should speak of his ‘servants’ as those who would sideline Pharaoh
and take over responsibility for giving the Israelites permission to
leave (interestingly in 11.1 it is Pharaoh himself who is said to
‘let Israel go’, another indication of the separateness of vv. 1-3a
from vv. 4-8). But in the event, although the ‘servants’ are still
mentioned in 12.30, Pharaoh himself (despite his earlier words in
10.28) is ready to take the lead and yield to what Moses has been
demanding all along. His submission makes the greater impact for
its having been initially presented as impossible. As parts of the
fuller plagues narrative that has been preserved, then, vv. 4-8 are
attributed by us to E.
There is no denying that vv. 9-10 are very closely related to 7.3-6,
the introduction to the Priestly plague-narrative. Once the iterative
use of the imperfect and the past reference of v. 9a are recognised,
the verses make good sense as a concluding summary which leads
into the Priestly Passover legislation in 12.1-20. The only question
which remains is whether they support or detract from arguments
that the Priestly strand is a supplement to the older narrative (as
11.1-10 7
Blum and Van Seters hold) or an originally independent source.

Van Seters has claimed that the reference to Pharaoh ‘not listening’
presupposes the non-Priestly accounts where Moses makes a succes-
sion of explicit demands of Pharaoh. But this claim is completely
undermined by the fact that exactly similar references to Pharaoh
‘not listening’ appear in the conclusions of the individual Priestly
episodes in 7.13 (before the first non-Priestly episode), 8.15 and
9.12 (as well as in what were once almost certainly parts of separate
Priestly episodes in 7.22 and 8.11). There, what Pharaoh ‘does not
listen’ to is what Moses and Aaron are commanded to say in 6.11
and 7.2, and the same was surely originally the case in 11.9, which
is the next Priestly section after 9.12. Blum has written that 11.9-10
were composed for their present position (see above), and as far
as their connection to 12.1 is concerned this is perfectly true. But
in the combined present text they come too late: such a summary
would be required before 11.1, where the story moves forward to
the next stage. Their present position is in that sense an argument
against P being a supplement to the older narrative. It is due to
EndR not wishing to disturb (further) the sequence of the older
narrative until he had no choice: P’s Passover regulations had to
precede the narrative of the final plague, and it also made better
sense for them, as an address of Yahweh to Moses and Aaron, to
precede Moses’ delivery of instructions about the Passover to ‘all
the elders of Israel’ in 12.21-27. The P conclusion to the plague-
narrative naturally enough remained attached to the beginning of
the Passover regulations as their point of departure.
In two different ways, and at two different points in the overall
plot, this passage marks a turning-point in the Exodus story. The two
literary units in vv. 1-8 together (and thereby with greater emphasis)
introduce Yahweh’s new plan to bring his conflict with Pharaoh to
a decisive conclusion with one final, terrible, plague. The words
(to be) spoken in them are directed first to Israel (vv. 1-3, for the
first time since chs. 3–6 and as they are more fully in 12.1-27) and
then to Pharaoh and his people (vv. 4-8), and they spell out the
different consequences of the plague for Israel (positively) and for
Egypt (negatively). The God who has ‘favoured’ Israel throughout
(3.7-8; 8.18; 9.4-7, 26; 10.23) now even makes the Egyptians into
their benefactors. This distinction is underlined in vv. 6-7, where
it also forms the content of what the Egyptians will learn (‘know’)
from the coming catastrophe.
8 EXODUS 1–18
Verses 9-10 were originally the introduction to a second

announcement of the coming plague in 12.1-20 (cf. vv. 12-13) and
by summing up a series of plagues which ended in 9.12 they showed
why further action was necessary. What they say has been largely
anticipated in 7.3-6 and they confirm that the whole sequence of
events has been under Yahweh’s sovereign control (compare also
10.1-2). In their present position they also serve to provide a belated
justification for the originally separate announcements in vv. 1-8.
1 [Yahweh said to Moses, ‘One more plaguea I will bring on

Pharaoh and on Egypt: afterwardsb he will let you go from herec.
When he lets (you) go, he will surely drive you out all togetherd
from here. 2 You speake in the hearing of the people, so that each
man may askf his companiong and each woman her friendg for
objectsh of silver and gold.’ 3 iYahweh gave the people favour
(lit. put the favour of the people) in the sight of the Egyptiansi.]
j
The man Moses was indeed very greatj in the land of Egypt in
the sight of Pharaoh’s servants and of the people. 4 Moses said,
‘Thus says Yahweh, At midnightk I am going forthl in the midst of
Egypt, 5 and all the firstborn in the land of Egypt shall die, from
the firstborn of Pharaoh who sits on his throne to the firstborn
of the maidservant who is behind the grinding stones and all the
firstborn of the animalsm. 6 There will be a great cry throughout
the land of Egypt, the like of whichn has never happened before
nor will there be anothero like itn. 7 But as for the Israelitesp, no
dog shall bark (?)q rat either man or animalr, so that you may
know thats Yahweh makes a distinction between Egypt and
Israel. 8 All theset your servants shall come down to me and bow
down to me, saying “Depart, you and all the people whom you
lead (lit. are at your feet)”. After this I will depart.’ Then Moses
went out of Pharaoh’s presence in a furious rageu. 9 Yahweh said
to Moses, ‘Pharaoh has repeatedly not listenedv to you (pl.), so
that my wonders may be multiplied in the land of Egypt’. 10 So
Moses and Aaron didw all these wonders before Pharaoh, but
Yahweh made Pharaoh’s heart stubborn and he did not let the
Israelites go from his land.
Notes on the Translation

a. The object is unusually placed at the beginning of the sentence (as again
in 18.23), to draw attention to it (GK §142f, noting that cases where the subject
is represented in the verbal form itself ‘are far more numerous’; Muraoka,
11.1-10 9
Emphatic Words, pp. 38-39; JM §155o). Heb. ‫( נֶ גַ ע‬cf. TWAT 5, 224-26 = TDOT
9, pp. 207-209) is used only here of the Exodus plagues, but in Gen. 12.17
it stands (with the Piel of the related verb) for the ‘great sicknesses’ inflicted
by Yahweh on Pharaoh and his household because of his violation of Sarah.
‫ נֶ גַ ע‬can mean a physical ‘blow’ inflicted by a human being (Deut. 17.8; 21.5;
Prov. 6.33), but it is more often used metaphorically of an illness seen as
sent by God and then specifically of the marks left by skin disease (‫ )צרעת‬or
similar-looking mould on clothes or house-walls (Lev. 13–14, passim). Here
(like ‫ מגפה‬in 9.14) it has a more general meaning, as in 1 Kgs 8.37 perhaps,
embracing both the varied earlier plagues and (though without revealing yet
its full severity) the final slaughter of the Egyptian firstborn.
b. The asyndeton of MT (for its avoidance in the Vss see Text and Versions)
makes the consequence unusually direct (cf. JM §177a).
c. Heb. ‫מזה‬, as again at the end of the verse, with ‫ זה‬used adverbially for
‘here’.
d. Heb. ‫כלה‬, rendered here ‘all together’, most often means ‘destruction’,
but the positive sense ‘completion’ is also attested at Gen. 18.21 (in the same
adverbial usage as here) and is entirely natural in view of the frequency with
which the cognate verb (Exod. 5.13-14; 31.18; 34.33; 39.32; 40.33 etc.)
and some nominal derivatives of the root are used for ‘completion’ as well
as ‘destruction’ (cf. Houtman, W.H. Schmidt). BDB, p. 478, recognises the
positive sense of ‫ כלה‬but regards it as ‘dubious’. Doubts about it were already
expressed in the nineteenth century and emendations proposed: among newer
dictionaries only DCH 4, pp. 418-19, accepts it. See further Text and Versions.
Syntactically the word is best connected to what follows: it precedes the
verb and so receives added emphasis, no doubt to highlight the contrast with
Pharaoh’s earlier unwillingness to let the whole people depart.6
e. Heb. ‫נא‬. There are other examples of the surprising use of ‘the particle of
entreaty’ when God is addressing a human: cf. 4.6 and see Note f on the trans-
lation of 4.1-9, where it is suggested that it is a sign of God’s condescension
or intimacy when dealing with his chosen human servants. But in such cases
(as more clearly perhaps in 10.11, where Pharaoh addresses[?] Israel) it is also
possible that the particle simply strengthens a command (cf. JM §105c: ‘Do
come!’) and this may even be the original use which then became specialised
as an element of deferential pleading.
f. Heb. ‫וישׁאלו‬. On the surface this is a case of simple waw indicating the
purpose of the preceding command (GK §165a), but since the command does
6
This is probably what the Masoretic accents intend: the Zaqeph magnum
is a lesser disjunctive then Zaqeph parvum on the preceding word (GK §15f;
Bergsträsser §12g). A different view is taken by Jacob, p. 294; Schmidt, p. 455;
Houtman, p. 131; Ges18, p. 547. A minority of scholars interpret by ‘mit Gewalt’
or ‘gewaltsam’, i.e. ‘by force’ (Gressmann, Anfänge, p. 41; Beer, p. 58; Ges18,
p. 547).
10 EXODUS 1–18
not specify what Moses is to say it may be correct to see the clause as doing
that, with the waw being (like ‫ )לאמר‬virtually an introduction to words of
direct speech. There are very similar examples of the same formulaic structure
in Numbers (e.g. 17.2; also 19.2, although there the use of the second person
m.s. suffix on ‫ אליך‬shows that the command to the people remains indirect):
see further JM §177h-k.
g. Heb. ‫רעהו…רעותה‬. The fem. nouns from this root are more specialised in
their use than ‫ ֵר ַע‬itself, and ‫ רעות‬is the word that regularly occurs in reciprocal
phrases with ‫( אשׁה‬cf. my ‘The Ethics of Friendship in Wisdom Literature’, in
K.J. Dell [ed.], Ethical and Unethical in the Old Testament: God and Humans
in Dialogue [LHBOTS 528; London, 2010], pp. 135-50 [138]). But the
context makes it clear that the sense cannot be, as in the weakened idiomatic
uses of both words, ‘(from) one another’ since it is from non-Israelites
(Egyptians) that the gifts are to be sought (cf. v. 3: also the related passages in
3.21-22 [where other words are used] and 12.35-36).
h. See Note r on the translation of 3.16-22.
i. Heb. ‫ויתן יהוה חן־העם בעיני מצרים‬. On this idiom see Note m on the transla-
tion of 3.16-22. Its full form occurs only here and in Gen. 39.21 and Exod.
3.21; 12.36 (two passages closely related to this one). Abbreviated forms of
it may occur in Prov. 13.15 and, less certainly, in Prov. 3.34 and Ps. 84.12: in
none of these places is a ‘third party’ named as the bestower of favour.
j. Heb. ‫גם האישׁ משׁה גדול מאד‬. The precedence of the subject in the nominal
clause here follows the S-P pattern which most grammarians would now
regard as normal (cf. Muraoka, Emphatic Words, p. 6; JM §154f). Those, like
S.R. Driver and F.I. Andersen, who regard(ed) P-S as the normal pattern in
descriptive nominal clauses would presumably see the ‘inversion’ as due to
the change of subject and the emphasis laid on it (in fact Andersen’s explana-
tion of such cases is that they are circumstantial [Verbless Clause, p. 43], if
there is an explanation at all). But both these factors are sufficiently signalled
by the particle ‫ גם‬and the phrase that follows it in the present text.
k. Heb. ‫כחצת הלילה‬. For ‫‘ = כ‬at’ cf. BDB, pp. 453 (1.a) and 454 (3.b: with
inf. constr. or verbal noun). ‫ חצת הלילה‬is found elsewhere only in poetry (Ps.
119.62; Job 34.20): ‫ חצת‬resembles an inf. constr. (cf. GK §83a). The more
usual expression is ‫( בחצי הלילה‬12.29; cf. Judg. 16.3 [2x]; Ruth 3.8); ‫בתוך‬
‫( הלילה‬1 Kgs 3.20) is probably less precise: on ‫( באישׁון הלילה‬Prov. 7.9; cf.
20.20) see HAL, pp. 43, 91; Ges18, pp. 51, 106.
l. See Note l on the translation of 10.1-20. ‘At midnight tonight…’ is
evidently meant: the narrative suddenly gathers pace.
m. Heb. ‫בהמה‬. The absence of the def. art. is due to the fact that ‫ כל‬itself
‘conveys a certain notion of determination’ (JM §138d; cf. 125h and GK
§117c), especially with collective nouns: hence the alternative Eng. tr. ‘every
firstborn of animals’. The SP variant (see Text and Versions) may be partly
due to the sense that in such an expression the ‫ כל‬should immediately precede
the collective noun.
11.1-10 11
n. Heb. ‫)אשׁר( כמהו‬, with the masc. suffix unusually after the fem. noun
‫צעקה‬, which does determine the gender of the verbal forms in this verse: for
other instances see GK §135o, or a regular formula may be involved (cf. 9.18,
24; 10.14). There is therefore no need to follow SP (see Text and Versions): its
fem. suffixes will be due to secondary ‘smoothing’ of the grammar.
o. Heb. ‫לא תסף‬, with the inf. ‫ להיות‬easily understood. Such omission of the
expected inf. occurs quite frequently (BDB, p. 415), especially in poetic and
later Heb. (for the latter cf. 2 Sam. 7.20 with 1 Chr. 17.10): for a probably
pre-exilic instance see Deut. 25.3.
p. Heb. ‫ולכל בני ישׂראל‬. The ‘fronting’ of this phrase expresses contrast, or
more precisely limitation. The ‫ ל‬denotes the indirect object, as in the other
instance of the idiomatic use of ‫( חרץ‬Josh. 10.21: see below).
q. Heb. ‫יחרץ‬. The same idiom appears in Josh. 10.21 with a human subject.
The verb occurs once (pass. pt.) in the meaning ‘cut’ (Lev. 22.22; cf. ‫כרות‬
in v. 24) and two derivatives point to this sense too (cf. 2 Sam. 12.31; Dan.
9.25 [dub.]). It is probably the basis for the more widespread idea of ‘decide,
decisive’. The present idiom seems to be related to the sense ‘sharp(en)’,
which elsewhere appears only in derivatives, and to the expression ‫שׁננו לשׁונם‬
‫ כמו־נחשׁ‬in Ps. 140.4, which Propp aptly compares (p. 344): cf. his other refer-
ences to the comparison of a malicious tongue to a weapon. These parallels
suggest that it is not a gesture (‘pointing’ or ‘sticking out’ the tongue) that
is meant, but a hostile sound, as most of the Vss already understood it. In
Josh. 10.21 ‘railed against’ is required, but here in relation to dogs ‘bark’ (or
‘growl’) is appropriate.7
r. Lit. ‘from man to animal’. ‫ למאישׁ‬prefixes ‫ ל‬to ‫ מן‬in a frequent idiom
where ‫ עד‬follows (see Note r on the translation of 9.13-35).
s. Heb. ‫אשׁר‬, which occasionally has this meaning (cf. Num. 32.23; Deut.
1.31), especially in LBH (BDB, p. 83 [8.a]).
t. Heb. ‫אלה‬, without the art.: see Note c on the translation of 10.1-20.
u. Heb. ‫בחרי־אף‬.ָ The prefix vowel is qamets hatuf, which is required before
the following hatef-qamets (GK §102d).
v. Heb. ‫לא־ישׁמע‬. Most likely an iterative imperfect (GK §107e: so Heinisch,
p. 96; L. Schmidt, Beobachtungen, p. 56; W.H. Schmidt, pp. 455-56; but
rejected by Houtman, p. 135 n. 212, and ignored by Dozeman, p. 258.8
7
A derivation from a Heb. cognate of Ar. ḥaraṣa (‫ חרץ‬II in HAL, p. 342:
‘sich mit Eifer an etwas machen’), which can mean ‘be eager for’ as well as ‘tear,
split’ (Ges18, p. 401), has been suggested by Zorell (p. 271: ‘hostiliter movit’) and
F.C. Fensham (‘The Dog in Ex. XI 7’, VT 16 [1966], pp. 504-507: ‘move eagerly’,
sc. to eat their dead bodies). This homonym may well be present in 2 Sam. 5.24,
but this seems less likely here than the explanation offered above.
8
Beer translates in the present tense, ‘hört nicht’ (p. 58), but does not comment
on the grammar. Others state that the reference is to the earlier plagues, but without
translating or commenting further (Smend, McNeile, Fohrer, Steingrimsson, Blum).
12 EXODUS 1–18
w. Or ‘had done’: Heb. ‫עשׂו‬, with the perfect representing a repeated action
in what JM §111e describes as a ‘global’ way, i.e. the successive episodes are
treated as a single act by way of summary: the fronting of the subject also
avoids the implication that a new action is involved (cf. JM §118d) and would
justify the rendering as a pluperfect.
Explanatory Notes
1-2. No change of scene is indicated for Yahweh’s new address to
Moses, so it is possible to understand these words as being spoken
to him in Pharaoh’s presence (so Houtman). But, unlike many
previous passages (e.g. 9.1-4[5], 13-19), they contain no message
specifically for Pharaoh and v. 2 in fact instructs Moses about what
he is to say to the Israelites (‘the people’). So the narrative takes a
sharp, if temporary (see the note on vv. 4-6) detour at this point (see
the introduction to this section for the possibility that this is due to
editorial activity). The Heb. word-order lays emphasis on the fact
that there is to be one (and only one) more plague (see Note a on the
translation). Nothing is said here about what form it will take, but
its effect on Pharaoh will, unlike the earlier plagues, be decisive:
he will not only let Israel go (Heb. šlḥ Piel, the constant demand of
Moses) but will drive them out (Heb. grš Piel: cf. 12.39). This verb
has already been used (with šlḥ) in the non-Priestly introduction to
the plague-story in 6.1, where Yahweh declared that through (his)
powerful intervention Pharaoh would be impelled to give way.9
‘All together’ shows that now there will no longer be any limitation
of the permission to leave to a part of the people, as Pharaoh had
insisted in negotiations with Moses (10.8-11, 24). With departure
imminently in prospect, it is time for Moses to inform the people
and activate the plan previously described in more detail (except for
the reference here to men) in 3.21-22 (cf. 12.35-36): see the notes
there. The ‘objects’ may well have included jewellery, but Heb. kly
has a wider range of use which includes tableware, for example (see
Note h on the translation).
3. The enjoyment of favour is frequently associated with the
granting of a request (cf. TWAT 3, 28-31 = TDOT 5, pp. 26-28).
9
As Dozeman points out (pp. 255-56), the force implied by grš is clearly
indicated by its use later in Exodus of the conquest of the land (23.28-31; 33.2;
34.11: cf. also 10.11).
11.1-10 13
The statement has often been seen as premature, since it appears

to anticipate 12.35-36 (this may be why the Samaritan Pentateuch
and the Vulgate put the statement into the future tense [see Text and
Versions]: cf. Propp’s proposed emendation of MT). But if those
verses are properly understood as a recapitulation of something
which has happened earlier, as both intrinsic probability and
Hebrew grammar would suggest (see the notes there and the use of
the pluperfect tense in e.g. NEB, NJPS and NRSV), then 11.3a can
be taken as implying both the execution of Yahweh’s instructions
and a favourable response by the Egyptians prior to the final plague.
W.H. Schmidt takes an intermediate position, suggesting that the
expression indicates only the creation of favourable conditions for
a request at this stage (p. 455). But that separates what properly
belongs together.
If the first half of the verse rounds off vv. 1-2, the recogni-
tion of the high regard in which ‘the man’ Moses was (now) held
by the Egyptians relates rather to what follows (especially what
is envisaged in v. 8) and should be seen as the introduction to it.
Two expressions in v. 3b require clarification. The phrase ‘the man
Moses’ is found only here and in Num. 12.3 in a similar state-
ment about Moses’ exceptional character. Despite speculation (see
Schmidt, p. 455, and Houtman 1, p. 7; 2, p. 131, for details) it
probably has nothing to do with any special status of Moses, as
the same kind of combination, with ‘the man’ preceding a proper
name, appears with Lot (Gen. 19.9: cf. RV ‘the man, even Lot’),
Micah the Ephraimite (Judg. 17.5), Elkanah (1 Sam. 1.21), Jero-
boam (1 Kgs 11.28), Mordecai (Esth. 9.4) and the angel Gabriel
(Dan. 9. 21: for a listing of the examples except the last see Driver,
Notes, p. 19; DCH 1, p. 232 includes them all but in a mixed list
of passages where a personal name is in apposition to ʾîš, ‘man’ or
vice versa). What all these examples have in common is that the
person named has already been mentioned in the context, so that
the usage is anaphorical in intention and equivalent to the fuller
phrase ‘this man, Coniah’ in Jer. 22.28 (it is itself amplified in Dan.
9.21).10 ‘The people’ here most likely means ‘the Egyptians’, rather
10
Cf. Montgomery, Daniel, p. 370: ‘called “the man” to identify him with the
being in 815ff’. None of the standard grammars (GK, Jouon, JM, IBHS) seems to
deal with this special case of apposition, although it may be related to phenomena
such as those treated in GK §131g and JM §131h-i.
14 EXODUS 1–18
than Israel (Nachmanides), to whom it refers in vv. 2-3a: mention

of Pharaoh’s servants naturally leads to the addition of the whole
Egyptian people (cf. 10.6).
4-6. It is not at first obvious to whom Moses speaks these
words, but in v. 7 and especially v. 8 (‘your servants’), which are
the continuation of vv. 4-6, Pharaoh is clearly in view (see Text
and Versions here for early additions to make this explicit). The
omission of the addressee recalls 10.29, where it is not a problem
because of the preceding dialogue with Pharaoh, and here it is one
sign that vv. 4-8 form the conclusion to that dialogue. The divine
speech is in several ways similar to those which have introduced
other episodes of the non-Priestly plague-narrative – the participial
formulation ‘I am going forth’, the fixing of a time, the unprec-
edented severity of the plague (cf. 9.18, 24; 10.14), the distinction
between the Israelites and the Egyptians, and the intention that the
plague will give instruction, ‘so that you may know…’ – but in
other respects the finality of the imminent event leads to differences
of formulation and content. For example, the word of Yahweh is
cited as it is delivered to Pharaoh and the Egyptians (like 10.3-6),
rather than when Yahweh speaks it to Moses; and the release of the
Israelites appears no longer as a demand addressed to Pharaoh but
as the certain outcome of this most dreadful visitation.
Its timing ‘at midnight’, without further qualification, suggests
that for this version of the Exodus story (in contrast to what is
implied by the instructions in 12.1-20; those in 12.21-27 are compat-
ible with what is said here) the warning of what was to happen
was given on the same day. The mention of midnight recalls wider
fears of the terrors of the night (Pss. 91.5-6; 121.6b), but another
of Israel’s origin traditions included the tale of a fearful struggle
of their ancestor Jacob in the night (Gen. 32.22-32) and Dozeman
makes much of the ‘one night’ in the Exodus story which is the
time of both disaster and deliverance (pp. 257-58; cf. his treat-
ment of 10.21-29). Yahweh’s ‘going forth’ (Heb. yṣʾ) at the time
of the Exodus is picked up in Ps. 81.6 and its significance is not
to be simply equated with his ‘passing through’ the land in 12.12,
23 (so Schmidt, pp. 454-55). The expression draws attention to
Yahweh’s departure from his dwelling-place in heaven or on earth
(so Knobel, p. 90; Dillmann, p. 96) and his appearance as a mighty
warrior on the field of battle, as in Judg. 5.4; Ps. 68.8; Mic. 1.3; cf.
11.1-10 15
Jeremias, Theophanie, pp. 7-12, 148; TWAT 3, 803-804 = TDOT 6,

pp. 232-33.11
The ‘grinding stones’ (Heb. rēḥayim, a dual form) in biblical
times did not yet take the form of two round horizontal stones, but
consisted of a base and a roller, both made of hard stone such as
basalt (cf. ABD 4, p. 831; IDB 3, pp. 380-81, with an illustration
of an Egyptian model which shows a woman ‘behind’ the stones;
further refs. to illustrations in Schmidt, p. 465). Grinding grain into
flour was a woman’s work (Eccl. 12.3), but might (especially in
royal palaces) be done by slaves (cf. Isa. 47.1-2; even by a captive
man in Judg. 16.21). In the account of the death of the firstborn in
12.29 the slave-girl is replaced by an imprisoned captive: it seems
that precise correspondences between prediction and fulfilment
were not a matter of great concern to the non-Priestly narrator(s).
The ‘great cry’ which the Egyptians would now utter (v. 6: cf.
12.30) cannot but recall the ‘cry’ of the Israelites at their oppres-
sion by the Egyptians which Yahweh several times says that he has
heard (and heeded) in chs. 1–6 (see also 14.10, 15). The verbal echo
is closest to 3.7 and 3.9, where the same word (ṣeʿāqāh) is used in
both strands of the non-Priestly narrative of Moses’ commissioning.
But words with a closely similar meaning (including the probably
related verb zāʿaq in 2.23) occur in the Priestly account in 2.23-24
and 6.5. The irony created by the coming reversal of roles is thus
bitterly underlined here, with a striking intensification in the addi-
tion of the adjective ‘great’ and the incomparability statement in
the second half of v. 6 which is quite unrestricted: the catastrophe
is presented as one that is totally unique. The language of the ‘cry’
itself on the lips of the powerful people of Egypt reinforces the
sense of transformation, since the great majority of biblical exam-
ples of it arise from those who are suffering or at risk from the
power of enemies or rulers (see Boyce, The Cry to God). As such
the ‘cry’ might be expected to awaken sympathy in the reader, but
that is clearly far from the narrative’s intention here: the cry is (or
will be) the expression of intense suffering, which in this case is
seen as fully deserved. There are at least a few other cases where
11
It is surprising that Schmidt (p. 455) rejects this view out of hand. Early
interpreters were more alive to the word’s implications (see Text and Versions).
16 EXODUS 1–18
this more detached and perhaps more basic use of cry-language is

found (e.g. 1 Sam. 5.10, 12; Isa. 65.14; Jer. 25.36; 48.3, 5; 49.21;
Zeph. 1.10), without the theological overtones which it has become
customary to associate with it.12
7-8. There is no formal transition in these verses from Yahweh’s
message (cf. v. 4a) to Moses’ own words, but in v. 8 ‘me’ and ‘I’
must refer to Moses (despite Dozeman’s suggestion [p. 254] that
Yahweh himself could be meant and still be the speaker). This
is possible but not so certain in v. 7: the use of the third person
for Yahweh might signify that the divine speech is over, but the
verse is closely linked to vv. 5-6 by two features of its wording.
The mention of the Israelites at the beginning of the verse in the
Hebrew is a way of underlining the contrast with the sufferings
of the Egyptians, which is also highlighted in the second half of
the verse; and both in v. 5 and in v. 7 the infliction of harm and
its non-occurrence take in animals as well as humans. As already
noted, by v. 8 Moses’ words are clearly being addressed to Pharaoh
himself (‘your [sing.] servants’; cf. also ‘out of Pharaoh’s presence’
at the end of the verse). But in v. 7 the ‘you’ is plural, including at
least Pharaoh’s courtiers and perhaps his whole people. The purpose
of the final plague is still didactic, like many of those which have
preceded, but now the scope of the learning embraces not Pharaoh
alone, as previously (cf. the singular ‘you’ in 7.17; 8.6, 18; 9.14,
29, which takes up Pharaoh’s own words in 5.2), but the Egyptians
more generally. The content of what is to be learned also changes
from Yahweh’s power (as in the verses just cited) to the special
place of Israel in his purposes, which has also been made evident
earlier in the non-Priestly plague narrative (8.18; 9.4-7, 26; 10.23).
It is in line with this wider educational purpose, as well as
Pharaoh’s exclusion of any further contact between him and Moses
(10.28), that Moses envisages that it will not be Pharaoh himself, but
his ‘servants’, or courtiers, who will ‘come down’ (presumably from
the palace) to give Moses and the Israelites permission to ‘depart’
(Heb. yāṣāʾ, here being used in its epic sense of the long-sought
12
R. Albertz reckons that approximately a third of the occurrences of ṣʿq / zʿq
are of this kind (THAT 2, 573 = TLOT 3, pp. 1091-92). The article in TWAT 2, 628-
39 = TDOT 4, pp. 112-22 (G. Hasel), regards ‘cry for help’ as the basic meaning
(p. 115): cf. THAT 2, 569-70 = TLOT 3, pp. 1089. The frequency of prepositions
like ʾel and ʿal after the verbs certainly shows the importance of this meaning.
11.1-10 17
departure from Egypt itself, as often later [e.g. 12.17, 42, 51] and
in the causative form ‘bring out’ already part of Yahweh’s promise
to the people [3.10-12 etc.]). The courtiers will ‘bow down’ (Heb.
hištaḥawāh) to Moses, a verb often used for the worship of a god
(4.31 [probably]; 12.27; 18.7; 20.5) but also sometimes of the rever-
ence paid to a human leader such as a king (cf. Gen. 43.28; 1 Sam.
24.9; 2 Sam. 14.4 and other examples in BDB, p. 1005 s.v. ‫שׁ ָחה‬, ָ 1).
So Moses is not being regarded as divine by the Egyptians, but he is
being seen as the equal (at least) of their king Pharaoh. Then, Moses
says, he will ‘depart’. It is a little confusing that in the Hebrew, when
the narrator goes on to speak of Moses leaving Pharaoh’s presence,
he uses the same verb yāṣāʾ in its more straightforward sense. But
since this has already appeared several times in the plague-narrative
(8.8, 25-26; 9.33; 10.6, 18) there is no real ambiguity. As nothing
is said at this point about any reaction by Pharaoh to what Moses
has said (one might well infer that he and his courtiers were dumb-
struck by it), Moses’ anger is at first sight unexplained. But if it is
recognised, as seems likely for other reasons, that Moses’ words in
vv. 4-8a are the conclusion of the dialogue begun in 10.24-26, 28-29,
Moses’ anger is fully explained by the enduring effect of Pharaoh’s
obstinacy and threats there. Some scholars (e.g. Van Seters, Life of
Moses, p. 108) seek to make the link even closer by proposing that
10.28-29 originally stood between v. 8a and v. 8b, but this is hardly
necessary and no convincing explanation has been given for why
10.28-29 should have been moved from here to where they are now.
9-10. It is likely that both these verses, which contain a short
explanatory word of Yahweh to Moses and the (or a) narrator’s
summary of the whole plague-story, refer back to and summarise the
preceding narrative. This is agreed to be the case for v. 10, but many
commentators (e.g. Houtman, Propp, Dozeman) and translations
(e.g. JB, NJPS, NRSV) understand Yahweh’s words in v. 9 to refer
to a still future refusal of Pharaoh to accede to Yahweh’s demands.
It is true that the Heb. imperfect tense used here commonly has a
future meaning, so that ‘will not listen to you’ is in itself an accept-
able translation and it could be seen as an immediate warning that
Moses’ words in vv. 4-8 will have no effect. The problem with this
is that on this occasion Pharaoh does relent and let the Israelites
go (12.30-31). In fact the Heb. imperfect can also refer to repeated
action in the past and present as well as action in the future (a clear
example in Exodus is in 33.7-11), and so Pharaoh’s not listening can
18 EXODUS 1–18
be a reference to his repeated not listening in the earlier narratives of

the plagues (see Note v on the translation). Such an understanding
is supported by the following v. 10, where the same word ‘wonders’
(Heb. môpetîm) is used, and there it clearly refers to the earlier
plagues. On this view the two verses provide a concluding summary
of the plague-narrative just as 7.1-7 serves as its introduction, and
there both the idea of ‘hardening Pharaoh’s heart’ (with a different
Heb. verb) and a reference to ‘multiplying [signs and] wonders’ also
occur. Those verses are part of the Priestly narrative and it is likely
that these are too, since their wording (including the word used for
‘hardening’ here) is also similar to that of the conclusions to the
individual Priestly plague-stories (esp. 7.13, 22; 8.11, 15; 9.12), but
not so close to the conclusions of the other episodes in the present
text, which come from the non-Priestly version.
Text and Versions

‫( עוד‬11.1) TgJ and Sy use ‫תוב‬/twb, normally ‘again’, so perhaps pointing
more to an additional plague rather than to ‘just one more’. But the difference
is slight.
‫( נגע אחד‬11.1) Most of the Vss use general words for ‘blow, plague’, but
Aq and Symm preserved the etymology by using ἁφήν (cf. Aq in Gen. 12.17).
Vulg did the same by rendering the verb by tangam, while retaining plaga. TgJ
added ‘which will be the harshest of all upon them’.
‫( אחרי־כן‬11.1) SP reads ‫ואחרי־כן‬, and LXX, Vulg, TgJ,N and Sy follow
suit. But the effective asyndesis of MT is supported by TgO,Nmg and is surely
original.
‫( ישׁלח‬11.1) Sy and TgNmg render in first person si., continuing the divine
subject from ‫אביא‬.
‫מזה‬1o (11.1) Vulg has no equivalent in its very free rendering of the latter
part of the verse, which seems designed to omit anything superfluous.
‫( כשׁלחו‬11.1) LXX introduces a second ὑμᾶς after this (similarly TgN). Sy
with šdrtkwn again has first person si., with Yahweh as the subject, possibly
from a variant Vorlage ‫כשׁלחי‬. But Sy seems to have rewritten much of this
verse to give it a stronger theological thrust (and maintaining the first person
si. subject of ‫ אביא‬throughout).
‫( כלה‬11.1) LXX συν παντι and Sy klkwn saw here a form of ‫כֹּל‬, which
has inspired some modern emendations of MT (see below). Vulg ignored it,
but TgO,J,Nmg used ‫גמירא‬, ‘completeness’, which is probably correct (see Note
d on the translation). TgJ added ‫יהי לה‬, ‘will happen to him’, seeing here a
reference to ‘complete destruction’ as in the uses of ‫ כלה‬in Jeremiah, but this
does not fit the wording of MT. TgN’s ‫בפרעיה‬, ‘quickly’, is probably a guess.
11.1-10 19
Emendations to a suffixed form of ‫ כל‬or a form of the verb ‫ כלה‬have been

proposed in modern times, but are not necessary.13
‫( גרשׁ יגרשׁ אתכם‬11.1) Sy pwqw lkwn, ‘you depart’ (imperative) is very
free and misses the main point: it is probably modelled on Pharaoh’s words
in 12.31.
‫מזה‬2o (11.1) Here not only Vulg but LXX has no equivalent to this
superfluous expression: Aq, Theod and the hexaplaric text correct to MT’s
reading. Lemmelijn’s view that the omission of ‫ מזה‬here is more original
(p. 190) overlooks the freedom with which the translator dealt with non-
essential features of the text.
‫( דבר־נא‬11.2) SP has a pl. verb (presumably assuming that Aaron is
addressed too, as SP’s longer additions often do: cf. 9.5, 19), while Sy
prefixes the copula. MT is to be retained. The renderings of ‫ נא‬here (LXX
οὖν; Vulg ergo; TgO,N ‫ ;כען‬TgJ ‫ ;כדון‬Sy nil) are typical, although LXX and
Vulg frequently have no equivalent and Sy occasionally uses hšʾ (in Exodus
only 10.11 and 34.9 [2o]). The indications therefore are that LXX and Vulg
took it to mean ‘therefore’ and the Aram. Vss understood it as ‘now’ (though
Sokoloff, Dictionary, who is followed by CAL, says that ‫ כען‬is in these cases
an enclitic particle). A precative understanding is only very rarely found: TgJ
in 33.13 (1o) and 34.9 (1o); Vulg in 34.9 (2o). See also Note e on the transla-
tion, which casts some doubt on the precative use being original.
‫( באזני העם‬11.2) The idiom is a frequent one, but only LXX renders it lit-
erally: TgJ,N come closest with ‘in the hearing of’, while TgO and Sy generalise
with ‫קדם‬, ‘before’. Vulg omni plebi turns it into an expression for the extent
of the instruction. LXX prefixes κρυφῇ, ‘secretly’, which Wevers suggests
‘makes explicit what is implicit in MT’ (Notes, p. 162: cf. Lemmelijn, pp. 190-
91), but it is certainly not implicit in the Heb. idiom: this should be recognised
as a place where the translator embroidered his Vorlage.
‫( וישׁאלו‬11.2) Sy has no equivalent to the copula here, but its evidence
alone is scarcely sufficient to make the shorter text original.
‫( אישׁ‬11.2) Sy ʾnš gbr, ‘each one, a man…’, gives expression to both the
uses of ‫ אישׁ‬that are involved here. TgJ adds ‘Egyptian’ after each of the words
for ‘friend’ to clarify the meaning. LXX does not represent the possessive
suffixes, but equivalents are restored in the Three and Vulg.
13
See McNeile, pp. 60-61 n. 1; and further refs. in Houtman, p. 130: so
most recently in Fuss, p. 255, and HAL, p. 455. According to Houtman, Ehrlich
proposed to read ‫ !לילה‬An ingenious change of the vowels alone to ‫ ְכּ ִשׁ ְלּחוּ ַכ ָלּה‬, ‘as
one sends a bride’ (sc. with gifts) was first contemplated by A. van Hoonacker and
taken up by J. Coppens, ‘Miscellanes Bibliques, XIII’, ETL 23 (1947), pp. 178-
79, comparing 1 Kgs 9.16. This found some support initially (see the refs. in
Childs, p. 130) and in part at least in NEB and REB (cf. Brockington, p. 10). But
the grammar is dubious (GK §155g; JM §174d) and the change, like the others
proposed, is in any case unnecessary.
20 EXODUS 1–18
‫( וכלי זהב‬11.2) In LXX and Vulg there is no equivalent to ‫ כלי‬this time,

since Greek and Latin idiom (and the use of adjectives for ‘silver’ and ‘gold’)
makes repetition unnecessary. SP adds ‫( ושׂמלות‬cf. LXX καὶ ἱματισμόν), in
agreement with 3.22 and 12.35, having earlier modified 3.22 to agree with this
verse about the persons involved. The readings of MT and the other Vss which
preserve the distinctiveness of each passage are primary and the harmonisa-
tions secondary.
‫( ויתן יהוה‬11.3) LXX, Tgg and Sy follow MT, but SP ‫ ונתתי‬makes v. 3a the
continuation of the direct speech of God in v. 2 and a promise of what is yet
to come (cf. 3.21, where this very wording occurs) instead of MT’s report of
what happens next. Vulg dabit achieved the same result with less departure
from the wording of MT here (cf. Propp’s repointing of MT to ‫וְ יִ ֵתּן‬, understood
as a future [p. 342]; but in normal BH prose this would be ‫)ונתן‬. TgNmg as often
prefixes ‘the Memra of’.
‫( את־חן העם‬11.3) SP adds ‫הזה‬, evidently harmonising again with 3.21. The
Vss all change the construction according to their native idiom, with LXX and
Vulg correctly understanding that the ‘genitive’ ‫ העם‬is equivalent to a dative,
TgN understanding ‫ ויתן‬in the sense ‘put’, and TgO,J and Sy making the ‘favour’
(‫רחמין‬, ‘love, mercy’) into an indirect object.
‫( בעיני מצרים‬11.3) SP adds ‫( והשׁאילום‬the mss almost all have the plene
form, against von Gall) to match the account in 12.36, as LXX καὶ ἔχρησαν
αὐτοῖς does even more precisely.
After v. 3a SP inserts its own version of vv. 4b-7 (see the notes below on
these verses for its variations from MT), thus extending the divine speech
further and ensuring (as in its addition after 10.2) that Moses only speaks to
Pharaoh words that have already been given to him by Yahweh. It is likely that
4QpalExm had the addition too, but it does not survive for this part of ch. 11.
Wevers’ edition this time cites no Greek equivalent to the plus from Syhex – all
it cites (p. 161) is a version of vv. 3b-7 with the addition from 4.22-23 (on
which see below, at the end of the notes on v. 3). The same omission appears
in Brooke-McLean, pp. 189-90, and in the Lagarde ed. of Syhex.
‫( גם‬11.3) SP has ‫וגם‬, avoiding the asyndesis of MT, no doubt finding
this the more necessary after its change and additions before. But the regular
‘weak’ renderings of ‫ גם‬in LXX (καί) and Vulg (-que) served a similar purpose.
‫( גדול מאד‬11.3) LXX, Vulg and Sy add an explicit ‘was’ as required by
native idiom. Most of the Vss left the precise sense of ‫ גדול‬undefined, but TgN
‫תקיף‬, ‘mighty’, and Sy yqyr, ‘honoured’, gave it more specific interpretations.
‫( בארץ מצרים‬11.3) From here to the end of the verse LXX renders by
three parallel phrases, ἐναντίον τῶν Αἰγυπτίων καὶ ἐναντίον Φαραώ καὶ
ἐναντίον πάντων τῶν θεραπόντων αὐτοῦ. The differences in detail from
MT are considerable and even Wevers, for once, concedes that a variant
Vorlage must be responsible (Notes, p. 163; cf. BAlex 2, p. 142; Propp, p. 309;
Lemmelijn, pp. 91, 192-94). This may not be necessary: LXX’s first phrase
may be a paraphrase shaped by what was to follow, the second may be the
11.1-10 21
translator’s inference from a passage such as 10.16-17, the addition of πάντων

is in line with earlier ‘heightening’ of the narrative in LXX (cf. 10.4, 15), and
MT’s ‘in the sight of the people’ may have seemed redundant once ‫בארץ מצרים‬
had been paraphrased. The omission of the final phrase was remedied by
Aq, Theod and, in consequence, the O-text (with ἐν ὀφθαλμοῖς rather than
ἐναντίον as a typical verbatim rendering of the Heb.: cf. O’Connell, Theodo-
tionic Revision, pp. 175, 283-84). A geographical understanding of ‫ארץ מצרים‬
seems clearly to lie behind 2QExa’s [‫[בתו]ך ארץ‬:
ׄ compare the similar variant
in 4QpalExm at 7.18 (and the plus there).14
‫( עבדי־פרעה‬11.3) TgN renders ‫ עבדי‬by ‫שׁליטוי‬, ‘his rulers’, reflecting (as
often) the reference rather than the sense of the Heb.
‫( ובעיני העם‬11.3) Vulg (coram, not repeated) omni populo, a natural
expansion after the preceding reference to a specific group.
Before v. 4 SP adds, with an appropriately expanded introductory formula,
the words that Moses was instructed to say to Pharaoh in 4.22-23, with
the substitution of ‫ יהוה‬for ‫ אנכי‬in v. 23 (presumably because Moses is now
delivering Yahweh’s message), evidently judging that this final encounter
was the appropriate time for such a warning (cf. vv. 4-7, which take up the
same theme). Again 4QpalExm does not survive, but is likely to have had
this harmonising addition. It also appeared in the Sam. Gk. version, which is
preserved in Syhex (including the variant ‫ יהוה‬from 4.23).
No evidence survives from Qumran for the division between vv. 3 and 4
which is marked in the medieval mss. One is restored in 4QpalExl and 2QExa,
but without any strong basis.
‫( ויאמר משׁה‬11.4) Vulg et ait found the repetition of the explicit subject
unnecessary after v. 3b. 2QExa (]‫)אל פרע[ה‬,TgJ and some mss of LXX make
the addressee explicit, in accordance with v. 8 (cf. SP’s expanded introduction
in its addition after v. 3).
‫( יהוה‬11.4) TgNmg has ‘the Memra of Yahweh’, as usual.
‫( כחצת הלילה‬11.4) SP writes ‫ כחצית‬for the first word, implying a different
vocalisation from MT here and a form that is not found elsewhere (except in
SP’s own secondary plus in the middle of v. 3). There is no evidence from
Qumran at this point. The form in MT is itself rare (see Note k on the transla-
tion), with ‫( חצי‬or ‫ מחצית‬in LBH) being preferred in such expressions. Possibly
SP was influenced by these more common forms (and cf. 12.29). The Vss,
which render idiomatically (for LXX περὶ μέσας νύκτας cf. LSJ, p. 1185,
and Xen., Anab. 7.8.12), are of course of no help in resolving the textual
issue here. TgJ renders ‘tomorrow night at/about this time’, presuming that it
is now midnight on the 13th of the month Abib, an interpretation also cited
in B.Ber. 4a.
14
DJD III, p. 51, regards 2QExa as supporting SP here, but it has ‫בארץ מצרים‬
like MT at this point: the occurrence of ‫ בתוך‬in SP comes later in the passage.
22 EXODUS 1–18
‫( אני יוצא‬11.4) LXX’s εἰσπορεύομαι (followed, in a future form, by OL)

disregarded the meaning of Heb. under the influence of the context: Vulg’s
egrediar restored the true sense (cf. Sy). Tgg have ‫מתגלי‬, using a frequent
paraphrase which is especially appropriate here where Heb. speaks of Yahweh
‘coming forth’ (so also Tg at Mic. 1.3). The same expression occurs in 12.12
(cf. Chester, Divine Revelation, pp. 145-46, 148-49; AramB 7, p. 29 n. 3
[comparing Gen. 11.5, 7]). TgN also has ‫ ממרי‬for ‫אני‬, producing a combination
which is most common in TgNmg, but appears in TgF,G at 19.11 and in TgJ at
Gen. 11.8 in additional material (see Chester, pp. 103-104).
‫( מצרים‬11.4) SP prefixes ‫( ארץ‬from v. 3 or v. 5?), clarifying the geograph-
ical interpretation of ‫מצרים‬, which most of the Vss follow (but TgJ’s ‫מצראי‬
takes it to mean the people). Both 4QpalExl and 2QExa preserve the word
‫ מצרים‬but not what precedes it.
‫( ומת‬11.5) TgN has ‘will be killed’, perhaps a deliberate echo of the
wording of 4.23, but also anticipating the personal intervention of Yahweh in
12.12, 29. Sy puts the whole clause into the pl., presumably only a stylistic
change.
‫( בארץ מצרים‬11.5) Vulg Aegyptiorum takes ‫ מצרים‬to refer to the people
(cf. TgJ at the end of v. 4). The Leiden ed. of Sy (following most early mss)
prints dʾrʿʾ dmṣryn, ‘of the land of Egypt’, a free rendering which 7a1 like
the other mss has in 12.12, 29, whereas here it corresponds closely to MT
(possibly through ‘interference’ from Sy’s rendering of the next verse rather
than consultation of the Heb.).
‫( הישׁב‬11.5) Most of the Vss take the natural view that the participle refers
to Pharaoh himself, but the Tgg linked it to ‫ בכור‬instead and understood it in a
future sense: ‘who is [TgN: was] destined to sit on the throne of his kingdom’
(so recently Houtman). For further references see AramB 7, p. 29 nn. 4-5.15
‫( כסאו‬11.5) LXX does not represent the suffix, as τοῦ θρόνου can imply
‘his throne’.
‫( עד‬11.5) MT is supported by 4QpalExl, Tgg and Vulg. SP has the
idiomatic ‫ועד‬, which could be behind Sy’s wʿdmʾ. The LXX mss are divided
between ἕως (= MT) and καὶ ἕως (= SP), but Wevers argues cogently that
the former conforms to the style of the translator in contexts such as this
(Notes, p. 164; THGE, pp. 164-65) and is the original Greek reading. The
reading with καί is scarcely natural Greek and could perhaps be attributed to
later scribes’ familiarity with ‘Hebraising’ translations (or to influence from a
similar phrase at the end of the verse). 2QExa preserves ‫[ועד‬, but according to
DJD III, p. 51, it is not clear whether the waw is the conjunction or the suffix
of ‫כסאו‬. MT’s shorter reading is probably preferable.
‫( השׁפחה‬11.5) TgJ heightens the contrast with Pharaoh’s firstborn by
adding that she was ‘the least in Egypt’.
15
TgNmg has in place of ‘to sit on the throne’ a variant which seems to mean ‘to
hate (‫ )למתעב‬all the thrones’. But see the editio princeps, p. 66 n. 2.
11.1-10 23
‫( אשׁר אחר הרחים‬11.5) ‫ אחר‬apparently troubled LXX and Vulg, who render
‘beside’, and TgO, with ‘who is in the mill-house’. TgN clarifies by adding ‘is
grinding’, while TgJ goes even further with ‘who was born to her while she
was grinding’. Both here and earlier in the verse the Tgg make the short state-
ment much more interesting to the hearer or reader.
‫( וכל בכור בהמה‬11.5) SP recasts this phrase in line with the previous
one, making a more elegant sentence: ‫( ועד בכור כל בהמה‬cf. also 13.15). The
Vorlage of LXX was evidently identical to this, a point which Wevers (Notes,
p. 164) overlooks. The more abrupt wording of MT, which the other Vss
reflect, is probably original.
‫( והיתה‬11.6) TgN has ‫והווה‬, which if correct would imply a past sense (‘and
there was’) in Aram. As later in the verse (see on ‫ )לא תסף‬it is most likely that
the scribe has been influenced by a similar passage that was more familiar, in
this case 12.30.
‫( בכל־ארץ מצרים‬11.6) MT’s reading is supported by the fragmentary
remains of 2QExa ([‫ )בכל‬and 4QpalExl ([‫)ב ׄכ[ל] ׄא ׄר ׄץ‬,
ׄ LXX, Vulg and Tgg. SP
has ‫במצרים‬, which is clearly due to assimilation to the later narrative in 12.30
(LXX made the opposite change). Sy’s bʾrʿʾ dmṣryn (i.e without ‫כל‬: the same
rendering appears at 12.30) may in some way be related to the shorter phrase
in v. 5.
‫( כמהו‬11.6) SP has the expected fem. form for both occurrences, and
most of the Vss naturally make the gender consistent throughout the verse
(TgJ’s ‫ כותיה‬is an exceptional retention of MT’s masc.). There is no surviving
evidence from Qumran. Odd as it is, MT can be retained (see Note n on the
translation). Some of the Vss expand their renderings for additional clarity,
Vulg by adding ante in the first subordinate clause and TgJ by making ‘a night’
and ‘a plague’ (‫ )מחתא‬the subject in the two clauses.
‫( לא תסף‬11.6) LXX’s οὐκέτι προστεθήσεται, ‘shall no more be added’,
supplements its usual rendering of ‫ יסף‬by προστίθημι with the alternative
equivalent used in 10.29. The woodenness of the rendering is partly due to
the unusual ellipse of the following inf. cons. (here ‫ )להיות‬in the underlying
Heb. Vulg (postea fuerit) and Sy (simply thwh, ‘shall be, happen’) found more
idiomatic equivalents.
‫( ולכל בני ישׂראל‬11.7) The choice of ἐν (LXX), apud (Vulg) and mn d (Sy)
to represent MT’s ‫ ל‬tends to suggest that dogs belonging to the Israelites are
meant (cf. Jub. 49.4), although this is clearly not what the text has in view.
‫( יחרץ‬11.7) The meaning ‘bark’ or ‘growl’ is given by LXX (γρύξει: some
mss read βρύξει, ‘bite, gnash’) and TgN: Vulg’s muttiet, ‘shall mutter’, perhaps
softens the meaning for greater effect. That this is a contextual interpretation
is shown by the other Vss, which start from the sense ‘hurt’ (TgO,J, Sy),16
16
Sy nhr is from hr, for which Payne Smith (p. 106) gives ‘bark’ as a possible
meaning. But it is clear, especially from the derivatives, that the real meaning of
the verb is ‘fight, strive, hurt’.
24 EXODUS 1–18
which is a natural extension of the common sense ‘cut’ for ‫ ;חרץ‬TgO,J then add
‘by barking’ as their understanding (because of ‫לשׁנו‬, which they like LXX,
TgN and Sy regard as instrumental) of what this means in the context. SP
supports the reading of MT (for which Josh. 10.21 provides a good parallel –
see Note q on the translation): the old ms. Camb. 1846 does seem to read ‫ירחץ‬,
but this will be a simple case of metathesis towards a more familiar verb. For
rabbinic discussion of the expression see the refs. in AramB 2, p. 45 n. 5.
‫( תדעון‬11.7) MT is supported by 4QpalExl, the only Qumran ms. to
preserve this word, Vulg, Tgg and Sy. But SP reads the sing. ‫תדע‬, presumably
with Pharaoh rather than all the Egyptians in mind, and LXX’s εἰδῇς agrees
with it. Both readings make sense, but the pl. is the more difficult reading in
the present context, where v. 8 indicates that Moses’ words are addressed to
Pharaoh, and also represents a departure from earlier indications of Yahweh’s
purpose in the plague narrative, where the focus is on Pharaoh alone (cf. 7.17;
8.6, 18; 9.14, 29; 10.7): the reading of SP and LXX could well have been
harmonised with these passages, whereas a change in the opposite direction
is more difficult to explain. The pl. is not impossible in words addressed to
Pharaoh, since it is equivalent to ‘you and your people’.
‫( אשׁר‬11.7) LXX ὅσα may have its distinctive sense here (‘the great things
which’: cf. Vulg quanto miraculo): the choice of a relative pronoun (rather
than a word for ‘that’ such as ὅτι) is probably a consequence of LXX’s under-
standing of the following word (see the next note).
‫( יפלה‬11.7) LXX again, as in 8.18 and 9.4, read this as a form of ‫פלא‬,
‘be wonderful’, in Hiph. ‘do wonders’, and SP (as in 9.4) spells the word
with a final aleph (see Text and Versions on 8.18 and 9.4). But Sam.Tg.
‫( יפרשׁ‬cf. 8.18; 9.4) shows that SP did not intend a sense different from MT
and simply displays its characteristic fluidity in the use of the gutturals. Tgg
and Sy follow the sense of MT here, while Vulg combines both alternatives
with quanto miraculo dividat, which is probably a partial correction of the
OL in the light of a text like MT and/or Jerome’s knowledge of contemporary
Jewish interpretation.
‫( בין מצרים ובין ישׂראל‬11.7) Sy byt mṣryʾ ldbyt ʾysrʾyl, where the first byt is
the common abbreviated spelling of bynt, ‘between’, the l corresponds to the
second ‫( בין‬in an idiom which is also found in later BH), and dbyt means ‘those
of the house of (Israel)’, a small departure from the unusual use of ‫ ישׂראל‬alone
in the other witnesses (cf. Sy’s substitution of byt for ‫ בני‬at the beginning of
the verse, and the list in Propp’s note on 3.11 [p. 185]).
‫( והשׁתחוו‬11.8) SP has ‫וישׁתחוו‬, presumably to be read as simple waw with
the imperfect to give the same meaning as MT, an isolated lapse into the
later syntax. No Qumran ms. survives at this point, and the future forms in
the Vss are inconclusive. The regular usage of Exodus makes MT’s waw
consecutive and the perfect more likely to be original: simple waw with
the imperfect and related PC forms elsewhere indicates either purpose (e.g.
11.1-10 25
2.7) or a modal sense (e.g. 3.3).17 The meaning of the verb is represented as
expected (‘bow down, worship’) in LXX (though the accusative με following
προσκυνήσουσί reflects classical Gk. idiom rather than normal Septuagintal
usage), Vulg and Sy, but the Tgg found this inappropriate to a human figure
and rendered either ‘ask (a favour)’ (TgO,J) or ‘salute, greet’ (TgN).
‫( לאמר‬11.8) Sy wnʾmrwn ly, using a finite verb for the Heb. inf. (cf.
13.19) and repeating ‘to me’, both examples of its sometimes ‘free but
faithful’ translation technique.
‫( העם‬11.8) LXX (‘your people’) and Sy (‘this people’) secondarily antici-
pate the specification which is to follow.
‫( אשׁר ברגליך‬11.8) LXX οὗ σὺ ἀφηγῇ, ‘whom you lead’ (the rel. pron. is
in the gen. as the obj. usually is with ἡγέομαι and compounds: contra Wevers,
Notes, p. 166; Lemmelijn, p. 93 n. 361), Vulg qui subiectus est tibi and TgNmg
‫דאית בידך‬, correctly recognise the sense of the Heb.; TgO,J,N and Sy have ‘which
is with you’, which is too imprecise.
‫( ואחרי־כן‬11.8) Sy hydyn, ‘then’, loses a little of the sharpness of Heb.
‫( אצא‬11.8) Vulg egrediemur gives the pl. which the situation implies, but
Heb. is more specific.
‫( ויצא‬11.8) LXX and Sy supply the subj. ‘Moses’ again, clearly second-
arily (cf. Lemmelijn, pp. 182-83).
‫( בחרי־אף‬11.8) LXX μετὰ θυμοῦ, ‘angrily’, is rather weak, and the other
Vss (and some mss of LXX which add μεγάλου) come closer to the Heb. with
‫( בתקוף רגז‬Tgg), iratus nimis (Vulg) and bḥmtʾ rbtʾ (Sy).
4QpalExl had a major break after v. 8, and 4QpalExm an interval within a
line. In 4QExc v. 9 seems to have begun a new column and DJD XII, p. 111,
reconstructs an interval at the bottom of the previous column.
‫( יהוה‬11.9) TgNmg prefixes ‘the Memra of’.
‫( ישׁמע‬11.9) LXX and Vulg use future verbs; cf. the imperfects of Tgg
and Sy.
‫( רבות‬11.9) Vulg and TgN follow MT, but the other Vss render as if the
verb were Hiphil like that in 7.3, with LXX πληθύνω and and Sy dʾsgʾ making
Yahweh the explicit subject as there, while the inf. Aphel of TgO,J leaves the
subject undefined (and could imply that it is Pharaoh who brings the conse-
quences on himself).
‫( מופתי‬11.9) TgNmg added ‘these’ as in v. 10. Several of the Vss rendered
as if ‫ אתתי‬were present, either additionally (LXX, TgN) or instead of ‫מופתי‬
(Vulg: cf. 7.9). The influence of the earlier narrative, esp. 7.3 but also 8.19 and
10.1, is no doubt responsible (cf. Lemmelijn, p. 195).
17
The only exceptions would seem to be 15.2 and 15.17 in the Song of Moses,
19.3 in a semi-poetic parallel clause, 23.8 similarly, and 24.7 in a clause synony-
mous to the previous one.
26 EXODUS 1–18
‫( המפתים‬11.10) LXX prefixes τὰ σημεῖα καί, as in v. 9, and TgN has

‫נסייה‬, using its usual equivalent for ‫ אות‬here for ‫מופת‬. See the previous note.
Vulg, however, reverts (for variety?) to its more precise rendering ostenta (a
Ciceronian word: cf. 4.21; 7.3).
‫( האלה‬11.10) Vulg quae scripta sunt makes the retrospective reference
of ‘these’ explicit. LXX again expands its translation, adding ἐν γῇ Αἰγύπτῳ
from v. 9.
‫( לב‬11.10) TgJ, as usual in this expression, prefixes ‫יצרא‬: see Text and
Versions on 4.21 and 7.3.
‫( ולא שׁלח‬11.10) LXX καὶ οὐκ ἠθέλησεν ἐξαποστεῖλαι psychologises
Pharaoh’s lack of response as in 8.28, with some support from statements in
4.23; 7.14, 27; 9.2; 10.3-4.
‫( מארצו‬11.10) LXX ἐκ γῆς Αἰγύπτου names Egypt again (for an earlier
case see 7.11).
After v. 10 4QpalExl has a clear vacat beneath, and DJD XII, p. 113,
conjectures the same for 4QExc. But in 4QpalExm there is no room for any
division before ch. 12 (DJD IX, p. 85).
C h ap t er 1 2 . 1 - 2 0
I nstr u cti o n s fr om Y ah weh a b ou t t he

P assover an d U n l eaven e d B r e a d
The boundaries of this section are clearly defined by the introduc-

tory speech-formulae in vv. 1 and 21 and are already represented
by an empty line before v. 1 in 4QpalExl and a likely reconstructed
division before v. 21 in 4QpalExm (where 4QpalExl does not
survive), as also by section-divisions in the medieval mss (both
Maimonides and BM Or. 4445 actually have a setumah before
12.1 where Codex L has a petuchah). But 4QpalExm evidently did
not have a division before 12.1, as there is no room for one, while
it may have had one before 12.17b (see Text and Versions). The
initial speech-formula presents the section as a divine speech to
Moses and Aaron, which makes it likely that there is an original
connection with v. 28 (‘as Yahweh had commanded Moses and
Aaron’): see also the introduction to 12.21-28. Within the section
the divine first-person pronoun is found only in vv. 12-13 and 17;
in vv. 11 and 14 Yahweh is referred to in the third person, but in
a special formula (‘for Yahweh’) which is frequent elsewhere in
divine speeches (e.g. 28.36; 29.18, 28; 30.12-13, 37; 31.15). It has
been noted that the instructions for Israel are formulated alternately
in the third person plural (‘they’ etc.: vv. 3b-4a, 6b-8) and in the
second person plural (‘you’ etc.: vv. 2, 4b-6a, 9-14, 15-20), which
might have implications for the earlier history of the passage (e.g.
von Rad, Priesterschrift, pp. 47-49; Rendtorff, Gesetze, pp. 56-
57; Laaf, Pascha-Feier, pp. 11-16; Kohata, Jahwist, pp. 262-66:
see further below). The major sub-division within the section is
between the instructions about Passover and those about the eating
of unleavened bread.1 There is disagreement over whether the break
1
This is fundamental to the recent examination of Exod. 12 by Gesundheit/
Bar-On (Three Times, pp. 44-95), who also holds that v. 22 was the original
continuation of v. 11 rather than belonging to a different source: on this see the
introduction to 12.21-27.
28 EXODUS 1–18
between these sections follows v. 13 or v. 14: see the Explanatory

Note on v. 14, where it is concluded that the break comes after
v. 14. If so, both Passover instructions and the first part of those
about unleavened bread end with the same formula, ‘as an everlast-
ing statute’ (vv. 14, 17). There is indeed a good deal of repetition,
and some significant variation, between vv. 15-17 and vv. 18-20
(see the Explanatory Notes), which suggests that these sections
have separate origins.
In more detail the passage may be divided as follows: (i) introduc-
tory speech-formula (v. 1); (ii) inauguration of calendar beginning
with the month of Passover (v. 2); (iii) commissioning formula and
instructions about the selection of the animal for Passover (vv. 3-
5); (iv) the guarding and slaughter of the animal (v. 6); (v) the
blood-ritual (v. 7); (vi) the cooking and eating of the animal and
the disposal of any remains (vv. 8-11); (vii) the announcement of
Yahweh’s imminent judgement on the Egyptians and their gods and
the sparing of the Israelite firstborn because of the blood smeared on
their doorways (vv. 12-13); (viii) the establishment of this festival
as a permanent observance (v. 14); (ix) rules about the removal
of leaven and the eating of unleavened bread for seven days as a
permanent observance (vv. 15-17); (x) further rules about the obser-
vance regarding unleavened bread (vv. 18-20).
It is noteworthy that the Israelites are referred to here for the first
time by some terms that will be frequently used later in Exodus and
in the rest of the Pentateuch: ‘assembly’ (Heb. qāhāl) and ‘congre-
gation’ (Heb. ʿēdāh). These expressions do not entirely displace the
representation of the Israelites as a ‘people’ or as ‘the children of
Israel’ (i.e. Jacob), and ‘congregation’ is frequently (though inter-
estingly not in this passage) combined with the latter expression.
But in slightly different ways (see the Explanatory Notes on vv. 3
and 6) they emphasise the identity of the people as an organised
unity, a feature which has already appeared in references to their
‘tribal divisions’ (Heb. ṣebāʾôt) earlier in the narrative (6.26; 7.4:
so also here in v. 17). As will appear below, the occurrence of these
designations for the people is characteristic of the later portions of
the Pentateuchal narrative and laws.
12.1-20 29
From Knobel onwards the whole of this section has been almost without
exception attributed to the Priestly material of Exodus (cf. Exod.-Lev., p. 91;
Wellhausen, Composition, pp. 72-73).2 The observation that the Priestly text
in Exodus 12 was not a coherent unity may have led to Dillmann’s view that
vv. 14-20 originally stood after v. 49 (pp. 98-100): soon afterwards Jülicher
(‘Exodus VII,8 – XXIV,11’, pp. 107-109) introduced a proposal that has
been much more influential, that the legislation about Unleavened Bread
in vv. 15-20 was a later addition to P (so also Carpenter/Harford-Battersby,
Holzinger, Baentsch, Smend [p. 137], Rudolph [possibly: p. 24], Beer-
Galling, R. Schmitt [p. 20], W.H. Schmidt [p. 508 (with v. 14 possibly also
secondary: 498-99)], Kohata [p. 266: cf. n. 29], Ahuis [pp. 106-107, 120:
ascribed to DtrT], L. Schmidt [Priesterschrift, pp. 29-31], Graupner [p. 67],
Gertz [pp. 31-32]). But it has not been universally accepted (cf. Gressmann,
McNeile, Eissfeldt, Noth, Fohrer, Hyatt, Childs, Houtman, Van Seters, Propp,
Dozeman: the views of Blum, Blenkinsopp and K. Schmid are unclear).
Further secondary additions were identified by Holzinger (vv. 2, 4) and
Baentsch (vv. 2, 14): Baentsch seems also to have been the first (and the last
for a long time!) to suggest that within the Unleavened Bread law vv. 18-20
were later than the rest. The analysis of vv. 1-14 was taken further by von Rad,
who drew attention to the variation between second- and third-person refer-
ences to the Israelites and attributed it to the interweaving of two originally
separate versions of the Passover law, within the wider context of two parallel
Priestly narratives (Die Priesterschrift [1934], pp. 47-49). This explanation
was accepted by Beer-Galling (pp. 60, 63: so also later for chs. 25–31 + 35–40
[pp. 129, 165]), but it was generally viewed with some scepticism (Fohrer,
p. 88 n. 17; cf. p. 49 n. 67), until it re-emerged in a modified form with Rend-
torff (see below). The view that v. 2 was intrusive was more widely adopted
(Noth, Fohrer, R. Schmitt, Kohata, L. Schmidt, Gertz). In addition Fohrer
saw the Begründung in vv. 12-13 as a later insertion, but without giving any
reason (pp. 87-88). In 1954 Rendtorff had offered a different explanation of
von Rad’s form-critical observations, preferring to see the grammatical vari-
ation as due to the reworking of an older third-person ‘ritual’ by the Priestly
writer, who used the second person (Gesetze, pp. 56-57). Kohata followed him
closely (pp. 262-66), and Gertz (pp. 32-35) and W.H. Schmidt (pp. 493-95)
have taken the same view. Others have seen either the third-person verses
(Ahuis, pp. 36-39: DtrT) or the second-person verses (L. Schmidt, p. 29;
Grünwaldt, Exil, pp. 84-88: PS; Laaf, Pascha-Feier, pp. 11-16, unconvinc-
ingly divides them between PG [vv. 1, 3aα, 12-13] and PS [the remainder])
as secondary. A more extreme view, originating with Eerdmans, is taken by
Levin (p. 336) and Knohl (pp. 19-23, 52), who see all of vv. 1-14 as post-
Priestly (so also J.-L. Ska in ‘Les plaies d’Égypte dans le récit sacerdotal’,
2
Wellhausen mentioned the view of Kayser that vv. 11-13 were from JE, but
gave good reasons for rejecting it; on Ahuis and Gertz see below.
30 EXODUS 1–18
Bib 60 [1979], pp. 23-35 [30-34]). Kohata also, following Rendtorff and
Laaf, revived Baentsch’s view that vv. 18-20 were a secondary supplement to
vv. 15-17 (p. 266 n. 29; cf. W.H. Schmidt, p. 508), and Knohl in similar vein
attributes them to a late stage of his ‘Holiness School’ (Sanctuary, pp. 19-21;
cf. Gesundheit’s description of them as an ‘appendix’ [Three Times, p. 93]).
Gertz observes the duplication too, but in his view it is vv. 15-17 which are
the later addition, with v. 14, partly because they (and not vv. 18-20) betray
the influence of the Holiness Code (pp. 36-37: similarly Otto, ‘Innerbiblische
Exegese’, pp. 155-57 [PS and PentR], and Grünwaldt, Exil, pp. 90-96 [with
detailed vocabulary comparisons]). The idea that v. 14 is (like v. 17b) a late
element in the passage is shared, with some hesitation, by Propp, but he attrib-
utes the rest of vv. 1-20 to P (pp. 374, 380, 406-407).
There is good reason to attribute the whole passage to the Priestly

section of the Pentateuch in a broad sense. The inclusion of Aaron
with Moses as the recipient of legislation (v. 1), the use of precise
date-formulae (vv. 3, 6, 18), the designations for the people (espe-
cially ‘congregation’ and ‘tribal divisions’: see above), and other
specific expressions such as ‘between the two evenings’ (v. 6), ‘acts
of judgement’ (v. 12), ‘throughout your generations’, ‘as an ever-
lasting statute’ (both in vv. 14 and 17), ‘cut off from Israel’ (vv. 15,
19), ‘holy occasion’ (v. 16), ‘on this very day’ (v. 17) and ‘both aliens
and natives of the land’ (v. 19) all point this way (for references
see the Explanatory Notes). But, as the review of earlier scholar-
ship has shown, there is long-established and growing support for
the view that the passage reached its present form by a process of
amplification of its original core. Of particular importance is the
question whether from the beginning P included regulations about
a seven-day ‘festival’ of Unleavened Bread, but linked with this is
the discussion about whether a distinction should be made between
vv. 15-17 and vv. 18-20 and if so which section is the earlier. We
shall consider these issues first and other possible additions to the
original core afterwards.
Consideration of vv. 15-20 may benefit from the fact that there
are other Priestly passages (in the broad sense) where Passover and
Unleavened Bread are treated.3 Somewhere in the background of
these passages is the ‘cultic calendar’ of Ezek. 45.18-25, in which
there is an overwhelming concentration on the holy days of the first
3
There are non-Priestly passages too, but these can more appropriately be
considered in relation to vv. 21-27.
12.1-20 31
month: only v. 25 deals, very briefly, with the ‘feast’ (Heb. ḥag) in
the seventh month. The first and seventh days of the first month are
days of purification and atonement for the temple (vv. 18, 20), then
on the fourteenth day is the Passover, apparently also described as
a ‘feast’ and combined with seven days of eating unleavened bread
and daily sacrificial offerings. Nothing is said about any connection
with the Exodus or the other distinctive features of the ritual. In the
Priestly sections of the Pentateuch themselves the Passover is treated
again in Numbers 9, one year on from its inauguration. After the
regular celebration in the first month (vv. 1-5) provision is made for
those who could not keep the Passover at the regular time (through
uncleanness or, with the longer term in view, absence from home
on a journey): they are permitted to celebrate the Passover in the
second month (vv. 6-14). The details of the celebration correspond
closely to the prescriptions of Exodus 12, including the additional
laws in vv. 43-49 (cf. Num. 9.12). But there is no mention of a
subsequent seven-day period of eating unleavened bread, only of
its inclusion in the Passover meal itself (v. 11: cf. Exod. 12.8). This
points to a stage in the Priestly legislation about Passover when
it was limited to a single day (or rather night), just like the other
festival which receives special treatment in the Priestly texts, the
Day of Atonement (Lev. 16). This strongly suggests that, quite apart
from other arguments that have been put forward (e.g. about the
contradiction between a prolonged festival and the need for a rapid
departure), originally the laws about Unleavened Bread as a seven-
day celebration did not appear in Exodus 12, and in fact that they
were not added until after the supplementary legislation in vv. 43-49
and the still later composition of Num. 9.1-14. On the other hand,
the festival calendar in Leviticus 23, the fullest and latest of the
cultic calendars in the Pentateuch, does envisage that Passover will
be followed by a seven-day Feast of Unleavened Bread, beginning
on the fifteenth day of the first month, with a ‘holy occasion’ on the
first and final day (vv. 5-8). The same combination is reflected in
the equally comprehensive sacrificial list in Numbers 28–29, which
clearly follows throughout the pattern of Leviticus 23 (for Passover
and Unleavened Bread see Num. 28.16-25). Leviticus 23 is a central
part of the Holiness Code, which is now generally agreed to be later
than the original Priestly Writing, which it modifies in important
respects (see e.g. Nihan, ‘The Priestly Covenant’; and more fully his
From Priestly Torah to Pentateuch). Exodus 12.15-20 need not all
32 EXODUS 1–18
be as late as the Holiness Code, but the mention of ‘holy occasions’

in v. 16 is probably best explained if vv. 15-17 were inserted under
the influence of H. Verses 18-20 seem to be of separate origin (see
the Explanatory Note on vv. 15-17) and could well be a somewhat
earlier addition. Their only new requirement is the inclusion of
resident aliens, which is already present in Exod. 12.43-49 and
Num. 9.14, and v. 20 still envisages that Unleavened Bread will
be observed ‘in all your dwelling-places’ and not as a centralised
feast. The date for the beginning of Unleavened Bread in v. 18 also
makes it coincide with Passover itself, rather than being a day later
as in Lev. 23.5-6. It may even be possible to date this addition quite
precisely, since it corresponds closely to the main instructions in the
‘Passover Papyrus’ from Elephantine (Cowley 21; TAD 1, A4.1),
which is dated to the fifth year of Darius (II), 419 B.C.
This conclusion, that vv. 15-17 are a later addition to the passage
than vv. 18-20, has the support of Otto and Gertz (see above), but it
does seem at first sight to go against the similar concluding formulae
in vv. 14 and 17, which suggest that vv. 1-17 were composed, or
at least edited, as a unit. However, it is possible to see this struc-
turing of the passage as a deliberate intervention of the redactor
who inserted vv. 15-17, to underline the eternal validity both of the
main Passover law and of his version of the supplementary ruling
about Unleavened Bread. Knohl may go too far in claiming that
the ‘everlasting statute’ formula is always a mark of the ‘Holiness
School’ (Sanctuary, pp. 46-55), but it was clearly popular in such
contexts (Lev. 17.7; 23.14, 21, 31, 41; 24.3) and its use here fits
well with other indications of affinity to the Holiness Code in
vv. 15-17. Provisionally then we may see vv. 18-20 as an initial
expansion of vv. 1-14a to include the traditional seven-day obser-
vance of Unleavened Bread and vv. 14b-17 as a later addition
designed to bring this aspect of the ongoing celebration by later
generations into closer conformity to the pattern in the Holiness
Code.
Returning now to the Passover instructions themselves in vv.
1-14(a), we may reasonably infer from the variation between second
and third person forms first noted by von Rad that the Priestly
authors built upon an existing ritual prescription, to which they
added more detailed prescriptions which were partly an assimila-
tion to traditional sacrificial practice (vv. 5a, 6a) and partly also a
reaction against new developments in the Deuteronomic law about
12.1-20 33
Passover (vv. 4b, 5b, 9a: cf. Deut. 16.1-8).4 The original omission of
a separate festival of Unleavened Bread would have been a depar-
ture not only from Deuteronomy but from older commemorations
of the Exodus (cf. 23.15; 34.18). At the same time the features on
which P agrees here with Deuteronomy 16 provide further evidence
of its grounding in older tradition.
There remains the problem of the calendrical ruling in v. 2, which
a growing number of scholars have regarded as a secondary element
in the passage (see above). The structure of vv. 1-3 is certainly
somewhat unusual: an introduction to a divine speech (as in v. 1)
is usually followed directly by an instruction to ‘speak’ to the
people (as here in v. 3), and as the text stands the repeated ‘for you’
seems to refer only to Moses and Aaron, the addressees specified in
v. 1, when the ruling is clearly intended for the people as a whole.
Moreover, both Holzinger and Baentsch found a ruling about the
calendar quite inappropriate to the narrative (‘historical’) context,
where a means of protection for Israel is needed (and subsequently
provided). But these objections are not as compelling as they seem.
Whatever the standard pattern for introducing instructions to be
passed on to the community, the Priestly writers were able to vary
it if a matter of sufficient importance demanded this, and it is not
at all difficult to believe that they regarded the establishment of the
calendar, according to which worship was to be regulated, as highly
important. In fact, earlier in Exodus there is a similar delay between
the speech-introduction and the instruction to Moses to speak to
the people in 6.2-6, where Yahweh tells Moses the basis for what
he is going to do before commissioning him to tell the people of
the coming deliverance. The claim that in the context ‘for you’ can
mean only ‘for Moses and Aaron’ rests on a pedantic reading of the
text which overlooks the representative function which Moses and
Aaron have: ‘for you’ can perfectly well mean ‘for you two and
all the rest of the Israelites’, and no reader would suppose other-
wise (cf. Schmidt, p. 475).5 The fact that a means of protection and
4
The view that all of vv. 1-14 are later than P (Eerdmans, Levin, Knohl) can
hardly be correct, as it leaves no connection between the impasse in 9.12 and the
Israelites’ departure in 12.40 (cf. L. Schmidt, pp. 29-30). Even suggestions that
one of the strands is post-Priestly (Ahuis, L. Schmidt) run into difficulty.
5
A similar, at first surprising, second person pl. pronoun occurs in Zech. 1.2,
with the expected instruction to speak to the people only in v. 3. In v. 2 ‘your (pl.)’
can also easily mean ‘your (sing.) and their (sc. the people’s)’.
34 EXODUS 1–18
deliverance for the Israelites is going to be provided need by no

means make the association of this with the fixing of the calendar
inappropriate: the combination of the two is probably precisely the
writers’ point. Moreover, as both Kohata and Gertz have observed,
eliminating v. 2 from the core of the Priestly text has to lead to
the removal as well of all the references to the calendar later in
the passage and this leaves a very ‘un-Priestly’ series of instruc-
tions, surprisingly so in a context which is concerned with a major
religious festival. Verse 2 is therefore very probably an original part
of the Priestly text: on its significance see further my essay ‘The
Passover as the New Year Festival in P’.
The original Priestly instructions to the Israelites were, then,
limited to the Passover ritual itself (vv. 1-14a) and it is initially a
ritual carried out in anxious haste (v. 11) and expectation of what
is to come: its purpose is grounded in the means of escape which it
provides from Yahweh’s imminent intervention in judgement against
the Egyptians because of their ruler’s immoveable recalcitrance in
the face of Yahweh’s demands (vv. 12-13: cf. 11.9-10). But in an
almost unique enlargement of the whole conflict between Yahweh
and Egypt, which is recalled elsewhere only in the summary of the
Israelites’ journey to Canaan in Num. 33.4, Yahweh’s judgement
falls also on the Egyptian gods. Whatever form this is imagined to
have taken (see the Explanatory Note on vv. 12-13), it recognises
that in the ancient Near East the power of kings was undergirded by
their patron gods. Only with the defeat or displacement of the latter
could a nation’s overthrow be complete. If the Priestly narrative
was indeed composed in Babylon and was designed as a retelling
of Israel’s origin traditions for those in exile (see especially 2.23-25
and the commentary on that section), this detail will have served a
similar purpose to Deutero-Isaiah’s more explicit polemic against
the Babylonian gods (Isa. 41.21-29 etc.).
The conclusion to the instructions about Passover in v. 14(a)
makes it clear that this ritual was meant to be observed as a
‘memorial’ or reminder by the Priestly writers’ contemporaries and
here, as rarely elsewhere in the Old Testament (34.25; Ezek. 45.21),
it is given the honorific title of a ‘festival’ (Heb. ḥag). As the only
annual festival apart from the Day of Atonement to be prescribed
in the original Priestly work, it gives particular prominence in
worship to its theme of Yahweh’s past deliverance of his people,
which is no doubt seen here as it is elsewhere (e.g. Isa. 43.16-17)
12.1-20 35
as an assurance that he can do the same again. The Exodus tradi-

tion is thus as important a focus of the people’s worship as the need
for atonement for uncleanness and sin, and the Priestly writers do
not hesitate to make it the basis for the liturgical calendar which
governs the timing of all the people’s worship. For the exiles this is
also a festival which they can celebrate even before their return to
the homeland and the rebuilding of the destroyed temple, for it is to
be held in their ‘houses’.
The very detail of the ritual prescriptions implies that they are
not simply part of a narrative of the past but a pattern for the present
and the future. The care taken to correct Deuteronomic innova-
tions in vv. 5 and 9 is further evidence of this. In their turn later
writers restored the Deuteronomic (and perhaps older) combination
of Passover with the seven-day festival of Unleavened Bread as a
commemoration of the Exodus (vv. 18-20; vv. 14b-17). Both parts
of the celebration now received added weight from their description
as ‘everlasting statutes’ (vv. 14b, 17): such permanence is a trade-
mark of the Priestly mind-set, whether in regard to divine covenants
(Gen. 9.16; 17.7, 17), the possession of the land (Gen. 17.8) or legal
enactments as here. Yahweh’s relationship with his people is secure
and essentially unchanging: only individuals who fail to abide by
the rules are excluded (vv. 15, 19; cf. Gen. 17.14). The same sense
of security, rooted in the memory of Yahweh’s ‘passing by’ his
people at the time of the Exodus, was also cherished in the older
account of the Passover (vv. 23, 27). The prophet Amos must have
shocked his hearers with his ‘never again’ (Amos 7.8; 8.2), which
the Old Testament preserves side-by-side with the continuing tradi-
tion of Passover (Schmidt, p. 526).
1 Yahweh said to Moses and Aaron in the land of Egypt, saying:

2 ‘This month is a beginning of months for you, it is the first of a
the months of the year for you. 3 Speak to the whole congrega-
tionb of Israel, saying: “On the tenthc of this month letd each
of them take an animal from the flocke for (their) familiesf, an
animal for each household. 4 If the household is too small to
require a (whole) animalg, then a manh and his neighbour who is
closest to his house shall take one by a reckoningi of the persons:
you shall reckon the division of the animal according to each
person’s share of the foodj. 5 You shall have an unblemished
male animal a year old: you may take it from the lambs or from
36 EXODUS 1–18
the goats. 6 You shall have it with you under observation until
the fourteenth dayk of this month, and then all the assemblyl of
the congregation of Israel shall slaughter it between the two
eveningsm. 7 They shall take some of n the blood and put it on the
two doorposts and the lintelo, on the houses in which they will
eat it (i.e. the animal). 8 They shall eat the meat on this night,
roasted by firep, with unleavened breadq, onr bitter herbs they
shall eat it. 9 You shall not eat any of its raw or boiled, that is
boiled in watert, but only roasted by fire with its head as well asr
its legs and its innards. 10 You shall not let any of its remain over
until morningu; what does remain over until morningu you shall
burn in the fire. 11 Thisv is how you shall eat it – wwith your waist
girded, your sandals on your feet and your stick in your handw
– and you shall eat it in (fearful?) hastex: ity is a passover for
Yahweh. 12 I shall pass through the land of Egypt on this night
and I shall strike down every firstborn in the land of Egypt, both
man and beastz, and upon all the gods of Egypt I shall perform
acts of judgementaa: I am Yahweh. 13 The blood shall be a sign
for you on the houses where you are, and when I seebb the blood I
will pass over youcc and among you there will be no plaguedd for
destructionee when I strike in the land of Egypt. 14 This day shall
be a memorialff for you, and you shall celebrate itgg as a festival
for Yahweh: [throughout your generations – as an everlasting
statutehh you shall celebrate itgg. 15 For seven daysii you shall
eat unleavened bread: be sure thatjj on the first day you removekk
leavenll from your houses, for whoever eats anything leavened,
that person shall be cut off mm from Israel, nnfrom the first day
until the seventh daynn. 16 On the first day (you shall have) a
holy occasionoo and on the seventh day you shall have a holy
occasion – on those days no work shall be donepp. Onlyjj what
is to be eaten byqq any one of you, that alone may be prepared
byqq you. 17 You shall keep the Unleavened Bread (Feast)rr, for
on this very dayss I brought out your tribal divisionstt from the
land of Egypt. So you shall observe this day throughout your
generations as an everlasting statute.] 18 [In the first month,
on the fourteenth day of the monthk in the evening, you shall
eat unleavened bread until the twenty-first daynn of the month in
the evening. 19 For seven days no leaven shall be found in your
houses, for whoever eats anything leaveneduu, that person shall
be cut off from the congregation of Israel, both aliens and natives
of the land. 20 Anything that is leaveneduu you shall not eat; in all
your dwelling-places you shall eat unleavened bread.” ’]
12.1-20 37

a. Heb. ‫ל‬, used as often (cf. BDB, p. 513, s.v. 5c) as an alternative way of
expressing the genitive relation: again, also in dates, in vv. 3 and 6.
b. Heb. ‫עדת‬. ‫( ֵע ָדה‬cf. vv. 6, 19, 47; 16.1. etc.) is the most common term
for the Israelite community in Priestly texts, with over 100 occurrences, and
appears here for the first time in Exodus. It is derived from the root ‫( יעד‬like
another important Priestly term, ‫[א ֶֹהל] מו ֵֹעד‬, on which see the notes on 13.10
and 27.21), apparently in the sense of those who gather or meet together, and
it is even used of a swarm of bees (Judg. 14.8). In non-Priestly texts it is used
of the council of the gods (Ps. 82.1, with a parallel usage in Ugaritic) and of
groups of men, both righteous and wicked (e.g. Pss. 1.5; 22.17: so also in P of
the supporters of Korah in Num. 16.5 etc.). This suggests that, unlike e.g. ‫עם‬
and ‫גוי‬, it lacks an ethnic or political aspect and is particularly appropriate to
a community defined by its religious allegiance. (See further in the Explana-
tory Note.)
c. Heb. ‫ה ָעשׂר‬,ֶ employing a different word from the usual words for ‘ten’ and
‘tenth’, one which is almost always (not in Num. 7.66) used for the tenth day,
whereas ‫ עשׂירי‬is used for the tenth month and year (so also in Heb. inscrip-
tions: see AHI 1, p. 467; 2, p. 208). ‫ ָעשׂר‬is also a set of ten things (Gen. 24.55;
Pss. 33.2; 92.4; 144.9). Its use in dates is mainly Priestly but not exclusively
so: cf. 2 Kgs 25.1; Jer. 52.4, 12. On the (common) omission of ‫ יום‬in dates see
GK §134n, p: the same happens in epigraphic Heb. (Gogel, pp. 25-51).
d. Heb. ‫ויקחו‬, but the waw simply indicates the apodosis after the date (JM
§176g).
e. Heb. ‫( שׂה‬and presumably also its Ugaritic cognate š: cf. DULAT, p. 794
for refs.) can mean either a sheep or a goat (cf. Deut. 14.4), an ambiguity
which is made use of here (cf. v. 5). There is no convenient English equiva-
lent: the archaeologists’ ‘geep’ (used mainly of bones) is unlikely to find wider
acceptance. The word is not restricted to young animals, so ‘lamb’ is doubly
wrong.
f. Heb. ‫ לבית אבת‬uses the normal pl. form in BH and so should mean ‘for
families’, not ‘for a family’. On the expression in general see Note j on the
translation of 6.10–7.5 and the Explanatory Note and Excursus on 6.14-15.
Here (cf. the equivalent ‫בית‬, ‘household’) it is used in its original more limited
sense of ‘a family’, rather than for a larger social unit. The use of the pl. is
logical after the pl. verb, but the absence of the pron. suffix is then surprising:
it is perhaps due to the influence of the nearer distributive expression.
g. Heb. ‫מהית משׂה‬, lit. ‘from/than being from/of a (whole) animal’. The
first ‫ מן‬is easily understood as an example of a special kind of comparative
‫מן‬: BDB, p. 582, s.v. 6d; GK §133c), but the second has no clear parallel,
and commentators have struggled to explain it (for attempts see Houtman,
pp. 170-71), even though it is clear enough what the expression must mean.
38 EXODUS 1–18
Propp’s observation that two different ways of saying it have been combined
(‫ מהית שׂה‬and ‫ משׂה‬alone: p. 388) is an attractive solution, whether we attribute
the ‘overkill’ to textual corruption or the clumsy style of the writer.
h. Heb. has ‫הוא‬, ‘he’ (or ‘it’ referring to the ‘household’ [cf. NRSV]; but
this is difficult to square with ‫ אל־ביתו‬afterwards), presumably meaning the
head of the household: the pronoun may refer back to ‫ אישׁ‬in v. 3.
i. Heb. ‫ מכסה‬occurs only here and in Lev. 27.23, where it refers to the
‘calculation’ of the value of land consecrated during the period of a jubilee.
The masc. form ‫ מכס‬is more common, but is used specifically of a tax (Num.
31.28ff.; for an epigraphic occurrence see AHI, 101.280). The only BH occur-
rence of the related verb comes later in this verse (‫תכסו‬, from ‫)כסס‬: cognates
in other languages generally have the sense ‘chew, break up’, but on MH see
Jastrow, p. 655 (CAL, consulted 2 December 2011, does not give any Aram.
exx. for ‘reckon’, but does cite nouns referring to taxation).
j. Heb. ‫אישׁ לפי אכלו‬, which is to be taken as the obj. of ‫תכסו‬. Some para-
phrase in translation is inevitable here.
k. Heb. ‫ארבעה עשׂר יום‬, with neither the numeral nor ‫ יום‬having the def.
art. according to widespread practice (GK §134o): so also below in v. 18a.
According to GK §134p the inclusion of ‫ יום‬is a late characteristic: pre-exilic
inscriptions omit it in dates (cf. AHI 1, pp. 297-98). On the other hand they
regularly have the article with the numeral (cf. AHI 2.007.4; 2.008.3).
l. Heb. ‫קהל‬. This is a word with even stronger connotations of ‘gather-
ing together’ than ‫( עדה‬cf. Note b), though without the same grounding in
an ‘appointment’ (cf. the related verbs). In Ezekiel (e.g. 17.17) and some
probably older texts it is used of an army. As a term for Israel as a whole
(apparently: ‫ )קהל יהוה‬it occurs in Mic. 2.5 in a legal context of land distribu-
tion. In relation to Israel’s early history it is used in Deuteronomy (e.g. 5.22;
9.10) and it is also found in P, though less often than ‫( עדה‬3x in Genesis, 2x in
Exodus, 5x in Leviticus, 11x in Numbers). ‘The assembly of the congregation’
as a phrase is found elsewhere only in Num. 14.5 (for looser collocations see
Lev. 4.13 and Num. 16.3), and it presumably lays even greater stress on the
united action of the whole community (cf. also the ‫ כל‬which precedes in both
verses). (See further the Explanatory Note and Houtman, pp. 168-69.)
m. Heb. ‫בין הערבים‬. Grammatically ‫ ערבים‬seems to be a dual. This is
disputed by GK §88c and BL §63b′ on the basis that not all words with the
‘dual’ ending are duals (cf. ‫ )צהרים‬and LXX (apart from Lev. 23.5) knows
nothing of any ‘twoness’ (cf. TWAT 6, 362 = TDOT 11, p. 337). But ‫ בין‬surely
implies two things. The expression is found elsewhere of Passover in Lev.
23.5; Num. 9.3, 5, 11; of the evening sacrifice in Exod. 29.39, 41; Num. 28.4,
8; of the evening incense offering in Exod. 30.8; and of the coming of the
quail in Exod. 16.12. Several times the expression ‫ בערב‬is used of the same
actions (e.g. Exod. 16.13), with less precision (this could explain the tendency
of LXX to ignore the dual ending). ‫ בין הערבים‬is in all cases except the last
used of a ritual action and must have been designed to specify exactly when it
12.1-20 39
was to be done (cf. ‫ כבוא השׁמשׁ‬in Deut. 16.6). Josephus places the slaughter of
the Passover sacrifice between the ninth and the eleventh hour (BJ 6.423), the
Mishna brings it back by implication to between 8½ and 9½ (Pes. 5.1; cf. 5.3,
which seems to allow any time after noon) and Rashi (on this verse) argues
that any time in the afternoon is permitted, which was evidently the traditional
view for the Passover references in his time. This cannot be what the expres-
sion originally meant: it was a pragmatic piece of exegesis to provide a period
long enough for the Passover rituals to be completed for the whole people, as
was very clearly seen by Ibn Ezra in his longer commentary, following Saadya
Gaon (Rottzoll 1, p. 392: so also Milgrom, Leviticus 23–27, pp. 1969-70). The
‘two evenings’ are most likely to be sunset and the time shortly afterwards
when darkness falls, so that ‘twilight’ is meant (this is still the Samaritan
understanding according to de Vaux, Institutions 1, p. 278 = ET p. 182, and
TWAT 6, 362 = TDOT 11, p. 337; and cf. Jub. 49.10-12). The ending of one
day and the beginning of the next in the evening is not a problem if it can be
assumed (as it probably can) that the text predates this late development (first
clearly attested in Neh. 13.19: see further de Vaux, Institutions 1, pp. 275-77
= ET pp. 180-82).6 Some possible support for the ‘twilight’ interpretation may
also be found in Arabic: GK §88c, contrary to its main argument, notes the Ar.
ʾel-ʿišâʾân, ‘the two evenings’, which Lane defines as ‘the time of sunset and
the darkness after nightfall (or: the first third of the night)’ (1/5, p. 2056a); and
W. Johnstone has added the bolder suggestion that the idiom of taǵlib would
allow the phrase to be understood as ‘between the evening and the morning’
(‘The Legacy of William Robertson Smith: Reading the Hebrew Bible with
Arabic-Sensitized Eyes’, in Johnstone [ed.], William Robertson Smith: Essays
in Reassessment [JSOTSup 189; Sheffield 1995], pp. 390-97 [esp. 392-94]).
n. Partitive ‫( מן‬BDB, p. 580).
o. Heb. ‫ המשׁקוף‬occurs only here and in vv. 22-23 below in BH, but the
contexts leave little doubt about the meaning (‘threshold’ is the only plausible
alternative and there are other words for that); MH ‫ משׁקוף‬and ‫ שׁקוף‬confirm this
(cf. Jastrow, pp. 858, 1621), as now does 11Q19 (the Temple Scroll) 41.14-15
(for other occurrences in 11Q19 see DCH 5, p. 566). It is surprising that BDB,
p. 1054, associates the word with Aram. ‫שׁקף‬, ‘strike’, rather than with BH
‫ שׁקף‬Niphal and Hiphil, ‘look down, overhang’ (cf. also Ar. saqfun, saqafa,
‘roof’), and Ges18, p. 761, cf. 1409, persists with this view, apparently on the
basis that it fits ‫ ֶשׁ ֶקף‬and ‫ ְשׁ ֻק ִפים‬better. But these words have a different meaning
and may not be connected with ‫משׁקוף‬. In HAL, pp. 616, 1518-19, and DCH,
loc. cit., only the possibility of a link to BH ‫ שׁקף‬is mentioned.
6
It appears from what de Vaux cites that only the latest layer of the cultic
legislation (Exod. 12.18; Lev. 23.32) presupposes the change from the (Egyptian)
morning demarcation to the (Mesopotamian) evening one. In fact (see the Explan-
atory Note) Exod. 12.18 can be understood in terms of the older system.
40 EXODUS 1–18
p. Heb. ‫צלי אשׁ‬. In Isa. 44.16 ‫ ָצ ִלי‬is a noun, but here and in v. 9 the use in
the constr. st. before a nomen rectum denoting the cause or the means corre-
sponds to a construction used with adjectives (GK §128x). An adjectival form
also fits its modifying role after ‫( הבשׂר‬the ‘adverbial accusative’ acc. to GK
§118n): cf. Deut. 15.18.
q. Heb. ‫ומצות‬. The etymology of ‫ מצה‬remains unclear: Propp compares
Ar. naḍā = ‘be thin, poor’ (p. 394). ‫ מצות‬here is apparently a (further) object
of ‫ יאכלו‬or perhaps, in view of the Masoretic accents, of ‫( יאכלהו‬but if so it
is curious that it precedes the main object): for waw virtually in the sense of
‘with’ cf. BDB, p. 253 (s.v. 1g; more briefly GK §154a n. 1, JM §150p, 151a),
though most exx. are of an additional subject.
r. Heb. ‫על‬, which can also mean ‘in addition to’ or ‘with’ (BDB, p. 755),
and is often so used in sacrificial laws (ibid.).
s. Heb. ‫ממנו‬. Cf. Note n: in a negative sentence the partitive sense of ‫ מן‬is
still present (‘[no] part of’) but requires a different translation.
t. Heb. ‫ומבשׁל במים‬. ‫ בשׁל‬alone can mean ‘boil(ed)’ (cf. 1 Sam. 2.13 and
probably Exod. 16.23, where it occurs alongside ‫)אפה‬, but it was also a general
word for ‘cook’ (2 Sam. 13.8) and the supplementary phrase (introduced
by explicative waw: BDB, p. 252) serves to remove ambiguity (perhaps
in relation to the use of ‫ בשׁל‬in Deut. 16.7: see the Explanatory Note): cf.
the similar amplification in Deut. 14.6 and Fishbane, Biblical Interpreta-
tion, pp. 58-60, though he later takes a different view of this particular case
(pp. 135-36: the fact that ‫ בשׁל‬alone was found insufficient here undermines his
view that it always meant ‘boil’).
u. Heb. ‫עד־בקר‬, without the def. art., whereas the prepositions ‫ ב‬and ‫ל‬
usually have it. The absence of the art. after ‫ עד‬is chiefly a feature of Priestly
style (though not in 23.18), and it is not universal even in P (e.g. 16.23-24;
29.34). For other (probably P) omissions of the art. with ‫ בקר‬see 16.7 and
Num. 16.5.
v. Heb. ‫ככה‬, here used cataphorically to refer to the prescriptions that are
to follow. Of the four ways of saying ‘thus, so’ in BH ‫( כן‬743x) and ‫( כה‬577x)
are by far the most frequent (vs. ‫ ככה‬37x; ‫כזה‬/‫ כזאת‬40x), but they have a wider
range of meaning (including also ‘here’ and ‘now’, esp. with prepositions).
‫ככה‬, although found in some probably early texts (Num. 11.15; 1 Sam. 19.17;
2 Sam. 13.4; 17.21; 1 Kgs 1.6, 48), became popular in LBH (cf. 2 Chr. 18.19
[vs. ‫ בכה‬in 1 Kgs 22.20]; Song 5.3 [2x]; Esth. 8.6 [2x]; 9.26)7, in legal texts
(Exod. 29.35; Num. 8.26; 15.11-13; Deut. 25.9; 29.23) and in the Jeremiah
prose (4x). It is used predominantly in discourse rather than in narrative and
may have been preferred not only for a somewhat emphatic force (BDB,
p. 462), but because it was less ambiguous than ‫ כן‬and ‫כה‬. In MH it was
shortened to ‫( כך‬so already in Sirach 3x), while ‫ כה‬was little used.
7
For occurrences at Qumran etc. see DCH 4, p. 393. In ancient Hebrew
inscriptions only ‫ כן‬and ‫( כזאת‬once each) are attested so far.
12.1-20 41
w. The three short phrases, coordinated like a series of nouns by waw

between the last two (cf. JM §177c for this possibility), could perhaps be
supplied with a verb in the future (‘shall be’) from the context, but they are
best grouped with a wide range of attached descriptive phrases such as ‫אפים‬
‫ארצה‬, which can be regarded as short circumstantial clauses (GK §156c; JM
§159a-b). On the realia see IDB 1, pp. 869-71; 4, pp. 213-14; BRL, pp. 185-
88; ABD 2, pp. 232-35. All three actions are associated with setting out on
a journey: cf. 2 Kgs 4.29; 9.1; Deut. 29.4; Josh. 9.5, 13; Gen. 32.11; Num.
22.27. ‫ מקל‬is a rarer word than ‫ מטה‬and ‫ שׁבט‬and can refer also to sticks or
branches, so ‘stick’ is perhaps better than ‘staff’ for it.
x. Heb. ‫ חפזון‬occurs only here, in Deut. 16.3 (also in Passover legislation
but of the departure itself) and in Isa. 52.12 (negatively, with a likely allusion
to the Exodus, and again of departure). The par. with ‘flight’ (‫ )מנוסה‬in Isa.
52.12 matches the similar association of the verb ‫ חפז‬in Ps. 104.7, but in
2 Sam. 4.4 and 2 Kgs 7.15 it is used of the ‘haste’ with which one flees (cf.
Ar. ḥafaza) and in Pss. 31.23; 48.6; 116.11 it seems to refer to the ‘panic’ that
the need to flee brings on (cf. the context in Deut. 20.3 and Job 40.23: TWAT
3, 98-100 = TDOT 5, pp. 90-91; Ges18, p. 379). It is possible that the mood
of anxiety is implied by the noun too (so e.g. Childs, p. 183).
y. Heb. ‫ הוא‬must, like the preceding pronouns, refer to the animal (or its
meat), so that ‫ ֶפ ַּסח‬will here mean the slaughtered animal (as it clearly does
in v. 21: cf. TWAT 6, 668 = TDOT 12, p. 9) and not the festival as a whole
as it sometimes does. In many passages either sense is possible. The original
meaning and etymology of ‫ ֶפ ַּסח‬have been extensively debated in connection
with the origin of the Passover itself: for a summary see TWAT 6, 660-69 =
TDOT 12, pp. 1-9 (with extensive bibliography). The text of Exodus associ-
ates the expression with the verb ‫( ָפּ ַסח‬12.13, 23, 27), which is normally taken
to mean ‘spare, protect’ (cf. Isa. 31.5). There are three occurrences of the verb
which appear to be related in meaning to the adjective ‫ ִפּ ֵס ַח‬, ‘lame’, and so
to mean ‘limp, be lame’ (2 Sam. 4.4; 1 Kgs 18.21, 26), but this is probably
a separate root (BDB, DCH). Two recent suggestions are that all the occur-
rences can be derived from a sense ‘hop, leap, jump’ (O. Keel, ‘Erwägungen
zum Sitz im Leben des vormosäisches Pascha und zur Etymologie von ‫’פסח‬,
ZAW 84 [1972], pp. 414-34 [428-33]; Houtman, p. 183; HAL; Ges18) or from
a sense ‘push, strike’ (E. Otto, ‘Zur Semantik von hebr. psḥ/pisseaḥ und akk.
pessû[m]/pessātu[m]’, BN 41 [1988], pp. 31-35; TWAT 6, 665-68 = TDOT 12,
pp. 5-8), but neither of these is very probable. Conjectures about the original
nature of ‫ ֶפ ַּסח‬have been based on all these proposed etymologies: for further
discussion see the introduction to this section and the Explanatory Notes.
z. Lit. ‘from man and until beast’, with the idiomatic ‫ ועד‬which MT also
has in 13.15 but not in the similar phrases in 11.5 and 12.29.
aa. Heb. ‫שׁפטים‬, as in 6.6 and 7.4: see Note m on the translation of 6.1-9.
bb. Lit. ‘I will see…and (I will pass over)’, but as often the coordination
implies a consequential relationship between the clauses. GK §124g includes
42 EXODUS 1–18
this among its examples of conditional clauses, but since no uncertainty is

implied a temporal understanding (according to GK §164b[4]; JM §166b) is
preferable.
cc. Heb. ‫ופסחתי עלכם‬. The uncommon verb ‫ פסח‬has undoubtedly been
chosen here to provide an explanation of the noun ‫ פסח‬in v. 11 (see also Note
y), as it is also in vv. 23 and 27: the connection is particularly clear in v. 27.
There is a close parallel to the meaning in Isa. 31.5, which is about the protec-
tion of Jerusalem from enemy attack, and the association with ‫ גנן‬there (as well
as ‫ הציל‬and ‫ )המליט‬suggests the sense ‘protect, spare’. Here, to bring out the
etymological connection that is being made, the translation ‘pass over’ is to
be preferred (and perhaps also in Isa. 31.5 if [as e.g. O. Kaiser, Der Prophet
Jesaja, Kap. 13–39 (ATD; Göttingen, 1973), p. 252, ET, p. 317, thinks] a
deliberate allusion to the Exodus story is being made there).
dd. Heb. ‫נֶ גֶ ף‬, an alternative to the more common word ‫( ַמגְּ ָפה‬26x BH), is
found almost exclusively in Priestly texts in the Pentateuch (Exod. 30.12;
Num. 8.19; 17.11-12; elsewhere only in Josh. 22.17 [related to P] and Isa.
8.14 [in a different sense]); compare P’s use of ‫ ֶשׁ ֶפט‬as an alternative to ‫ִמ ְשׁ ָפט‬
(above, Note aa).
ee. Heb. ‫למשׁחית‬. ‫ משׁחית‬also appears in the non-Priestly Passover prescrip-
tions (in v. 23, where with the definite article [and the verb ‫ נגף‬following]
it seems to refer to a personal ‘Destroyer’ [see further the notes on v. 23]).
Here (pace Noth, pp. 66, 71, 75, ET, pp. 91, 96, and Houtman) Yahweh acts
alone and ‫ משׁחית‬is used as a nomen actionis, ‘destruction’ (so Jer. 5.26; 51.25;
Prov. 18.9; 28.24; Ezek. 9.6; 21.36; 25.15; Dan. 10.8; 2 Chr. 20.23; 22.4).
The combination with ‫ל‬, perhaps significantly, is attested elsewhere only in
Ezekiel, Daniel and 2 Chronicles.
ff. Heb. ‫לזכרון‬. ‫ זכרון‬probably originally meant ‘remembering’ (cf. BL
§61cθ), although a purely mental use is only attested in late texts (Eccl.
1.11; 2.16) and generally ‘memorials’ or aids to remembering in the form of
observances, objects or records are meant (e.g. 13.9; 17.14; 28.12, 29; 39.7).
The word is frequent in P but also elsewhere. The occurrences are part of the
important larger theme of remembering, whether by God or by man, in the
OT. A ‫ זכרון‬is often also described as an ‫אות‬, ‘sign’ (cf. 13.9; TWAT 2, 586-87 =
TDOT 4, pp. 77-78). But here there is a distinction: the ‫( אות‬v. 13) is the blood
on the doors and it shows Yahweh which houses are to be spared, whereas
the ‫ זכרון‬is ‘this day’, the observances as a whole which are to remind Israel
of their great deliverance. This points to a further difference, that here the ‫אות‬
relates to the one past event whereas the ‫ זכרון‬is the observance that will be
continued, ‘throughout your generations’.
gg. Heb. ‫וחגתם אתו‬. ‫ חגג‬is basically intransitive (cf. 5.1; 23.14), meaning
‘celebrate a pilgrim-feast’ (cf. Ar. ḥajja, ‘make a pilgrimage’), and so it can
properly take only an internal or cognate accusative (‫ חג‬in Lev. 23.39; Num.
29.12; Nah. 2.1; Zech. 14.16, 18, 19), according to GK §117p-r. Only here
(twice in the verse) and in Lev. 23.41 (again twice, about the Feast of Booths)
12.1-20 43
is an ‘it’ the object. In Lev. 23.41 the ‘it’ is probably ‘the feast of Yahweh’ in
v. 39 (though it might be ‘the fifteenth day’ earlier in the same verse), and so
there is no departure from normal usage. But here the ‘it’ can only be ‘this
day’ at the beginning of the verse (‫ חג‬then follows as an adverbial accusative,
‘as a pilgrimage feast’: GK §118m, q), and so ‫ חגג‬becomes genuinely transitive
(for some different cases of such variation in BH see JM §111h).
hh. Heb. ‫חקת עולם‬, a very frequent Priestly expression (cf. v. 17; 27.21;
28.43; 29.9 etc.) which occurs elsewhere only in Ezek. 46.14 (‫ חק־עולם‬is
likewise always Priestly except for Jer. 5.22). The regular translation ‘an
everlasting statute’ is supported by the reference to future generations here.
ii. Heb. ‫ שׁבעת ימים‬is an ‘accusative’ of duration (GK §118k).
jj. Heb. ‫אך‬. Muraoka (Emphatic Words, p. 130) includes the occurrences
in vv. 15-16 in a group of mainly cultic texts in which the adversative and
emphatic uses of ‫ אך‬are combined: ‘But, mark!’ But in v. 15 there is no
contrast with the preceding clause: ‫ אך‬introduces the prior action which
enables its requirement to be fulfilled and so must (as Childs, p. 183, saw) be
purely emphatic: against Houtman too, but like 31.13 where Muraoka recog-
nises this use. 31.13 also indicates that emphatic ‫ אך‬can occasionally appear
with commands as well as statements. In v. 16, on the other hand, the adversa-
tive, or better restrictive, use is to the fore: whether ‫ אך‬conveys emphasis there
is at least doubtful.
kk. Heb. ‫תשׁביתו‬, lit. ‘cause to cease’: for this rare use (as an alternative to
‫ )הסיר‬cf. Ezek. 23.27, 48 (with ‫ מן‬as here: Isa. 30.11 and Ps. 89.45 are less
certain instances).
ll. Heb. ‫ ְשּׂאֹר‬is the ‘leaven’ itself, while ‫ חמץ‬and its derivatives (as later in
this verse and in vv. 19, 20, 34 and 39) are the dough or bread with which
it has been mixed (Lev. 2.11 is not an exception, despite IDB 3, p. 105).
According to TWAT 2, 1063 = TDOT 4, p. 489, ‫ שׂאר‬was a mixture of yeast and
lactic acid (sour milk). Old dough was often used as a source of leaven (IDB,
loc. cit.). For the dagesh forte see GK §20g (apparently regarding this and the
similar case in v. 31 as conjunctivum): according to the apparatus in BHS it is
present in L (so also Dothan, but not BH3) but it is omitted in many mss and
edd. (including the early 16th-cent. Rabbinic Bibles).
mm. Heb. ‫ונכרתה‬, with waw of apodosis following the casus pendens ‫כל‬
‫אכל חמץ‬, which is resumed not by a retrospective pronoun but by a fresh
nominal expression, ‫( הנפשׁ ההוא‬JM §156e-f). One view is that ‫ כרת‬in such
expressions means the death penalty (so BDB, p. 504; THAT 1, 858 = TLOT
2, p. 636; Ges18, pp. 574-75). But others see exclusion from the community
as the sense in all or most of the legal texts (HAL, p. 476; TWAT 4, 362-63
= TDOT 7, pp. 347-48: the death penalty is mentioned only in Exod. 31.14;
Lev. 20.2). According to the traditional rabbinic view and some moderns it
is a comprehensive term for divine punishment (Milgrom, Leviticus 1–16,
pp. 457-60; Propp, p. 404). For discussion see W. Zimmerli, ‘Die Eigenart
der prophetischen Rede des Ezechiel. Ein Beitrag zum Problem an Hand
44 EXODUS 1–18
von Ez. 14 1-11’, ZAW 66 (1954), pp. 1-26 (6-19); cf. his Ezechiel, 1 (BKAT;
Neukirchen, 1969), pp. 302-306, ET pp. 302-305; W. Horbury, ‘Extirpation
and Excommunication’, VT 35 (1985), pp. 13-38 (esp. 31-34).
nn. Heb. ‫מיום הראשׁן עד־יום השׁבעי‬. Here, though not earlier in the verse or
in the next verse, ‫ יום‬twice curiously lacks the def. art. when it is present with
the following numeral adj. (so also in v. 18b). This implies that it is in the
constr. st., with the numeral as a specifier: cf. the expression ‫ יום השׁבת‬in 20.8
and more generally GK §128k, m.
oo. Heb. ‫מקרא־קדשׁ‬. ‫ מקרא‬is to be associated with ‫ קרא‬I, ‘call’ (so BDB,
etc.), because of several cases where noun and verb occur together (Lev. 23.2,
4, 21, 37; Isa. 1.13; Neh. 8.8): in addition the occurrence in Num. 10.2 clearly
has the verbal sense ‘(for) summoning (of the congregation)’ (also at Qumran:
see DCH 5, p. 471). There is a question about the sense(s) of ‫ קרא‬I to which
it is related: in Neh. 8.8. it is exceptionally the sense ‘read’ (BDB s.v. 4), but
elsewhere it might be ‘proclaim’ or ‘invite’ (BDB s.v. 3 and 5). In one case it
refers to a place or places (Isa. 4.5), but elsewhere, apart from Num. 10.2 and
Neh. 8.8, an occasion or time is meant. The occurrence in Isa. 1.13 attests its
occasional use for a cultic celebration in pre-exilic times (possibly also in Ps.
68.27), but ‫ עצרה‬seems to be more common then. In most of its occurrences
it is combined (as here) with ‫קדשׁ‬, but the other examples are concentrated
in two related passages, Leviticus 23 and Numbers 28–29, which must be
determinative for its meaning here. There it is used for special occasions in
the Priestly liturgical calendar on which no work was to be done. Leviticus
23 additionally associates it with the term ‫מועד‬, ‘festival’ (vv. 2, 4, 37), and
in its present form (which is probably secondary) applies it to the Sabbath
(v. 3), presumably because it too was a day (the paradigmatic day indeed) on
which no work was to be done. The application to the first and last days of the
Feast of Unleavened Bread is found there too (Lev. 23.7-8; Num. 28.18, 25).
It seems most likely (Milgrom, Leviticus 23–27, pp. 1952-59; cf. Ges18, p.
732, and already E. Kutsch, ‘‫’מ ְק ָרא‬,
ִ ZAW 65 [1953], pp. 247-53) that it means
a ‘holy occasion’, which is proclaimed, rather than a ‘holy assembly’ (BDB,
p. 896; HAL, p. 594; DCH 5, pp. 470-71), to which the people are called or
‘invited’: the object of ‫ קרא‬is regularly the event and not the people, and the
point may well be that, being dependent on the lunar calendar, these occasions
had to be officially announced (Milgrom, pp. 1958-59). The evidence of the
Vss (see Text and Versions) broadly supports this view.
pp. The third person m.s. verb unusually agrees in gender with the
preceding ‫ כל‬rather than with ‫מלאכה‬, apparently because of emphasis on ‫כל‬,
‘(not) any at all’ (GK §146c).
qq. ‫ ל‬in ‫( לכל־נפשׁ‬lit. ‘by every/any person’) and ‫ לכם‬in v. 16b represents
the agent, as most frequently in the formula ‫ברוך ל‬, but also occasionally
elsewhere in both older and later texts (BDB, p. 514). This usage seems not
to be found again in P.
12.1-20 45
rr. Heb. ‫את־המצות‬. In 23.15 and 34.18 the feast is called ‫חג־המצות‬, as it is
also in Lev. 23.6 (cf. Num. 28.17), where there is a similar emphasis to here
on the first day. Although unique the ellipsis of ‫ חג‬is understandable enough
towards the end of a passage which has been focussed on the festival. On the
(secondary) variant text of SP and LXX see Text and Versions.
ss. Heb. ‫בעצם היום הזה‬, lit. ‘on the bone/body of this day’, occurs (apart
from two verses in Joshua [5.11; 10.27]) only in P (12x: e.g. Gen. 17.23, 26)
and Ezekiel (4x).
tt. Heb. ‫את־צבאותיכם‬. For the sense see Note t on the translation of 6.10–7.5.
uu. Here the Heb. is not ‫ חמץ‬as in v. 15 but ‫מחמצת‬, as also in v. 20 but
nowhere else in BH (or anywhere else apparently). The word looks like a fem.
Hiphil part. (cf. GK §53o): but in MH ‫ חמץ‬Hiphil can be intransitive, ‘become
sour, ferment’, so its meaning need not be different from ‫חמץ‬. It seems to be a
short-lived neologism of the later legal strata: compare Notes aa and dd on ‫שׁפט‬
and ‫נגף‬. Did it perhaps have a more general application (‘anything leavened’?),
like that seen by later rabbis (see Text and Versions), while ‫ חמץ‬simply meant
‘normal bread’?
Explanatory Notes
1. The inclusion of Aaron with Moses as a recipient of Yahweh’s
instructions about Passover and the feast of Unleavened Bread
follows a pattern already used in some Priestly passages earlier
in Exodus (cf. 6.13; 7.8; 9.8). Later there are further examples
(Num. 14.26; 16.20; 20.12, 23), especially as here to introduce
legal material (12.43; Lev. 11.1; 13.1; 14.33; 15.1; Num. 2.1; 4.1,
17; 19.1). The explicit location of these regulations ‘in the land of
Egypt’ (cf. 6.28) is also in line with a frequent Priestly practice to
mention a place in introductory formulae, even where it is obvious
from the narrative context (Lev. 25.1 [cf. 26.46; 27.34], ‘on Mount
Sinai’; Num. 1.1, 19; 9.1, ‘in the wilderness of Sinai’; Num. 33.50;
35.1 [cf. 36.13], ‘in the plains of Moab’). This is not necessarily to
distinguish one location from another (though in the present case the
regulations do have a specific relation to their location): in relation
to later readers all of these locations served, like the references to
Moses (and Aaron), to anchor the origin of the laws in the era of the
formation of Israel.
2. Here for the first time in the biblical narrative an indication
is given of when a new year began, although Priestly passages
in the Flood story (e.g. Gen. 7.11) had already used numbers to
46 EXODUS 1–18
identify the different months of the year. There is no verb in the

original Hebrew, and Cassuto argued from this that an already
existing calendrical system was being referred to, not a new one
(p. 137; so also Childs, p. 206). Even if this were so, the choice
of the month of Passover with its associations with the Exodus
and the solemn declaration at this particular point in the narra-
tive would say something very significant about P’s conception
of the basis of the calendar. Most modern scholars, however, do
see in these words the grounding of a new calendar, with the year
beginning in the spring, in Israel’s own salvation history, even
though it had probably been adopted from the Babylonian impe-
rial power in the late monarchy period (so recently Houtman,
Propp).8 A detail of the wording of the Hebrew, reflected in our
translation, confirms this: ‘a beginning of months’ is not how one
would speak of an established custom (see Text and Versions for
the Samaritan Pentateuch’s ‘the beginning…’, which reflects later
acceptance of this system). Perhaps the repeated ‘for you’ (which
must have a wider application to all the Israelites) points in the
same direction. No name is given to the month here or anywhere
in the chapter: the old, probably Canaanite, Abib (cf. 13.4; 23.15;
34.18; Deut. 16.1) has been discarded and the Babylonian name
Nisan is perhaps deliberately not used to avoid an anachronism in
the supposedly ancient festival laws (cf. Lev. 23; Num. 28–29).
In the Targums ‘Nisan’ was sometimes added to match what had
become accepted terminology (see Text and Versions on vv. 2, 3,
6 and 18): here the Fragmentary Targum (ms. P) has a long addi-
tion in which the claims of all the other months to be the first are
considered and rejected, and similar poems occur in some mss
from the Cairo Genizah.
3-5. The delay in the command for Moses and Aaron to deliver
Yahweh’s instructions to the Israelites serves to give additional
prominence to the opening statement about the calendar. Unlike
earlier passages which use ‘the children of Israel’ (e.g. 1.7: Heb.
benê yiśrāʾēl), ‘the Hebrews’ (e.g. 2.6) or simply ‘Israel’ or ‘the
people’ (e.g. 5.2; 1.20), v. 3 introduces for the first time the expres-
sion ‘the (whole) congregation [Heb. ʿēdāh] of Israel’, which
will occur again with minor variations in vv. 6 and 19 and then
8
For fuller discussion see my ‘The Passover as the New Year Festival’.
12.1-20 47
frequently in the Priestly sections throughout the Tetrateuch (so

in v. 47; 16.1-2 etc.).9 In historical narrative it is rarely used after
the beginning of the monarchy and it is not found in clearly post-
exilic contexts. It appears to be an expression coined by the Priestly
writers to refer to the original Israelite community of the time of
Moses, which was then occasionally picked up in later sections of
the books of Joshua and Judges (especially Josh. 9 and 22; Judg.
20–21). In non-technical contexts Heb. ʿēdāh usually refers to
‘groups’ of people with a common character or cause (see Note b on
the translation for examples), but for P its etymological relation to
the verb yāʿad, ‘appoint’, may have been significant as an indicator
of Israel’s divine election, especially as another derivative of the
same root, môʿēd, was used in P for ‘appointed’ festivals (so also
earlier in 13.10; 23.15; 34.18; Hos. 9.5) and the designation of the
desert shrine as ‘the tent of meeting’ (27.21, again following older
usage as in 33.7, 11). Chapter 12 as a whole contains a number of
features of community life which fit what is known of the exilic
and post-exilic periods (cf. Blenkinsopp, Pentateuch, pp. 156-57,
referring to J. Weinberg, ‘Das bēit ʾaḇōt im 6-4 Jh. v.u.Z.’, VT 23
[1973], pp. 400-14 for the terminology), and this setting very likely
played a part in shaping what the Priestly writers imagined to be
the social structure of their ancestors in the earlier time of ‘exile’
in the Exodus period.10 It is striking that the term is almost entirely
absent from Ezekiel and Chronicles–Ezra–Nehemiah, despite their
affinities in other ways with P: Ezekiel prefers the term ‘house
(of Israel)’, while Chronicles-Ezra-Nehemiah accounts for over a
third (43) of the 123 occurrences of qāhāl, ‘assembly’, in biblical
Hebrew (on which see the Note below on v. 6). ʿēdāh is juxtaposed
directly with ‘Israel’ four times in the Priestly Passover legislation
(cf. vv. 6, 19, 47) and only seven times elsewhere (Lev. 4.13; Num.
9
Of possibly earlier occurrences referring to Israel (BDB, p. 417) only Ps.
74.2 (in an [early?] exilic lament) and perhaps 1 Kgs 12.20 (of the assembly
which appointed Jeroboam to rule the Northern Kingdom) are likely to be real
antecedents: in Hos. 7.12 and Jer. 6.18 the text is probably corrupt, 1 Kgs 8.5 is in
a passage which has been worked over in the light of P and/or Chr, and Jer. 30.20
belongs to one of the later layers of the book.
10
On the characteristics attributed to the ʿēdāh see more fully TWAT 5,
1079-89 = TDOT 10, pp. 469-77, although the authors of this article write as if it
was an actual institution from Israel’s early history.
48 EXODUS 1–18
16.9; 32.4; Josh. 22.18, 20; 1 Kgs 8.5 par. 2 Chr. 5.6): much more
often (27 times) P’s preferred designation of the people as ‘the
children of Israel’ follows it (e.g. 16.1): see also Text and Versions
on this verse.
The Masoretic accents seem to divide v. 3 in such a way that
the tenth of the month is the day of the command, not the day of
the ‘taking’, a view that was already considered, and rejected, in
MRI (Lauterbach 1, pp. 23-24) and by Rashi: for both the decisive
argument was that v. 2 located the command on the first day of the
month (taking the first occurrence of Heb. ḥōdeš as ‘new moon’ rather
than ‘month’). Modern commentators (e.g. Propp, p. 387) lay more
stress on ‘saying’ in v. 3 (Heb. lēʾmōr), which though grammatically
an infinitive, ‘to say’, almost always functions to introduce direct
speech, which begins immediately after it. For the grammatical link
between the date and what follows it see Note d on the translation.
Selection of the animal in good time would allow time to check its
suitability, but other explanations have been suggested (cf. Propp,
ibid.; Houtman, pp. 170-71), the most plausible of which notes the
importance of the tenth day of the month elsewhere (Lev. 16.29;
23.27; 25.9; Josh. 4.19; Ezek. 40.1) and sees it as enjoying special
significance in priestly circles.11
The use of the broad expression ‘an animal from the flock’ (Heb.
śeh: see Note e on the translation and TWAT 7, 718-21 = TDOT
14, pp. 46-49) prepares for the option in v. 5b to use a young goat
instead of a lamb. In Deut. 16.2 a different expression is used and the
option of an ox is also allowed. The family setting is underlined by
the preferred ruling for one animal per household: only in the case
of small households is sharing of an animal with a close neighbour
allowed (v. 4). The ‘reckoning’ would no doubt relate to the cost of
the animal as well as the distribution of its meat. The decision to
share is here left to individual discretion: later a minimum of ten
persons became the norm (Jos., BJ 6.9.3 [423]: see further Text and
Versions).12
11
Gesundheit even proposes that the tenth day was the original date for
Passover itself, reviving a suggestion of H. Ewald (Three Times, pp. 91-92).
12
The 1962 NJPS translation had ‘the neighbour closest to his household in the
number of persons’, and Childs was persuaded that the context made this prefer-
able to the translation given here (p. 182: cf. Ehrlich, Randglossen 1 [Leipzig,
1908], pp. 304-305). But the context is surely against it: it is ‘the neighbour’ who
12.1-20 49
‘Unblemished’ (v. 5: Heb. tāmîm) is a requirement for sacrificial

victims too (e.g. Lev. 1.3): the central sense is ‘complete’, but it is
often used of what is qualitatively ‘perfect’, including persons and
their behaviour (see further THAT 2, 1045-51 = TLOT 3, pp. 1424-
28; TWAT 8, 688-701 = TDOT 15, pp. 699-711). It is entirely
understandable that an animal for use in a religious ritual (cf. v. 11)
should need to be a perfect specimen, whether it was technically
regarded as a sacrifice (as in the parallel passage in v. 27: cf. Deut.
16.2, 5, 6) or not (as probably in vv. 1-14). The demand for a male
animal fits other ceremonial rituals in the Priestly system, such as
the ordination of priests (29.1), the Day of Atonement (Lev. 16.5)
and sin-offerings for a priest, a ruler or the whole people (Lev. 4.3,
14, 23), but for some other purposes a female was allowed (Lev. 3.1)
or even specified (Lev. 4.28, 32). The precise significance of ‘a year
old’ (Heb. ben šānāh, lit. ‘a son of a year’) has been much discussed
(already in MRI [Lauterbach 1, p. 29]; for fuller details see de Vaux,
Studies, pp. 4-5; Houtman, pp. 172-73; Propp, pp. 389-90). In the
case of the animal for Passover, given its timing in the spring when
young animals are born, the reference is probably simply to an
animal of the previous year’s birthing season.13
Whereas vv. 3-4a speak of the people in the third person plural
and of individuals within it in the third person singular (‘they’, ‘he’),
vv. 4b-5 address them directly in the second person plural (‘you’).
This variation continues in vv. 6-14 (in vv. 15-20 direct address
predominates) and it has been seen as a pointer to two different
stages in the composition of the passages (see further the introduc-
tion to this section).
6-7. The chosen animals are to be slaughtered by all the people
on the fourteenth day of the (first) month, a precise date corre-
sponding to the middle of the lunar month (full moon: cf. Ps. 81.4)
which is also specified in the liturgical calendars in Lev. 23.5;
is closest (cf. Prov. 27.10), not the household/house, and v. 4b indicates that ‘by a
reckoning of the persons’ goes with ‘shall take’. Nowhere else in BH, in addition,
does Heb. qārôb mean ‘numerically close’. The 1985 NJPS revision reverted to
the generally accepted interpretation that is followed here. Despite what Childs
says (and cf. Sarna, p. 244 n. 12; Propp, p. 389), it is not clear that the Masoretic
accents support one view rather than the other.
13
The more precise expression ben šenātô, lit. ‘son of its year’ (e.g. Lev. 12.6),
was perhaps needed to specify a ‘yearling’ in the case of rituals which could be
required at any time.
50 EXODUS 1–18
Num. 28.16; Ezek. 45.21 and in Num. 9.3-5, but not in the probably
earlier legislation in Exod. 12.21-27 and Deut. 16.1-8. Verse 6 is
also very specific about the timing of the slaughter ‘between the two
evenings’ (i.e. ‘at twilight’: see Note m on the translation), using an
expression that also regulates the timing of the evening sacrifice in
the tabernacle/temple in 29.39, 41; Num. 28.4, 8, instead of ‘sunset’
as found in Deut. 16.6. Thus here, although the domestic setting of
Passover is very clear, it is also correlated by the Priestly school
with activities in the tabernacle/temple. The specification of the
time of the ritual as in the evening is also unusual but is obviously
dictated, like its interpretation, by the narrative context here (v. 12:
cf. 11.4; 12.29-32; Deut. 16.1, 6, though the latter connects the
timing of Passover with a nocturnal departure from Egypt, not the
slaying of the Egyptian firstborn). The involvement of the whole
people together is emphasised by the use of the additional word
‘assembly’ (Heb. qāhāl) for them (see Note l on the translation).
This word is more widely used than ‘congregation’ (Heb. ʿēdāh),
occurring especially in Deuteronomy (and passages influenced by
it), Psalms and Chronicles as well as in P (cf. 16.3): for an attempt
to distinguish different nuances of meaning in these texts see TWAT
6, 1210-19 = TDOT 12, pp. 551-59. The contexts in which it is
used tend to be cultic or at least religious in a wider sense (so very
clearly in the Psalms), in contrast to ‘people’ (Heb. ʿām): cf. THAT
2, 615-18 = TLOT 3, pp. 1122-25.
The blood ritual, also mentioned in vv. 22-23, is focused on the
doorway as the route by which external danger could be expected
to approach. The omission of the threshold itself is surprising and
has led to the suggestion that the animal was slaughtered there, so
that its blood had already run on to the threshold: possibly v. 22
contains a reference to this in the word usually translated ‘basin’
(see the notes there and Houtman, pp. 192-93). The blood ritual
is only mentioned in the Exodus narrative in the Old Testament,
where its explanation is tied to the specific narrative context (cf.
vv. 13, 23): in the Mishnah (Pes. 5.6), which assumes that Passover
animals will be slaughtered at the temple, it was replaced by the
general sacrificial practice of pouring out the blood at the base of
the altar (cf. 24.6; 29.12). 2 Chronicles 30.16-17 seems already to
presuppose this practice. The blood ritual in Exodus 12 is clearly
apotropaic and reflects very ancient beliefs about threats to human
life and how they may be averted, like the power attributed to the
12.1-20 51
blood of circumcision in 4.25-26 (but not in the Priestly instruc-

tions about circumcision in Gen. 17.9-27 or in the narrative in
Josh. 5.2-9, where no mention is made of the blood shed). In ritual
texts, especially in P, similar blood rites are used in other contexts,
probably with different meanings, as in the old account of covenant-
making in Exod. 24.6, in the ordination of priests (29.20-21) and in
the cleansing of the tabernacle/temple (Lev. 16.14-15, 18-19; Ezek.
45.18-20).14 Since the nineteenth century and especially since the
much-cited study of L. Rost (‘Weidewechsel und altisraelitischer
Festkalender’, ZDPV 66 [1943], pp. 205-16) close parallels to the
blood-ritual (and other features of this passage) have been found
in pre-Islamic Bedouin practices, especially in springtime rituals
connected with the change of pasture by semi-nomads, with the
implication that the Exodus account gives a specific historical
origin to a practice that was originally much more widely current
(for a summary see R. Schmitt, Exodus und Passa [2nd ed., 1982],
pp. 25-29); this approach was strongly criticised by B. Wambacq,
‘Les origines de la Pesaḥ israélite’, Bib 57 [1976], pp. 206-24,
301-26, who is followed by Van Seters, Life, pp. 114-19. But
Albertz, Religionsgeschichte, p. 61, ET 1, p. 35 (see also p. 253
n. 47), continued to find it persuasive.
8-10. The mention of eating at the end of v. 7 leads into the
detailed prescriptions for the Passover meal itself. ‘On this night’
must mean the night of the day mentioned in v. 6, not the day
when the command was given: its reference is determined by the
co-text rather than the narrative con-text, to use the terminology
of discourse analysis. Verse 8 gives the basic requirements about
when the animal is to be eaten and how it is to be cooked and
served; vv. 9-10, in the reverse order, amplify these requirements
(with another change from the third person to the second) and add
a ruling about how any left-overs of the meat are to be disposed
of. In the present context the mention of unleavened bread (v. 8)
serves to reinforce the link between vv. 1-14 and vv. 15-20: there is
an overlap between the day of Passover and the festival of Unleav-
ened Bread. There are, however, other circumstances in which
unleavened bread is mentioned, with no ritual significance: in the
narratives about Lot’s angelic visitors (Gen. 19.1-3) and Saul’s
14
Ezekiel’s version of temple-cleansing specifically mentions ‘doorposts’
(mezûzōt, as here) and so is the closest parallel to the practice in Exod. 12.
52 EXODUS 1–18
visit to the medium of Endor (1 Sam. 28: cf. v. 24) such bread is
served to unexpected guests when there was presumably no time
to bake bread in the normal way. The narrative later in the chapter
(vv. 29-39), which dissociates the eating of unleavened bread from
the Passover ritual by placing it after the Israelites had departed
from Egypt, provides a similar explanation based on the need for
haste (v. 39). Unleavened bread (like ‘bitter herbs’) was (and is) a
standard part of Bedouin diet (cf. Schmitt, p. 29, with references),
so that it could have been a part of the ritual which went back to
very early times. The Mishnah defined acceptable ‘bitter herbs’ as
‘lettuce, chicory, pepperwort, snakeroot and dandelion’ (Pes. 2.6, in
Danby’s translation, p. 138; cf. the renderings of Pseudo-Jonathan,
the Vulgate and possibly the Septuagint in Text and Versions). In the
Old Testament they appear elsewhere only in the related legislation
in Num. 9.11, which adds no clarification, and in Lam. 3.15, where
(unless the verse is entirely metaphorical, as some think) plants that
would only be eaten (by city-dwellers at least) in dire circumstances
seem to be meant.15 In the Priestly context the mention of them
provides an apt reminder of the ‘bitter’ oppression (Heb. mrr) of the
Israelites (1.14: Schmidt, p. 504).
‘Raw’ meat could not be eaten, because of the blood in it (cf.
Gen. 9.4; Deut. 12.23). Since this was so well known, it could be
that here ‘raw’ means ‘rare, half-cooked’, as it does in later Heb.
(Schmidt, p. 476: cf. HAL, pp. 620-21; see also Text and Versions).
The requirement that the animal be roasted over a fire, which is
accentuated in v. 9, may be another ancient feature of the ritual.
If so, the corresponding ruling (for later times) in Deut. 16.7 that
allows it simply to be ‘cooked’ or possibly ‘boiled’ (Heb. biššēl,
as here in v. 9) involves an assimilation, which would be natural
enough in a temple context, to normal sacrificial procedure (see
Note t on the translation). In v. 9 the Priestly law seems explic-
itly to repudiate such ‘modernism’.16 An important consequence of
15
P.D. King, Surrounded by Bitterness (Eugene, 2012), pp. 339-40, under-
stands Lam. 3.15 metaphorically, but apparently sees real experience behind
it (cf. Job 9.18). On the other hand Beer’s comment suggests that these herbs/
vegetables were a selection from those that were regularly eaten (Die Mischna,
vol. 2/3, pp. 120-22).
16
In 2 Chr. 35.13, a later text, the requirements of Exod. 12.8-9 and Deut.
16.7 are conflated (so Fishbane, Biblical Interpretation, pp. 134-37; H.G.M.
Williamson, 1 and 2 Chronicles [NCB; London, 1982], p. 407).
12.1-20 53
roasting as opposed to boiling was that the animal could be cooked

whole, which seems to be the concern in v. 9b (see Houtman,
pp. 180-81, for further discussion). Importance seems also to have
been attached to the eating of the whole animal at a single meal,
hence the provisions in v. 10 (also found, in part, in Deut. 16.4b)
and the special measures for small families in v. 4.
11. The dress and equipment required (still in the second
person plural) for the meal all reflect preparedness for immediate
departure and anticipation of it, as is appropriate in the narrative
context. Deuteronomy 16.3 uses the same word (Heb. ḥippāzôn)
for ‘(fearful?) haste’ (see Note x on the translation for its connota-
tions) but, like what is probably an allusion to it in Isa. 52.12, of
the departure itself rather than the manner of eating the Passover
meal. The same idea is expressed in different words in the departure
narrative in Exod. 12.33, 39. It may be that the Priestly writers have
dramatised the ritual here to make the recollection of the Exodus
story more vivid.
‘Passover’ (Heb. pesaḥ) here, as more clearly in v. 21 (cf. Deut.
16.5), where NRSV’s ‘the Passover lamb’ renders this single word,
is the name for the animal which is slaughtered and eaten, rather
than the festival as a whole, and Schmidt suggests that the state-
ment may have been the conclusion of the old underlying ritual
text (p. 494). The expression ‘a Passover for Yahweh’ is another
allusion to standard sacrificial and more generally cultic language:
compare ‘a festival for Yahweh’ in v. 14 and ‘a Passover sacrifice
for Yahweh’ in v. 27. Although not limited to Priestly passages
(cf. Gen. 4.3; 8.20; Exod. 32.5), such expressions are naturally
particularly frequent there (e.g. 29.18; 30.10, 37). The one (other)
observance which is not cultic in the narrow sense to be designated
in this way is the Sabbath (20.10: cf. 16.23, 25; 31.15), but this
may be a way of drawing it into the cultic realm, like its description
as ‘holy’ (a description which is never applied to the Passover in
Exodus: contrast the ‘holy’ occasions associated with the festival of
Unleavened Bread in v. 16).
12-13. As sometimes happens elsewhere in Priestly legislation
(e.g. Num. 19.1-20), the explanation for a ritual only follows after
the details are complete. Here the transition is especially marked by
the use, for the first time in the chapter, of the divine first person: ‘I
will pass…’ There is even a minor formal tension here, since in vv. 11
and 14 Yahweh is spoken of in the third person. The first part of the
54 EXODUS 1–18
explanation corresponds closely but not exactly to the non-Priestly

announcements of the final plague in 11.1, 4-6 and especially to
the briefer recapitulation in v. 23a below, where the words ‘pass
through’ (Heb. ʿābar) and ‘strike down’ (Heb. hikkāh) both recur.
A distinctive element in the Priestly version is the judgement on
the Egyptian gods, a feature which is picked up elsewhere only in
Num. 33.4 (in 6.6 and 7.4 the object of Yahweh’s acts of judgement
is not specified). Here Yahweh is punishing not only Pharaoh and
his people but the gods whom they worship. The implication seems
to be that they are real beings, but totally under Yahweh’s control,
as in passages like Pss. 29.1-2, 97.7, 9 and especially 82.1-2, 6-7,
where Yahweh (here referred to as ‘God’, ʾelōhîm) proclaims judge-
ment against other gods more generally in the ‘divine council’. In
prophecy of the sixth century B.C. and later the gods of Egypt are
sometimes mentioned as bearing their share of divine judgement on
the nation (Isa. 19.1; Jer. 43.12-13; 46.25; Ezek. 30.13 [cf. v. 19]).
There is also some similarity to the polemic against Babylonian
religion in Deutero-Isaiah, especially in Isa. 46.1-2. Here in Exodus
too the idea that the gods of Egypt are the guarantors of Pharaoh’s
power, but unable to protect him against Yahweh’s intervention,
may be present. But the theme is left unelaborated and it is not
even clear whether the judgement on the Egyptian gods involves
the destruction of their temples and images (as in the Mekhilta of
Rabbi Ishmael and the Palestinian Targums: see Text and Versions)
or simply the humiliation of Pharaoh and the killing of the firstborn
(see further Houtman, p. 184). For ‘I am Yahweh’ compare 6.2, 6
and especially 9, where it also concludes a divine assurance.17
Verse 13 takes up the blood ritual of v. 7 (just as v. 23b does
with v. 22) and gives it the Priestly designation of ‘a sign for you’,
recalling the use of the same word (Heb. ʾôt) of the rainbow after
the story of the Flood and the mark of circumcision (Gen. 9.12, 17;
17.11). In both of those cases the sign is a reminder of a covenant,
and although it is not explicitly mentioned here the Priestly account
has already made clear twice that the deliverance of Israel from
17
S.R. Driver had already seen verses such as this as perhaps being extracts
from a body of ‘Holiness legislation’ because of this formula (Introduction, pp. 59,
151) and such ideas have recently been taken much further by Knohl, Sanctuary,
esp. p. 16: see the note below on v. 14. Here and in ch. 6, however, the formula
can be seen as an element of P’s own style.
12.1-20 55
Egypt is based on Yahweh’s remembering of his covenant with

Abraham, Isaac and Jacob (2.24; 6.5-6). Here, as the sequel shows,
the blood is not so much a sign for Israel to learn from as a mark
which distinguishes their houses from those of the Egyptians and so
ensures that Yahweh will spare them from the judgement: ‘a sign
for you’ rather than ‘a sign to you’ (for the latter see 3.12 and, by
implication, 4.1-9). This language lifts the rationale of the present
text (whatever may have been true of the ritual in earlier times)
out of the realm of magical thought into a theology of election and
covenant.
It is in this verse (as also in vv. 23 and 27) that an explanation
for the name ‘Passover’ (Heb. pesaḥ) is given, by the use of the
rare verb pāsaḥ, here translated for the sake of the connection ‘pass
over’. This is not the earliest rendering of the verb that is known
nor the most widespread in the ancient translations (see Text and
Versions). Its sense, to judge from its use in Isa. 31.5, is probably
‘protect’ or ‘spare’ (see Notes y and cc on the translation). The
result, in any case, is that ‘among you there will be no plague [or
“blow”: see Note dd on the translation] for destruction’: ‘to destroy
you’ would be more idiomatic English (cf. NRSV), but Heb. mašḥît
is a noun. In fact it is the same Hebrew word that is translated ‘(the)
Destroyer’ in v. 23, where it is probably to be parsed as a participle
of the verb šāḥat. There the mašḥît has a personal character: it seeks
to ‘enter your houses’ and ‘to strike you down’ (see further the notes
on v. 23), but here that is not the case: mašḥît is the effect of the
plague and the personal agent is Yahweh alone (‘When I strike’).
With such similar, but significantly different, wording some connec-
tion between the two verses is certain, and there can be little doubt
that in this case the Priestly writer knew v. 23 (or perhaps the litur-
gical tradition from which it derives) and modified it to eliminate
the traces of an angelic or demonic figure which earlier generations
had seen to be active along with Yahweh in the Exodus story and/or
in older celebrations of the Passover (‘protection’) ritual.18
18
The extensive dependence of vv. 12-13 on the older text (including vv. 29-
30) and the direction of dependence have been strongly argued by P.N. Tucker,
‘The Priestly Grundschrift: Source or Redaction? The Case of Ex 12:12-13’, ZAW
129 (2017), pp. 205-19. Contrary to his reasoning, however, this would not decide
the question of whether PG was a source or a redaction in favour of the latter. The
duplication earlier in Exodus clearly points to the combination of two separate
56 EXODUS 1–18
14. There is some disagreement over whether this verse belongs

more with what precedes (so most recent commentators, at least
on the continent of Europe) or with what follows (so most English
translations, but significantly not JB; also Dillmann, Baentsch,
McNeile, Hyatt [p. 134], Childs, Dozeman[?]): for further discus-
sion see the introduction to the section. The transition from the
unique setting of the Exodus to a permanent institution might
seem to support a break after v. 13, but both ‘this day’ (which it is
difficult to take cataphorically, as v. 15 begins with a ruling about
a seven-day period) and the (added) reference to ‘an everlasting
statute’ (Heb. ḥuqqat ʿôlām) as in v. 17 favour the view that v. 14 is
the conclusion to a law: so R. Hentschke, Satzung und Setzender:
ein Beitrag zur israelitischen Rechtsterminologie (BWANT 53;
Stuttgart, 1963), p. 65, followed in TWAT 3, 153 = TDOT 5, p. 144.
In fact none of what Hentschke saw as exceptions to this rule (Lev.
3.19; 16.29; Num. 19.10) need be such. This means, first, that it is
the day of Passover (the fourteenth of the first month, later Nisan)
which is seen as a ‘memorial’, that is an occasion when the events
of the Exodus and perhaps especially the sparing of the Israelites
from the death of the firstborn are to be remembered, an idea which
in general terms is found also in Deut. 16.1 (cf. v. 6b), although
the specific reference to ‘remembering’ (v. 3) occurs in the context
of the seven-day abstinence from unleavened bread (as in Exod.
13.3). Secondly, here Passover is treated as a ‘festival’ (Heb. ḥag;
so also in the Hebrew verbs here translated ‘celebrate’: see Note
gg on the translation) in its own right. This is unusual, although
the phrase ‘festival (ḥag) of the Passover’ occurs again in 34.25
(cf. 23.18?),19 for it seems incongruous for a word that is generally
thought to mean a ‘pilgrimage festival’ (e.g. BDB, p. 290; cf. de
Vaux, Institutions 2, p. 366, ET, p. 470) to be used in the context
of a family celebration such as appears here. Two explanations are
sources and originally PG probably had its own distinctive account of the death of
the Egyptian firstborn, traces of which can be found in 12.29a and in Num. 33.3.
Equally the use by PG of a Vorlage distinct from vv. 21-23 in the third-person
sections of vv. 3-11 remains a definite possibility: the correspondence with the
verbs in vv. 21-23 is not in fact complete (cf. vv. 7-8), and why would the Priestly
writer have turned the second-person forms in vv. 21-23 into the third person when
his own contributions are in the second person?
19
The two words are conjoined in the opposite order in Ezek. 45.21, perhaps
with a similar sense to here.
12.1-20 57
possible. The Priestly writers were probably aware of recent moves

to make Passover into a national festival at the Jerusalem temple
(Deut. 16.5-6; 2 Kgs 23.21-23; Ezek. 45.21-24) and could here be
indicating their approval of such a development. Alternatively, it
might be suggested that in discussion of the meaning of ḥag too
much weight has been given to the use of the Arabic cognate ḥajjun
to designate the pilgrimage to Mecca and that Hebrew ḥag simply
means a solemn festival (cf. Segal, The Passover, pp. 128-29;
TWAT 2, 730-44 = TDOT 4, pp. 201-13). The old ‘festival calen-
dars’ in Exod. 23.14-18 and 34.18, 22-25 probably antedate any
moves towards the centralisation of national religion in Jerusalem
and so attendance at the ḥaggîm mentioned in them need not origi-
nally have involved a pilgrimage at all. The final words of the
verse are very similar, not only to v. 17b, but to Lev. 23.41 (cf.
Gertz, pp. 35-36) and were very probably added with vv. 15-17: the
original Priestly law thus ends with v. 14a.
15-17. The Priestly law about the ‘festival’ (P does not use the
expression here: see below on v. 17) of Unleavened Bread seems to
have been preserved in two different forms in vv. 15-17 and 18-20.
Each is self-contained and the key features are found in both of
them: the seven-day abstinence from leavened bread (vv. 15 and 18;
cf. v. 20) and the punishment of anyone who transgresses by exclu-
sion from the community (vv. 15 and 19). Different words are used
for what is leavened: the usual ḥāmēṣ in v. 15 and the otherwise
unattested maḥmeṣet in vv. 19 and 20 (see Notes ll and uu on the
translation). In addition to the core requirements, vv. 15-17 demand
specifically the removal of all leaven on the first day, and ‘holy
occasions’ on both the first day and the seventh, when no work is
to be done except for the preparation of food. The seventh day was
already hallowed in earlier laws on the subject (13.6; Deut. 16.8;
not, however, specifically in Exod. 23.15 and 34.18). The first day
will be the day following Passover, when the departure from Egypt
was thought to have taken place: the passage elevates this day to
the same status as the final day and confirms this by the use of the
same concluding formula as now appears in v. 14: ‘as an everlasting
statute’ (v. 17). The special status of both the first and the seventh
days is maintained in the laws of Lev. 23.7-8 and Num. 28.18, 25,
where similar regulations apply to the festivals of Weeks (Lev.
23.21; Num. 28.26), ‘autumn New Year’ (Lev. 23.24-25; Num.
29.1), Atonement (Lev. 23.27-28; Num. 29.7) and Tabernacles
58 EXODUS 1–18
(Lev. 23.35-36; Num. 29.12, 35). It is most probable that vv. 15-17
(and v. 14b) were added here to conform to Leviticus 23 (see the
introduction to this passage).
Unlike the law in 13.3-10, which can probably be understood
to be attached (whether originally or not) to the narrative account
in vv. 33-34 and 39, and Deut. 16.1-8, where both ‘the bread of
affliction’ and ‘in great haste’ (v. 3) recall implicitly the Exodus
narrative, P provides no explanation for why unleavened bread is
to be eaten, except for the temporal coincidence with the day of
departure from Egypt (v. 17). The formula ‘on this very day’ (Heb.
beʿeṣem hayyôm hazzeh) is a standard Priestly formula (see Note ss
on the translation).20 For P it was the ritual of the Passover lamb,
apparently, which was the chief symbol for the Exodus deliverance,
and this may be why the beginning of v. 17 refers literally only to
the observance of ‘unleavened bread’, without using the word ḥag
here. This would be a departure from tradition (23.15; 34.18; Deut.
16.16; Ezek. 45.23) and one that was not to endure (cf. Lev. 23.6;
Num. 28.17; 2 Chr. 30.13, 21; 35.17), but the same tendency to
elevate Passover over Unleavened Bread in P can be seen in Num.
9.1-14 and below in vv. 43-49.
18-20. Compared to vv. 15-17, the main distinctive features of
these verses are (i) the correlation of the seven days of Unleavened
Bread with exact dates in the first month (v. 18); (ii) the require-
ment that leaven is to be excluded from the Israelites’ houses for the
whole seven-day period (v. 19); and (iii) the extension of the ruling
to cover ‘resident aliens’ (Heb. gērîm) as well as native Israelites
(v. 19). As in v. 8, it is laid down here that the eating of unleav-
ened bread must begin with the Passover meal itself (cf. the date in
v. 6) and the specification ‘in the evening’, although using different
terminology, actually allows time to elapse after the slaughtering of
the animal at twilight and so fits well with v. 6b. According to de
Vaux, who rightly places the change from a day beginning in the
morning to a day beginning in the evening in the late Old Testa-
ment period (Institutions 1, pp. 275-77, ET pp. 180-82: Neh. 13.19
provides a terminus ante quem), the expressions used here imply the
later system (ibid., 1, p. 277, ET p. 182). This may not be the case:
20
The reference to the Exodus in the past tense here is anachronistic and
shows, even more clearly than the description as ‘an everlasting statute’, that the
wording of much later parenesis is being drawn on (cf. 13.8; Deut. 16.3).
12.1-20 59
the evening of Passover may be an addition to the seven full days

here, and if ‘until the twenty-first day of the month in the evening’
is inclusive, it would put the beginning of the twenty-second day on
the following morning. As de Vaux observed, the different dating of
the beginning of Unleavened Bread on the fifteenth in Lev. 23.6 and
Num. 28.17 follows the older system.
The inclusion of resident aliens and the actual terminology used
in v. 19 link this section to the additional laws about Passover in
vv. 43-49 (esp. vv. 48-49: see further the Explanatory Note there)
and to Num. 9.1-14 (esp. v. 14), as well as to other laws which seem
not to be an original part of P (Lev. 16.29; 17.15; 18.26; 19.34;
24.16, 22; Num. 15.29, 30; see also Josh. 8.33; Ezek. 47.22). The
same spirit of inclusiveness is reflected also in, and may have been
inspired by, some of the later prophets (esp. Isa. 56.3-8). The expres-
sion ‘natives of the land’, as in v. 48 and Num. 9.14, which scarcely
occurs outside the Priestly and Holiness legislation (the only excep-
tions are Josh. 8.33, Ezek. 47.22 and [?] Ps. 37.35), provided a
useful identity marker in later times for true-born Israelites in
contrast to foreign residents. It clearly has in view the occupation
of Canaan and, indirectly, the restored community in Judah centred
on Jerusalem. On the ‘resident aliens’ see Note dd on the translation
of 2.11-22: the same status is earlier attributed by P to Israel’s own
ancestors in Canaan (see the notes on 6.3-4).
Text and Versions

Before 12.1 4QpalExl has an empty line, but there is no room for a division
in 4QpalExm.
‫( ויאמר‬12.1) Vulg dixit quoque: quoque only occasionally (as in 10.25)
reflects ‫ גם‬in the Heb. Here (as in 7.19; 8.16; 9.13) it is added to mark a new
word from Yahweh soon after an earlier one (cf. 11.9).
‫( יהוה‬12.1) TgNmg adds ‘the Memra of’, as do the early TgG mss AA, HH
and KK.
‫( ואל־אהרן‬12.1) LXX and Vulg have no equivalent to the second ‫אל‬, but
the ‘Three’ and Syh follow MT.
‫( בארץ מצרים‬12.1) LXX as usual renders by an appositional expression
ἐν γῇ Αἰγύπτῳ (cf. Wevers, Notes, p. 168) but the genitive Αἰγύπτου is well
attested and was taken over into (OL and) Vulg.
‫( לאמר‬12.1) Sy has no equivalent, as in other places where ‫ אמר‬has
preceded in the same clause (e.g. 5.10; 7.8).
‫( החדשׁ הזה‬12.2) TgN and TgG (HH) add ‘Nisan’, as do TgF and TgG (AA) after
the next word. TgG (GG), like TgO,J, does not have it at all.
60 EXODUS 1–18
‫( לכם‬12.2) Sy nhwʾ supplies a verb in the future tense, as TgG (AA) and
Vulg also do in the next clause. TgJ adds ‫מקבעיה‬, ‘to fix it’, emphasising like
its addition later in the verse (see the next note) the human regulation of the
calendar.
‫( ראשׁ חדשׁים‬12.2) SP reads ‫ראשׁ החדשׁים‬, adding the def. art. to correspond
to the determined expression at the end of the verse. The remains of 4QpalExl
are faint at this point, but it probably agrees with MT (DJD IX, p. 83), as
does LXX ἀρχή and TgN ‫רישׁ ירחין‬. The ‫ ירחיא‬of TgO,J may reflect a similar
concern to SP, but not necessarily a different Vorlage. MT is the more difficult
reading and probably original. After this phrase TgJ added ‘and from it you
shall begin counting the feasts and the times and the seasons’. TgF (ms. P)
has a much longer addition of a poetic ‘Dispute of the Months’ (ET in AramB
2, p. 46: cf. S.P. Brock, ‘A Dispute of the Months and Some Related Syriac
Texts’, JSS 30 [1985], pp. 181-211 [185-86, 209-11]). Similar poems appear
in some Genizah mss (cf. Klein 1, pp. 186-207, with additions in M. Klein,
‘Complementary Fragments from the Cairo Genizah’ [Heb.], in M.V. Fox et
al. [eds.], Texts, Temples and Traditions [FS M. Haran; Winona Lake, 1996],
pp. 95*-117* [96*-98*]).
‫( לחדשׁי השׁנה‬12.2) LXX ἐν τοῖς μησὶν τοῦ ἐνιαυτοῦ and Vulg in
mensibus anni paraphrase the genitival lamed of MT, while TgJ and TgG(HH)
insert ‘for counting’ to clarify the meaning. Sy adds ‘all’, and this may also be
the intention of TgNmg’s ‫לכל‬, which is now oddly attached to ‫ לכון‬earlier in the
v. TgN itself has ‘all the beginnings of…’ (likewise in TgF and TgG (AA; cf. HH)),
perhaps implying that it understands ‫ חדשׁ‬throughout the verse in its sense of
‘new moon’ (although this would make its earlier addition of the month-name
Nisan anomalous).
‫( דברו‬12.3) LXX and Sy have sing. imperatives, conforming to the
common pattern in Exodus where Moses alone speaks on Yahweh’s behalf
(and cf. esp. v. 21). For the pl. cf. vv. 1 and 28 and e.g. Lev. 11.1; 15.1. SP,
Vulg and Tgg have the pl. here like MT: the Qumran mss do not survive at
this point. SP adds ‫ נא‬afterwards, probably conforming to the similar expres-
sion in 11.2, which is also not at the beginning of a speech: see Note e on the
translation of 11.1-10.
‫( אל־כל‬12.3) TgO,J,N imply ‫ ִעם כל‬as expected, but the early TgG mss
read ‫( ַעם‬i.e. ‫( לבת )לות‬HH) and ‫( אל־כל ַעם‬AA), with ‫ ַעם‬added from v. 6 and
displacing ‫ כל‬in HH.
‫( עדת‬12.3) Here and generally elsewhere LXX renders ‫ עדה‬by
συναγωγή(ν). The use of this word for a community of people (rather than
the action of gathering or a collection of things) is apparently a new develop-
ment in Hell. Gk. (cf. LSJ, p. 1692), but it is not restricted to Jewish sources
(or in LXX to references to Israel [cf. Muraoka, p. 533]).
‫( ישׂראל‬12.3) So MT and TgO,J,G(AA), and cf. vv. 6 and 19: but the prefixing
of ‫( בני‬as commonly elsewhere after the constr. form of ‫עדה‬: e.g. 16.1-2) has
strong support (SP, 4QpalExl, LXX, Vulg, Sy, TgN,G(HH)) and is characteristic
12.1-20 61
of P’s style. Still, the consistent use of ‫ ישׂראל‬alone throughout vv. 1-20 in MT
(cf. also v. 15) speaks in favour of the originality of the unusual reading (so
also Propp, p. 358).
‫( לאמר‬12.3) TgG(HH) has no equivalent, probably a careless omission.
‫( בעשׂר‬12.3) TgN and both the early mss of TgG have ‫בעשׂרה יומין‬, lit. ‘in ten
days’, reflecting the use of BH ‫ ָעשׂ ֹר‬for a period of ten days in Gen. 24.55 and
perhaps a more flexible view of when the Passover animal was to be ‘taken’
(for a variant of the latter see on TgJ in the next note).
‫( לחדשׁ הזה‬12.3) Both mss of TgG add ‘Nisan’ here. TgG(AA) has no equiva-
lent to ‫הזה‬. After the date TgJ adds ‘Its time is fixed on this occasion but not
for the generations (to come)’, a ruling found in M.Pes. 9.5 and elsewhere.
‫( ויקחו להם‬12.3) LXX ignores the idiomatic initial waw and ‫ להם‬in its freer
rendering of the Heb. TgN has ‫לכון‬, but the mg corrects to ‫להון‬.
‫( אישׁ‬12.3) TgJ has no equivalent, simplifying its rendering.
‫( שׂה לבית־אבת שׂה לבית‬12.3) Most of the versional renderings for ‫ שׂה‬limit
the reference to sheep, thus missing the wider meaning of this term (see Note
e on the translation). But LXX πρόβατον had a wider meaning in earlier
Gk., and in LXX it is often the equivalent of ‫שׂה‬. 4QpalExl and Sy, perhaps
independently, invert the order of the two phrases, either accidentally or to
put the more specific phrase second. TgN and Sy add the expected pron. suffix
(sing. or pl.) to ‫אבת‬, while TgN and both mss of TgG add ‫ כל‬before the second
‫בית‬. TgO and TgG(AA), probably wrongly, treat ‫ בית־אבת‬as sing. A distinction
between families and households is generally assumed, most clearly in Vulg’s
paraphrase per familias et domos suas and in TgJ’s addition after ‫ אבת‬of ‘and
if they are more numerous than the counting’: on the basis of the alterna-
tive in v. 4, the possibility is envisaged that individual households in a large
family might each have a separate animal, a variation that already appears in
Tannaitic midrashim (see AramB 2, p. 190 n. 9).
Both LXX and Vulg translate v. 4 quite freely, but the meaning of the Heb.
is mostly preserved and there is little if any evidence of a Vorlage divergent
from MT.
‫( הבית‬12.4) Most of the Vss vary their rendering to make clear that the
members of the ‘household’ are meant here: only TgO and Sy do not.
‫( מהית משׂה‬12.4) LXX, Vulg and TgJ paraphrase with expressions for
‘sufficiency’, while Tgg all introduce the idea of ‘being counted’ which is
found also in MRI (Lauterbach 1, p. 26): TgJ adds the minimum of ten persons
which Jos., BJ 6.9.3 (423) already has (but M.Pes. 9.11 seems to allow for
companies of five persons). Sy has dnpwq bh ʾmrʾ, lit. ‘so that the lamb should
go out in it’, i.e. ‘be fully consumed’ (cf. the uses of npq for ‘spend, expend’
in Payne Smith, pp. 345-46). TgN,G additionally specify that the Passover lamb
is meant.
‫( ולקח הוא‬12.4) So also most SP mss (4QpalExl does not survive at this
point), but Camb. 1846 reads ‫ ולקחו‬without the ‫הוא‬, evidently an aural error, as
the continuation with ‫ ושׁכינו‬does not make sense.
62 EXODUS 1–18
‫( ושׁכנו‬12.4) LXX τὸν γείτονα and Vulg vicinum suum ignore the waw
and oddly take ‫ שׁכנו‬as the object rather than a second subject, probably
because of the sing. ‫ולקח‬, though the other Vss had no problem with this.
‫( במכסת נפשׁת‬12.4) So also 4QpalExl, but SP has ‫במכסות נפשׁות‬, which BHS
takes as a pl. form (for the ‘numbers’ in each household?). It might, however,
be a sing. noun in ‫וּת‬- (so GSH §144b). A sing. form is reflected in LXX, Vulg,
Tgg and Sy.
‫( לפי אכלו‬12.4) LXX τὸ ἀρκοῦν αὐτῷ closely resembles TgJ ‫לפם מיסת‬
‫מיכליה‬, but the same ‘unpacking’ of the Heb. phrase could have arisen
independently in the translators’ minds. The mst mʾklh of Sy might more
plausibly be seen as related to one of the other two.
‫( תכסו‬12.4) συναριθμήσετε has been plausibly reconstructed as the
original LXX reading, even though it occurs in no ms. (cf. Walters, Text, pp.
61, 105; Wevers, Notes, p. 169; THGE, p. 230): the middle form in mss A,
B etc. is only attested in a late papyrus acc. to LSJ, p. 1699. TgJ has ‫תיכסון‬
(followed by ‫)ית אימרא‬, from ‫‘ = נכס‬slaughter’, which it presumably saw as the
verb behind the Heb. form because of the wider context (cf. v. 6). But it does
not fit the ‫ על‬of the Heb. or the preceding ‫מכסת‬, which TgJ correctly rendered
by ‘number’ (‫)סכום‬.
‫( על־השׂה‬12.4) TgN,G again add ‘of the Passover’. TgG has ‘lambs’ in the
pl. (as it did also earlier in the verse) and TgNmg has picked this variant up.
TgG’s ‫ עם‬for ‫( על‬again as earlier in the verse) is another sign of its frequent
carelessness.
‫( תמים‬12.5) LXX has τέλειον, a word used since Homer for valid
sacrifices. TgN,G and Sy make clear the sense in which ‫ תמים‬is being used by
adding (or, in the case of Sy, substituting) ‘without blemish’ (mwm). Cf. MT
at Lev. 22.21 and Num. 19.2. Such clarification is widespread in TgN (Exod.
29.1; Lev. 1.3, 10) and TgG (cf. Lev. 23.12, 18; Num. 28.19) and can be traced
back to MRI (Lauterbach 1, p. 28).
‫( בן שׁנה‬12.5) The original text of TgO read ‫בר שׁתיה‬, ‘the son of his year’
(cf. Lev. 12.6), which comparison with MRI shows to have been a way of
confirming that an animal up to a full year old was meant (Lauterbach 1,
p. 29).
‫( מן־הכבשׂים‬12.5) SP and 4QpalExl read ‫מן־הכשבים‬. The spelling found in
MT (where it is the prevalent form [107x; plus a fem. form]) corresponds to
the cognates in related languages and so is older. But the metathesised spelling
of SP and 4QpalExl is found 13x in MT, including 4x in Genesis 30, which
may have led to the variant spelling here: in general SP has the same spelling
as MT for this word. At Qumran the spelling ‫ כבש‬predominates, except where
a cited biblical text has ‫ כשב‬and at 4Q251 f12.1.
‫( ומן־העזים‬12.5) TgO,J insert ‫ בני‬to show that a young goat is meant, and Sy
has mn gdyʾ with the same intention. TgN,G combine the two. Vulg apparently
misunderstood the waw (which here means ‘or’) as requiring an additional
12.1-20 63
offering: iuxta quem ritum tolletis et hedum probably means ‘alongside this
ritual you shall also take [or “kill”] a young goat’.
‫( תקחו‬12.5) TgN (but not TgG) adds ‫לכון‬: cf. MT at v. 3.
‫( למשׁמרת‬12.6) LXX and Vulg understandably render with some freedom,
cf. TgJ and Sy nṭyr, ‘kept’. TgJ prefixes ‫קטיר ו‬, ‘tied up and’, indicating the
means of ‘keeping’ the animal.
‫( יום‬12.6) LXX and Sy have no equivalent, following the idiom of the
target languages (and sometimes also Heb.: cf. Note c on the translation).
‫( לחדשׁ הזה‬12.6) TgG has ‘Nisan’ in place of ‘this’, as in vv. 2-3. TgJ adds
‘so that you may know that you are not afraid of the Egyptians who see it’, an
explanation found also in MRI (Lauterbach 1, p. 39) and Exod.R. (cf. 8.26).
‫( אתו ושׁחטו‬12.6) A fr. of 4QpalExm which seems well located here reads
]‫אתם ו[ושׁחט‬, making the obj. pl. to fit the pl. subject. But as the easier reading
it must be secondary. The other witnesses all read the sing. After these words
TgJ adds ‫כהילכתא‬, ‘according to the law/rule’.
‫( כל קהל עדת־ישׂראל‬12.6) TgG originally omitted all the remainder of the
verse and replaced it with v. 3b. A subsequent scribe added the missing words
and deleted the intrusive ones. The original mistake was probably due to
parablepsis arising from the similarity of ‫ ויכסון‬here and ‫ ויסבון‬in v. 3 (Klein 2,
p. 57). The most widespread variant is the addition of ‘children of’ before ‫ישׂ‬
‫ ראל‬as in v. 3 (so SP, 4QpalExl [almost certainly to be restored in a lacuna],
LXX, Vulg, TgN,G and Sy here): but again this is probably due to assimila-
tion to the more common formula. For ‫ קהל‬LXX unusually has πλῆθος
(only elsewhere at 2 Chr. 31.18), probably because its normal equivalent in
Genesis–Numbers, συναγωγή, is being used for ‫ עדה‬here (Vulg has multitudo
alone for both words): cf. Barr, Semantics, pp. 128, 253. The Tgg mainly
use the Aram. cognate (TgO ‫קהלא‬, without a following ‫ד‬, seems to place it in
apposition to ‫עדת־ישׂראל‬, perhaps a sign of bafflement at the double description
of the people), but TgG (like Sy) has ‫עם‬. TgNmg has both, either by conflation
or to show that ‫כנישׁתא‬, its rendering for ‫עדת‬, is being used of the community
rather than a building.
‫( בין הערבים‬12.6) LXX πρὸς ἑσπέραν and Vulg ad vesperam are
sometimes treated as not distinguishing this phrase from ‫( בערב‬cf. GK §88c)
but their renderings of ‫ בערב‬are generally distinct and these phrases appear
to mean ‘towards evening’, so earlier than ‘in the evening’. Sy bmʿrby šmšʾ
(also in 16.12), ‘at sunset’, will be based on Deut. 16.6. Tgg ‫ביני שׁמשׁיא‬/‫בין‬, a
phrase whose MH equivalent occurs in Aboth 5.9 and is precisely defined in
B.Shabb. 34b: ‘from the time that the sun sets as long as the eastern horizon is
red’ (cf. J.Ber. T. 2b). ‫ בין שׁמשׁא‬was specifically used of Sabbath eve, perhaps
without such a precise sense (see refs. in Jastrow, p. 1602): the exact sense
of ‫ שׁמשׁ‬here remains uncertain. The first hand of the early SP ms. Camb. 1846
omitted the phrase and it was restored above the line.
‫( הדם‬12.7) Sy dmh, ‘its blood’, sc. of the animal, for precision.
64 EXODUS 1–18
‫( ונתנו‬12.7) Sy wnrmwn, ‘and they shall put’; later mss wnrswn, ‘and they
shall sprinkle’, as fitting a liquid.
‫( המשׁקוף‬12.7) Some of the equivalents used in the Vss could also mean
‘doorpost’ (LXX φλιάν) or ‘threshold’ (TgO,N,G ‫ ;שׁקפא‬TgJ, Sy ‫)אסכופא‬: to avoid
ambiguity Aq substituted ὑπέρθυρον (an ancient word) and TgJ added ‫עילאה‬.
TgJ also specified that the daubing was to be done ‘on the outside’, perhaps
a trace of the view (which TgJ’s translation does not follow) that ‫ סף‬in v. 22
means ‘threshold’: see Text and Versions there.
‫( על הבתים‬12.7) LXX ἐν τοῖς οἴκοις probably has the same meaning as
Heb., but in correcting it the Three prefixed ‘and’, as if the whole house was
to be daubed with blood. The presence of καί (but not ἐπί) in LXXO shows
that this was an early addition (cf. Sy wʿl btʾ), and the widespread attestation
of this reading possibly derives from the elusive Lucianic recension.
‫( אתו‬12.7) LXX αὐτά, the pl. at this point probably being due, as Wevers
suggests (Notes, p. 171), to the fact that ‘houses’ are now being spoken of. TgN
replaces with the specific ‫פסחא‬, but this does not appear in TgG. TgJ adds ‘and
sleep’, perhaps with reference to those who went to share Passover in another
house: the issue was already a concern for MRI (Lauterbach, p. 45), where
v. 13 is cited in support.
‫( בהם‬12.7) According to BHS a Geniza Heb. ms. reads ‫ שׁם‬here, but this
is probably only assimilation to v. 13.
‫( בלילה הזה‬12.8) Vulg nocte illa, reflecting the distance of the readers
in time from the events.21 TgJ adds ‘of the fifteenth [sic] of Nisan’ and, in
line with M.Zeb. 5.8 (and other rabbinic sources: cf. AramB 2, p. 190 n. 20),
‘until midnight’. TgN,G substitute ‘of Passover’ for ‘this’ (cf. their additions in
vv. 4 and 7).
‫( צלי־אשׁ‬12.8) Sy has kd mṭwy bnwrʾ. In Syr. kd with a part. can indicate
an ongoing process (‘while…’: Payne Smith, p. 204), but this is clearly
inappropriate here and kd often means simply ‘when’ or ‘after’ (cf. its use in
v. 9 in TgO).
‫( מצת‬12.8) TgO,J have the sing. ‫( פטיר‬cf. Sy), which could presumably
have a collective meaning like, e.g., ‫לחם‬. Propp (p. 360) prefers the reading
of Num. 9.11, ′‫על־מצות ומ‬, which is certainly easier, but that very fact together
with the absence of a well-attested variant here suggest that MT was regarded
as intelligible.
‫( על מררים‬12.8) For ‫ על‬TgN,G have ‫עם‬, which may well capture the
meaning intended by the Heb. (cf. Note r on the translation). Several of the
Vss give very specific renderings of ‫מררים‬: TgJ ‘chervil and endives’ names
two of the five options listed in M.Pes. 2.6, and Vulg’s lactucis agrestibus,
‘wild lettuce’, is also in the list. LXX’s πικρίδων looks generic (so Lust et al.
21
Or is it that of Moses’ speech? The Heb. ‘this’ might seem to suggest the
day when Moses was speaking, but it is ‘co-text referential’ rather than ‘con-text
referential’.
12.1-20 65
and BAlex), but LSJ, p. 1404, followed by Muraoka, p. 457, gives two specific
equivalents, ‘ox-tongue’ and ‘chicory’, the latter of which is also in the
Mishnaic list (cf. BAlex). Vulg may perhaps be based on this understanding
of πικρίδων.
‫( יאכלהו‬12.8) LXX and TgG (but not TgN) ignore the suffix, perhaps to
ease the syntax, with ‫ מצת‬as the obj. of ‘eat’ (cf. Wevers, Notes, p. 172; BAlex).
But it is attested by SP as well as MT and the other Vss.
‫( נא‬12.9) 4QpalExl reads ‫נו‬, reflecting a common orthographical/phonetic
variation at Qumran (cf. Qimron, pp. 39-40; more fully N. Mizrahi, ‘Linguistic
Change through the Prism of Textual Transmission: The Case of Exodus
12:9’, in A. Moshavi and T. Notarius [eds.], Advances in Biblical Hebrew
Linguistics: Data, Methods and Analyses [Winona Lake, 2017], pp. 27-52).
The meaning is not affected. TgO,J ‫( כד הי‬cf. Sy) use ‫ כד‬in its common sense
of ‘when’, but without a verbal clause, and ‫ הי‬for ‘raw’, a sense it has only
once in BH (1 Sam. 2.15), apart from the references to ‘raw [human] flesh’
in Leviticus 13. TgN,G,F have ‫מהבהב‬, ‘lightly roasted, rare’, perhaps rightly
reasoning that no one would consider eating meat raw.
‫( ובשׁל מבשׁל במים‬12.9) The abbreviated renderings of LXX ἡψημένον ἐν
ὕδατι and Vulg coctum aqua have the main point but seem uninterested in the
intricate construction of the Heb. phrase (the Three and LXXO restore the full
expression). The Tgg and Sy, on the other hand, show awareness of various
rabbinic interpretations of it. TgO and Sy take ‫ בשׁל‬as an inf. abs., giving
the sense ‘thoroughly boiled in water’ and the ‫שליק‬, ‘overboiled’, of TgN,G
produces the same sense, which is discussed and rejected in MRI (Lauterbach,
p. 48). Some important SP mss (cf. Tal, Sadaqa, Camb. 713 and Camb. 1846)
read the first word as ‫בשול‬, but in view of the infrequency of the inf. abs. in
SH (cf. GSH §178) this is more likely intended as a Qal pass. pt., without
difference in meaning from MT. TgJ assumes that since the following phrase
mentions water ‫ בשׁל‬alone must refer to other liquids in which meat might be
cooked: ‘boiled in wine or oil or (any) liquid’ – another view that is mentioned
but rejected in MRI (ibid.). The reading ‫ ומבשׁל‬in 4QpalExl may indicate that
it too saw a general ban on boiling followed by a specific one here. In all
the Tgg except TgG ‫אף לא‬, ‘not even’, is inserted before the whole phrase to
draw attention to the fact that a method of cooking allowed in other laws was
forbidden for Passover here.
‫( על־כרעיו‬12.9) The ‫( על‬like the following one) is translated by ‘with’ in
LXX, Vulg, TgJ,N,G: compare TgN,G in v. 8. The rendering ‘its feet’ for ‫כרעיו‬
(LXX, Vulg, TgJ,N) may derive from the fact that in JAram ‫ כרע‬sometimes
means ‘foot’. Sy replaces both occurrences of ‫ על‬with w and adds a w before
‘its head’, suggesting that it (like Vulg more clearly by its addition of vorabitis)
mistakenly presumed that all these parts of the animal were to be eaten.
‫( ולא‬12.10) LXX does not represent the waw, but SP (apart from Sadaqa’s
edition), 4QpalExl and the other Vss support MT, which is probably original
here.
66 EXODUS 1–18
‫( תותירו‬12.10) The original LXX reading ἀπολείψετε is preserved only

in miniscules and citations, but the readings of A and B are clearly corrup-
tions of it. Vulg remanebit is a typically free reformulation and not significant
textually.
‫( עד־בקר‬12.10)1o After this LXX (or perhaps its Vorlage) added ὀστοῦν
οὐ συντρίψετε ἀπ’αὐτοῦ from v. 46, presumably because the disposal of
the carcass was the point at which its bones might easily have been broken
and the regulation might have been ignored if included only in the ‘appendix’
at the end of the chapter. The same addition is made in MT at Num. 9.12.
The wording is identical to v. 46: the citation in John 19.36 differs slightly
(συντριβήσεται), so that could not be the source of the addition.
‫( עד־בקר‬12.10)2o Here TgJ adds ‘you shall put it aside and on the even-
ing(?) of the 16th’ and then, after ‘you shall burn it with fire’, explains that
‘it is not possible for the remainder of an offering of holy things to be burned
on a holy day (‫’)ביומא טבא‬. The same concern is evident in MRI (Lauterbach,
pp. 50-51) and in other rabbinic texts (see AramB 2, p. 191 n. 24).
‫( תשׂרפו‬12.10) Most SP mss and edd. agree with MT, but Tal prints ‫תשרפ‬
(sic): this is almost certainly a printing error. TgG (which had ignored an obj.
suff. in v. 8) and Sy supply ‘it’ here.
‫( ככה‬12.11) TgJ expands with ‘according to this rule’, similarly TgN,G in
different words.
‫( תאכלו אתו‬12.11) TgJ adds ‘at this time but not for (future) generations’:
cf. v. 3 and M.Pes. 9.5.
‫( חגרים‬12.11) TgF spiritualises by adding ‘with the commandments of
the law’, and this is picked up by TgNmg (among many other cases of such
Targumic references to the law see e.g. TgJ on Gen. 2.15; 49.10, 14-15; TgF
on Gen. 49.7, 10).
‫( נעליכם‬12.11) LXX prefixes ‘and’, in line with Greek idiom, as does Sy.
‫( ומקלכם בידכם‬12.11) MT’s sing. nouns with the 2 pl. suff. (cf. JM
§135b-c) are not surprisingly rendered as plurals in the Vss (only TgO ‫בידכון‬
keeps to MT), and SP likewise exhibits the pl. forms. In 4QpalExl only ‫[כם‬
survives from this phrase, so its reading for the nouns cannot be determined
(DJD IX, p. 33, inclines to the MT reading on the basis of very faint traces).
‫( ואכלתם אתו‬12.11) Sy has ʾwklwhy, without the initial waw of MT,
creating a more staccato effect (for the form, as again in v. 14, cf. Brockel-
mann §186 n. 3).
‫( בחפזון‬12.11) LXX (probably), Symm, Vulg and Sy understood the
meaning to be ‘in haste’ and Tgg ‫ בבהילו‬can mean this (cf. Ezra 4.23), though
it could refer to agitation or alarm. The latter is the view of Theod and Aq (ἐν
θάμβῳ). TgJ attributes the ‫ בהילו‬to ‘the Shekinah of the eternal Lord’, a view
considered possible in MRI (Lauterbach, p. 52), though it prefers to apply it
to the Egyptians (cf. v. 39).
‫( פסח הוא‬12.11) Sy introduces the clause with an artificial connection, mṭl
d, ‘because’. ‫ פסח‬is transcribed in its Aram. form πάσχα in LXX and Theod:
12.1-20 67
Symm φάσεχ and Vulg phase use forms that are closer, but not identical,
to the Heb. as transmitted by MT. Each adds an interpretation, Symm
ὑπερμάχησις, ‘defence, protection’, probably deduced from the occurrence
of the related verb in Isa. 31.5 (cf. Salvesen, Symmachus, pp. 83-85; Walters,
pp. 174, 248, 344), and Vulg id est transitus. The latter was Aq’s view
(ὑπέρβασις), based on the use of ‫ עבר‬in the following verse and in v. 23, and
it has given rise to the standard Eng. expression ‘Passover’. TgO,G and Sy have
simply ‫פסחא‬/pṣḥʾ, but TgN prefixes ‫מכס‬, ‘sacrifice’, to accord with MT in v. 27.
TgJ picks up a common view that ‫ פסח‬is related to ‫חוס‬, ‘spare, have pity’ (cf.
Salvesen, pp. 84-85) and renders ‘because it is a sparing (‫ )חייסא‬for you…’:
cf. its renderings in vv. 13 and 27 and rabbinic pars. in AramB 2, p. 191 n. 28.
‫( ליהוה‬12.11) The Tgg. mostly have ‘before the Lord’, as elsewhere, but
TgG has ‘for the name of the Lord’, a periphrasis found widely in TgN and
elsewhere (cf. Chester, Divine Revelation, pp. 343, 345-46).
‫( ועברתי‬12.12) Some SP mss read ‫( והעברתי‬only the Cambridge ms.
713 among the earliest), presumably implying the interposition of some
unspecified agent (perhaps the ‫ משׁחית‬of v. 23). The Tgg show a similar concern
to avoid any suggestion that God moves (cf. 11.4 and the note there), either
by rendering with ‫( ואתגלי‬O,J) or by involving the Memra (as subject in G, as
agent in N): TgNmg combines the two.
‫( בארץ מצרים‬12.12)1o TgJ adds ‘in the Shekinah of my glory’ and a refer-
ence to ‘90,000 myriads of destroying angels’, which goes far beyond the
single intermediary in TgJ’s renderings of 4.25 and 12.13: cf. J.Sanh. 2, 20a.
The foundation for the idea could be found in Ps. 78.49.
‫( הזה‬12.12) Vulg illa as in v. 8. TgG has ‘of Passover’, TgN combines this
with MT (cf. v. 8).
‫( והכיתי‬12.12) The Tgg use ‫ קטל‬to make the slaughter explicit (with the
Memra as subj. in TgG): Sy ‘(they) shall die’ has the same effect and avoids
direct responsibility of God.
‫( כל־בכור‬12.12) TgN,G and Sy have ‘all the firstborn’ in the pl. and Sy
continues with ‘of the land of Egypt’ instead of ‘in…’ (cf. 11.5), in both cases
rendering MT with a little freedom.
‫( מאדם‬12.12) TgN,G and Sy again agree in using an explicitly pl. form, this
time ‘sons of men’.
‫( ועד‬12.12) All SP mss (except Sadaqa’s) read ‫עד‬, as SP seems to prefer
for this phrase (see the note on 9.25). The absence of ‘and’ in LXX and Vulg
need not imply a Vorlage different from MT.
‫( אלהי‬12.12) The Tgg have ‫טעות‬, ‘errors’, i.e. false gods, Sy dḥlthwn,
‘what they worship’: both are ways to avoid implying that the ‘gods’ other
than Yahweh actually exist (for details see AramB 6, p. 20 and n. 21).
‫( שׁפטים‬12.12) LXX has τὴν ἐκδίκησιν, ‘vengeance, punishment’ (for the
sing. cf. Sy dynʾ) as in 7.4 (see the note there) and Num. 33.4, apparently a
deliberate variation towards a word more often used to represent the root ‫נקם‬
(in 6.6 LXX has the expected κρίσει [again sing.] to convey the purpose of
68 EXODUS 1–18
God’s action). The Pal. Tgg and TgJ follow the view, already recorded in MRI
(Lauterbach, p. 55), that the pl. points to between two and four different kinds
of judgement on images made of different materials: TgJ in full detail, TgN,G
by just adding the word ‘different’ (which is how MRI begins its comment on
the phrase).
‫( אני יהוה‬12.12) TgN has ‘says Yahweh’, an interesting view of the
meaning of this common formula: but the mg corrects to MT (and to TgG).22
‫( והיה הדם‬12.13) 4QExc definitely and 4QpalExl possibly did not have
an interval after v. 12: SP has one, but MT does not. Sy has ‘this blood’;
TgJ expands its explanation to include not only ‘the blood of the Passover
sacrifice’ but the blood from circumcision mixed with it, having already
indicated the power of the latter to ward off ‘the destroying angel’ in 4.24-26.
For numerous rabbinic references to this mixture (inc. MRI on v. 6 [Lauter-
bach, pp. 33-34]) see AramB 2, p. 191 n. 31.
‫( לאת‬12.13) LXX ἐν σημείῳ, with ἐν = ‘as’, a possible sense in Hell. and
LXX Gk. (cf. LSJ, p. 552, s.v. I.10; Muraoka, p. 182, s.v. 14 and 20).
‫( אתם‬12.13) TgJ,N expand with ‘(are) dwelling’; Vulg uses the future
eritis. TgN has ‫אנון‬, ‘they’, but the earlier TgG has ‫אתון‬, so ‫ אנון‬is probably a
scribal error.
‫( וראיתי את־הדם‬12.13) TgJ has ‘the merit of the blood’, an idea implied in
MRI (Lauterbach, p. 56). TgG attributes the seeing to the Memra: see also the
next note.
‫( ופסחתי‬12.13) TgG uses the third person, with the Memra still the sub-
ject; TgN adds ‘by my Memra’ here. Both combine a transcription of MT with
forms of the verb ‫נגן‬, ‘protect’: for this interpretation of ‫( פסח‬on which see
Brock, ‘An Early Interpretation’, p. 29) compare LXX σκεπάσω and see the
notes on vv. 11 and 23. TgO,J have the related understanding of ‫ פסח‬as ‘spare’,
which is specifically related to Isa. 31.5 in MRI on this verse (Lauterbach,
pp. 56-57), while Vulg again follows Aq’s lead with transibo. The exact
sense to be given to Sy wʾpṣḥ (a dialectal variant of psḥ) is uncertain, since
it is used only in this context (cf. vv. 23 and 27), but in Isa. 31.5 Sy renders
‫ פסוח‬by nsyʿ, from syʿ = ‘help’, which seems closer to ‘protect’ and ‘spare’
than ‘pass over’.
‫( נגף למשׁחית‬12.13) TgO,N and Sy use the stronger ‘death’ for ‫נגף‬, which is
comparable to its use for ‫ דבר‬in 9.3 and 9.15: see the notes there and AramB
7, p. 23 n. 4. TgJ elaborates this understanding into a reference to ‘the angel
of death, who has been given power for destruction’, but will have no access
to the Israelites: TgNmg has a variant of this.
22
TgN’s rendering is found in other places where the formula concludes a state-
ment: e.g. Lev. 18.5 and frequently in the following chapters; also in TgF at Lev.
18.21; 19.16 and TgG at Lev. 23.22. But TgN does not use it in Exod. 6.8, perhaps
because that is the end of a divine promise and not an ordinance.
12.1-20 69
‫( בהכתי‬12.13) The Tgg and Sy again use qṭl to render Heb. ‫נכה‬, as in
v. 12. The absence of an object in the Heb. is dealt with either by repeating
‘all the firstborn’ from v. 12 (TgN and mg) or by ignoring the ‫ ב‬before ‫ארץ‬
‫( מצרים‬Vulg). TgNmg adds a further reference to divine self-revelation.
‫( לזכרון‬12.14) LXX has simply μνημόσυνον; the Three, followed by the
hexaplaric witnesses, prefixed εἰς to represent the ‫ ל‬of MT. TgN,G add ‫טב‬,
‘good’, to distinguish this from a ‫ זכרון‬which is a reminder of something evil,
as in Num. 5.15, 18: for other examples of this, also in TgJ, see AramB 2,
p. 48 n. 12. Already Jub. 49.15 refers to the Passover as ‘a reminder accept-
able to the Lord’.
‫( וחגתם אתו חג‬12.14) Vulg usually has sollemnitas for ‫חג‬, but occasion-
ally it uses sollemnis dies (1 Kgs 8.12; 12.32; Neh. 8.14) and its celebrabitis
eam sollemnem here (after hanc diem just before) seems to draw on that
phraseology.
‫( ליהוה‬12.14) Tgg as usual have ‘before Yahweh’. TgNmg adds ‘your God’,
a formula more characteristic of Deut. than Exodus (and especially its Priestly
sections); the Passover law in Deuteronomy 16 may perhaps have been in
mind.
‫( תחגהו‬12.14) Sy hwytwn ʿbdyn lh, using the perf. of hwy idiomatically to
express a command (Brockelmann §208) ‘to perform a durative or repeated
action’ (J. Joosten, pers. comm.). The use of the colourless ʿbd for ‫ חגג‬is
compensated for by the repetition of ʿdʿʾdʾ, ‘feast, festival’, before ldrykwn.
‫( תאכלו‬12.15) TgNmg records a third person pl. reading ‫יאכלון‬, but this is so
out of keeping with the rest of the context that it must be a mistake, perhaps
from vv. 7-8.
‫( אך‬12.15) The Vss divide between adversative (LXX, TgO,J) and restric-
tive (TgN,G) interpretations of the particle, with Sy (‘and’) and Vulg (no
equivalent) effectively ignoring it. In MRI (Lauterbach, p. 64 and n. 6) the
interpretation of ‫ אך‬forms the basis for the halakah added in TgJ (see the next
note).
‫( ביום הראשׁון‬12.15) Both LXX and Sy render loosely ‘from’ rather than
‘on’ (possibly influenced by the phrase later in the verse). TgJ specifies ‘from
midday on the day before the feast’, following the rulings in M.Pes. 1.1-4 and
MRI loc. cit.
‫( תשׁביתו‬12.15) LXX’s ἀφανιεῖτε can mean ‘remove’ or ‘destroy’ (the
later requirement was that the leaven should be burnt), but Aq and Symm
replaced it respectively with διαλείψατε and παύσατε, probably both liter-
alising with the sense ‘cause to cease’ (for διαλείπω in the sense ‘cease’ cf.
LSJ, s.v. II.2). Vulg’s non erit is very free.
‫( חמץ‬12.15) LXX ζύμην, as several times elsewhere, although it really
means ‘leaven’ rather than ‘leavened’. The other Vss (inc. Vulg fermentatum
here) preserve the distinction, and so does LXX (ζυμωτόν) in v. 19. Sy
repeats ‘from your houses’ here, where it makes little sense: as Propp saw
(p. 360), it will be an inner-Syriac corruption, presumably an early one.
70 EXODUS 1–18
‫( ונכרתה‬12.15) LXX and Tgg ‘shall be destroyed’ and Sy and Vulg ‘shall
perish’ lose the specific connotations of the Heb. (as in v. 19).
‫( הנפשׁ ההוא‬12.15) SP as usual reads the expected fem. form ‫ההיא‬: none of
the Qumran mss preserves the text at this point. TgO,J interpret correctly with
‫אנשׁא הוא‬.
‫( מיום הראשׁון‬12.15) TgJ pedantically adds ‘of the feast’.
‫( עד‬12.15) SP ‫ ועד‬adds the idiomatic but not indispensable waw. TgJ,G and
Sy also have a waw, whereas LXX, Vulg and TgO,N do not; but in most cases
the reading is just as likely to be due to the idiom of the target language as to
the Vorlage. Again, none of the Qumran mss preserves this word. It is more
likely that the waw would have been added to conform to idiom than removed,
so MT is probably original here.
‫( וביום הראשׁון‬12.16) A number of SP mss, including two early ones (F
and Tal’s), omit the waw (cf. Vulg), but the other Vss support MT. Given the
tendency of scribes to add waw, the text without it might be more original (cf.
Propp, p. 360). 4QExc has only ]‫ [הראשׁון‬for this verse (probably for its first
occurrence), so its reading cannot be determined. LXX, Sy and Vulg do not
render the ‫ ב‬and so make ‘the first day’ the subject of the clause (likewise in
the next clause): this will be connected with their interpretations of ‫מקרא־קדשׁ‬.
‫( מקרא־קדשׁ‬12.16) Tgg render ‫ מקרא‬by words for ‘festival’, with the
addition of ‫יום טוב‬, ‘feast day’ in TgN,G, as often elsewhere: so also MRI ad
loc. (cf. AramB 2, p. 48 n. 13). Vulg’s renderings, though free, have a similar
intention: sollemnis is frequently used with dies to represent ‫ חג‬and eadem
festivitate for the second occurrence presumably uses festivitas (as always
elsewhere in Vulg) of a religious festival rather than in its wider classical
sense. Sy qrʾ, ‘proclaimed’ (pass. part.), is probably dependent on LXX
κληθήσεται…κλητή and a clue to the true meaning of the latter word, which
Walters rightly saw as being ‘proclaimed’, with ‫ מקרא‬being (mis)understood
as a Pual part. (pp. 244-46; Lust et al., p. 258).23
‫( מלאכה‬12.16) LXX ἔργον λατρευτόν; λατρευτόν in this expression
probably has the ancient association of the word-group with slavery, so it
will mean ‘servile labour’. In Leviticus 23 and Numbers 28–29, in similar
contexts, the phrase stands for ‫מלאכת־עבדה‬, where the use of λατρευτόν
can be seen to be based on the association of ‫ עבדה‬with ‫ע ֶבד‬,ֶ ‘servant, slave’
23
κλητή is commonly but wrongly taken as a noun meaning ‘assembly’
(sc. ἐκκλησία: cf. LSJ, p. 960; Lee, p. 51; Muraoka, p. 320). But Wevers recog-
nised that this does not fit here, though he thought that it did in Lev. 23 (Notes,
p. 177), and BAlex (p. 148) cited Philo, De spec. leg. 2.157 in support of an
adjectival parsing ‘called’ here (its own appel, ‘call’, reflects an inability to shake
off the dominant nominal understanding and is scarcely an improvement). This is
supported by the fact that in Lev. 23 κλητή is three times associated with the verb
καλέω (vv. 2, 21, 37), as here, and in Num. 28.18, 26; 29.1, 7, 12 ‫ מקרא‬is rendered
by ἐπίκλητος, ‘called, designated’.
12.1-20 71
(an association that may well not have been originally intended in the Heb.:
cf. the widespread rendering ‘laborious work’). Here LXX is introducing a
specification from other passages: the language of MT (which is followed by
the other Vss and SP) is that of the Sabbath commandment of the Decalogue,
qualified by the specific exception of food preparation (see the next note but
one).
‫( לא יעשׂה‬12.16) Vulg nihil…facietis, TgG and some later mss of Sy render
by second p. active forms, assimilating to the language of Lev. 23.7 etc.
Many LXX mss, including LXXB, likewise read ποιήσετε and this should
be regarded as the original LXX reading, with Rahlfs: after the addition of a
word from Leviticus 23 it would be entirely natural to continue with its verbal
form.24
‫( יאכל‬12.16) LXX ποιηθήσεται probably deliberately eliminates the
specification in the Heb. (or at least MT: it is possible that LXX’s Vorlage
already had a revised text) to maintain the agreement with Leviticus 23’s more
general ruling: the result is a clause that has no real point. TgJ ‫דיתעביד למיכל‬
combines ‘doing’ with ‘eating’: cf. the fuller (and wider?) expression of TgN.
Vulg and Sy paraphrase, but are clearly based on MT.
‫( את־המצות‬12.17) The oddness of this phrase gave rise to a number of
variants. Vulg, TgN and Sy agree with MT. SP reads ‫את המצוה‬, ‘the command-
ment’, and LXX τὴν ἐντολὴν ταύτην is evidently based on the same text, to
which it has added ‘this’ to provide the specification which such instructions
normally have (cf. Deut. 11.22). TgG,Nmg provided specification by rendering
‘the commandment of unleavened bread’. The derivation of ‫ המצות‬here from
‫ המצוה‬was known in rabbinic Judaism: it is attributed to R. Josiah in MRI
(Lauterbach, p. 74), and recently it was adopted in NEB (Brockington, p. 10),
though not REB. It can scarcely be the original meaning, as Propp agrees
(p. 361). The reading ‫ המצוה‬of SP and LXX’s Vorlage, if original, could never
have given rise to the ‫ המצות‬of MT, because it is unambiguous, whereas ‫המצות‬
in an unvocalised text could potentially be read in two ways: presumably the
sing. ‫ המצוה‬arose from the specific law in the context (and LXX’s ταύτην took
this process further). A general reference to ‘the commandments’ would be out
of place in this context. A different kind of variation, also with attestation in
MRI (p. 73), is found in TgJ, which reads ‘(you shall watch) the dough of the
unleavened bread’.
‫( בעצם‬12.17) LXX has simply ἐν, as in v. 51: the other Vss use appro-
priate equivalents.
24
Wevers oddly follows the A reading ποιήσεται here (likewise in 30.32, 37:
cf. THGE, pp. 230-31, and Notes, p. 177 – his own wording has evidently suffered
textual corruption!), but the middle form is out of place and it (not ποιήσετε)
must be due to the common scribal confusion between -ε and -αι.
72 EXODUS 1–18
‫( הוצאתי‬12.17) Most witnesses support MT’s past tense, despite its

inappropriateness at this stage of the narrative: it was the latter which presum-
ably led LXX (ἐξάξω), Vulg (educam) and possibly TgNmg (‫ )אפק‬to use the
future. The reading of MT (and SP: no Qumran evidence survives) should be
preferred as the difficilior lectio.
‫( את־צבאותיכם‬12.17) LXX and Vulg render in the sing., but the pl. is
regular in Exodus (cf. 6.26; 7.4; 12.41, 51) and the sing. translation may
simply represent the unity of the ‘tribal divisions’. TgJ,N add ‘redeemed’ to
reinforce the theological interpretation of the event.
‫( ושׁמרתם‬12.17)2o The waw seems to be preserved at the end of a line in
4QpalExm, after an empty space, and DJD IX, pp. 59, 86, suggests that it is
a further example of one of the scribe’s methods of marking a division in the
text. No verse division appears in MT or SP at this point (in MT the word
does follow athnach), but ‘this day’ does fit v. 18 more closely than v. 17a and
makes a division at this point more appropriate. SP adds ‫ ועשׂיתם‬and LXX reads
καὶ ποιήσετε instead of repeating καὶ φυλάξεσθε from the beginning of the
verse. SP recalls a number of passages in Deuteronomy where ‫ שׁמר‬is immedi-
ately followed by ‫( עשׂה‬e.g. 4.6) with reference to the commandments. ‫ עשׂה‬can
be used of ‘celebrating’ a feast but is never, apparently, used of a day (‫)יום‬.
This may be why Wevers suggests that LXX means ‘you will make this day
an eternal ordinance…’ and that SP is a conflation of MT and (the Vorlage of?)
LXX (Notes, p. 178). But Wevers’s translation, while theoretically possible
(and NETS agrees), produces an idiom which is (apparently) unparalleled
elsewhere and, even if it is what LXX intended, it is more likely that ‫ועשׂיתם‬
was originally added because of the Deut. parallels (as in SP) and that LXX,
having the amplified text in its Vorlage, then ignored ‫ ושׁמרתם‬because it had
already been rendered earlier in the verse. For the closeness of LXX and SP
in this verse compare the note above on ‫את־המצות‬.
‫( חקת‬12.17) Five SP mss, including Camb. 1846 and two cited by Crown,
read ‫חקות‬: whether this is meant as a pl. or a variant pronunciation of the sing.
it belongs with some other exx. noted in GSH §146j, 165c. In v. 43 it can
only be a sing. The formula with ‫ עולם‬is always found in the sing. elsewhere
(except for Ezek. 46.14 where the text is doubtful), as in v. 14, and should be
so read here.
‫( בראשׁון‬12.18) TgO,J specify that it was Nisan, while TgN,G, Sy, Vulg add
‘month’; but the adjective alone is found several more times (e.g. Gen. 8.13;
Num. 9.5) and is doubtless original. The name Nisan is first attested in this
connection in the ‘Passover Papyrus’ from Elephantine, dated to 419 B.C.
(Cowley 21; TAD 1, pp. 54-55 [A4.1]: ‫ לניס[ן‬in line 7). LXX ἐναρχομένου
will mean ‘of the beginning (month)’ (cf. BAlex 2, p. 149), an unusual periph-
rasis for ‘first’ which is probably an echo of v. 2: it occurs elsewhere only
in Num. 9.5, another passage about Passover/Unleavened Bread, where it is
presumably dependent upon the rendering here. LXX also adds the normal
equivalent τοῦ πρώτου to ‘month’ later in the verse.
12.1-20 73
‫( בערב‬12.18) TgG has ‫ביני שׁימשׁתא‬, its rendering for ‫ בין הערבים‬in v. 6, both
times (so also TgN the first time, alongside ‫ברמשׁא‬, the regular Targumic equiv-
alent to ‘in the evening’), despite the fact that v. 6 refers to a earlier stage of
the ritual. TgJ emphasises the distinction by additions which locate the eating
on the evening of the 15th (with a different view of when the new day begins);
and then expands its rendering of the second ‫ בערב‬to ‘on the evening of the
22nd you may eat leavened bread’ (compare its additions in vv. 8 and 10).
‫( עד יום האחד ועשׂרים‬12.18) SP reads ‫אחד‬, without the article, despite the
fact that in such dates the article is common (also in Heb. inscriptions: see
Note k on the translation) and SP is more inclined to add the article than to
remove it (GSH §170h [cf. 166c]). Both these factors might suggest that ‫אחד‬
is the original reading, but SP may have removed the article to accord with its
apparent absence in ‫ בארבעה עשׂר יום‬earlier in the verse.
‫( שׂאר‬12.19) Sy renders with ḥmyʿʾ, ‘leavened bread’, although it had
recognised the distinction between ‫ שׂאר‬and ‫ חמץ‬in v. 15. Likewise in 13.7 it
merged the separate commands about ‫ שׂאר‬and ‫ חמץ‬into one, as though it saw
no significance in the difference. For similar indifference to details of ritual
elsewhere in Sy see Weitzman, Syriac Version, pp. 210, 217-19.
‫( ימצא‬12.19) TgG (cf. TgNmg) renders ‘shall be seen’, as in MT of 13.7
(and Deut. 16.4).
‫( כי‬12.19) ‫ כי‬has no equivalent in LXX and Vulg (as also in v. 15), so that
the two instructions remain independent of each other.
‫( מחמצת‬12.19) TgO,J introduce a new word for ‘leavened’, ‫מחמעא‬, here
(and in v. 20) to reflect the difference from ‫ חמץ‬earlier in the passage. LXX
ζυμωτόν marks the distinction from ‫( שׂאר‬ζύμη) that it ignored in v. 15.
‫( הנפשׁ ההוא‬12.19) SP as usual reads ‫ההיא‬, without the curious consonantal
text of MT. TgO,J render ad sensum ‫(בר) אנשׁא‬, while the other Vss keep to the
wording of MT and Vulg’s anima eius introduces a soul/body distinction that
distorts the meaning.
‫( מעדת ישׂראל‬12.19) TgG has ‘from the people (‫ )עם‬of the assembly of
Israel’, amplifying the phrase (as in v. 3) to agree with the fuller expression
in v. 6.
‫( בגר ובאזרח‬12.19) All the Vss except TgN, Aq and Symm understandably
use pl. forms to render the generic singulars of the Heb. (likewise TgNmg and
in the first case only[?] TgG). LXX γιώραις for ‫ גר‬is an exceptional use of a
loan-word from Aram. (cf. TgO,N,G here and often), found in LXX only here,
in Isa. 14.1 and in the mg. of a few mss at Lev. 19.34. Elsewhere LXX uses
πάροικος when Israelites are referred to (as at 2.22 and 18.3) and otherwise
generally (c. 70x) προσήλυτος (as in vv. 48-49 and 20.10). The latter, which
has long been supposed to be a Jewish coinage but now has a third-century
B.C. pagan attestation (for which see D.M. Moffit and C.J. Butera, ‘P.Duk.
inv.727r: New Evidence for the Meaning and Provenance of the Word
Προσήλυτος’, JBL 132 [2013], pp. 159-78; earlier J.A.L. Lee, ‘Equivocal
and Stereotyped Renderings in the LXX’, RB 87 [1980], pp. 104-17 [112
74 EXODUS 1–18
n. 27]), was used here by Aq and Symm. LXX evidently quickly decided that a
word intelligible to Greek-speakers and increasingly used in a technical sense
for Gentile converts to Judaism (so already in Philo acc. BAG, p. 722) was
preferable to other options where foreigners were meant. But here γιώρας,
like the words chosen in the other Vss (Vulg advena, TgJ ‫ דיורא‬and Sy ʿmwrʾ),
probably maintains the purely social connotations of the Heb. Sy inverts its
equivalents for the two words here (agreeing with v. 49), to give priority to
the ‘normal’ case.
‫( כל־מחמצת‬12.20) A few SP mss prefix waw, including two from the early
thirteenth cent., but the rest of the tradition exhibits the asyndeton of MT.
For ‫ מחמצת‬TgN has ‫( אחמע‬after ‫ חמיע‬in v. 19), an unattested form: the mg has
corrected to ‫מחמע‬, (close to) the reading of TgO,J. TgJ has ‫כל עיבובין דמחמע‬,
‘any mixtures of what is leavened’, an expansion which alludes to an issue
discussed in MRI (Lauterbach, p. 79): M.Pes. 3.1 gives a list of forbidden
‘mixtures’.
‫( בכל־מושׁבתיכם‬12.20) LXX ἐν παντὶ κατοικητηρίῳ again experiments
with a rendering which it will hardly ever use for ‫ מושׁב‬again (only twice in
Ps. 106[107]): compare the note on ‫ בגר‬in v. 19. Other derivatives of the same
stem (esp. κατοίκησις and κατοικία) or a periphrasis (as in 10.23) are much
more frequent equivalents. TgJ inserts ‫אתר‬, ‘place of’, which might also allude
to a discussion in MRI (Lauterbach, p. 80), and Sy inserts byt as it did in 10.23:
TgN,G have both, which Klein (2, p. 59) understandably regards as excessive.
‫( תאכלו מצות‬12.20) In 4QpalExm a very little of these two words survives
at the end of a line. Space considerations suggest that there was an interval
of about 10 spaces at the beginning of the next line to mark the division
found also in MT and SP. This would be the only possible case of this kind
of division in what survives of 4QpalExm and several other Exodus mss from
Qumran have no examples of it at all (of course occasions for its use would
be limited). There is one certain case (before 26.15) and several reconstructed
cases in 4QNumb (DJD XII, pp. 208-10) and as many as 16 clear cases in
1QIsa. For further examples see Tov, Scribal Practices, p. 146. All these
examples are in mss using the square script, so Tov (ibid.) doubts whether this
method of division was used in palaeo-Hebrew mss. The alternative way of
dealing with a break after a full line of text was to leave a completely empty
line, and Tov gives two examples of this in 4QpalGen-Exl, and there are more
(cf. DJD IX, pp. 19-20). But there are none in the more extensive remains of
4QpalExm, and after 12.20 there was certainly not an empty line. So either
no division was marked at this point or the indentation method was used.
The second seems intrinsically more likely, and if there was no indentation a
longer reading than MT would be needed in v. 21 to fill the line.
C h ap t er 1 2 . 2 1 - 2 7
M oses I n s tr u cts th e I s r a e li t e s
a bou t th e P as s ove r
The section comprises (except for a short narrative about the

people’s response in v. 27b) a speech of Moses in which he gives
instructions to ‘the elders of Israel’ about certain aspects of the
Passover ritual. The boundary between it and the preceding speech
of God to Moses is unambiguous and was represented by divisions
in MT and SP and probably in 4QpalExm. At the end MT and SP
place the division after v. 28. This is readily understandable, since
v. 28 is more closely related in content to vv. 1-27 than it is to
v. 29ff. But its connections are with vv. 1-20 rather than vv. 21-27
(see further the notes on the next section), and once these two
sections are separated it is less natural to associate v. 28 with vv. 21-
27. Moreover, as a verse of narrative v. 28 does resemble what
follows it more than what precedes it from a formal point of view.
Internally the unit can be divided readily into six sub-sections
(cf. Coats, pp. 82-83): (i) introduction to the speech (v. 21a); (ii)
instructions to the elders (vv. 21b-22); (iii) reason/consequence
(v. 23); (iv) future observance of Passover (vv. 24-25); (v) instruc-
tion of the children (vv. 26-27a); (vi) conclusion: response of the
people (v. 27b). The progression from one sub-section to the next is
entirely logical once it is recognised that the narrative is written for
the benefit of later generations of Israelites. Within each sub-section
there are no signs of literary disunity except in v. 24, which overlaps
with v. 25 and contains a surprising shift to second person singular
forms in its second half (see the Explanatory Note on vv. 24-25).
At first sight these verses look like a straightforward account of
Moses’ transmission of Yahweh’s instructions to the people (vv. 1-
20) in an abbreviated form, and that is no doubt how the canoni-
cal text is meant to be understood. But the contents of vv. 21-27
make it most unlikely that this is what their author intended. There
are many differences between the two passages, such as the group
to which Moses speaks, the sub-division of the people into ‘clans’
rather than households, the addition of instructions about the bunch
76 EXODUS 1–18
of marjoram and not leaving the house before morning, variations

in vocabulary and, perhaps above all, the introduction in v. 23 of a
personal ‘Destroyer’ who must be kept from the Israelites’ houses.
The instructions about the teaching of children also greatly amplify
what is said earlier about the Passover being a ‘memorial’ (v. 14)
(see in more detail the Explanatory Notes on the whole passage). In
addition much of what is prescribed in vv. 1-20 is omitted here. Even
if the total omission of any reference to the Feast of Unleavened
Bread is disregarded, on the grounds that the people are instructed
about this later, in 13.3-10 (though again in very different words
from 12.15-20), it is remarkable that none of the instructions about
the criteria for selecting animals, the timing of different stages in the
ritual, and the way in which the animal is to be cooked and eaten, are
repeated here. It is scarcely conceivable that vv. 1-20 and vv. 21-27
were originally intended to go together (cf. Schmidt, pp. 481-83).
This has been generally recognised almost from the beginning
of modern critical scholarship on Exodus and is rarely disputed
today except by conservatives (Van Seters [see below] is a rather
surprising bedfellow for them).
The dominant critical view has been that vv. 21-23 and v. 27b come from J
and that vv. 24-27a are a Deuteronomistic addition (with some recognition of
glosses here and there). But Knobel, followed closely by Nöldeke in 1869, had
maintained the traditional view that vv. 21-23 were the continuation of the (for
him early) Priestly instructions in vv. 1-20, while then attributing vv. 24-27
to the Jehovist.1 This view of vv. 21-23 was rejected by Wellhausen, who saw
that they were closely connected to vv. 24-27 and exhibited many differences
from P (Composition, pp. 72-75). But he found it impossible to attribute the
passage to J, because it would interrupt the narrative link between 11.8 and
12.29 and also the emphasis on the sparing of Israel was quite different from
the JE narrative. He considered the possibility that vv. 21-27 might be a
subsequent addition to P (cf. Kuenen’s view in his Hexateuch, p. 331 [cf. 168
n. 4], that the section was from the Endredaktor), to which the content of the
passage is close, but favoured the view that (like 13.3-16) it was either from
RJE or a later supplement to JE (p. 75).2
1
So Exod.-Lev., pp. 91, 105. In Num.-Jos. his summary of the sources of the
Hexateuch assigns vv. 24-27 to his Rechtsbuch (= E; p. 532).
2
In the Prolegomena the twice-repeated statement that there are no refer-
ences to Passover prior to Deuteronomy (ET, pp. 86-87; cf. 4th German ed.,
pp. 84, 86) must presume that Exod. 12.21-27 is later than Deuteronomy. The
context is Wellhausen’s well-known argument that it was only in Deuteronomy
that the originally agricultural festivals began to acquire a historical element.
12.21-27 77
It was first Dillmann and then especially Karl Budde who laid the founda-
tion for the majority view. Dillmann, having criticised Knobel’s view for
ignoring the omissions in vv. 21-23 and the non-Priestly terminology there,
argued that 13.3-16 must be J because 12.34, 39 needed to be complemented
by a law for the continued observance of the Feast of Unleavened Bread.
Since 12.21-27 presented the same pattern of an ancient action which was
turned into a regular practice, these verses too should (all) be assigned to
J. Eleven years later (with more emphasis on their own style and character)
Budde also attributed vv. 21-23 and v. 27b to J, but for linguistic reasons he
regarded the intervening verses as a Deuteronomic insertion (‘Die Gesetzge-
bung der mittleren Bücher des Pentateuchs, insbesondere der Quellen J und
E’, ZAW 11 [1891], pp. 193-234 [199-200]). He was followed by B.W. Bacon,
Holzinger (with some modifications) and Baentsch, with the last-named
appealing, in an almost Gunkelian way, to the combination of fixed older
traditions as the reason for the apparent ‘intrusiveness’ of the passage in the
J narrative (p. 100).3
Shortly afterwards a different way of reconciling such tensions within the
non-Priestly material with an early origin for some of its Passover legislation
was to be proposed, based not on oral tradition but on a more thoroughgoing
use of source analysis. In 1912 Rudolf Smend assigned 12.21-23, 27b to
the earlier of the two Yahwist sources which he distinguished (J1), while the
surrounding older narrative was attributed to the later one (J2). In other words,
for Smend the tensions that had concerned Wellhausen, and have continued
to require some defensive explanation by those who have seen 12.21-27 (or
some of it) as part of a single J narrative, became one of the arguments for the
idea of a ‘fourth narrative source’ in the Pentateuch, which was taken up in
turn by Eissfeldt, Beer and Fohrer.4 For vv. 24-27a they, like more mainstream
critics, envisaged a Deuteronomic author. With the (partial) exception of this
distinguished but uninfluential minority, scholars throughout the twentieth
century (and beyond: cf. Schmidt, pp. 517-21) were in general well satisfied
with the analysis of the passage proposed by Budde.5
A few scholars have, like Dillmann, not held to a Deuteronomic origin
for vv. 24-27a. One group places it earlier than Deuteronomy (Lohfink, [Das
3
A little curiously Gressmann, who might have been expected to favour such
an approach, found the inconsistency too great and reverted to Wellhausen’s view
(Anfänge, pp. 41-42; cf. Mose, pp. 97-98 n. 1).
4
See further the introduction to 12.28-42, 50-51. To anticipate, Eissfeldt
attributed to L (his name for J1) vv. 21-23, 27b, 33-39, while 11.4-8 and 12.29-30,
32 were from J and v. 31 from E.
5
This widespread agreement is at first ignored by Gertz (Tradition, pp. 38-39:
‘nun seit langem wie kaum eine andere Textanalyse innerhalb des Pentateuch
umstritten’: but later recognised, p. 45 n. 75 [die Mehrheitsmeinung’]), despite the
fact that his own (partly tentative) conclusion is quite close to it. For the (impor-
tant) differences in his view see below.
78 EXODUS 1–18
Hauptgebot, pp. 121-22]; Clements, Blenkinsopp, Propp). Some others have

argued that these verses are from P, along with vv. 21-23 or most of them (first
apparently H.G. May, ‘The Relation of the Passover to the Festival of Unleav-
ened Cakes’, JBL 55 [1936], pp. 65-82; then B.N. Wambacq [‘Les origines de
la Pésaḥ israélite’, Bib 57 (1976), pp. 206-24, 301-26 (316-19)], Norin, Van
Seters [‘The Place of the Yahwist in the History of Passover and Massot’, ZAW
95 (1983), pp. 167-82; Life, pp. 114-19], Coats [pp. 82-83], Gesundheit [Three
Times, pp. 58-75; earlier (as S. Bar-On) ‘Zur literarkritischen Analyse von Ex
12, 21-27’, ZAW 107 (1995), pp. 18-30, with vv. 22-27a being the original P
continuation of v. 11] and apparently J.S. Baden [cf. Composition, pp. 76, 173
with n. 18]).6 This view has been strongly criticised by Blum, Studien, p. 39 n.
149, and Gertz, Tradition, p. 45 n. 74, but a Priestly or post-Priestly origin just
for v. 24 or part of it has, understandably, been more popular (so Carpenter/
Harford-Battersby [as the original conclusion of vv. 1-20], McNeile, Noth [for
v. 24b], Blum, [ibid.], Propp: cf. Baentsch, R. Schmitt [Exodus und Passa,
p. 21], Gertz [Tradition, p. 40]). Levin judges the whole of vv. 21-27 to be
very late (‘wahrscheinlich…nachendredaktionell [RS]’, p. 336), since they
interrupt the natural sequence in both J and P, but Gertz retains the distinc-
tion between vv. 21-23 and vv. 24-27a, regarding only the latter as very late
(because of its mix of Deuteronomic and Priestly language) and eventually
concluding that vv. 21-23 are most probably a pre-Priestly supplement to the
non-Priestly Exodus narrative (Tradition, pp. 39-50).
Within vv. 24-27a it does seem necessary, as some others have

seen, to distinguish between v. 24 and vv. 25-27a, for the reasons
given below in the Explanatory Note on vv. 24-25, and to attribute
v. 24 not to P itself but to a redactor who sought to bring vv. 25-
27a within the Priestly view that the continuing celebration of
the Passover could be observed anywhere, even outside the land
of Canaan. This is very likely the redactor who combined P with
the non-Priestly Exodus narrative, who also placed v. 28 where it
is now to make vv. 21-27 appear as a summary of vv. 1-20. The
reasons given by Gertz for seeing the influence of Priestly style in
vv. 25-27a are, however, inconclusive, because they concern mainly
common expressions which are not distinctively Priestly. The idea
of the gift of the land is found already in Gen. 12.7; 13.15, 17. The
proposal to view all or most of vv. 21-27 as Priestly is founded
6
Gesundheit, Three Times, pp. 66-67, for no good reason separates vv. 21 and
27b from the rest of the unit as a ‘redactional frame’, conveniently sidelining some
of the passage’s most clearly non-Priestly elements.
12.21-27 79
partly on the difficulties of seeing it as part of J, to which we must

return, and partly (this is ‘the principal argument’ according to Van
Seters, Life, p. 116) on the occurrence in vv. 21-24 of vocabulary
that is typical of P. Blum (see above) has reasonably objected that
the technical language in question cannot be presumed to be the
sole property of one biblical writer or school, and it should also
be recalled that in several ways, including the use of some of the
terminology in question, vv. 21-27 are so different from vv. 1-20
that they cannot have been designed to follow that passage (see the
Explanatory Notes). It is in fact not characteristic of P to follow a
divine instruction to Moses (and Aaron) with an account of how
this instruction was handed on to the people. The transition from
v. 20 to v. 21 (and from v. 27 to v. 28) must be due to the use by
the redactor of material from two different versions of the Exodus
story.
To return to vv. 25-27a, there are undeniably some affinities in
their form and language to Deuteronomy. The formula ‘When you
come into the land which Yahweh will give you’ (v. 25) recalls both
the theology of the land and the association of legal requirements
with entry into the land that are characteristic of Deuteronomy
and brought together in passages like Deut. 17.14; 18.9; 26.1;
and 27.3. The Kinderfrage motif in vv. 26-27 also has a parallel
in Deut. 6.20-25. Some connection between Deuteronomy and this
passage is highly likely, but the common view that this passage is
therefore influenced by Deuteronomy is not necessarily correct. As
noted already, the theology of the land as Yahweh’s gift is already
found in non-Priestly passages in Genesis, to which the words ‘as
he has promised’ (v. 25) might be referring. Secondly, this passage
and Deut. 6.20-25 are only two out of six passages in which such
Kinderfragen appear.
Excursus on the Kinderfrage Passages

A series of passages in the Hexateuch display a similar pattern to Exod. 12.26-
27a and Deut. 6.20-25, in which children’s questions, stated or assumed,
provide the impulse for instruction about the early traditions of Israel. In
Exod. 13.14-15 the killing of firstborn animals is explained as a commemora-
tion of the slaying of the Egyptian firstborn at the time of the Exodus. Verse
8 provides for a similar explanation of the Feast of Unleavened Bread to be
given to a child (without an explicit question, but one is almost implied).
80 EXODUS 1–18
Another pair of such formulae occurs in Joshua 4, for the explanation of

twelve stones, placed either in the midst of the river Jordan (vv. 6-7; cf. v. 9) or
at Gilgal near Jericho on the west side of the river (vv. 21-22[24]; cf. vv. 2-3,
8) to commemorate Israel’s miraculous crossing in the time of Joshua.
Such questions were seen by Gunkel as universal and as the basis for what
he called ‘aetiological legends (Sagen)’: ‘Es gibt eine Fülle von Fragen, die
ein antikes Volk beschäftigen. Das Kind schaut mit grossen Augen in die
Welt und fragt: Warum?’ (Genesis, p. xx; cf. Legends, p. 25). Not surpris-
ingly Gunkel observed that such questions were cited in the Old Testament
specifically in relation to aspects of the cult (ibid., p. xxiii, citing Exod. 12.26;
13.14; Josh. 4.6; cf. Legends, pp. 30-31, also L. Köhler, Der Hebräische
Mensch: eine Skizze [Tübingen, 1953], pp. 66-68, ET pp. 79-80). The two
Joshua passages (mainly in fact the second) were brought into H.-J. Kraus’s
argument that Joshua 3–5 were closely modelled on the liturgy of an early
festival at the Gilgal sanctuary (‘Gilgal. Ein Beitrag zur Kultusgeschichte
Israels’, VT 1 [1951], pp. 181-99 [esp. 186, 196]; cf. Gottesdienst in Israel.
Grundriss einer Geschichte des altisraelitischen Gottesdienstes [Munich,
2nd ed., 1962], pp. 183-86, ET pp. 157-58, and J.R. Porter, ‘The Background
of Joshua 3–5’, SEÅ 36 [1971], pp. 5-23 [14]). But it was J.A. Soggin who
noted that the Exodus passages too (and Deut. 6.20-25, on the rather dubi-
ous grounds that it cited an ancient ‘cultic credo’) were associated with
cultic observances and had a formal character which made it possible to see
the Kinderfragen as a whole as fragments of an old catechetical system for
teaching about the great traditions (‘Kultätiologische Sagen und Katechese
im Hexateuch’, VT 10 [1960], pp. 341-47, spec. 345-46). Although Soggin
accepted the Deuteronomistic authorship of all the passages, he adopted the
view that was widespread at the time, that the Deuteronomic tradition had
ancient roots, and saw the catechetical form as going back to the Shechem
and Gilgal sanctuaries in the premonarchic period. In his later commentary
on Joshua (ET Joshua: A Commentary [London, 1972]) Soggin was able to
say: ‘For a very long time, the sight of the stones in the river, in the middle
of the ford, and of the twelve pillars in the sanctuary, was the occasion of a
liturgical dialogue between the community taking part in the celebration and
the priests…’ (p. 67).
Soggin’s article was mentioned, but not relied upon, in an Excursus in
Lohfink’s Das Hauptgebot, which challenged the common view that Exod.
12.24-27a and 13.3-16 were modelled on Deut. 6.20-25 by making a close
examination of the vocabulary and style of the Exodus passages. Lohfink
concluded that Exod. 12.24-27a was ‘not Deuteronomic’ (pp. 121-22), while
13.3-16 was ‘proto-Deuteronomic’ (a term used here for the first time), repre-
senting a stage in the development of Deuteronomic style in which the classic
formulations found in Deuteronomy were not yet present (pp. 122-24): in both
cases the similarities to Deuteronomy were outweighed by the differences.
Lohfink was not so sure that a catechetical practice lay behind the biblical
12.21-27 81
passages – it was possible ‘dass die Einführung der Kinderfrage in unseren

Texten nur ein rhetorisches Mittel ist’ (p. 116) – and preferred to see their
origin in older ‘preaching traditions’ (p. 118).
M. Weinfeld has seen very clearly that a distinction is to be made between
the parenetical use of the Kinderfrage in Deut. 6.20-25 and the real ritual
situations envisaged in the other passages: in Deuteronomy 6 ‘the author
transformed a collocution attached to a ritual ceremony into a liturgical
oration irrelevant to the ceremony’; ‘the son–father collocution’ was ‘divorced
from its actual setting and converted into an instruction’ (Deuteronomy,
pp. 34-35). In his more recent commentary on Deuteronomy 1–11 (AB 5; New
York, 1991) Weinfeld has argued for a specific dependence of Deut. 6.20-25
on Exod. 13.1-16 (pp. 328-30, following M. Caloz, ‘Exode, XIII, 3-16 et son
rapport au Deutéronome’, RB 75 [1968], pp. 5-62), while ‘the Gilgal tradition’
in Joshua 4 lies behind Exodus 12–13 as well as Deuteronomy 6: ‘Gilgal was
the birthplace of religious education and dramatisation of salvation-history
in ancient Israel… [T]he fourth catechism in Deuteronomy is freed from any
connection to a ritual… Education in Deuteronomy does not depend on ritual
ceremonial media…, but is formal and abstract’ (ibid.).
A similar distinction was drawn by J. Loza (‘Les catachèses étiologiques
dans l’Ancien Testament’, RB 78 [1971], pp. 481-500), who analysed the
common structure of all the passages in detail. Loza made a further distinc-
tion between the Kinderfragen in Joshua 4, which do have a cultic origin, and
the rest, which probably do not. But Deut. 6.20-25 is a separate case from all
the others, since it is dependent on the terminology of ancient Near Eastern
treaties as well as the catechetical tradition (p. 500).7 In Exod. 12.24-27a Loza
sees contacts with P in vv. 24-25, so that the old, pre-Deuteronomic, nucleus
is restricted to vv. 26-27 (pp. 486-87).
The most recent review of all the passages is provided by Thomas
Dozeman in his God at War (pp. 48-60), which should probably be seen as
the key to understanding the corresponding section of his commentary, where
it is not always clear which view he favours in the complex debate which
he describes (pp. 248-52, 271-79). Dozeman’s earlier discussion comes in a
chapter on the Deuteronomistic edition of Exodus: it opens with a reference to
‘Deuteronomistic tradents’ and Exod. 12.24-27 ‘look to be Deuteronomistic’
(p. 49). Against Lohfink he believes that we must recognise a ‘flexibility’ in
the language of the Deuteronomists, so that minor variations in vocabulary
do not undermine the attribution of sections to them. He places more weight
on ‘literary arguments’ such as the parenetic style, the role of the elders, the
instruction of children, and the relationship to the older narrative of Exodus
7
The parallels cited are in the Vassal-Treaties of Esarhaddon §25 (ANET,
p. 537: already noted by Lohfink, p. 115 n. 7) and in the Sefire treaty 1C, 1-9
(ANET, p. 660): but in neither case is there any question and answer pattern.
82 EXODUS 1–18
(pp. 49-50). The Passover ritual, like the despoiling of the Egyptians, lays
weight on ‘active participation by Israel’ in contrast to the purely theocen-
tric older narrative (cf. pp. 46-47). It is recognised that ‘older cultic roots’
probably underlie the Kinderfrage (p. 49), but Dozeman’s emphasis falls on
what he sees as the Deuteronomistic use of this catechetical tradition, which
he traces from Exodus through Deuteronomy to Joshua (pp. 54-59). But since
‘Joshua 4 is the oldest version of the catechism’ (p. 60), the tradition has been
extended backwards from the Deuteronomistic History into the (later) Deuter-
onomistic revision of the Tetrateuch and Deuteronomy itself, to define a
theology of the Exodus which is closely tied in to the whole salvation-history.
This provides what is perhaps the most appealing synthesis of the
consequences of a thoroughgoing Deuteronomistic interpretation of the
Kinderfrage passages. There is no doubt that, both in Deuteronomy itself
and in the Deuteronomistic History, a prime concern is to uphold the central
theological importance of the Exodus, alongside and in conjunction with the
covenant at Horeb and the Deuteronomic law. This builds in a particular way
on the prominence of the Exodus tradition in earlier times, particularly in
the northern kingdom. But more needs to be said, and can be said, about a
pre-Deuteronomic use of the Kinderfrage pattern in this connection. This is
certainly evident in Joshua 4, where the clearly Deuteronomistic rewriting of
the chapter includes the second Kinderfrage section in vv. 20-24 (which is the
one that includes the Exodus) but not, it would seem, the first such section
in vv. 6-7, which will be older. It and the Exodus texts represent the stage
at which such instruction was still closely linked to cultic situations, which
preceded its adaptation to a specific law-centred application in Deuteronomy
6, as Lohfink and Weinfeld saw.8
There is thus an important difference between Deut. 6.20-25 and

the other five passages. Deuteronomy 6.20-25 uses the Kinderfrage
motif as one of many ways in which the Deuteronomists seek to
persuade their hearers/readers of the need to obey the laws contained
in the book. All the other passages relate to objects or observances
which were designed (or at least so understood) to remind future
generations of Yahweh’s great saving actions in the past. In Exod.
13.8 and 13.14-15 (where very similar issues arise to here) the
instruction concerns the Feast of Unleavened Bread and the killing
of firstborn males. In Josh. 4.6-7 and 4.21-24 it is the significance of
twelve stones set up close to the river Jordan, near Gilgal, that is at
8
Dozeman refers to Weinfeld in some other connections, but does not pick
up his very important observations about the changed use of the Kinderfrage in
Deut. 6.
12.21-27 83
issue. The success of such reminders can be seen in a passage like

Ps. 80.9-12, where the Exodus and conquest traditions are drawn
upon as support for a prayer for help which most likely dates from
the late eighth century B.C. Contrary to the general view today that
all the Kinderfrage passages are Deuteronomistic in origin, a strong
case can be made for seeing Deut. 6.20-25 as a typical adaptation of
an originally cultic motif to promote the acceptance and observance
of the whole collection of laws in Deuteronomy. Exodus 12.25-27a
at least could well belong to the non-Priestly narrative of the Exodus
itself (the case for saying this also of 13.3-16 will be considered in
the introduction to 13.1-16), along with vv. 21-23 and 27b, which
have some strong linguistic affinities especially with ch. 4 (see the
Explanatory Notes on vv. 21-22 and 26-27).
In the introduction to 4.18-31 we attributed the related verses to
J (vv. 24-26 and 29-31). We have observed above that a number of
tensions have been noted between 12.21-27 and 11.1-8 + 12.29ff.,
which have led to the view that 12.21-27 are not part of the main
narrative of the plagues but come either from a redactor or from a
‘third early narrative source’. Specific differences include the need
for a mark to show Yahweh which houses belong to the Israelites,
the use of an intermediary (‘the Destroyer’) and the timing of the
Israelites’ departure. Although explanations can be given for these
variations, they sound a little like special pleading (so most recently
Gertz, pp. 47-48, who lays particular weight on the absence of any
reference back to the Passover and the sparing of the Israelites in
vv. 29-30). Our attribution of the main plague-narrative to E (see
most recently the introduction to 11.1-10) removes the difficulty at
a stroke, without any need to assume an additional early narrative
source. 12.21-27 (apart from v. 24) can readily be attributed to J,
since the narrative into which it appears to intrude is not from J but
from E, and the affinities of the passage to the J narrative earlier in
Exodus do not need to be ignored or explained away.
Closer to hand, 12.21-23, 25-27 will belong with other frag-
ments of J’s plague narrative which we have detected in 7.15-17,
20*; 9.22-23a, 35; 10.12-15*, 20 and 10.21-23, 27. The feature
that is common to these passages (and distinctive from the main
body of the plague-story) is that Moses must stretch out his staff
(or his hand) to bring on the plague in question. This reflects the
same requirement for human action to initiate (subject to a prior
divine command) Yahweh’s intervention, whether in judgement or
84 EXODUS 1–18
in deliverance, as Dozeman has observed to be characteristic of the

Passover ritual. In this respect the surviving fragments of J hang
together theologically. Further evidence of such coherence can be
seen in the involvement of an angelic figure (as in 3.2) and in the
apotropaic use of a blood ritual (as in 4.25-26).
According to the traditional dating of the Pentateuchal sources,
which this commentary broadly shares (see the general Introduc-
tion), the assignment of these verses to a non-Priestly narrative
source also implies that they are older than the Priestly instructions
in vv. 1-20 and come from pre-exilic times. This is, for example,
the best explanation for the different uses of the same Heb. word
(mašḥît) in vv. 13 (‘destruction’) and 23 (‘the destroyer’): see further
the Explanatory Notes. The question then arises about the basis for
the close similarity of wording between parts of vv. 1-20 and 21-27
respectively, which will also be treated in the Explanatory Notes.
Recent scholarship has tended to explain such similarities by direct
borrowing from the older account by the authors of the later one (see
e.g. n. 18 on 12.1-20). This may be correct, but the alternative possi-
bility that knowledge of the same words was widespread in popular
or priestly (oral) tradition is especially likely in the case of the ritual
for an annual ceremony such as Passover, to which both vv. 14a and
vv. 24-27a refer (cf. Schmidt, p. 483). The larger questions about
the origins and history of the celebration of Passover in ancient
times cannot be handled fully here, not least because no one biblical
passage provides a basis on its own for discussing them: the present
passage, for example, deals only with certain aspects of the ritual
and the later instructions in vv. 1-20 very probably preserve ancient
features of it too. For orientation in the wider debates reference may
be made to such works as de Vaux, Institutions; Albertz, Religions-
geschichte; Schmitt, Exodus und Passa; and the overviews in ABD
6, pp. 755-65; Propp, pp. 427-52; and Schmidt, pp. 483-92 (cf.
469-72). Very probably Passover began, as both accounts in Exodus
12 suggest, as a domestic celebration and was already practised
before its association with the Exodus as a spring rite for the protec-
tion of families and their livestock. At some point (no later than the
reign of Josiah) it was amalgamated with the agricultural festival
of Unleavened Bread and became part of a national festival at the
temple in Jerusalem (Deut. 16.1-8; 2 Kgs 23.21-23; Ezek. 45.21-24;
M.Pes. 5).
12.21-27 85
Whereas the other component of the non-Priestly narrative moved

straight on to the fulfilment of Yahweh’s declaration in 11.4-6 that
there would be one terrible, final plague to force Pharaoh’s hand,
this one delays that intervention to incorporate a passage which
draws attention to the celebration of the regular Passover festival
as having its origin (at least in the way that it was now observed)
at a crucial point in the Exodus story. Just as Moses himself had
earlier been preserved from a divine attack by his wife’s applica-
tion of the blood of his son’s circumcision, so the Israelites’ houses
had been marked and preserved from the onset of the angelic
‘Destroyer’ who carried out Yahweh’s vengeance on the Egyptians
(as he had perhaps, in earlier versions of this ritual, been warded
off from attacks on the newborn animals of the flock) by the blood
of a slaughtered animal. And so for ever hereafter, and especially
in the land which was Yahweh’s gift, this ritual was to be repeated
and explained to succeeding generations as a memorial of that first
great act of Yahweh’s deliverance of his people. Like them, future
generations could be expected to ‘bow down’ and ‘worship’ at the
recollection of it, and no doubt in hope too of Yahweh’s future care
for his people.
21 [Moses called all the elders of Israel and said to them: ‘Go
(along?)a and take animals from the flockb for your clansc and
slaughter the Passoverd. 22 Take a bunche of marjoramf and dip
it in the blood which is on the threshold/in the bowlg and apply
some of the bloodh which is on the threshold/in the bowlg to the
lintel and the two doorposts. Not one of you shall go out of the
door of his house until morning. 23 Yahweh will pass through
to smitei the Egyptians/Egypt, but whenj he sees the blood on
the lintel and on the two doorposts he will “pass over”k the door
and he will not allowl the destroyerm to come into your houses
to smite.] 24 [You shall maintain this practicen as a statute for
you and your descendants for ever.] 25 [When you come into
the land which Yahweh will give you, as he has promised, you
shall maintaino this service/worshipp. 26 When your children say
to you, “What is this servicep to you?”q, 27 you shall say, “Itr is a
Passover sacrifice for Yahweh, whos ‘passed over’ the houses of
the Israelites when he smotei the Egyptians/Egypt and kept our
houses safe”.’ The people bowed down and worshippedt.]
86 EXODUS 1–18

a. Heb. ‫משׁכו‬. The normal meaning of ‫משׁך‬, ‘drag, pull’, would fit with an
animal as obj. here (so DCH 5, p. 523) but is inappropriate for an action prior
to the ‘taking’ of the animal (‫)וקחו‬. Hence ‫ משׁכו‬is generally taken as a case of
the rare intransitive use of the verb for ‘go (along), proceed’ (compare then the
frequent use of ‫ הלך‬before ‫לקח‬, e.g. in 5.11), as in Judg. 4.6; 5.14; 20.37; Job
21.33; Sir. 14.19 (cf. BDB, p. 604; HAL, p. 610). But the form in Sir. 14.19
may well be passive (DCH 5, p. 525; Ges18, p. 753) and the three occurrences
in Judges can all be explained differently (Ges18, ibid.), so that only Job 21.33
provides a strong parallel in BH for a meaning parallel to ‫הלך‬, if indeed its
own interpretation is clear enough to do so. There are possible examples of it
in post-biblical Heb. and JAram., but they may in fact be passives. Perhaps,
therefore, ‫ משׁך‬here has its common meaning ‘pull’ or, as F. Perles suggested
(Review of Gesenius-Buhl, Handwörterbuch [14th ed.], JQR 18 [1906],
pp. 383-90 [385]), the sense ‘acquire’ which it has in MH (Jastrow, p. 853: cf.
MRI, Lauterbach, p. 83) and ‫ וקחו‬is a gloss to explain the unusual word or an
assimilation to v. 3.9 See further Text and Versions.
b. Heb. ‫ צאן‬is the collective noun corresponding to ‫ שׂה‬in vv. 3-5 and shares
its ability to refer to goats as well as sheep and to both together (Gen. 30.31-
32: cf. BDB, p. 838).
c. Heb. ‫ משׁפחה‬is normally the next kinship group below a tribe (so in the
P genealogy in 6.14-25), though it appears sometimes to refer to both a wider
and a narrower group (cf. TWAT 5, 87-89 = TDOT 9, pp. 80-82). ‘Each clan
was ruled by the heads of its families, the zeqēnîm or “elders” ’ (de Vaux, Insti-
tutions 1, p. 22, ET p. 8): so here it is envisaged that the selection of animals
was organised at the level of clans rather than individual families as in vv. 3-5.
d. Here as in v. 11 (and v. 27) ‫ פסח‬refers to the slaughtered animal, not the
festival as a whole: see Note y on the translation of vv. 1-20 and the Explana-
tory Note on v. 11. Schmidt suggests (p. 477) that the absence of ‫ את‬might be
due to older usage.
e. Heb. ‫אגֻ ָדּה‬,ֲ a rare word which only has this meaning here in BH. It is
more frequent in post-biblical Heb., where the related verb also occurs (e.g.
5Q13 f2.7: see also Jastrow, pp. 10-11).
f. Heb. ‫אזוב‬. The familiar Eng. equivalent ‘hyssop’ is taken from LXX and
Vulg, but it is now generally agreed that ‫ אזוב‬cannot mean Hyssopus officinalis
L., as it is not native to Palestine, and probably refers to ‘Syrian marjoram’
(Origanum maru L.: so already BDB, p. 1119, and L. Baldensperger and G.M.
Crowfoot, ‘Hyssop’, PEFQS 63 [1931], pp. 89-98: see further Houtman 1,
p. 164; Propp, p. 407; Dozeman, p. 273; NEB, REB; and the recent lexica
s.v.): on its characteristics see ABD 2, p. 812.
9
Perles cited Job 28.18 and Ezek. 32.20 as other possible instances of the
sense ‘acquire’ in BH. In the latter case it is hardly likely.
12.21-27 87
g. Heb. ‫בסף‬. There are two Heb. words ‫סף‬,ַ one meaning a vessel for liquids
(cf. Zech. 12.2) and the other, which is more common, ‘threshold’. Usually
the former is thought to be involved here, with the def. art. being explained
as prolepsis (cf. ‫ הסנה‬in 3.2 and the note there: GK §126q). But ‘threshold’
is also possible and would explain why no mention is made of the applica-
tion of blood to the threshold later in the verse. This was the interpretation of
Vulg (in limine) and probably LXX (παρὰ τὴν θύραν: see further Text and
Versions): Houtman 2, pp. 175, 193, gives additional references. One might
expect ‫ על־הסף‬if ‘threshold’ were meant (cf. the use of ‫ על‬in v. 23), but ‫ ב‬does
occasionally mean ‘on’ (Gen. 8.20; Num. 23.2; Judg. 8.21; 1 Kgs 2.5; Isa.
59.17).
h. Heb. ‫מן־הדם‬. For the partitive use of ‫( מן‬as in v. 7) see BDB, p. 580.
i. Heb. ‫לנגף‬. The verb (again in v. 27) has already appeared in 7.27 and the
related nouns ‫ מגפה‬and ‫ נגף‬in 9.14 and 12.13. There is considerable overlap in
meaning and usage between the verb ‫ נגף‬and the Hiphil of ‫( נכה‬for which see
vv. 12-13 and 29 in the nearer context: also e.g. 3.20; 7.17, 25; 9.15), but in
the Qal the former generally refers to divine action, whereas ‫ הכה‬is frequently
used of human action as well, especially in the context of war (cf. TWAT 5,
228-29, 446-52 = TDOT 9, pp. 211-12, 416-22). This perhaps gives ‫ נגף‬a more
distinctively theological ‘colouring’ than ‫הכה‬.
j. See Note bb on the translation of vv. 1-20 for the identical construction
in v. 13.
k. For the rendering of ‫ פסח‬here see Note cc on the translation of vv. 1-20.
From the original meaning ‘protect’ one might infer the sense ‘stand guard’
in v. 23.
l. Heb. ‫ולא יתן‬. For the use of ‫ נתן‬for ‘allow’ see 3.19 and BDB, p. 679.
m. Heb. ‫המשׁחית‬. For the different senses of ‫ משׁחית‬here and in v. 13 see Note
ee on the translation of vv. 1-20. Here ‫ משׁחית‬is the nominalised Hiph. part.
of ‫שׁחת‬, and the reference is to an angelic ‘destroyer’ like the one in 2 Sam.
24/1 Chr. 21, where ‫ (ה)משׁחית‬appears as an attribute of the ‫ מלאך‬who brings a
plague. The other occurrences of the nominalised part. (1 Sam. 13.17; 14.15;
Isa. 54.16; Jer. 22.7; possibly Jer. 4.7 and 51.1) refer to human ‘raiders’, in the
first two cases perhaps as a technical military term for a unit of an army. See
further the Explanatory Notes on vv. 13 and 23.
n. Heb. ‫את־הדבר הזה‬. The translation of ‫ דבר‬as ‘practice’ (cf. NRSV ‘rite’)
is based on its indefinite use for a ‘thing’ and especially its use after ‫עשׂה‬
in e.g. Gen. 22.16 (cf. BDB, p. 183). But since the ‘practice’ has just been
commanded by Moses, a rendering closer to the core meaning ‘word, saying’
is also possible: for ‫ דבר‬specifically of commands see BDB, p. 182, and for
its use in that sense elsewhere with ‫( שׁמר‬and no intervening form of ‫ )עשׂה‬cf.
Deut. 12.28; 13.1; 17.19; 29.8; Ps. 119.17, 57, 101; 1 Chr. 10.13; 2 Chr. 34.21.
In view of the continuation in v. 25, where ‫ שׁמר‬is construed with ‫ עבדה‬as its
object, ‘practice’ is nevertheless preferable here.
88 EXODUS 1–18
o. Heb. ‫ושׁמרתם‬. Waw of the apodosis, after the preceding temporal clause
(GK §112oo).
p. Heb. ‫(את־)העבדה הזאת‬. The sense ‘service, worship of God’ is mainly
limited to P, Ezekiel, Chronicles and Nehemiah according to BDB, p. 715,
but it appears again in the non-Priestly law about the Feast of Unleavened
Bread in 13.5. In view of the more widespread use of the verb ‫ עבד‬for the
service and worship of God (cf. esp. Exod. 3.12; 7.16, 26; 10.26), it is not
surprising to find some occurrences of the noun in worship contexts outside
the main places where it occurs. The sense is also somewhat different,
referring to a particular cultic practice (cf. ‘this’) rather than to the service
of God in general. J. Milgrom pointed out that in most, perhaps all, of the
occurrences of ‫ עבדה‬in P and Ezekiel it has the sense ‘physical labour’,
not ‘worship’ (Studies in Levitical Terminology, I: The Encroacher and
the Levite. The Term ʿAboda [Berkeley and Los Angeles, 1970], pp. 60-82,
83-86): so certainly in Exod. 1.14; 2.23; 6.6, 9. This sense was developed in
Neh. 10.33 and Chronicles into a general use for ‘cultic service’ (ibid., pp. 82-
83, 86-87), which is quite different from the use here (and in 13.5) for a
specific ‘act of worship’.
q. Questions of a similar structure to this one occur in Gen. 33.5, 8; Josh.
4.6; 2 Sam. 16.2; Ezek. 12.22; 37.18. Where an answer is given (in Ezek.
12.22 the question is rhetorical), it shows that, apart from Gen. 33.5 (where
‫ מי‬introduces a straightforward question about identity), the question seeks
to elicit the meaning of something, in terms of its purpose or cause, to the
addressee. Here, as in the very similar case of Josh. 4.6, the reason for the
practice is being sought. The briefer formulations in 13.14 and Deut. 6.20
(without ‫ )לכם‬probably have a similar intent.
r. Heb. ‫הוא‬. R.D. Holmstedt and A.R. Jones (‘Tripartite Verbless Clauses in
Biblical Hebrew: Resumption for Left-Dislocation or Pronominal Copula?’,
JSS 59 [2014], pp. 53-89), include this ‫ הוא‬in their ‘best-case corpus’ of
examples of a ‘copular pronoun’ in BH (p. 83), but it is hard to see why, as
the clause does not exhibit the ‘tripartite verbless clause’ structure whose
analysis they are discussing. ‫ הוא‬is simply the subject of the clause and the
words preceding it are the predicate, placed first as is natural in the response
to a question (cf. Muraoka, Emphatic Words, pp. 18-20; JM §154g).
s. Heb. ‫ אשׁר‬could mean ‘because’ here (BDB, p. 83, s.v. 8c: cf. NJPS and
Houtman and ‘for’ in NEB, REB, NRSV), though scarcely ‘when’ (cf. Vulg
quando). But the straightforward ‘who’ of LXX, Tyndale (‘which’), AV, JB
and NIV is preferable (cf. Propp, p. 410).
t. Heb. ‫וישׁתחוו‬. Codex L has the strange pointing ‫וַ יִּ ְשׁ ַתּ ֲחוּוּ‬, as also in 4.31 (cf.
11.8). Dotan ad loc. (cf. p. 1230) omits the dot in the penultimate waw (as did
BH3) and describes it as a mappiq (p. xiv), designed to show that the waw is
consonantal (see in general GK §14d). Many mss and earlier edd. also omit
the mappiq here.
12.21-27 89
Explanatory Notes
21-22. The new introductory formula signals a change of speaker,
which is also reflected in the divisions marked in MT, SP and
probably one Qumran manuscript (see Text and Versions). Moses
now gives instructions to ‘the elders of Israel’ which continue to the
middle of v. 27. Verse 3 in the previous (Priestly) section had antici-
pated that Yahweh’s instructions given there would be transmitted
to ‘the whole congregation of Israel’, and the elders here are seen by
some commentators as the natural intermediaries for the fulfilment
of that task. In Deuteronomy and the Deuteronomic History the
elders do receive the law from Moses and exercise various leader-
ship roles, as they do also in non-Priestly sections of the Tetrateuch
(see Dozeman, p. 273, for a summary of the evidence: more
fully J. Buchholz, Die Ältesten Israels im Deuteronomium [GTA
36; Göttingen, 1988]). But they never do so in Priestly texts (the
mention of them in Lev. 9.1 does not lead anywhere and is probably
secondary: cf. Buchholz, p. 36), where different titles are employed.
Moreover, the details of the instructions here, while similar to
certain features of vv. 1-20, differ from that passage in important
ways (see below) and were probably not originally designed to be
its sequel. The role of the elders as leaders of the community here
fits well, however, with other references to them in the non-Priestly
Exodus narrative (cf. 3.16, 18; 4.29; 17.5-6; 18.12; 19.7; 24.1, 9,
14; Num. 11.16, 24-25, 30; 16.25: in Exod. 10.9 Heb. zāqēn refers
to ‘old’ people in general), and these verses are therefore best seen
as part of it too, embodying an alternative account of the Passover
ritual which was once the continuation of 11.1-8.
The first word of Moses’ instructions, Heb. miškû, is not the
usual word for ‘Go’ and it could mean ‘Draw’ or ‘Get’ (see Note a
on the translation and Text and Versions). ‘Animals from the flock’
(Heb. ṣōʾn), as in vv. 3-5, allows for a goat as well as a lamb to be
chosen (but not a bull as in Deut. 16.2; cf. 2 Chr. 30.24; 35.7-9, 12).
The people are seen here as a body of ‘clans’ (for this meaning,
which is preferable to ‘families’ [NRSV], see Note c on the trans-
lation), a unit that comes between the tribe and the household in
Israelite society (cf. the genealogy in 6.14-25 and the notes there).
Here too the language is different from that used in vv. 3-4 and the
problems associated there with large and small households do not
90 EXODUS 1–18
arise. The mention of ‘the Passover lamb’ (Heb. happesaḥ, with the
definite article) can be explained in the present context from the
inclusion of the word in the preceding prescriptions (v. 11), though
even in the combined narrative it would be the first that the elders
had heard of it. Another possibility, especially when vv. 21-27 are
read apart from vv. 1-20, is that this writer knew that Passover was
an ancient custom, not one that was first invented at the time of the
Exodus, and so could assume that it would make sense to the early
Israelites portrayed in the narrative. What is added in v. 23 is then
a new explanation for the practice when it was associated with the
Exodus.
In v. 22 some details are included in the instructions that were
not specified in Yahweh’s words to Moses in vv. 1-20, a further
sign that vv. 21-27 were not originally designed to follow on from
them. The use of ‘a bunch of marjoram’ (Heb. ʾēzôb: see Note f
on the translation) to spread the blood of the slaughtered animal
over the doorway is similar to aspects of the rituals prescribed for
the cleansing of individuals who have suffered from skin disease
and houses with fungal growths (Lev. 14.4, 6, 49, 51-52: cf. Ps.
51.9). In both cases this is no doubt a practical measure where a
comparatively large area was to be smeared.10 The location of the
blood is also specified here, by a Heb. word (sap) which is unfortu-
nately ambiguous (see Note g on the translation). Most translations
and commentators render it ‘bowl’, comparing the practice in 24.6
(where a different Heb. word is used). But some early interpreters
(see Text and Versions) took it to mean ‘threshold’, presumably
as indicating the place where the animal had been slaughtered,
and this understanding has been revived in modern times (e.g.
Houtman 2, p. 193, with refs. to other scholars: see also NEB mg.):
it is also satisfying from a ritual point of view, as the whole of the
door opening is then protected. Finally the elders are warned that
no one should leave their protected houses till the morning, when
the danger will be past (cf. 11.4; 12.12, 29). There is some tension
between this requirement and the non-Priestly narrative in 12.31,
where Moses and Aaron are summoned by Pharaoh ‘in the night’
and apparently receive his permission to leave at the royal palace.
In the narrative context it is clear that the slaughter is now past
10
By contrast putting blood on a more restricted area is done with the finger
(e.g. Lev. 4.6-7).
12.21-27 91
(vv. 29-30), but the tension suggests that v. 22 and its context may
originally have been composed (or this requirement was added) for
a separate purpose, perhaps as a guide to the ongoing ritual of the
Passover (cf. vv. 24-27).
The word used for ‘applying’ the blood here is neither the word
used in v. 7 (Heb. nātan, ‘give, put’) nor one of the words used in
other ritual contexts (e.g. hizzāh, ‘sprinkle’, as e.g. in Lev. 4.6), but a
word that literally means ‘cause to touch’ (Heb. higgîaʿ, from nāgaʿ),
which is used of the application of blood only once elsewhere, in
Exod. 4.25 (cf. Blum, Studien, p. 12 n. 14). This confirms that these
verses belong to the non-Priestly strand of the Exodus narrative,
and the two passages actually describe two very similar, apotropaic,
rituals (on some non-biblical parallels see Propp, p. 408).11
23. In most respects this verse is very close to features of
vv. 12-13: see the relevant notes there on these points. The formula-
tion is similar enough for some literary relationship to be probable,
most likely with vv. 12-13 borrowing from here (see further the
introduction to this passage). At one point a verbal similarity and
a difference of religious conception coincide. ‘The destroyer’ here
is the same Heb. word (mašḥît) as is used for ‘destruction’ in v. 13,
except that the definite article is not present there. Both mean-
ings are also found in other passages. The personal interpretation
(which has long been accepted: see Text and Versions) is appropri-
ate here because of the definite article, Yahweh’s permission and
the mašḥît’s (normal) role to ‘smite’. In other places ‘destroyers’
are sometimes human foes (see Note m on the translation), but the
reference here will be to an angelic figure like the one described
in 2 Sam. 24.16 = 1 Chr. 21.15 as ‘the destroying angel’ (Heb.
hammalʾāk hammašḥît), who brought a deadly plague in the time
of David. The concept here, as there, is of a supernatural being who
is thought to be used by Yahweh, and therefore under his control, to
perform his will on earth. This bears some resemblance to a polythe-
istic world-view (and no doubt that is why the Priestly regulations
subtly reformulate the expression at this point), though in the Old
Testament such beings are not usually regarded as gods (see further
ABD 1, pp. 248-53; DDD, 81-90, 456-63). The problems are similar
11
‘Lintel’ and ‘doorposts’ are, as in v. 23, in the reverse order to v. 7. It is
unlikely that this has any profound significance, but it may be another pointer, in
this case linguistic, to the different origins of vv. 1-20 and vv. 21-27.
92 EXODUS 1–18
to those raised by mention of ‘the angel of Yahweh’ in 3.2-3 (see

the Explanatory Note there) and elsewhere in Exodus (14.19-20;
23.20-21, 23; 33.2). Psalm 78.49-51, in referring to this episode (in
its non-Priestly form) enlarges the single ‘destroyer’ to ‘a company
of harmful [Heb. rāʿîm] angels’, but appears to equate them with
Yahweh’s own fierce anger, rage and fury and the distress that they
bring, so playing down the idea of a separate agent. Psalm 105.36
avoids the expression altogether.
24-25. The provision for future observance of Passover (unlike
vv. 1-20 there is no mention of Unleavened Bread) in these two
verses is somewhat repetitive (note the repeated ‘you shall maintain’
[Heb. ûšemartem] with two different words for the observance),
and they define it in two different ways: v. 24 simply temporally,
as ‘a statute…for ever’ and v. 25 spatially, from the arrival in the
land which Yahweh has promised (Heb. dibber) and will ‘give’
to the Israelites.12 In addition, while v. 25 uses the second person
plural forms that are found throughout the rest of vv. 21-27 (and in
vv. 1-20), the second half of v. 24 stands out by its use of second
person singular forms in ‘for you and your descendants’.13 Such
references to the people by the singular form, whether understood
collectively or with reference to each individual, are frequent
elsewhere, especially in legal texts such as the Decalogue (e.g.
Exod. 20.2-3). But here the variation jars and suggests that at least
v. 24b and, since v. 24a on its own has little point, perhaps the whole
verse is a secondary addition to the passage.14 The terminology ‘a
statute…for ever’ is very similar, though not identical, to Priestly
expressions which are used in vv. 1-20 (cf. vv. 14 and 17) and it is
also in the Priestly ‘appendix’ to the Passover rules in vv. 43-49 that
12
The ‘promise’ could mean passages like Gen. 12.7 or Yahweh’s words to
Moses in Exod. 3.8 and 6.8. Similar language is widespread in Deuteronomy (e.g.
27.3).
13
The closest parallels to the whole formula occur in provisions for the
Aaronide priests (29.28; Lev. 7.34; 10.13-14). Other passages include later genera-
tions without using the word ‘statute’: Lev. 10.15; Num. 18.8, 9, 11, 19; Deut.
4.40; 12.25, 28; Josh. 14.9 (Propp, p. 409).
14
Houtman, p. 194, following Strack, cites some other examples of such
variation within a passage (e.g. Exod. 23.25), but even if these are not also due
to secondary redaction the additional evidence mentioned regarding Exod. 12.24
makes this a very probable case of editorial expansion.
12.21-27 93
several more second person singular forms occur (vv. 44, 46, 48:
alongside, in MT, just one second person plural form, in v. 46b). By
contrast, the location of future Passover celebrations specifically in
the land of Canaan in v. 25 is found nowhere in P, and in a Priestly
passage in Num. 9.1-14 Yahweh instructs the Israelites to keep the
Passover while they are still in the wilderness. Verse 25 is linked by
its use of the key word ‘service/worship’ (Heb. ʿabōdāh: see Note
p on the translation) with the verses that follow (cf. v. 26), which
have different affinities (see the notes on these verses and the intro-
duction to the section as a whole). If v. 24 is indeed a redactional
addition (cf. Blum, Studien, p. 39 n. 149), its use at the beginning of
the same plural form ‘you shall maintain’ as in v. 25 was probably
designed to make its alien origin less apparent. By placing such
a general introduction before the narrower restriction to the land
of Canaan, the redactor will have been aiming to make the latter
only a particular case of a celebration that could potentially be held
anywhere, an important concession for a time when a large number
of Jews had, through exile and emigration, come to live outside the
homeland. Both the technique and the aim find some parallels in
Num. 15.22-29, where wording is taken over from an underlying
law (in this case from further away, in Lev. 4) and the application
of the law is widened to include ‘resident aliens’ (gērîm) as well as
native Israelites (see Fishbane, Biblical Interpretation, pp. 189-92).15
26-27. The preservation of the memory of the Exodus deliver-
ance by means of the Passover ritual is, in general terms, a further
example of continuity with the preceding Priestly regulations (cf.
v. 14: the following verses are preoccupied with the week-long
Festival of Unleavened Bread), but the terminology and the proce-
dure differ. Here the occasion is elaborated into an opportunity for
the teaching of children, sparked off by the expectation that they will
15
This kind of widening (which is of course quite distinct from that which
seeks to include Israelites outside the land as well as those within it) also appears
in the Passover laws of P (Exod. 12.19, 48-49, the latter verses being a clear
modification of the principle stated in v. 43). A further example of the use of
a repeated phrase to ‘key in’ additional material is noted in Lev. 23.39-42 by
B. Levinson, ‘The Right Chorale’ (Tübingen, 2008), pp. 220-21: the date formula
‘the fifteenth day of the seventh month’ is repeated in v. 39 from its appearance
at the beginning of the original section of the calendar which dealt with the Feast
of Booths (v. 34).
94 EXODUS 1–18
question their parents about the meaning of the ‘service’.16 The form
of these verses has attracted comparison with five other passages
where children are to be instructed about their ancestral traditions:
Exod 13.8 (about the Exodus, at the Feast of Unleavened Bread);
13.14-15 (about the Exodus, at the sacrifice of firstborn animals);
Deut. 6.20-25 (about the Exodus, in connection with the laws of
Deuteronomy); Josh. 4.6-7 (about the crossing of the Jordan, at the
stones set up in Gilgal); and 4.21-24 (the same). On these passages
and their likely origin see the Excursus in the introduction to this
section. A more general instruction for the teaching of the Exodus
story, and specifically the plague-narrative, appears in Exod. 10.1-2:
see the notes there.
The description of the ritual as a ‘service’ (Heb. ʿabōdāh) takes
up the language of v. 25 (with which v. 26 is also linked by the initial
‘And it shall be’ in the Heb. [cf. AV, RV: in common with many
other modern translations we have omitted this formula from our
rendering above]). This word is widely used in P, Ezekiel, Chroni-
cles and Nehemiah for the service and worship of Yahweh (but not
apparently, as here, for a particular ‘service’ like Passover).17 The
closest parallel is in 13.5, where it refers to the Feast of Unleavened
Bread, in a non-Priestly passage, and the related verb ʿābad, which
can also (like ʿabōdāh) refer to the ‘service’ of human masters, is
used several times of worship in non-Priestly sections of Exodus
(3.12; 7.16, 26; 10.26). The reply to be given by the parents makes
explicit the etymological connection seen between the noun pesaḥ,
‘Passover’, and the verb pāsaḥ, ‘spare, pass over’, which is already
suggested in the Hebrew in vv. 21 and 23 (cf. vv. 11 and 13). It
16
This is the origin of an important section of the Jewish Passover Seder
or Haggadah which is still used. The question and answer structure is already
mentioned in M.Pes. 10.4 along with other central features of the service, but the
wording there is different. The ‘instruction of the four sons’, where Exod. 12.26-
27 is cited with three other biblical passages (Deut. 6.20; Exod. 13.14; 13.8), is
referred to in MRI on 13.14 (Lauterbach, pp. 166-67) and in J.Pes. 34 (text and ET
in H.W. Guggenheimer, The Jerusalem Talmud: Second Order, Moʿed. Pesachim
and Yoma [Studia Judaica 74; Berlin, 2013], pp. 368-70) and may be of a slightly
later origin. The question in Exod. 12.26 is assigned to the ‘Wicked Son’ because
‘What is this service to you?’ is understood as the words of a wilful outsider.
17
An important distinction has been made by Milgrom between the uses of
ʿabōdāh in P and Ezekiel on the one hand and Chronicles and Nehemiah on the
other (see further Note p on the translation).
12.21-27 95
also designates the animal as a ‘sacrifice’ (Heb. zebaḥ), a term

which seems deliberately to be avoided in the Priestly version,
even though some of its detailed features correspond closely to
the tabernacle cult (see the notes on vv. 5, 6 and 10). This seems
to reflect a situation in which domestic sacrifice, even if unusual,
was not prohibited (e.g. Judg. 13.19; 1 Sam. 20.29), which is
quite different from the legislation of P (e.g. Lev. 3.2; cf. 17.2-7)
and Deuteronomy (12.8-14). Deuteronomy 16.5-7 very explicitly
legislates that Passover is not to be offered anywhere except at
the central sanctuary, just because it was a sacrifice, and 2 Kgs
23.21-23 is remarkably frank in its acknowledgement that this is
not how Passover had been observed in earlier times. The response
prescribed here, in other words, must antedate the centralising
legislation of Deuteronomy. Its emphasis falls not so much on the
slaughter of the Egyptian firstborn as on the sparing of the Israel-
ites and their houses, which is mentioned twice – on deliverance
rather than judgement (for ‘kept…safe’ [Heb. hiṣṣîl] cf. 3.8; 5.23;
6.6 and later 18.8-10).
The response of the people is, as in 4.31, to an earlier announce-
ment of Yahweh’s imminent intervention, one of prostration and
worship, an anticipation of the hymns of praise that they will sing
when the deliverance is complete (15.1-21). Since it is again here in
hope of coming deliverance that they so respond, v. 27b may just as
well be the continuation of v. 27a as of v. 23.
Text and Versions

There is strong indirect evidence that 4QpalExm began the section with an
indented line, after v. 20 had filled the previous line; this corresponds to
divisions in both MT and SP. No other Qumran mss survive at this point.
‫( זקני ישׂראל‬12.21) On the renderings of ‫ זקני‬in LXX and TgN see Text and
Versions on 3.16. TgG has the literal rendering of the other Tgg here. Some
LXX mss, Vulg and Sy insert ‘the sons of’ (cf. 4.29 MT) in line with the
designation of the Israelites that is frequent elsewhere (e.g. v. 28).
‫( משׁכו‬12.21) Probably none of the Vss presupposes a different Vorlage
from MT (and SP), but their varied renderings indicate uncertainty over its
meaning. A majority see ‫ משׁך‬here as an intrans. verb of movement (cf. Note
a on the translation: the Vss need not point to the [in any case easier] reading
‫[ לכו‬Ehrlich, Beer]), whether in a general way (LXX ἀπελθόντες; Vulg ite)
or at speed (TgO,G, with different verbs; cf. Sy bʿgl, ‘quickly’): the latter has
been taken up in the proposal to read ‫ מהרו‬instead (cf. DCH 5, p. 523). ‫ מהר‬is
96 EXODUS 1–18
certainly often used as an auxiliary before other verbs, as it would be here (cf.
TWAT 4, 713-14 = TDOT 8, pp. 138-39; DCH 5, p. 166), but these Vss are
probably guided by an Aram. meaning of ‫ משׁך‬rather than a different Vorlage.
TgJ (cf. TgNmg) ‫ נגודו ידיכון‬presumes an ellipse of ‘hand’ as the object of ‫משׁכו‬,
giving the sense ‘withdraw, cease’ and relating it to an abandonment of the
worship of Egyptian gods (cf. Josh. 24.14; Ezek. 20.4-8). This interpretation
is found already in the early midrashim (cf. AramB 2, p. 192 n. 39) and is
connected to TgJ’s understanding of Exod. 6.9. MRI (ed. Lauterbach, p. 83)
also records the view that ‫ משׁך‬is here a technical legal expression for acquiring
a small animal, which is used in M.Kidd. 1.4 (cf. M.Sheb. 10.9), so that ‫צאן‬
would be its object. Finally TgN ‫אתמנון‬, ‘count yourselves’ (not ‘appoint [the
participants]’ as in AramB), finds the key in v. 4, where it also uses the verb
‫מני‬: this has no etymological basis.
‫( וקחו‬12.21) SP omits the copula. None of the Vss certainly supports it,
though the participial constructions of LXX and Vulg could be based on a
Vorlage with or without the copula. Asyndeton is certainly frequent with two
imperatives (GK §120g-h) and the copula seems often to be added second� -
arily, so the reading of SP deserves serious consideration, especially if ‫ משׁכו‬is
interpreted as a verb of motion here.
‫( למשׁפחתיכם‬12.21) LXX κατὰ συγγενείας ὑμῶν takes up its first
equivalent for ‫ משׁפחה‬in 6.14-25, which it uses again in Lev. 20.5 and Num.
1.2 before abandoning it for δημος: on the various equivalences see Text and
Versions on 6.14-19, 24-25.
‫( הפסח‬12.21) TgJ (cf. TgNmg) prefixes ‘the lamb’, correctly specifying the
sense which ‫ פסח‬has here; TgG, ‘sacrifice’, in the sense of ‘sacrificial animal’,
likewise, with the phrase borrowed from v. 27.
‫( ולקחתם‬12.22) 4QpalExm preserves what may be the final letter of this
word, but nothing else until v. 31. Vulg has no equivalent and takes ‫אגדת אזוב‬
with ‫וטבלתם‬. BHS cites a Genizah Heb. ms. which has ‫ולקחתם לכם‬, no doubt a
secondary expansion based on expressions like ‫ וקחו לכם‬in v. 21.
‫( בדם‬12.22) LXX ἀπὸ τοῦ αἵματος, suggesting ‫ מן־הדם‬as later in the
verse, but this is less appropriate here.
‫בסף‬bis (12.22) LXX παρὰ τὴν θύραν, i.e. presumably ‘outside’, is prob-
ably based on the understanding of ‫ סף‬here as ‘threshold’ (cf. Vulg in limine).
This is the view taken in MRI (Lauterbach, p. 84), which sees it as a reference
to a hole dug near (‫ )בצד‬the threshold, where the animal would have been
slaughtered. But MRI also records R. Akiba’s opinion that ‫ סף‬here means
‘vessel’, which is found in all the Tgg (and approved by Rashi). TgJ adds that
it should be a ‘clay’ vessel; TgNmg uses a more specific word meaning ‘bowl,
cup’. Sy has ‘of the lamb’ the first time, but ‘vessel’ for the second.
‫( והגעתם‬12.22) The LXX reading preferred by Wevers καὶ θίξετε (cf.
19.18) is the most literal of the versional equivalents, both lexically and in
the retention of καί for waw when the previous verb has been rendered by an
12.21-27 97
aor. part.18 TgN has ‫ותתנון‬, reflecting the wording of v. 7:19 the other Tgg have
‘sprinkle’, which often renders ‫ ִהזָּ ה‬in cultic contexts and is followed here by
Sy and Vulg. This freer rendering was no doubt favoured because of the rare
use of ‫ הגיע‬with a liquid.
‫( אל־המשׁקוף ואל־שׁתי המזוזת‬12.22) SP has ‫ על‬for ‫ אל‬both times (cf. perhaps
LXX ἐπ’ἀμφοτέρων τῶν σταθμῶν and more clearly TgN,G and Sy ‫ על‬both
times), whereas TgO,J and Aq follow MT and Vulg is inconclusive. The variant
is most likely due to assimilation to the wording of v. 7 or v. 23: ‫ אל‬or ‫ ל‬is
much more common after ‫הגיע‬, although ‫ על‬is occasionally found. On the
versional renderings of ‫ המשׁקוף‬and ‫ המזוזת‬see Text and Versions on v. 7. TgJ
again specifies that the sprinkling or daubing was done on the outside of the
lintel: this reflects an interpretation that was known to, but rejected by, MRI
(Lauterbach, p. 84).
‫( לא‬12.22) SP reads ‫אל‬, which is the more common negative with a
specific prohibition (JM §113m). But BH seems to allow some flexibility of
usage (cf. Deut. 19.14; 23.22 compared with Prov. 22.28; Qoh 5.3 and 1 Kgs
3.26-27, cited in JM): perhaps SP reflects a more rigid view of grammatical
precision (compare the opposite variation in 5.9).
‫( ועבר‬12.23) TgN,G, as well as LXX and Vulg, render straightforwardly,
but TgO,J avoid the anthropomorphism with ‘will be revealed’, as in 11.4,
12.12 and often elsewhere. Chester, Divine Revelation, pp. 146-48, notes that
this practice is not universal with ‫עבר‬, which, with the lack of any periph-
rasis for ‫ וראה‬later in the verse (as also in v. 13), indicates that more than
anti-anthropomorphism is involved. Chester identifies ‘theophany’ as the key
element, which is reinforced in TgJ by the addition of ‫יקרא‬, ‘glory’, to the
divine name (cf. TgN ‫איקר שׁכינתה‬, perhaps a secondary development in the Pal.
Tgg, as TgG [cf. TgNmg] has ‫ ממרה‬here).
‫לנגף‬1o (12.23) Here and at the end of the verse the Tgg and Sy use
words for ‘blot out’ and ‘kill’ for ‫נגף‬, to match the coming reference to ‘the
Destroyer’ (cf. also 4.23).
‫( ופסח‬12.23) There is some shifting in the choice among possible render-
ings by the Vss compared with v. 13 (see Text and Versions there). LXX
παρελεύσεται this time equates ‫ פסח‬with ‫עבר‬, while Vulg transcendet makes
a slight variation with a clearer element of ‘pass over’. Among the Tgg it
is TgJ,N,G2 which use ‫ גנן‬Aph, ‘protect’, while TgG simply uses the Heb. ‫פסח‬
18
But against this see Desilva, ‘Five Papyrus Fragments’, p. 10. The uncials
and most other mss (and editions) read καθίξετε (papyrus 866 καθίξεται).
Wevers recovered his reading from a few miniscules (inc. the b-group) and the
daughter versions (cf. OL et linietis in citations; also Theod καὶ θίξεσθε): see
THGE, p. 263; Notes, p. 180.
19
The preceding ‫ ואתון‬is an error arising from its occurrence later in the verse
after the second occurrence of TgN’s equivalent for ‫בסף‬.
98 EXODUS 1–18
(which also appears alongside ‫ גנן‬in TgG2 and in the text of TgN: the mg of the
latter records ‫ויחוס‬, ‘spare’, the reading of TgO).20 The influence of Isa. 31.5
thus remains strong, if not total, in the Tgg here. LXX shows how early the
sense ‘pass over’ or ‘pass by’ was adopted, and Brock (‘An Early Interpreta-
tion’, p. 27) adds further evidence from Ps-Ezekiel (Exagoge, 159, 187), Jub.
49.3-4, Philo and Josephus. Brock has provided a valuable comprehensive
review of the evidence for the use and connotations of Aram. ‫גנן‬, in which he
concludes that it was ‘richer in overtones’ than its Heb. cognate and could
be ‘a technical term for divine activity of a salvific character’ (ibid., p. 34).21
Even if its sense in Exodus 12 does not go beyond ‘protect’, it is striking that
its use there ‘brings the very specialised pāsaḥ of the Hebrew into the wider
context of the covenantal theophanies of Gen xv and Exod xxxiii’ (ibid.).
‫יהוה‬2o (12.23) TgJ,N,G2 prefix ‘the Memra of’; TgG, having done so at the
beginning of the verse, does not repeat it here.
‫( על־הפתח‬12.23) TgN adds ‘of the fathers of the Israelites’ (cf. 17.12 for a
similar reference to the patriarchs added in TgN) : it can scarcely mean ‘of the
fathers’ houses’ (Le Déaut, Exode et Lévitique, p. 90). AramB 2, p. 50 n. 19,
plausibly sees an explanation for it in a rabbinic association of the ‘protection’
of the Israelites with Abraham’s ‘standing over’ (‫ )על‬his heavenly visitors in
Gen. 18.8 (T.Soṭa 3.1).
‫( ולא יתן‬12.23) TgN and its mg variously bring out the sense of ‘permis-
sion’ more explicitly.
‫( המשׁחית‬12.23) All the Vss treat the ‫ משׁחית‬here as personal. Vulg percus-
sorem softens the force of the Heb. to match its precise rendering of the first
‫ לנגף‬by percutiens. TgJ (cf. TgNmg) specifically identifies the Destroyer as an
angel, as in v. 13.
‫לנגף‬2o (12.23) Vulg laedere is a notably mild rendering of the Heb., but
has the effect of excluding even slight harm being done to the Israelites, in line
with 11.7. TgNmg adds ‘you’ as the object.
‫( ושׁמרתם‬12.24) Vulg custodi, sing. imper. to match the sing. ‘you’ later in
the verse, which is in fact the outlier in the passage as a whole. Presumably
Vulg assumes an address to the individual Israelite here.
‫( לחק‬12.24) LXX’s νόμιμον, originally an adjective and an equivalent
which is particularly favoured in phrases for an ‘everlasting’ statute (cf.
vv. 14 and 17), must be the basis for Vulg’s legitimum (presumably via the
OL), as the latter is not normally nominalised in classical Latin.
‫( לך ולבניך‬12.24) TgN,G have second person pl. suffixes to conform to
‫ ושׁמרתם‬and the remainder of the passage. TgJ adds ‫זכורייא‬, ‘male’, to make
explicit what MRI (p. 89) also saw as implicit in the specification of ‘your
sons’: the men bear responsibility for the observance of the rite.
20
‘TgG2’ refers here to the second transcription of vv. 21-34 in ms. AA, which
varies only slightly from the main text.
21
An additional attestation of this rendering appears in T.Soṭa 3.1 (AramB 2,
p. 50 n. 19).
12.21-27 99
‫( והיה‬12.25) LXX and Vulg understandably have no equivalent. TgG has

whwh (as also in v. 26), an unexpected perf. due to assimilation to the Heb.
‫( יתן‬12.25) Instead of a simple future LXX has (ἣν) ἂν δῷ (subjunctive),
Vulg daturus est and TgJ ‫עתיד למיתן‬. The two latter renderings (which may be
connected through Jerome’s consultation of Jewish advisers) could express
the imminence of the gift of the land, but the LXX is more puzzling, as the
construction normally has an indefinite sense (cf. NETS ‘whichever’ here and
in 25.16, 21, where the Greek is very similar): so Wevers, Notes, p. 398, on
25.16. But this cannot be what the translator meant, either here or in ch. 25.
BDF §380(2) observes that in a indefinite relative clause the future could be
used instead of the subjunctive with ἀν, so perhaps the translator mistakenly
thought that the reverse was also possible, as in the kind of temporal clause
(with ἐάν) with which the verse begins. Wevers, Notes, p. 182, thinks the
expression here shows ‘divine intentionality’, but that would be more appro-
priate to Vulg and TgJ.
‫( יהוה‬12.25) TgG (and TgNmg) adds ‘the Memra of’.
‫( את־העבדה הזאת‬12.25) Vulg caerimonias istas, with the pl. highlighting
the individual aspects of the ritual. SP adds ‫ בחדשׁ הזה‬from 13.5, where it
follows the same phrase; TgJ adds ‘from the time you arrive there’, probably
(like MRI, Lauterbach, pp. 89, 94) underlining that Passover was only to be
celebrated after the entry into the Promised Land (which creates [or exposes]
a contradiction with Num. 9.1-14 in the Priestly legislation). Eventually this
ruling was relaxed, but probably only after the destruction of the Second
Temple in A.D. 70 (cf. ABD 6, pp. 760-64): at any rate the exceptional case of
the Elephantine community seems to provide the only definite evidence of an
earlier celebration outside the land.
‫( והיה‬12.26) Here, unlike v. 25, LXX has καὶ ἔσται, but Vulg again has
no equivalent. On TgG see on v. 25.
‫( בניכם‬12.26) Traces of the final letter and the first letter of the next
word may be preserved in 2QExb. The location of this small fragment is not
completely certain, but the only possible alternative is at 14.13. TgJ adds ‘at
that time’, presumably meaning ‘in the land’: the phrase does occur twice
in MRI’s comment on v. 26 (Lauterbach, p. 94), but not in such a way as to
suggest that TgJ is alluding to its exegesis here.
‫( לכם‬12.26) There is no equivalent in LXX* (the O and C texts have
ὑμῖν from the Jewish revisers), Vulg and Sy, but it could have been omitted
because of its awkwardness and absence from the similar question in 13.14.
SP and TgO,J,G support MT: TgN lacks the whole verse through homoeoteleu-
ton.
‫( ואמרתם‬12.27) Vulg and Sy dispense with ‘and’ in the apodosis, but it
is retained in the Tgg and, more surprisingly, in LXX against normal Greek
grammar. But the Hebraism is quite frequent (cf. Wevers, Notes, p. 182),
although less so proportionately in Exodus than in the other books of the
Pentateuch (cf. Aejmelaus, ‘The Significance of Clause Connectors’, in On
the Trail, pp. 49-64 [57]). For a related example see Text and Versions on
100 EXODUS 1–18
‫ והגעתם‬in v. 22. After the verb here LXX, Vulg and Sy added ‘to them’, a
popular kind of versional addition in dialogues.
‫( זבח־פסח הוא ליהוה‬12.27) LXX has θυσία τὸ πάσχα τοῦτο τῷ κυρίῳ,
presumably ‘This Passover is a sacrifice to the Lord…’, with ‫ פסח‬as the
subject and ‫ הוא‬probably taken mistakenly as attributive rather than as an
independent pronoun. Both Vulg (domini) and Sy (dmryʾ) take ‫ ליהוה‬as a
genitive after ‫פסח‬: ‘(the sacrifice of) the Lord’s passover’, a possible construal
of the Heb. and only a little different in sense from a literal rendering. In their
treatment of ‫ פסח‬the Vss generally repeat their equivalents in v. 11 (see Text
and Versions there), but TgO exhibits the interpretation ‘sparing’ here (like TgJ
here and in v. 11), to match its rendering of the verb later in the verse, while
Symm and Vulg are briefer, the former now omitting its explanation (so far
as our evidence goes) and the latter (as in v. 21) the transliteration of the Heb.
(so perhaps also Aq). 2QExb (if the identification is correct) preserves ‫[ליה]וה‬
in palaeo-Hebrew script (on this unusual practice see Tov, Textual Criticism,
p. 220: 4QExj should be added to his list, see DJD XII, p. 150). TgO,J,N render
‫ ליהוה‬by ‘before the Lord’ as often, but TgG has ‘for the name of the Lord’ as
in v. 11 (see Text and Versions there).
‫( אשׁר‬12.27) 2QExb has ‫אשׁר‬, which would support MT and SP. Vulg
quando, ‘when’, is perhaps dependent (via OL) on the Greek: the major LXX
witnesses (followed by Rahlfs) have ὡς. Wevers (Notes, p. 182), perhaps
correctly, regards this as an error by homophony for ὅς, the reading of
numerous miniscules and the Ethiopic.
‫( פסח‬12.27) Here too (even if they diverged in v. 23) the Vss follow
closely their renderings earlier in the chapter (v. 13: see Text and Versions
there). ‘By his Memra’ is added in TgJ this time (cf. TgNmg), but not TgN,G.
‫( במצרים‬12.27) TgG (like TgG2, which resumes with this word after a
lacuna) inserts ‘the land of’ as later in v. 39. Such additions are frequent and
serve to distinguish the geographical and ethnic meanings of ‫מצרים‬.
‫( בנגפו‬12.27) On the renderings of Tgg and Sy see Text and Versions on
v. 23.
‫( ויקד העם וישׁתחוו‬12.27) TgO reflects the variation from sing. to pl. forms
in MT, but the other witnesses harmonise, either in the pl. (SP, TgJ,N,G, Sy) or
in the sing. (LXX, Vulg). The inconsistent text of MT is to be preferred as
the difficilior lectio (on the grammar cf. GK §145g, citing Exod. 1.20; 33.4).
Sy adds ‘to the Lord’ at the end for completeness, as it did in 4.31. TgJ has
a longer addition before these words (‘when the house of Israel heard this
word from the mouth of Moses’), perhaps to clarify that they do not belong
to Moses’ speech; while TgN,G,G2 replace ‫( וישׁתחוו‬as often elsewhere) with the
combination ‫ואודו ושׁבחו‬, ‘and they gave thanks and praise’ (so also at 4.31 in
TgN; cf. 15.1, 21 for ‫)שׁיר‬: the interpretation with ‘praise’ is found already in
MRI (Lauterbach, p. 95).
C h ap t er 1 2 . 2 8 - 4 2 , 5 0-51
Th e Fi n a l P lagu e an d t h e Isr aelites’

R eleas e
The narrative of the Israelites’ departure from their settlement(s)

in Egypt naturally includes some references to their fulfilment
of the instructions about the Passover (vv. 28, 50), even if these
references are less specific than might have been expected. No
doubt for the biblical narrators it was more important to record
the instructions themselves and the event which gave them their
meaning for later generations. For the purposes of this commentary
we have separated the additional instructions in vv. 43-49 (which
will be dealt with in the next section) from their narrative context.
The Masoretic section-divisions and the triennial reading-cycle
recognised the boundaries of the main unit but placed its beginning
at v. 29 rather than v. 28, thus associating the latter closely with
the preceding instructions to which it refers. In addition there are
divisions before vv. 37, 43 and 51 (so that, understandably, v. 50
too is connected with the verses to which it refers). From Qumran
there is evidence for the divisions before v. 37 (4QpalExm, 4QExc
and probably 2QExa), 43 (4QpalExl) and possibly 51 (4QDeutj).
No Qumran evidence survives for the beginning of the unit, but the
end of it is marked by an interval in the phylacteries and probably
in one similar manuscript (4QDeutj). In addition there are divisions
before v. 39 (4QpalExm; but not in 4QExc) and v. 40 (4QExc).
The number of sub-divisions marked by the tradition is a
reflection of the number of, mostly short, sub-units out of which
the passage is composed: (i) a summary note of the fulfilment
of the instructions about Passover (v. 28); (ii) an account of the
death of the Egyptian firstborn and the expulsion of Israel from
Egypt (vv. 29-34); (iii) a recapitulation of the ‘plundering of the
Egyptians’ (vv. 35-36); (iv) details about the departure of Israel
(vv. 37-38) and the making of unleavened bread (v. 39, which is the
direct sequel to v. 34); (v) another account of the Israelites’ depar-
ture and its association with the regular observance of Passover
102 EXODUS 1–18
(vv. 40-42); (vi) a further summary note of obedience and depar-

ture (vv. 50-51), which belongs with vv. 43-49.
Critical scholarship has at least since Knobel been in substantial
agreement about the distinction between a Priestly and a non-Priestly
version of the episode: vv. 28, 40-42 and 50-51 belong to P (in the
broad sense) and vv. 29-39 derive from another source or sources
(so recently both Albertz [pp. 212-19], with an older [amplified]
Exodus narrative [vv. 29-39] and two Priestly Bearbeitungen, and
Schmidt [pp. 540, 545]).
At one time it was common to attribute v. 37a to P, particularly because

of its formal resemblance to other, more obviously Priestly, itinerary-notes
later in the narrative (e.g. 16.1: so Knobel, Wellhausen, Holzinger, Baentsch,
Gressmann [the whole verse]). But Wellhausen already observed that some
such annotation was needed in JE (Composition, p. 72) and, probably because
of the links between the names in v. 37a and 1.11 and 13.20 (on the latter
cf. Smend, Erzählung, p. 133), first Dillmann and then the overwhelming
majority of subsequent scholars attributed it to the non-Priestly narrative.
Indeed for some recent critics it is the only or almost the only part of the
passage that they regard as relatively early (see below).1
Finer sub-divisions within the two main strands have also been made. In
the Priestly strand v. 42, with its reference to ‘night’, has been thought to be
inconsistent with ‘day’ in v. 41, and Noth’s observation that it ‘falls lamely’
after vv. 40-41 (p. 72, ET, p. 92) has also been influential. The verse has
therefore been ascribed to Ps or an even later redactor (Carpenter/Harford-
Battersby [with vv. 40-41; likewise McNeile], Noth, Kohata, Levin, Ahuis,
Graupner [p. 67 with n. 214], Gertz).2 Where vv. 43-49 were seen as an
addition to the main Priestly account, a similar view was naturally taken of
vv. 50-51 (Holzinger, Baentsch, Rudolph [but he saw v. 51, with an emen-
dation proposed by Ehrlich, as originally the introduction to 13.1-2], Noth,
Childs [pp. 201-202], Kohata, Houtman, Levin, Ahuis, Graupner, Gertz).
Within the non-Priestly material a larger number of alternative sub-divisions
have been envisaged. An early view, taken by Wellhausen and Dillmann, was
that while vv. 29-30 were from J (cf. 11.4-8), the following verses must be
from E because they contradicted earlier statements that Moses would not
meet Pharaoh again and it would be his courtiers who would come and beg the
Israelites to leave (10.28-29; 11.8). The majority of subsequent scholars have
thought that the shift could be explained by a change of mind on Pharaoh’s
1
Propp attributes the whole of vv. 37-38 to a redactor (p. 375).
2
Following Wellhausen, a number of scholars saw v. 42 (Dillmann, Holzinger,
Beer: Rudolph was unsure) or part of it (Baentsch) as associated with one of the
non-Priestly sources.
12.28-42, 50-51 103
part (so already Holzinger, p. 34), but Baentsch maintained Wellhausen’s view
for v. 31a, and it was also taken up for the whole of v. 31 by those who saw
in vv. 33-39 further evidence of their ‘third early source’ (Smend, Eissfeldt,
Beer, Fohrer), a view backed up initially by the shift from night to day as the
time of departure, the unlikelihood, for different reasons, of the aetiology of
Unleavened Bread being from J or E and the earlier attribution of 3.21-22 to J1
(Smend, Erzählung, p. 133; cf. Eissfeldt, p. 35). Not surprisingly, these argu-
ments did not carry much weight with other scholars, who continued to regard
most or all of vv. 29-39 as coming from J (so still Schmidt, p. 541). The main
exception was vv. 35-36, which like 11.(1)2-3 came to be widely ascribed to E
(Holzinger, Baentsch, Gressmann, McNeile, Hyatt, Childs [‘may be’ (p. 184);
cf. Schmidt, Exodus, Sinai und Moses, p. 56]) or, especially in more recent
analyses, to a later redactional layer (so already Carpenter/Harford-Battersby,
then Rudolph, Noth, Kohata, Blum [Kd], Ahuis, Graupner, Gertz, Schmidt
[p. 537]). For Propp, as usual, the attributions are reversed, with E being the
main contributor here and vv. 35-36 possibly from J (p. 375).
Most recently there has been a polarisation of opinion between those who
see vv. 29-39 as a unity (Van Seters, Life, pp. 98-99; Dozeman, pp. 250-51),
albeit from a quite late period, and an extreme redaction-critical approach
which takes up older arguments that most had discarded. Ahuis is the closest
to the older views, for he attributes vv. 29-33 and 37-38 to J, with his Deuter-
onomistic redactor (DtrT) adding vv. 34 and 39 as well as vv. 35-36 (and also
vv. 42, 50-51: Exodus 11,16–13,16, pp. 67, 70-72). At least here there is an
account of the slaying of the Egyptian firstborn in the older narrative. This
is no longer the case in the analyses of Levin, K. Schmid and Gertz. Levin’s
exilic Yahwist constructs a departure narrative consisting of vv. 35-37, 38b
around the older itinerary-notice in v. 37a: the implication is apparently that
Pharaoh has readily agreed to Moses’ request as stated in 3.18 and did not
need any plagues to persuade him (p. 326). The basis for this reconstruction is
not clear, unless it lies in vv. 35-36, which (with 3.21-22) Levin clearly sees
as of fundamental importance. A first stage of amplification is the addition of
vv. 30aα, 31-33, which for Levin embody the Egyptians’ response, not to the
slaying of the firstborn, but to a repetition of the first four plagues, which is
according to him threatened in 9.13-16 (pp. 337, 339: see the introduction to
9.13-35). The slaying of the firstborn (vv. 29, 30aβb: like its announcement
in 11.4-7) was only added at the next stage, when the sequence of plagues
was extended by additions in chs. 9–10. Verses 34+39 and 38a are also seen
as redactional (pp. 339-40), on the one hand because of an association with
13.3-10 and the connection (to v. 33 presumably) made in v. 39b, and on the
other because of the introductory ‘In addition’ (Heb. wegam).
Konrad Schmid deals with verses from this section only in passing, and
not every verse is commented on. But it is clear that he too regards the whole
plague-narrative (including the slaying of the firstborn) as a later addition to the
non-Priestly, once independent Exodus-story, accepting the superficial argu-
ment of Van Seters that Deuteronomy knows nothing of the plagues (p. 146:
104 EXODUS 1–18
cf. Van Seters, ‘The Plagues of Egypt: Ancient Tradition or Literary Inven-
tion?’, ZAW 98 [1986], pp. 31-39; Life, pp. 80-81).3 Particular verses are
noted to have links with Genesis (v. 32 [pp. 64, 196]; vv. 35-36 [p. 249]) or
to Exodus 3–4 (v. 39, via 5.22–6.1 [p. 251]) and the ‘life–death’ theme is seen
as a reversal of ch. 1 (p. 338), which in Schmid’s view of the literary history
of the Pentateuch only confirms their late origin. Only v. 33a and perhaps
v. 36 (but see above: and Schmid does not discuss v. 37) remain as possible
evidence of an older version of the story in which the Israelites were driven
out without any pressure from plagues (p. 150: presumably v. 33b is regarded
as a later addition after the inclusion of the slaying of the firstborn or some
other plague[s]).
Gertz’s discussion of the passage is also rather scattered, but he does have
a view about the origin of every (half-)verse (cf. p. 396). For him only v. 37a
certainly belongs to the old Exodus-story: it comes too early (i.e. before
vv. 40-41) to be from P and is ‘unverzichtbar’ for the non-Priestly account
(p. 203 n. 62; p. 208 n. 77, where v. 37b is regarded as dependent on the
other passages which give a similar number). In places Gertz puts vv. 35-36
on a similar plane (its theme is older than the plague-narrative: pp. 303-304;
cf. p. 187), but elsewhere he is clear that it comes from the Endredaktion
(pp. 183, 184, 186). Unlike Levin he does not defer the slaying of the firstborn
to a second stage of the plague-narrative,4 but he does regard the allusions to
Pharaoh in vv. 30aα, 31-32 as a secondary development, mainly because of
the supposed conflict with 10.29 (pp. 176, 183-84).5 Verses 33-34 are part
of the (expanded) non-Priestly account, but curiously not v. 39, for similar
reasons to Levin (p. 59 n. 130), or v. 38, which is said to depend on v. 37b
and on 10.24ff., which Gertz has already assigned to the Endredaktion (p. 208
n. 77).
The attribution of vv. 28, 40-41 to P is generally agreed, for good

reasons of continuity, language and duplication of the preceding
narrative account (see the Explanatory Notes). A possible problem
arises in relation to the chronology in vv. 40-41. It has been claimed
that this forms part of a chronology that makes world history last
for 4000 years, ending with the purification of the Jerusalem temple
in 164 B.C.: if this were correct it would mean that vv. 40-41 were
3
The counter-evidence (at least some of it) is actually listed in Schmid, p. 146
n. 533, and not at all convincingly dealt with.
4
This because he holds that ‘all my plagues’ in 9.14 can refer to it (alone): see
again the introduction to 9.13-35.
5
An additional reason for him is that Pharaoh’s request for a blessing and his
permission for the Israelites’ animals to leave take up earlier passages which he
has already (questionably!) ascribed to the Endredaktion (p. 176: cf. 8.4, 24; 9.28;
10.17, 24-26).
12.28-42, 50-51 105
only composed (or at least modified) after that date, much later
than any plausible date for P (cf. G. Larsson, ‘The Chronology
of the Pentateuch: A Comparison of the MT and LXX’, JBL 102
[1983], pp. 401-409; Schmid, Erzväter, pp. 20-22: for discussion
see J. Hughes, Secrets of the Times: Myth and History in Biblical
Chronology [JSOTSup 66; Sheffield, 1990], esp. pp. 234-35).
However, it is impossible for the figure of ‘430’ to be this late,
since it was already being modified in the Septuagint translation a
century earlier (see Text and Versions on v. 40): this modification is
presupposed by Demetrius in the late second century B.C. (Hughes,
p. 241). If the figures in MT as a whole were designed to point
to a climax in 164 B.C., it is most likely that this was done by
modifications to the ages at which Jared, Methuselah and Lamech
begot their sons according to Gen. 5.18, 25, 28, and the original
‘Priestly’ chronology (of which the figures in the P source in the
Pentateuch, it is suggested, were only a part) envisaged the era of
4000 years continuing for another 720 years after the beginning of
work on the post-exilic temple c. 538 B.C. (cf. Hughes, pp. 10-14,
44-45, 233-35, 265).6 On this view the figures could have been
established at any time after 538 B.C. (Hughes, pp. 51-54, suggests
‘late 6th or early 5th century’). A somewhat earlier date might also
be possible. Even supposing that the original chronology of P was
based on the suggested 4000-year era (which is not completely
certain, as the data for the history of Israel in Canaan come not
from the Pentateuch but from the Deuteronomistic History), one
might well conclude that the schematic fifty-year gap between the
destruction of the first temple and its rebuilding was in origin an
expectation (like the ‘seventy years’ in Jer. 25.11-12; 29.10, which
relate to a different if overlapping period), which happened, at least
if the date in Ezra 3.8 (cf. 5.13-16) is reliable, to be historically
close to the truth.
On the plausible assumption that vv. 43-49 are a later supplement
to P (see further the introduction to the next section of the commen-
tary) vv. 50-51 are best ascribed to the same stage of composition.
6
Schmid, who knew Hughes’s work, accepts this point (cf. p. 20 n. 116)
and in a later article he sees no difficulty in viewing vv. 40-41 as part of P
(‘The So-Called Yahwist and the Literary Gap between Genesis and Exodus’,
in Dozeman and Schmid [eds.], Farewell, pp. 29-50 [31, 46]). But in Erzväter,
pp. 21-22, he seems to maintain that the verses are part of a much later ‘Gen-2Kön
insgesamt überblickende Redaktionstätigkeit (vgl. v.a. Ex 12,40f; 1 Kön 6,1)’.
106 EXODUS 1–18
They are clearly based on vv. 28 and 40-41 and simply provide a new
finale to the Priestly Passover pericope.7 There is no reason to date
them any later. There has been more discussion about v. 42, but it is
evidently meant to connect to v. 41 (cf. the connective ‘It’ [for Heb.
hwʾ]) and it ends with a distinctively Priestly phrase (‘throughout
their generations’). The objections to the connection being original
are not strong: the shift back from ‘day’ to ‘night’ conforms with the
focus of the original Priestly instructions in vv. 6-14, where a similar
shift in the opposite direction occurs in vv. 12-14; and the impres-
sion of ‘lameness’ may be due to a failure to recognise its function
as a powerful finale to the original Priestly account of Passover and
the departure, bringing together the event commemorated after its
occurrence and the enduring celebration of it by a neat play on the
different senses of Heb. šimmurîm (‘watching’ and ‘observance’:
see the Explanatory Note).
One other question that has been raised is whether P origi-
nally gave a fuller account of the process of departure, which was
suppressed by the redactor who combined P with the non-Priestly
account: so e.g. Holzinger, p. 35, with the specific observation that
‘on this day’ in v. 41 needs something to refer back to. Such a fuller
account could (in part at least) be reconstructed from vv. 12-13 and
from Num. 33.3-4, which is certainly based on some (surviving)
features of the Priestly narrative (such as Exod. 14.8: see also the
Explanatory Note on vv. 29-30). On the other hand, P may have
deliberately curtailed the narrative at this point, to maintain attention
on the central place of Passover in Israelite worship. The expression
‘on this day’ in v. 41 could easily be taking up the similar expression
in v. 14, which in the original independent Priestly text would have
been separated from it by only two verses; and Gertz has recently
pointed out that the sequence of obedience and a date in vv. 28
and 40-41 is actually paralleled at some other points in P (p. 58 n.
126: see also the Explanatory Note). It was once popular to assign
the itinerary-unit in v. 37a to P (see above), but it seems premature
before v. 41 (cf. Gertz, p. 203 n. 62) and Holzinger’s proposal that it
once stood after v. 41 (p. 31) is a counsel of desperation (see further
below on the origin of this and similar texts).
7
On the device of Wiederaufnahme see C. Kuhl, ‘Die “Wiederaufnahme” –
ein literarkritisches Prinzip?’, ZAW 64 (1952), pp. 1-11.
12.28-42, 50-51 107
As for the remainder of the passage (vv. 29-39), it is necessary

first to discuss the recent fragmentation of it in the studies of Levin
and Gertz (and less fully Schmid). It is, as earlier sections of the
commentary have argued, based on some very questionable interpre-
tations of this and other passages of Exodus. The attribution of the
references to Pharaoh to late redactors relies on a pedantic, unimag-
inative understanding of the tension between them and 10.28-29
(and 11.8), as many commentators have recognised.8 Levin’s view
that even the idea of the slaughter of the firstborn is secondary to
the narrative is based on his interpretation of 9.14, which involves
the unsupported conjecture that there was once a different ending
to the non-Priestly plague-narrative which was displaced by what
is now there. This is both unconvincing and unnecessary (see the
introduction to 9.13-35).9
The further view of these scholars that even the first introduc-
tion of the plagues motif into the Exodus-story occurred only in the
exilic period relies heavily on an argument of Van Seters which has
not so far been addressed in detail (for references see the summary
of Schmid’s views above). Van Seters argues that Deuteronomy
knows nothing of any plagues narrative, so it must be a later devel-
opment of the Exodus tradition. According to him, Deuteronomy
speaks only of ‘the diseases of Egypt’ (7.15; 28.60) and they refer
to ‘afflictions suffered by the Israelites’ (Life, p. 81): they are there-
fore quite different from the plagues inflicted on the Egyptians,
which Van Seters considers to be the result of a drastic reinterpreta-
tion of the ‘disease’ tradition by the Yahwist, seen in Exod. 15.25,
where the word ‘diseases’ is used to refer to the plagues. But is
it so clear that Exod. 15.25 was wrong to understand the diseases
of Egypt as (some of) the plagues suffered by the Egyptians? Van
Seters’s statement that ‘the Israelites experienced’ them in Egypt is
actually ambiguous: it might mean that they suffered from them or
it might mean that they observed them. The two verses that mention
them say that the Israelites ‘knew’ them (Deut. 7.15) or that they
8
It is significant that in outlining this argument Gertz has to begin with the
word ‘Strenggenommen’ (‘Strictly speaking’), which betrays the preference for
simple logic over literary sensitivity.
9
Gertz, who takes a similar view of 9.14 to Levin, evidently finds the impli-
cation unacceptable, but his way of reconciling 9.14 with the presence of the
slaughter motif is also far from convincing.
108 EXODUS 1–18
‘were in dread of them’ (Deut. 28.60). Both these expressions can

mean that they observed them; there is nothing in either that neces-
sarily implies that they actually suffered from them. The evidence
appealed to is at best inconclusive.
But there is also strong evidence that Deuteronomy did know
the plagues tradition in something like the form which we find in
the non-Priestly narrative. The ‘signs and wonders’ which Deuter-
onomy repeatedly says accompanied the Exodus from Egypt could
very well be a reference to it (4.34 [not 4.32 as Van Seters has];
6.22; 7.19 [not 7.9]; 11.2-3; 26.8; 29.2). Taken alone the phrase
could theoretically mean something else, for example a miraculous
healing or provision of food. But Van Seters is positively misleading
when he says, ‘There is no indication that within Deuteronomy this
has specific reference to the plagues’ (Life, p. 81). In the passages
mentioned, and several others, expressions are used which much
more clearly point to repeated events which demonstrated Yahweh’s
power against the Egyptians. Deuteronomy 6.22 actually speaks of
‘great and awesome [Heb. rāʿîm, ‘evil’] signs and wonders against
Egypt, against Pharaoh and all his household’. 11.2 prefixes the
reference to Yahweh’s ‘signs and deeds’ in v. 3 by speaking of ‘his
greatness, his mighty hand and his outstretched arm’ (cf. 5.15; 6.21)
and v. 3 particularises the signs and deeds as those ‘that he did in
Egypt to Pharaoh the king of Egypt, and to all his land’ (cf. 34.11).
This cannot be a reference to the drowning of the Egyptian army
in the ‘Red Sea’, since the latter is mentioned separately in v. 4.
26.8 adds the expression ‘with great terror’, which might well be
a specific reference to the slaying of the firstborn and its impact as
they are described in Exod. 12.29-33. Deuteronomy 4.34, which
uses this expression too, also adds the description of Yahweh’s
‘signs and wonders’ as ‘trials’ or ‘tests’ (Heb. massôt), a term that
occurs in similar lists in 7.19 and 29.2 and probably points to the
way that the plagues in the non-Priestly narrative in Exodus seek
to shift Pharaoh from his obstinate unwillingness to ‘let Israel go’.
In the face of all this evidence there can scarcely be any doubt
that the plague-tradition is pre-Deuteronomic. More general
arguments based on the history of Yahwism, such as the denial of
belief in Yahweh as a god who could challenge a powerful foreign
king in this way before the exile (Schmid, pp. 143-45), are no more
persuasive, in the light of the ideas of prophets such as Amos and
12.28-42, 50-51 109
Isaiah and indeed the old Exodus tradition itself. It is noteworthy

that not only a more ‘traditional’ German scholar like Graupner and
an American such as Propp but also Erhard Blum (cf. Studien, pp.
13-19) have no hesitation in assigning the slaughter of the firstborn
motif, like most of the non-Priestly plague-narrative, to a pre-exilic
author.10
The close connection between vv. 29-32 and 11.4-8 should lead
to their attribution to the same underlying source, in our view E.
This need not be the case for vv. 33-39, and the duplication between
v. 33 and vv. 31-32 perhaps supports its ascription to a different
account, as envisaged by Smend, Eissfeldt, Beer and Fohrer. It is
closely associated (by the emphasis on haste) with the aetiology of
the Feast of Unleavened Bread in vv. 34 and 39. Since we have not
hitherto found any need for the hypothesis of a ‘third early source’,
the obvious candidate for these three verses is J. Verses 35-36 are
a retrospective parenthesis with a similar content to 11.2-3 and
‘according to Moses’ word’ is probably a reference back to that
passage, which we have found reason to attribute to a redactor who
knew the combined non-Priestly (‘JE’) narrative. These verses
cannot be any earlier. There is no more sign here than there of
Deuteronomic affinities, so Blum’s derivation of these verses from
Kd lacks any basis. In a previous study (‘The Wilderness Itineraries
and the Composition of the Pentateuch’, VT 33 [1983], pp. 1-13) we
have attributed v. 37a to a Deuteronomistic ‘itinerary-redaction’,
in contrast to most scholars since Baentsch, who have seen it as
part of an early narrative source (see the Excursus below). If this
is correct, then v. 37b lacks a connection to the preceding context,
but it is easy to suppose that a phrase like ‘They were’ (Heb.
wayyihyû) was displaced when the itinerary-note was inserted. If
so, then vv. 37b-38 could belong to the same parallel account as
vv. 33-34 and 39. But it is also possible that they were originally
linked to vv. 29-32, perhaps with a verb like ‘They departed’ (Heb.
wayyēṣeʾû).
10
The arguments for a late date for particular verses in 12.29-39 brought
forward by Levin, Schmid and Gertz are generally based on their association
with other verses or passages to which they have already assigned a late date on
grounds which we have contested elsewhere. In some cases the direction of any
dependence is not clear, which further undermines the argument.
110 EXODUS 1–18
Excursus on the ‘Wilderness Itinerary’

Exodus 12.37a is the first of a series of statements (‘itinerary-notes’) which
define the stages of the Israelites’ journey out of Egypt and through the
wilderness (cf. 13.20; 14.9 [cf. v. 2]; 15.22-23a, 27; 16.1; 17.1; 19.1-2).
This is part of a larger body of material which takes the journey on to the
borders of Canaan (Num. 25.1; cf. Josh. 3.1; 4.19), with some duplication
and variation (cf. Coats, ‘Wilderness Itinerary’). A comprehensive account
of the journey is given in Num. 33.1-49, where the characteristic two-part
formula for departure and arrival is used repeatedly, with occasional brief
amplification to note conditions or events en route. It has been shown that
itineraries of this and other forms were a type of administrative document
that was widespread in the ancient Near East and the classical world (in my
‘The Wilderness Itineraries: A Comparative Study’; additional examples in
A.R. Roskop, The Wilderness Itineraries: Genre, Geography and the Growth
of Torah [HACL 3; Winona Lake, 2011], Chapter 3). It is also clear that such
documents were used and imitated in more literary accounts of journeys in
Mesopotamia, Egypt and the classical world (cf. Davies as above, pp. 57-70;
Roskop, Wilderness Itineraries, Chapter 4). Within the OT Claus Westermann
observed a similar use of itinerary-formulae in the patriarchal narratives
(Genesis 12–36, pp. 49-51, 69-73).
Such evidence naturally has a bearing on the interpretation of the ‘itinerary-
notes’ in the narratives of Exodus and Numbers in particular. Do these
sections comprise a single ‘wilderness itinerary’ or parts of several, which
have been combined together like the other source-material in the Pentateuch?
And what is the relationship between them and the continuous and apparently
complete list of journey-stages in Numbers 33? Is Numbers 33 a compilation
from the data in the main narrative of Exodus and Numbers, supplemented
with other place-names which do not appear there? Or is it the source from
which some at least of the data in the main narrative were drawn? At one time
it was common to distribute the itinerary-notes among the main Pentateuchal
sources, especially J and P. But this sometimes led to disagreement and to the
allocation of adjacent itinerary-notes to different sources. It is therefore best
to begin by associating the different notes into a series of connected ‘strings’,
before considering the sources or redactional layers to which they belong.
According to F.M. Cross and G.W. Coats the majority of the notes form part
of the Priestly Work or a very late redactional layer (cf. Cross, Canaanite
Myth, p. 308; Coats, ‘Wilderness Itinerary’, pp. 143, 146-47). Close examina-
tion of the duplicate accounts of the arrival at Mount Sinai (Exod. 19.1-2) and
the departure from there (Num. 10.12, 33) suggests that the Priestly material
runs parallel to two other strands of itinerary-notes.11 By extrapolation it is
11
For what follows see my unpublished dissertation ‘The Wilderness Itiner-
aries in the Old Testament’ (Cambridge, 1975), pp. 47-119, and ‘The Wilderness
Itineraries and the Composition of the Pentateuch’. An analysis along similar
12.28-42, 50-51 111
possible to distinguish: (a) a limited number of Priestly itinerary-notes (Exod.

19.1; Num. 10.12; 20.1aα; 22.1); (b) a series of passages related in some
way to Num. 33.1-49 (Exod. 12.37a; 13.20; 14.2, 9; 15.27; 16.1; 17.1; 19.2a;
Num. 11.35; 12.16; 20.1b, 22; 21.4a, 10-11; 25.1); (c) other itinerary-material,
belonging either to one of the non-Priestly sources (Exod. 15.22-23a; 19.2b;
Num. 10.33) or to a redactional layer (Num. 21.12-20; Deut. 10.6-7).
The most likely explanation for the origin of group (b), in my opinion, is
that it was taken from an earlier form of Num. 33.1-49, before it underwent
supplementation by a redactor familiar with the Priestly Work.12 This original
form of Num. 33.1-49 was probably a pre-exilic document modelled on
Mesopotamian itineraries and making use of knowledge of routes, perhaps
used by traders, in eastern Egypt, the Sinai peninsula and southern Transjor-
dan.13 A study of the way in which extracts from Num. 33.1-49 have been
incorporated into a passage which makes the Israelites travel to the east of
Edom and Moab (Num. 21.10-20) reveals an affinity to parts of the speech of
Jephthah in Judg. 11.14-18 and also to one strand of Deuteronomy 2–3, which
suggests that the incorporation of the main sequence of itinerary-notes into
Exodus and Numbers was the work of a Deuteronomistic redactor.14
This short narrative comprises one of the central episodes in the

book of Exodus, along with the final defeat of Pharaoh at the sea in
chs. 14–15 and the encounter with Yahweh at Mount Sinai in chs.
19–20 and 24. Now at last the tension set up in ch. 1 by Pharaoh’s
measures to subdue and even eliminate Israel, Yahweh’s people, is
resolved and they are free to leave Egypt and recover their liberty.
The significance of this narrative is underlined by the regulations
for the commemoration of the event which precede and follow it
(Exod. 12.1-27, 43-49; 13.1-16) and are even hinted at within it
(vv. 34, 39, 42). Two features of the deliverance are emphasised in
both the Priestly and the non-Priestly accounts, as they are also in
the regulations earlier in ch. 12: it is all Yahweh’s doing (vv. 29, 42:
lines, but with different conclusions, was published by J.T. Walsh, ‘From Egypt
to Moab: A Source-Critical Analysis of the Wilderness Itinerary’, CBQ 39 (1977),
pp. 20-33.
12
On the special problems of Exod. 14.2, 9 see the notes on those verses in
this commentary.
13
For an important recasting of the order of the list see my ‘The Wilderness
Itineraries and Recent Archaeological Research’, in J.A. Emerton (ed.), Studies in
the Pentateuch (VTSup 41; Leiden, 1990), pp. 161-75 (172-74).
14
Cf. Davies, ‘The Wilderness Itineraries and the Composition of the Penta-
teuch’, pp. 9-12. For a different view of the relationship of the itinerary-notes to
Num. 33.1-49 see Roskop, Wilderness Itineraries, Chapters 5–7.
112 EXODUS 1–18
cf. v. 36) and it takes place at night (vv. 29-31, 42). In various ways
the narrative also looks to the future. The Israelites do not leave
empty-handed: they take the flocks and herds that they had brought
with them (vv. 32, 38: cf. Gen. 46.6) and also the ‘plunder’ which
the Egyptians have freely given them (v. 36). The family which
came into Egypt departs as a great multitude, joined by others who
are glad to leave Egypt behind (vv. 37b-38). And in words which
now ironically come from Pharaoh himself, they are going in order
to worship Yahweh and they are bidden to call down his blessing on
the very foreign ruler who has oppressed them (vv. 31-32: cf. Gen.
12.1-3).
28 The Israelites went and did as Yahweh had commanded

Moses and Aarona—so they didb. 29 At midnight Yahweh, he
struckc every firstborn in the land of Egypt, from the firstborn of
Pharaoh who sat on his throne to the firstborn of the prisonersd
who were in the dungeone and all the firstborn of the (farm)
animals. 30 Pharaoh got up in the night, he and all his servants
and the whole of Egypt, and there was a great cry in Egypt,
because there was no house where there was not someone dead.
31 He summonedf Moses and Aaron in the night and said, ‘Arise,
departg from among my people, both you and the Israelites, and
go, worshipg Yahweh as you have said. 32 Also your flocks
and herds,h take them as you said and go—and give me too
your blessing.’ 33 [The Egyptians pressedi the people, quicklyj
sending them away from the land, for they said, ‘We are all dead
men’. 34 The people carried their dough before it was leavened,
with their kneading-bowlsk tied up in their garments, on their
shoulders.] 35 [The Israelites had donel according to Moses’
word and asked the Egyptians for objects of silver and gold and
for garments, 36 and Yahweh had givenl the people favour in
the sight of the Egyptians: they had granted their requestm, and
so they plunderedn the Egyptians.] 37 [The Israelites journeyedo
from Ramessesp to Succoth,] [about six hundred thousand on
footq, (counting) the menr without their dependantss. 38 In addi-
tion a large body of foreignerst went up with them, and flocks and
herds, very many animals. 39 They baked the dough which they
had brought out from Egypt into unleavened cakesu, because it
was not leavened, since they were driven out of Egypt and they
could not delay or even get provisionsv for themselves.] 40 The
timew that the Israelites dwelt in Egypt was four hundred and
thirty years. 41 At the end of four hundred and thirty yearsx, on
12.28-42, 50-51 113
this very dayy, all Yahweh’s tribal divisions departed from the
land of Egypt. 42 It was a night of watchingz for Yahweh to bring
them out of the land of Egypt. That is this nightaa for Yahwehbb, an
observance for all the Israelites throughout their generations…
50 [All the Israelites did as Yahweh had commanded Moses and
Aaron—so they didcc, 51 and on this very daydd Yahweh brought
the Israelites out from the land of Egypt in their tribal divisions.]

a. Heb. ‫ואהרן‬, without repetition of ‫ את‬as in v. 50 (cf. ‫ ואל‬in v. 1). Such
exceptions to the normal rule are rare, and this seems to be the only exception
in Exodus. There are apparently none in Leviticus and only one in Numbers
(32.35). In Chronicles the variation is a little more common, both in the
genealogies (1 Chr. 1.32; 2.13; 5.41; 8.12) and in the narratives (1 Chr. 18.1;
2 Chr. 13.15; 28.18; 33.9; 34.3).
b. The sentence structure, with resumption of the original statement at the
end, is typically Priestly: see Paran, Forms of the Priestly Style, pp. 47-97
(‘the circular inclusio’: cf. summary, pp. viii-x), where he notes other exam-
ples in 12.8 and 12.14, and especially pp. 234-35 on its use in concluding
formulae, as here and in v. 50.
c. Heb. ‫ויהוה הכה‬. The S-P word-order here may simply introduce a new
stage in the story as in 1.7 (3.1 is more complicated: see Note a on the trans-
lation of 3.1-12). But after the opening temporal expression it is more likely
meant to highlight the intervention of Yahweh: for similar examples (but
mainly in discourse) see Muraoka, Emphatic Words, p. 35; JM §155ne. The
verb ‫ נכה‬Hiph. last appeared in the P section 12.12-13 and does not appear in
the otherwise closely parallel passage 11.4-8, but see 3.20; 7.17, 20, 25; 9.15
in the non-Priestly narrative.
d. Heb. ‫השׁבי‬, which is generally collective when it refers to people rather
than a condition.
e. Heb. ‫בבית הבור‬. ‫בור‬, lit. ‘pit, cistern’, alone can mean ‘dungeon’ (Jer.
38.6-13 [7x]), but the combination with ‫ בית‬also occurs at Jer. 37.16, where the
context makes clear that a building had been converted into a prison.
f. Taken alone ‫ ויקרא‬might have an indefinite subject and be equivalent to
‘(Moses and Aaron) were summoned’, but the following ‫ ויאמר‬and the suffixes
referring to Pharaoh in vv. 31-32 make this less likely.
g. In both cases of asyndeton here the first imperative (‫קומו‬, ‫ )לכו‬is one
which GK §120g describes as equivalent to an interjection, but in the context
of the departure narrative they probably retain their normal verbal force.
h. The fronting of the objects in this clause draws attention to the contrast
with Pharaoh’s previous refusal to let the Israelites’ animals accompany them
(10.24).
114 EXODUS 1–18
i. Heb. ‫ותחזק‬. Names of countries which are used for the population can
retain their fem. gender (GK §122i), but the m.pl. is much more common in
Exodus (Propp, p. 413). For ‫חזק‬, ‘be strong’, in the sense ‘put pressure on (‫’)על‬
cf. Ezek. 3.14. Another possibility is ‘be determined’, as in Deut. 12.23 with a
following inf. But the infinitives here can be rendered gerundivally.
j. Heb. ‫למהר‬. Whichever sense is given to ‫ותחזק‬, ‫ מהר‬may here have its
adverbial or auxiliary sense, as in Gen. 27.20 (cf. JM §102g and the Hiphil
verbs treated in GK §114n, with n. 2). Alternatively (see Text and Versions)
we may follow TgO,J and render ‘to hurry, so as to send them away’.
k. Heb. ‫משׁארותם‬. The word is rare (elsewhere only in 7.28; Deut. 28.5,
17) and its root uncertain (see Note g on the translation of 7.26–8.11 and
BDB, HAL). The present context gives the clearest indication of its meaning.
It introduces an asyndetic circumstantial clause in which the part. gives a
superficial resemblance to the genitive absolute construction in Greek and the
ablative absolute in Latin (for the absence of waw cf. GK §156a,c).
l. The S-P word-order here most likely indicates a pluperfect sense for the
verbs (GK §142b; cf. 106f). On the idiom at the beginning of v. 36 see Note
m on the translation of 3.16-22.
m. The Hiphil of ‫ שׁאל‬occurs elsewhere in MT only at 1 Sam. 1.28, of
Samuel’s dedication to Yahweh by Hannah, in close connection with the Qal
pass. part. of ‫ ;שׁאל‬cf. the use of the Qal perf. and the noun ‫ שׁאלה‬in 1 Sam. 2.20,
where 4QSama reads [‫ השׁאיל‬for the verb, which is surely to be completed to
the ‫ השׁאילה‬that was first conjectured by K. Budde in 1902 (BDB, p. 982; cf.
BHS) rather than the misformed ‫ השׁאילת‬proposed on the basis of the Vss in
DJD XVII, pp. 39-42. The sense ‘lent’ has sometimes been favoured in these
places (so still NRSV at 1 Sam. 1.28), on the basis of MH usage (for which see
Jastrow, p. 1507) and occurrences in Nabataean and Palmyrene (BDB, p. 981;
DNWSI, p. 1097). But the context in 1 Sam. 1.28 is strongly against the idea
of a temporary loan (cf. P.K. McCarter, 1 Samuel [AB 8; Garden City, 1980],
pp. 63, 80, 84: ‘dedicate’) and it is likely that the meaning is ‘grant a request’
and then more generally ‘give’ (cf. BDB, p. 982; H.P. Smith, The Books of
Samuel [ICC; Edinburgh, 1904], pp. 14, 19). HAL, pp. 1278-79, accepts this
for Exod. 12.36 but retains ‘lend’ for 1 Sam. 1.28; likewise Ges18, p. 1308
(both with ref. to an Ar. idiom in the former case). DCH 8, p. 216, is similar,
but adds ‘perh. make over, consecrate, dedicate’ for 1 Sam. 1.28 and 2.20
(with reference to McCarter for support on p. 711).
n. On the meaning of the Piel of ‫ נצל‬here see Note t on the translation of
3.16-22.
o. Heb. ‫ויסעו‬, with ‫ נסע‬in its secondary but common sense of ‘make a
journey’, marked by the association of a destination as well as or instead of
the place of departure (cf. Gen. 12.9; 20.1).
p. MT vocalises the name ‫ ַר ְע ְמ ֵסס‬here (as in Gen. 47.11; Num. 33.3, 5), in
contrast to ‫ ַר ַע ְמ ֵסס‬in 1.11.
12.28-42, 50-51 115
q. Heb. ‫רגלי‬, used in the same expression in Num. 11.21. With one excep-
tion (Jer. 12.5, which may perhaps not be an exception: horses were mainly
used in warfare in the ancient Near East [ABD 6, pp. 1136-37]) all the other
occurrences of the term are in military contexts, so it may contribute to the
picture of the Exodus as an army on the march (cf. my ‘The Wilderness Itiner-
aries: A Comparative Study’, p. 80 and the note below on 13.18), though one
that lacked a chariot-force like that sent to pursue them in 14.6-7 (cf. Schmidt,
p. 539).
r. Heb. ‫הגברים‬, as in 10.11 to designate the males clearly: ‫ אישׁ‬is not neces-
sarily so specific.
s. Heb. ‫מטף‬. On the meaning of ‫ טף‬see Note y on the translation of 10.1-20:
the case for ‘dependants’, whose definition may vary according to the other
terms mentioned in the context, is made strongly by P. Swiggers, ‘Paradig-
matical Semantics’, ZAH 6/1 (1993), pp. 44-54 (45-47).
t. Heb. ‫ע ֶרב‬.ֵ The word occurs elsewhere in Jer. 25.20; 50.37; Neh. 13.3;
less certainly in 1 Kgs 10.15; Jer. 25.24; Ezek. 30.5. It is related to ‫ ערב‬II,
‘associate with, share in’ (HAL, p. 830; Ges18, p. 1008) at Ps. 106.35; Prov.
14.10; 20.19; Ezra 9.2 (and at Qumran: DCH 6, pp. 546-47), which BDB
wrongly includes under the homonym for ‘give a pledge’ (p. 786). In Neh.
13.3 ‫ ֵע ֶרב‬may mean the children of mixed marriages (H.G.M. Williamson,
Ezra, Nehemiah [WBC; Waco, 1985], p. 488) or more generally people
of non-Israelite descent (e.g. NRSV), who were to be excluded from the
community in future. In Jer. 25.20; 50.37 it seems to refer to non-Egyptians
and non-Babylonians who had attached themselves to these peoples. Here
it presumably means non-Israelites who also took the opportunity to leave
Egypt. A similar recognition of the presence of such people among Israel’s
ancestors is found in Num. 11.4 (Heb. ‫)אספסף‬, but not in Priestly texts. See
also Text and Versions. The S-P word-order may again imply a pluperfect
sense for the verb (cf. Note l) or it could simply be drawing extra attention
(like ‫ )גם‬to the (unexpected?) presence of foreigners in the Exodus group.
u. Heb. ‫עגת מצות‬. The phrase (which occurs only here) is a second accusa-
tive of the product (cf. GK §117ii: after ‫ אפה‬also in Lev. 24.5; 1 Sam. 28.24
[here also ‫ )]מצות‬and ‫ מצות‬is in apposition for closer specification of ‫( עגת‬GK
§131b).
v. Heb. ‫וגם צדה‬. This time the object is fronted to underline the people’s
haste (GK §142f[a]), again with ‫גם‬: not only could they not wait to bake bread,
they could not even collect food that was ready to eat. The disturbance of the
normal word-order is unusually widespread in this lively passage.
w. Heb. ‫ מושׁב‬does not mean ‘period of residence’ (BDB, HAL, Ges18,
DCH) elsewhere, so the abstract meaning ‘residence’, as in Lev. 25.29; Num.
15.2; Ps. 107.4, should perhaps be preferred here, with the temporal expres-
sion then understood not as a simple complement but as an ‘accusative of
duration’ (GK §118k), ‘for 430 years’.
116 EXODUS 1–18
x. Lit. ‘thirty years and four hundred years’, as also in v. 41: the repetition
of ‘years’ (‫ )שׁנה‬is characteristic (apart from 1 Kgs 6.1) of the Priestly writings
and the placing of the smaller number before the larger one is also normal
there, with some occurrences also in Kings and Ezekiel (cf. GK §134h-i).
The origin of these distinctive patterns remains unclear: they do not seem to
correspond to epigraphic, Aramaic or Babylonian practice.
y. Heb. ‫בעצם היום הזה‬: see Note ss on the translation of 12.1-20.
z. Heb. ‫ליל שׁמרים‬. ‫ליל‬,ֵ from the shorter form ‫( ַליִ ל‬as at Isa. 16.3), is the only
construct form used for ‘night’ (elsewhere only at Isa. 30.29) and sometimes
serves for the absolute state as well (Isa. 15.1; 21.11). ‫( שׁמרים‬for the pl. form
see JM §136i) occurs in BH only here and in the second half of the verse, with
a play on two different senses of ‫שׁמר‬, ‘keep’: first ‘watch, keep vigil’ (as in the
nominalised part. ‫שׁ ֵֹמר‬, ‘watchman’, Ps. 127.1 etc.) and then ‘observe’, as of a
festival (e.g. v. 17; 31.13). The past tense (RSV, JB, NJPS, NRSV) is prefer-
able to a present (LXX, Vulg, Luther, AV, RV), which would duplicate v. 42b.
aa. Heb. ‫הוא הלילה הזה‬. The exact grammar of this verbless clause is
difficult to penetrate, although the overall sense seems clear enough: the
correlation in identity and meaning between the night of Israel’s deliverance
and Passover night throughout the ages, as in vv. 17, 41 and 51. EVV. tend
to paraphrase. BDB curiously groups ‫ הוא‬here with cases like Ezek. 3.18 [it
mistakenly has 3.8] and 33.8 where it is unusually an adjective placed before
a noun (p. 215) but none of the other examples has another demonstrative after
the noun. 6.27 (which BDB also includes there) and 6.26 are also different:
there ‫ הוא‬is a neuter ‘it’ (cf. Note s on the translation of 6.10-7.5, and the exx.
in BDB, p. 216, s.v. 5) and ‘it’ or better ‘that’ also fits here well, probably best
with the copula supplied immediately after it, rather than later in the sentence
as in NRSV (cf. JB). The ‘this’ seems to imply either a close association with
the Passover laws in vv. 1-14 or the intention that these words should be read
on Passover night (or both).
bb. Heb. ‫ליהוה‬, here with the same implication of something set aside ‘for’
Yahweh as in vv. 11 and 14 (in v. 42a the sense is different). The Masoretic
accents place a minor break in the verse here, which the paraphrases of RSV,
JB and NRSV ignore (NJPS, NEB and REB observe it).
cc. See Note b above.
dd. See v. 41 and Note ss on the translation of 12.1-20.
Explanatory Notes
28. A short section of narrative begins, before the introduction
of further legal material from v. 43 onwards. First the Israelites’
fulfilment of the instructions about Passover is recounted in summary
form. The mention of Yahweh and Aaron (cf. v. 1) indicates an orig-
inal connection with vv. 1-20 (and see Note b on the translation
12.28-42, 50-51 117
for the Priestly style of v. 28), but the placing of the verse after
vv. 21-27 now integrates the two sets of instructions and presents
Moses’ words to the elders as the way in which Yahweh’s instruc-
tions were transmitted on this occasion (cf. Childs, Houtman). Such
summary statements of obedience commonly follow the divine
instructions immediately (cf. 12.50; Num. 1.54; 2.34; 5.4) and no
doubt this was originally the case here too.
29-30. Now the promised intervention of Yahweh and the
immediate reaction of the grieving Egyptians are described, in words
that follow closely Moses’ warning in 11.4-6: but the wording at the
beginning of v. 29 is actually closer to 12.12 than 11.4-5, which
may indicate that the compiler preserved a few words of the Priestly
narrative here. Again it is emphasised that every Egyptian family
was affected, though with a different counterpart at the opposite
extreme from Pharaoh’s family (the narrator may have combined
variant versions of the story to heighten the picture of widespread
distress) and an explicit statement at the end of v. 30.
31-33. Still ‘in the night’ (mentioned for the third time here in
v. 31) Pharaoh and the Egyptians take urgent action to let the Israel-
ites depart at last, indeed they command them to leave (cf. 11.8).
Pharaoh’s own intervention is contrary to what he had said (10.28)
and Moses had expected (10.29; 11.8); but, as argued earlier (in the
introduction to 11.1-10), this need not be a sign of different versions
of the story: it is one way in which the impact of the slaying of the
firstborn is accentuated. It is only by Pharaoh’s own involvement
that the complete turnaround from earlier episodes in the narrative
(5.1-4 and the earlier plague-stories) can be shown. This is also
accomplished by the repeated statement that it is the words of Moses
and Aaron (‘as you said’ in vv. 31 and 32: cf. 7.16 etc.) that are to
be determinative of the future, not Pharaoh’s own.15 What Pharaoh
now says contrasts directly with the limitations which he had earlier
placed on who would be permitted to take part in the journey to
worship Yahweh: instead of excluding the Israelites’ families
(10.10-11) or their animals (10.24) so that he would have hostages
to guarantee their return, Pharaoh’s willingness to let everyone go
is expressed by no fewer than four occurrences of the Heb. particle
15
The wording that Pharaoh uses here, however, is also very close to his own
earlier qualified agreement in 10.8, 11, 24.
118 EXODUS 1–18
gam, ‘also’, in vv. 31-32 (cf. Propp, p. 411). There is some debate
over whether Pharaoh is still expecting the Israelites to return after
‘three days’ as Moses had said (8.23) or is letting them go for ever.
The reference to worship (v. 31) and the twofold ‘as you said’ could
be taken to suggest the former (cf. Houtman, p. 199; Propp, p. 411),
but it is probably an unreal alternative: Pharaoh and his people now
just want to be rid of the Israelites. Since blessing was commonly a
kind of farewell (cf. Gen. 24.60; 47.10; 2 Sam. 13.25; 1 Kgs 8.66),
Pharaoh’s final words to Moses may be understood in this way. No
doubt there is more to them than this: Pharaoh had earlier asked
Moses and Aaron to ‘pray’ for the removal of the plagues (e.g.
10.17), and now in an even greater crisis he has to recognise that it
is with Moses’ God that the power to bless and curse really lies. It
is a total capitulation. It also recalls Jacob’s blessing of an earlier
Pharaoh in the Joseph story (Gen. 47.7, 10). On the language of
blessing and cursing in general see J.K. Aitken, The Semantics of
Blessing and Cursing in Ancient Hebrew (ANESSup 23; Leuven,
2007).
34. The verse belongs closely with v. 39, which might once have
followed it immediately, before the insertion of the parenthesis in
vv. 35-36 and the details in vv. 37-38 (see the note below). The
haste of the Israelites’ departure (cf. v. 33, to which this verse is also
connected by the designation ‘the people’) is, in a homely touch,
taken to have interrupted the preparation of food for the journey,
so that at the earliest opportunity bread is baked from unleavened
dough. The point is undoubtedly to provide an aetiology for the
association of the festival of Unleavened Bread with the Exodus (for
which cf. 13.3-10 and the notes there), although no explicit connec-
tion with that observance is made here. The legal texts that deal
with worship in the Book of the Covenant (23.15) and the second
collection of ‘covenant laws’ (34.18) make the same association of
the Exodus with this festival, rather than with the Passover, and it is
likely that this was the practice at one or more of the major Israelite
sanctuaries for much of the monarchy period.16
16
On the celebration of the Exodus at Bethel cf. 1 Kgs 12.28 and other
evidence discussed by J.F. Gomes, The Sanctuary of Bethel and the Configuration
of Israelite Identity (BZAW 368; Berlin and New York, 2006), pp. 173-80.
12.28-42, 50-51 119
35-36. The Heb. word-order marks these verses as an interrup-

tion of the narrative relating an earlier event (see Note l on the
translation). The episode is more briefly narrated in 11.2-3a, to
which ‘according to Moses’ word’ refers back. As was argued in
the introduction to 11.1-10, these related passages seem to repre-
sent an addition to the non-Priestly account of the Exodus which
draws on both its major sources. Compared with 11.2 there is one
notable difference here: the Israelites request and receive ‘garments’
from the Egyptians (this feature is also picked up in Yahweh’s
words to Moses in 3.22). It can scarcely be coincidental that this
is related immediately after the same Heb. word has been used
for the ‘wrapping’ of the dough in v. 34. The recapitulation of the
‘plundering of the Egyptians’ (on this phrase see the Explanatory
Note on 3.21-22) at just this point and with this addition could well
be designed to explain the origin of the additional ‘garments’ which
the Israelites had available to carry their dough.17 On the (improb-
able) view that the Israelites were simply ‘borrowing’ what they
took see Note m on the translation.
37-38. These verses specify more precisely three kinds of detail
about the Israelites’ departure from Egypt: their route, their number
and their ethnic composition. Verse 37a is the first of a sequence
of ‘itinerary-notes’ which stretches all the way to Mount Sinai (cf.
13.20; 14.2, 9; 15.22, 27; 16.1; 17.1; 19.1-2) and indeed beyond
(Num. 10.12, 33 etc.). The classic form, which has close parallels
in ancient Near Eastern descriptions of journeys, including military
campaigns, is represented in 13.20: ‘they set out from Succoth and
camped at Etham, on the edge of the wilderness’. Here the second
verb is omitted and the point of arrival is marked by a directional
ending rather than the preposition ‘in’ or ‘at’ (Heb. be). On such
documents and their use in the composition of the Exodus and
wilderness narrative see the introduction to this section and works
referred to there. The first of the names mentioned here, Rameses,
is presumably identical to the store-city of that name in 1.11: it was
the Egyptian capital in the 19th and 20th Dynasties, during which
the events lying behind the biblical Exodus-story are generally
17
The different spelling of ‘garments’ in 4QExc in vv. 34 and 35 (see Text
and Versions) might, if original, give additional support to a redactional origin for
vv. 35-36.
120 EXODUS 1–18
thought to have taken place (see further the Explanatory Note on

1.11; Bietak and Forstner-Müller, ‘Topography’, give an extensive
synthesis of current archaeological knowledge). The second name,
Succoth, which appears again in 13.20 (and like Ramesses, in the
summary of the whole wilderness journey in Num. 33.5-6), has the
appearance of being a Heb. (f. pl.) form, ‘booths’, as in one name for
the Israelite autumn festival (Deut. 16.13 etc.), and the name is so
explained where it is applied to a different place in the Jordan valley
(Gen. 33.17). But since 1875 Succoth in Exodus has been associ-
ated with a place-name Ṯkw that occurs several times in Egyptian
texts (H. Brügsch, ‘Geographica’, ZÄS 13 [1875], pp. 5-13 [7-9]; cf.
E. Naville, The Store-City of Pithom [London, 1885], pp. 5-6, 23).18
The occurrences include two well-known texts relating to the eastern
border of Egypt (cf. ANET, p. 259) and objects inscribed with the
name have been found at two sites suggested for identification with
Pithom in 1.11, Tell Maskhuta and Tell er-Retabeh (on these see the
Explanatory Note there: for full reviews of the Egyptian evidence
for Ṯkw see LexAeg 6, p. 609; E. Bleiberg, ‘The Location of Pithom
and Succoth’, Ancient World 6 [1983], pp. 21-27; Hoffmeier, Israel
in Egypt, pp. 179-81). The evidence points to a location for Ṯkw
in the Wadi Tumilat, west of Ismailiya: it was the name both of a
region and of a city (either is possible here), the latter certainly from
the 22nd Dynasty (c. 945–715 B.C.), possibly earlier. It was impor-
tant militarily and the existence of a route between it and Ramesses
is no surprise, only the fact that the departing Israelites are supposed
to have travelled on it. It could have been another name for Pithom,
but this is not certain.
The number of ‘about six hundred thousand’ correlates broadly
with the figures from the two Priestly censuses in Numbers 1, 3
and 26, where again only adult males are counted (except for the
Levites), but the direction of dependence between Exodus and
Numbers is not entirely certain. There is no sign of Priestly vocabu-
lary in the present verse or its immediate context (for aspects of
the wording cf. 10.10-11 and Num. 11.21), and it is perhaps most
18
The spelling Ṯkwt seems to occur only in texts from the last centuries B.C.:
cf. H. Gauthier, Dictionnaire des noms géographiques (Cairo, 1925–31) 6, p. 83,
where an extensive list of occurrences is given. Gauthier, like some others at the
time, rejected any connection of Ṯkw with biblical Succoth.
12.28-42, 50-51 121
likely that the very precise census figures in Numbers were created
so that the totals would match the round figure in Exodus, which
could itself be much older: already in 2 Sam. 24.9 a much larger
figure was being given for the population of David’s kingdom. The
figures are in any case impossibly high (and they cannot plausibly
be reduced by suggestions that, for example, Heb. ʾelep here means
not ‘thousand’ but ‘clan’ or ‘military unit’: see for references Propp,
p. 414, to which add E.W. Davies, ‘A Mathematical Conundrum:
The Problem of the Large Numbers in Numbers i and xxvi’, VT 45
[1995], pp. 449-6919) and are meant to magnify the wonder of the
events of the Exodus and the wilderness journey, and perhaps also
the growth of the people during their time in Egypt (Dozeman, cf.
1.5).
The mention of ‘a large body of foreigners’ who accompanied
the Israelites on their departure from Egypt is surprising and finds
no parallel in the Priestly version of the story, where only ‘the tribal
divisions’ of Israel/Yahweh are involved (vv. 41, 51: cf. the layout
of the wilderness camp in Num. 2.1–3.39).20 But such a group is
referred to again in Num. 11.4 (Heb. hāʾsapsup, lit. ‘the gathered
company’), Deut. 29.10 and Josh. 8.35 (both times haggēr, ‘the
resident aliens’). Thus the non-Priestly texts trace back to the
Exodus period the mixed ethnic composition of Israel which is
evident at various points in the later biblical narrative (e.g. Uriah
the Hittite in David’s army in 2 Sam. 11.3 etc.) and has been empha-
sised much more by modern historians, who have come to view
the later Israelite population as composed partly or even largely of
former ‘Canaanites’ (so already W.F. Albright, Archaeology and the
Religion, p. 99; more recently W.G. Dever, Who Were the Early
Israelites and Where Did They Come From? [Grand Rapids, 2003],
chs. 9–11). The Heb. word used here (ʿēreb: see Note t on the trans-
lation) generally refers to foreign minorities in large states, and
Egypt throughout most of the second millennium B.C. had group-
ings of this kind, especially in the Delta region (cf. Aldred, The
19
The subsequent exchanges between C.J. Humphreys and other scholars in
VT 48 (1998), pp. 196-211; 49 (1999), pp. 131ff., 262ff.; 50 (2000), pp. 250-52,
323ff.; 51 (2001), pp. 392ff. do nothing to make these suggestions more plausible.
20
Some Priestly legal texts do include the ‘resident alien’ (Heb. gēr: e.g. Exod.
12.19, 43-44, 48-49), but these are generally held to be secondary additions to P
which refer to a later time.
122 EXODUS 1–18
Egyptians, pp. 139-40). It is not made clear whether the ‘flocks and
herds’ belonged to the Israelites, the foreigners or both, but passages
earlier in Exodus assume assume that the Israelites had their own
animals (9.4, 6; 10.9, 24-26; 12.3ff., 21: cf. Gen. 45.10; 46.6, 32-34;
47.1, 4; 50.8).
39. See the notes on v. 34. This verse clearly also presupposes the
narrative in v. 33. On the textual problem of ‘they were driven out
of Egypt’ see Text and Versions. The same verb ‘driven out’ (Heb.
grš) is used in 6.1 and 11.1 (and, in a more restricted sense, in 2.17
and 10.11).
40-42. The story of the Exodus is brought to a conclusion here
by a note of the duration of the Israelites’ residence in Egypt
(vv. 40-41) and a carefully constructed ‘diptych’ which relates
Yahweh’s deliverance of Israel to Israel’s ongoing observance of
Passover night (v. 42).21 The figure of ‘four hundred and thirty
years’ is not a computation from figures already given in the narra-
tive. It presumably begins with the arrival of Jacob in Egypt in
Genesis 46 and he is said to have lived in Egypt for seventeen years
(Gen. 47.28; cf. v. 8). At the end of the period Moses, who was
born in Egypt, is said to have been eighty years old at the beginning
of his confrontation with Pharaoh (Exod. 7.7). The length of the
intervening period is nowhere specified with any precision: it was
simply long enough for the family of Jacob to grow into a numerous
people (Gen. 47.27; Exod. 1.7, 9). The genealogy in Exod. 6.14-25
includes some lifespans which between them (more than) cover
this period, but despite the longevity of the persons concerned they
could not have given rise to the total figure of 430 years. In fact
the genealogy was responsible, together with the reference to ‘the
fourth generation’ in Gen. 15.16, for some early attempts to reduce
that figure by a half (for details see Text and Versions on v. 40).22
There is more likely some connection with the round figure of ‘four
hundred years’ in Gen. 15.13, but it is not certain whether that is the
basis for the figure here (as Propp supposes, p. 415) or derived from
21
The sense of finality is now rather spoiled by the partial repetition of these
verses in vv. 50-51, not to speak of the further conclusions of a different kind in
14.31 and 15.1-21.
22
The genealogy seems in any case to be a secondary insertion into the
Priestly narrative (see the introduction to 6.10–7.5), so it may well not even have
been available to the author of vv. 40-41.
12.28-42, 50-51 123
it, and if the former what its own rationale is (four generations of a
hundred years each is a possibility that would match Gen. 15.16).
A different approach relates the figure of 430 to the idea that there
is a larger chronological scheme underlying various eras in the Old
Testament (cf. 1 Kgs 6.1 and the more coherent series of dates in
Genesis). According to one such view an era of 4,000 years from
creation culminated in the rededication of the Jerusalem temple in
164 B.C. (Larsson, ‘The Chronology of the Pentateuch’: see the
introduction to this section). It is known that chronological calcula-
tions were being made around that time (cf. Dan. 9 and the book of
Jubilees), but it is unlikely that all the figures in the Old Testament
are this late.23 A more plausible suggestion is that the figure ‘430’
is modelled on significant figures in earlier biblical literature. Two
possibilities are the combined period symbolised by Ezekiel’s lying
on his side for first 390 days and then 40 days (Ezek. 4.4-8) and the
time that Solomon’s temple stood in Jerusalem as calculated from
figures given in Kings (which might even be connected): see briefly
Blenkinsopp, Pentateuch, pp. 47-50, and Propp, pp. 415-16, where
other suggestions are noted.
Such chronological interest is chiefly characteristic of the Priestly
source-material in the Pentateuch, and there are other features of
vv. 40-41 which confirm such an origin for them. ‘On this very
day’ and ‘Yahweh’s tribal divisions’ are both distinctively Priestly
phrases (see respectively 12.17 [cf. Gen. 7.13; 17.23, 26] and 7.4
[cf. 6.26; 12.17]): see also Note x on the translation. The formal
structure of the verses, with the duration of a period followed by a
statement of what happened ‘at the end of’ it, also matches closely
a sequence in the Priestly version of the Flood story (cf. Gen. 7.24;
8.3b-4: between these verses is the statement that ‘God remembered
Noah’, which recalls Exod. 2.24; 6.5).24 This implies an original
literary connection between vv. 40-41 and vv. 1-20 (and 28), so
that ‘On this very day’ will pick up the extended focus there on the
23
The figure in question here was already being corrected in the Septuagint
translation of Exodus a century earlier.
24
On the original reading at Gen. 8.3b see BHS. A similar sequence appears in
the non-Priestly version of the Flood story (Gen. 7.12; 8.6; cf. also Deut. 9.9, 11),
and the formula ‘At the end of…’ (Heb. miqqēṣ or miqqeṣēh) followed by a period
of time is itself quite widespread.
124 EXODUS 1–18
Passover night and its rituals.25 This is certainly how v. 42 under-

stands vv. 40-41: its initial ‘It’ must refer back to ‘this very day’, and
it very clearly (cf. ‘that is this night’) equates the night of Yahweh’s
deliverance of Israel with the night to be observed by the Israelites
as a memorial throughout their generations. On the play on words
between ‘watching’ and ‘observance’, which are the same word in
Heb. (šimmurîm) see Note z on the translation. The necessity for
the ‘observance’ of Passover gains its grounding here from the fact
that Yahweh himself had ‘watched’ for Israel on that night long ago.
50-51. Just as vv. 43-49 add further regulations for the obser-
vance of Passover (see the next section of the commentary), so
they too receive added authority from the statement that they were
carried out by the Exodus generation itself (v. 50, which is practi-
cally identical to v. 28) and from Yahweh’s deliverance of his people
‘on this very day’ (v. 51, which draws all of its wording, with small
variations, from vv. 41-42). The repetitions are a sure sign that the
preceding regulations were added by a later writer in the Priestly
tradition.
Text and Versions

‫( יהוה‬12.28) TgNmg adds ‘the Memra of’.
‫( ואהרן‬12.28) Von Gall erroneously records no variation from MT (cf.
Baillet, ‘Corrections’, p. 28) but ‫ ואת־אהרן‬is read by Tal, Sadaqa, all Crown’s
mss and Camb. 1846, which amply justifies regarding this as the standard
SP reading. It may still of course be secondary, whether by ‘grammatical
improvement’ or by assimilation to v. 50. No Qumran evidence is available.
TgJ,G have ‫ וית‬and Sy wl-, but they may merely be following standard Aramaic
practice. LXXA,B omit the καὶ Ἀαρών of the other mss, but Wevers (Notes,
p. 183) simply regards this as ‘odd’ and follows the majority, probably rightly
(so also Rahlfs): the omission may perhaps be due to the fact that Moses
25
Cf. Num. 33.3-4. This is more likely than the early Jewish interpretation
which saw ‘On this very day’ as locating the deliverance from Egypt exactly ‘four
hundred and thirty years’ to the day from some previous episode(s) in the biblical
narrative (see Text and Versions) and the view that the final day of the Feast of
Unleavened Bread (cf. v. 18) is meant (so Noth, p. 78, ET, p. 100; cf. Blenkinsopp,
p. 157). The verses about Unleavened Bread relate to later times and are probably
a secondary addition to the passage (see the introduction to 12.1-20). The phrase
‘on this very day’ in v. 17 must also refer to the Passover night, which is seen
there as the beginning of the feast (cf. v. 8 and the Explanatory Note on vv. 15-17).
12.28-42, 50-51 125
has just been speaking. On the other hand the omission of ‫ את‬in MT is very
unusual (see Note a on the translation) and it is at least possible that the refer-
ence to Aaron was only added by a careless scribe.
‫( כן עשׂו‬12.28) Vulg has no equivalent, probably because of the repetitive-
ness of the statement. TgJ prefixes ‫( אזדרזו‬from ‫)זרז‬, which means ‘hastened’
(AramB) or ‘were zealous’: the latter at least might be an attempt to explain
the repetition (MRI [Lauterbach, p. 96] mentions a view that it refers to Moses
and Aaron specifically).
‫( הלילה‬12.29) TgJ adds ‘of the fifteenth’: compare its additions to vv. 8,
10 and 18 and Text and Versions there.
‫( ויהוה‬12.29) Tg J,G,Nmg add ‘the Memra of’.
‫( הכה‬12.29) Tgg and Sy render ‘killed’, making the outcome clear as
earlier in vv. 12-13, 23 and 27.
‫( כל־בכור‬12.29) TgN,G and Sy render according to the sense in the pl.
(TgNmg restores the sing. of MT).
‫( בארץ מצרים‬12.29) Most Sy mss have ‘of the land of Egypt’, which might
be related to discussions about whether Egyptian firstborn living abroad were
affected (as in MRI [Lauterbach, p. 97]).
‫( מבכר‬12.29) TgJ ‫מביר בוכריה‬, ‘from his firstborn son’, as in 11.5, probably
to accord with the full wording of 4.23.
‫( הישׁב‬12.29) Tgg as in 11.5 relate this to Pharaoh’s son and understand it
of his future accession to the throne.
‫( על־כסאו‬12.29) On LXX’s reading, which Aq, Symm and Theod,
followed by the O-text and Vulg, assimilate to MT, see Text and Versions on
11.5. Most Tgg have as there ‘upon the throne of his kingdom’, but TgN reads
here ‘all the throne(s?) of the kingdoms’ (for ‘all’ instead of ‘upon’ cf. TgNmg
at 11.5), a phrase which occurs (without ‘all’) in MT at Hag. 2.22.
‫( עד‬12.29) On SP’s reading ‫( ועד‬cf. LXXmss, Sy) see Text and Versions
on 11.5.
‫( בכור־השׁבי‬12.29) TgO,N,G and Sy render with Aram. equivalents to Heb.
‫ שׁבי‬which can also have a collective as well as an abstract meaning (see
Note d on the translation). LXX αἰχμαλωτίδος and Vulg captivae, ‘woman
prisoner’, are probably harmonising with 11.5 (cf. BAlex, p. 151). TgJ has
a long expansion: ‘the firstborn of the sons of the kings who were captured
and kept as hostages by Pharaoh in the pit [for ‫]בבית הבור‬, and because they
rejoiced at the enslavement of Israel they suffered punishment too’. The
midrash combines the idea that foreigners in Egypt were included and the
justification given here, both of which appear separately in MRI (Lauterbach,
pp. 97-98), with a practice commonly used by powerful states to ensure
the loyalty of their vassals (cf. 2 Kgs 14.14, with the comment on Assyrian
practice in M. Cogan and H. Tadmor, II Kings [AB; New York, 1988], p. 157;
also Caesar, B.G.1.9).
‫( אשׁר‬12.29) TgG (AA, first version) amplifies with ‫דהוה יתיב‬: cf. TgNmg
‫דאית‬.
126 EXODUS 1–18
‫( בבית הבור‬12.29) LXX ἐν τῷ λάκκῳ and TgJ ‫בי גובא‬, ‘in the pit’, give
literal renderings of ‫ בור‬and ignore ‫( בית‬or take it to mean ‘within’), and Aq
and Theod extend such literalism with ἐν οἴκῳ τοῦ λάκκου. Symm ἐν τῷ
δεσμωτηρίῳ takes up a word used by LXX for another expression in Genesis
39–40 (cf. Vulg in carcere), providing an apt interpretation of the Heb. here
which is also found in TgO,N,G and Sy. The alternation between ‫ בור‬and other
clearer expressions in Genesis 39–41 should have left no doubt about the
meaning.
‫( וכל בכור בהמה‬12.29) LXX has καὶ ἕως for the initial waw, a Hebraising
variation. Sy and TgNmg have a (logical) pl. form for ‫בכור‬, which must be
collective here, as it is at the beginning of the verse. TgJ adds at the end ‘died,
whom the Egyptians worship’, providing a justification found also in MRI
(Lauterbach, p. 99) and perhaps based on the association between firstborn
and gods in v. 12.
‫( הוא‬12.30) LXX and Vulg have no equivalent, deeming it unnecessary to
reproduce exactly the Heb. idiom.
‫( וכל‬12.30) Both times TgJ adds ‫שׁאר‬, ‘the remainder of’, to take account
of the Egyptians who had been killed (so again in 32.12). There is some LXX
evidence for the omission of ‫ כל‬before ‫עבדיו‬, which (with two Heb. mss) leads
Propp to delete it here (p. 363) in line with other passages that do not have it.
But the other witnesses all support MT.
‫( ותהי‬12.30) Vulg ortus est, varying its equivalents for ‫ היה‬as it will again
with non iaceret for the second ‫ אין‬later in the verse. orior is used several times
in this way: e.g. Gen. 2.5; 17.16; 26.1; Exod. 10.6.
‫( במצרים‬12.30) LXX ἐν πάσῃ γῇ Αἰγύπτου, emphasising the totality
as elsewhere and following the wording of 11.6: SP read ‫ במצרים‬there. The
variation between the two passages in MT is likely to be more original: the
inclusion of ‫ כל‬here may have seemed unnecessarily jarring after ‫וכל־מצרים‬
just before. The expression is omitted altogether in both witnesses to TgJ,
probably by accident at an earlier stage of the tradition, unless it is connected
with the addition of ‘of the Egyptians’ in the next clause (see the next note: for
occasional omissions in the Tgg cf. Flesher and Chilton, Targums, pp. 51-52).
TgG and Sy have ‘in the land of Egypt’, conforming to the fuller phrase found
in v. 29.
‫( בית‬12.30) All the Tgg add ‘there’, i.e. in Egypt. TgJ and TgNmg also add
‘of the Egyptians’, since the Israelites were spared.
‫( שׁם‬12.30) So TgJ,N, but the other Vss and and TgNmg have ‘in which’, ‘in
it’ or (LXX) both, since the antecedent is a specific noun (‫)בית‬. The second
version in TgG combines both renderings.
‫( מת‬12.30) TgJ, again anxious for total precision, prefixes ‫בכור‬. TgN has
the pl. form ‫מיתין‬, which also appears in the second version in TgG along with
TgJ’s ‫ בכור‬in the pl. This version had evidently suffered from more contamina-
tion by other texts than the first one (compare the previous note). The pl. is
12.28-42, 50-51 127
surprising, but it is perhaps due to acceptance of the interpretation of v. 33 in

MRI (Lauterbach, p. 103), which infers that all the Egyptians’ children were
firstborn, due to adulterous relationships, the double negative being read as
equivalent to ‘in all houses there were dead people’. TgNmg corrects to the sing.
Before its rendering of v. 31 TgJ has a long addition about the size of the
land of Egypt and the locations within it of Goshen and Pharaoh’s capital,
which implies that there was a great distance between them. On TgJ’s location
of Ramesses see Text and Versions on 1.11. His estimate of the size of Egypt
as 400 parasangs (about 1400 miles) also appears in the Talmud (see AramB
2, p. 193 n. 50); this is over twice the actual distance from Aswan, usually
regarded as the southern boundary of Egypt, to the Mediterranean. Josephus,
on the other hand, greatly underestimated the distance as 2000 stades (about
230 miles): BJ 4.10.5(610). Herodotus’s figures (2.7, 9) add up to about 1000
miles; Strabo (17.1.2) cites Eratosthenes for 5300 stades (c. 650 miles), which
is close to the true figure (c. 550 miles as the crow flies).
‫( ויקרא‬12.31) LXX and Sy add ‘Pharaoh’ as the subject.
‫( לילה‬12.31) TgJ specifies that it was the night of Passover (cf. TgN in
vv. 8 and 12). TgJ then inserts ‘his voice was heard as far as the land of
Goshen. Pharaoh was pleading in a anguished voice’ (cf. v. 33). No exact
parallel to these motifs is known. MRI (Lauterbach, p. 100) records the view
that Pharaoh went all over Egypt looking for Moses and Aaron.
‫( ויאמר‬12.31) LXX adds αὐτοῖς (cf. Sy) to specify the unnamed
addressee, as it has already done at Exod. 2.18; 3.6-7, 12; 5.17 (but not every-
where). TgJ prefixes ‘So’.
‫( קומו צאו‬12.31) LXX inserts ‘and’ here, but not later in the verse. Wevers
(Notes, p. 184) suggests that its Vorlage may have read ‫ וצאו‬by dittography.
‫( מתוך עמי‬12.31) LXX ἐκ τοῦ λαοῦ μου and Vulg a populo meo ignore
‫תוך‬, as they sometimes do elsewhere (e.g. 2.5; 9.24).
‫( ולכו‬12.31) LXX has just βαδίζετε, creating an even more staccato
sequence than MT.
‫( את־יהוה‬12.31) Tgg and Sy as usual have ‘before the Lord’. LXX adds
τῷ θεῷ ὑμῶν to conform to the fuller wording used earlier in MT (8.22 etc.)
and already introduced in LXX at 10.9, 26b.
‫( כדברכם‬12.31) Most of the Vss have finite verbal forms, either present
(LXX, Vulg) or past (TgO,J,N, Sy), which are likely to represent MT’s inf.
constr. SP has ‫כדבריכם‬, ‘according to your words’, and both versions in TgG(AA)
give the same sense. 4QpalExm preserves enough of the word to show that it
read the same consonants as MT. In theory they could be vocalised ‫ ִכּ ְד ַב ְר ֶכם‬,
‘according to your word’, and this may have given rise to the reading attested
in SP and TgG, which has a precedent in Gen. 44.10. It is less easy to see how
the SP reading would have been changed to that of MT, and LXX provides
early evidence for the MT vocalisation. 4QExc preserves only ‫[כם‬, so adds
nothing relevant (cf. DJD XII, p. 115).
128 EXODUS 1–18
The beginning of v. 32 is omitted in TgN due to homoeoteleuton in its

renderings of ‘you said’. The other Pal. Tg. texts (in TgG) are unaffected
because their renderings differ (see the previous note).
‫( גם־צאנכם גם־בקרכם‬12.32) LXX and Vulg each ignore one of the pron.
suffixes for reasons of style. TgJ adds ‘and of mine’ in accordance with Moses’
request in 10.25, which MRI specifically sees as the point of ‫ כאשׁר דברתם‬here
‫( כאשׁר דברתם‬12.32) LXX has no equivalent, probably deliberately to
avoid repetition after the end of v. 31 (Wevers, Notes, p. 185). The omission
was rectified by Aq, Symm and Theod, and in the O-text. Vulg ut petieratis
responds to the same stylistic problem by using a more specific verb in the
pluperfect rather than the present.
‫( וברכתם‬12.32) LXX, Vulg, Sy and TgG (first version) follow MT, but
TgO,J,N and the second version in TgG have ‘pray (for me)’, the interpretation
given in MRI (Lauterbach, p. 102).
‫( אתי‬12.32) The first version in TgG adds ‫ אנא‬for additional emphasis; it
is not certain whether the second version had this too, as the ms. is damaged.
At the beginning of v. 33 TgJ has an addition to say that the Israelites
initially ignored Pharaoh’s command, presumably to explain why (as it must
have seemed) the Egyptians now urged them again to depart.
‫( ותחזק‬12.33) Most of the Vss take ‫ חזק‬in the sense ‘pressed’, but the use
of the Ithpaal of ‫ תקף‬in TgN,G suggests the alternative interpretation ‘were
determined’ (cf. Note i on the translation).
‫( העם‬12.33) TgJ adds ‘the house of Israel’ to avoid any ambiguity.
‫( למהר לשׁלחם‬12.33) LXX, Vulg and Sy all take ‫ למהר‬as an auxiliary verb
and render it by an adverb, but through failure to recognise that the infinitives
may have a gerundival character (see Notes i and j on the translation) they
are left with a difficult grammatical connection, which Vulg eases by a free
rendering of ‫ לשׁלחם‬by exire. TgN,G’s different rendering of ‫( ותחזק‬see above)
avoids this problem, but it is less likely to be correct. TgO,J do not render ‫למהר‬
by an adverb and apparently view ‘the people’ as its subject and the Egyptians
themselves as the subject of ‫( לשׁלחם‬so AramB 7, p. 33; 2, p. 193), which is a
possible understanding of the sequence.
‫( כי אמרו‬12.33) LXX adds a ὅτι recitativum, as it also does in 4.1 and
13.14 (cf. Sy dkln: 7a1 follows MT). TgF and TgNmg supply ‘the Egyptians’ as
an explicit subject. TgJ,N,F (but not TgG) then insert a clause, with slight varia-
tions in the wording (TgNmg is virtually identical to TgF), which makes any
further delay in the Israelites’ departure lead to the even worse catastrophe
taken to be implied by ‫כלנו מתים‬.
‫( כלנו מתים‬12.33) Rather than treating the statement in the present tense,
as we have done, the Vss render it as present continuous (LXX; Sy?) or future
(so also, e.g., JB, NRSV). TgN,F (like TgJ) introduce the statement with an
exclamatory ‫הא‬, ‘look’, and replace ‘all of us’ with ‘all the Egyptians’: the
latter reflecting the liturgical practice of not reproducing words that might be
an ill omen (cf. Michael Klein on the Targums, p. 194).
12.28-42, 50-51 129
‫( וישׂא‬12.34) Vulg igitur makes an explicit link with the haste required in
v. 33, in line with the recapitulation in v. 39.
‫( את־בצקו‬12.34) LXX follows the Greek idiom of indicating possession
by the def. art. alone, as it does with ‫ על־שׁכמם‬later: in both cases the O-text
adds αὐτῶν, following the Three in the latter case (and probably in the
former). Vulg too does not represent the suffix here, as also in the next clause,
but it has umeros suos at the end. Its equivalent to ‫ בצק‬is conspersam farinam,
lit. ‘sprinkled flour’, presumably with water: compare its rendering of Jer.
7.18. TgJ adds ‘upon their heads’, to distinguish the dough from what it takes
to be the meaning of the following word (see the next note).
‫( משׁארתם‬12.34) Neither LXX nor Vulg uses the (genitive or ablative)
absolute construction, which would have been the most idiomatic equivalent
to this clause: LXX has it in e.g. 2.10; 5.20 and 40.32 (LXX 38.27), cf. BDF
§423, but here both noun and participle are in the nominative; Vulg neatly
turns the clause into the active and can then attach the participle grammati-
cally to the main subject of the sentence. On the interpretation of ‫ משׁארת‬see
Note k on the translation and Text and Versions on 7.28. Here TgO,N have the
probably correct meaning ‘kneading-bowls’ (cf. Ibn Ezra [longer commen-
tary], Rashbam), but TgO combines it with the interpretation ‘remainder’ (cf.
‫ )שׁאר‬which also appears in TgJ,Nmg (with the amplification ‘of [their] unleav-
ened bread and bitter herbs’: cf. TgF, MRI [Lauterbach, p. 104] and Rashi)
and TgG. LXX and Sy again have ‘dough’, which Vulg seems to presume here
by its omission of any equivalent, although in the three other occurrences of
‫ משׁארת‬it follows the ‘remainder’ view. One does wonder whether the
Masoretic pointing of the second letter is not just a (mistaken) result of this
peculiar interpretation (cf. Propp, p. 326): if originally read with śin it would
allow a much more plausible association with the practice of baking (cf. ‫שׂאר‬, ְ
‘leaven’, and ‫ ַמ ְשׂ ֵרת‬in 2 Sam. 13.9).
‫( בשׂמלתם‬12.34) So also SP, but 4QExc has the alternative (and appar-
ently secondary) spelling ‫[ב]שׂלמתם‬, as in MT at 22.8, 25 and in fourteen other
places. SP has standardised the spelling as -‫מל‬- everywhere: in the one place
where von Gall reads -‫למ‬- (Deut. 24.13) this is a clear error (cf. Sanderson,
Exodus Scroll, p. 61: also BHS ad loc. and the editions of Sadaqa and Tal).26
The reading of 4QExc has no obvious explanation and is the more surprising
since it reads ‫ ושׂמלות‬in v. 35 with MT and SP. It may have the original reading
here (cf. Propp, p. 364): J.E. Sanderson has suggested (ibid., pp. 61-62) that
in another place where the two forms occur in adjacent verses (MT at 22.25-
26), this could be the result of the redactional use of ‫שׂמלה‬. The same might
be the case here: see the Explanatory Note and the introduction to the section.
‫( על־שׁכמם‬12.34) TgJ,F (cf. TgNmg) add ‘they carried’ or ‘were placed’,
probably as a result of their taking ‫ משׁארתם‬to refer to something unconnected
with the ‫( בצק‬cf. above).
26
This error in von Gall is apparently overlooked by Baillet in his ‘Corrections’.
130 EXODUS 1–18
‫( כדבר משׁה‬12.35) LXX renders a little freely with καθὰ συνέταξεν

αὐτοῖς Μωυσῆς: on the divergences from MT see Wevers, Notes, p. 187; and
v. 31 for the use of a verb instead of a noun (also Vulg). There is therefore
no need to envisage a Vorlage different from MT, with which SP, Tgg and
Sy agree. In 4QExc ‫ משׁה‬is the first word preserved in a line and there seems
to be more space before it than would be required for the preceding words in
MT. Possibly 4QExc had a longer text at this point, more like v. 28 perhaps
(cf. DJD XII, p. 115). TgN,G have ‘the words of Moses’, but TgNmg records the
sing. rendering of TgO,J.
‫( ממצרים‬12.35) 4QExc omits, apparently by carelessness, which is
apparent elsewhere in the ms. (cf. DJD XII, pp. 102-103).
‫( כלי כסף וכלי־זהב‬12.35) On the idiomatic renderings of LXX and Vulg
see Text and Versions on 3.22 and 11.2. Here, as in 11.2, all the earliest mss
of Sy follow the order of MT.
‫( ושׂמלת‬12.35) Both witnesses to TgJ have no equivalent to this word (cf.
Text and Versions on v. 30), probably through accidental omission in trans-
mission. Vulg vestem plurimam seems an over-correction of OL vestem, itself
based on LXX’s collective sing. ἱματισμόν. But classical Latin prose did not
use the pl. of vestis (LS, pp. 1981-82), so this may have seemed the best way
for Jerome to do justice to the pl. of the Heb. However, elsewhere he uses the
pl. (including 3.22) as well as the sing.
‫( ויהוה‬12.36) TgG adds ‘the Memra of’ (cf. TgNmg).
‫( נתן את־חן העם‬12.36) The Vss follow MT with the same small variations
of idiom as in 11.3 (see Text and Versions there). TgJ has ‫ חן וחסד‬here instead
of ‫רחמין‬.
‫( וישׁאלום‬12.36) Tgg and Sy employ the cognate Aram. verb, for which the
meanings ‘borrow’ and ‘lend’ are much more prominent than in BH (22.13 is
one example, but see Note m on the translation of this section) and ‘lent’ is
probably intended. This is certainly the meaning given by LXX ἔχρησαν (for
the distinction from giving cf. Arist., EN 1162b33) and Vulg ut commodarent
(with this as the consequence of the divine intervention), presumably on the
basis of familiarity with later Heb. and Aram. usage. This widespread inter-
pretation gave rise to ethical embarrassment (cf. Childs, Exodus, pp. 175-77),
and at least some Hellenistic Jews did not share it (Ezek.tr., Exag. 162-66;
Jos., AJ 2.14.6 [314]: see further Jacobson, Exagoge, pp. 126-27).
‫( וינצלו‬12.36) The Vss generally render as in 3.22 (see Text and Versions
there), but Sy ḥlṣw lines up with the interpretation ‘despoil’ (cf. CAL). TgG,
presumably by accident, omits ‘the Israelites’.
‫( את־מצרים‬12.36) TgJ,N add ‘of their possessions’ here (not in 3.22), to
clarify the meaning.
After v. 36 2QExa, 4QExc and 4QpalExm all had a long space and v. 37
began on a new line – early evidence of the division marked in the much later
Masoretic and Samaritan mss.
12.28-42, 50-51 131
‫( ויסעו‬12.37) TgN accidentally omitted the equivalent ‫ וינטלו‬which is

present in all the other Targum texts.
‫( מרעמסס‬12.37) LXX, Vulg, Sy and TgO reproduce MT. TgJ,N,G have ‘from
Pelusium’, erroneously identifying the site with an important Delta city of
Greco-Roman times as TgJ,N do in 1.11 (cf. my Way of the Wilderness, pp. 12-
13, 19).
‫( סכתה‬12.37) It seems likely that 4QExc read ‫[סכו]ת‬, without the direc-
tional ending (DJD XII, p. 115); cf. 13.20 and Num. 33.5-6. 2QExa and SP
agree with MT. The abbreviated expression here really requires the fuller
form, which appears (with reference to a different place of the same name) in
Gen. 33.17. The LXX reading preferred by Rahlfs and Wevers, Σοκχώθα, also
attests the ending, though many mss lack the -α, perhaps for the same reason
as 4QExc. The straightforward transcription as a place-name is also found in
Vulg, Sy, TgO,J,N and is one of the interpretations cited in MRI (Lauterbach,
p. 108, cf. 182). But MRI also records Akiba’s view that the meaning is ‘the
clouds of glory’ over Mount Zion mentioned in Isa. 4.5-6 (cf. Exod. 13.20-21,
where ‫ סכת‬is mentioned again) and TgJ has a long addition here to the same
effect. Symm εἰς σκηνάς and TgG ‫ למטלין‬gave the etymological meaning
‘booths’, as in MRI’s opening comment, ‘there were real booths’. TgJ also
adds that it was a distance of 130 miles from Ramesses/Pelusium to Succoth
(cf. MRI’s figure of 160 miles, a long distance which it seeks to justify
[Lauterbach, p. 107]). It is unclear where such figures came from, unless they
are perhaps based on the distance from Pelusium to the other Succoth in the
Jordan valley.
‫( כשׁשׁ־מאות אלף‬12.37) It is possible that 4QExc did not have the initial
‫כ‬, so conforming with Num. 11.21 (and cf. Sir. 16.10; 46.8), where ‫ רגלי‬also
occurs. TgN ‘about sixty myriads’ reflects an interpretation which correlates
the figure with Song 3.7-8 (cf. MRI [Lauterbach, p. 109]).
‫( רגלי‬12.37) The correct SP reading is ‫( רגלאי‬not ‫רגלי‬, as in von Gall’s
text), in line with the Sam. pronunciation (GSH §16a) which is close to the
Aram. form (GSH §139a: cf. TgO). TgG omits. In TgJ the exclusion of the
‘children’ is related to this word, not the total number, and ‘five per family’ is
added in accordance with TgJ’s interpretation of 13.18 (see Text and Versions
there).
‫( מטף‬12.37) The Vss render ‫ טף‬as they do in 10.10, 24 (see Text and
Versions there).
‫( וגם‬12.37) Vulg sed et, introducing an element of contrast.
‫( ערב רב‬12.38) A few SP mss (including Tal’s and Camb. 1846) write the
phrase as a single word, a reduplicated form from the root ‫‘ = ערב‬mix’, which
occurs in MH and later Aram. but not in BH. Albright independently favoured
this reading (cf. Archaeology and the Religion, p. 201 n. 9 [for the form see GK
§84bn]), but the textual support for it is very slight: on the other side, besides
MT, 4QpalExm has ‫ ערב ר[ב‬and the Vss all have ‘large’. LXX ἐπίμικτος (sc.
132 EXODUS 1–18
ὄχλος: the word seems to be attested as a noun only in LXX) associates ‫ערב‬
with the sense ‘mix’ (cf. Vulg, TgF,G,Nmg, Sy), perhaps mistakenly: TgO,J ‫נוכראין‬
and TgN ‫ גיורין‬point instead to the sense ‘foreigners’ (see further Note t on the
translation). For ‫ רב‬TgJ has ‘more than them [sc. Israel], 240 myriads’ (cf. MRI
[Lauterbach, p. 109] for this and other large figures), and Vulg’s innumerabile
(a word which elsewhere usually has some basis in the Heb. text: e.g. 10.28)
is probably based on knowledge of this interpretation.
‫( מקנה‬12.38) LXX, Vulg, TgJ and Sy prefix ‘and’, apparently not seeing
the appositional connection: Vulg even adds diversi generis to distinguish
these animantia from those just mentioned. TgG’s ‫ נכסין‬helps to identify one
source of TgNmg as a variant Pal. Tg. ms., perhaps a more authentic one than
TgN itself, which here has the rendering of TgO.
‫( מאד‬12.38) 4QpalExm has ‫( מאדה‬4QExc ‫מאד‬, likewise SP), followed by
an interval within the line which is not attested elsewhere. The longer form in
‫ה‬- appears to be a feature of Qumran Hebrew (cf. Qimron, pp. 69, 117; HAL,
p. 511; DCH 5, p. 103), which can be compared to the use of final he as an
adverbial ending in later Aram. (Stevenson, p. 25; Rosenthal §88[4]).
‫( ויאפו‬12.39) TgJ rewrites the beginning of the verse so that the dough
laid on the Israelites’ heads (as it understands v. 34) was baked by the heat of
the sun—a logical deduction perhaps from the statement that ‘they did not get
provisions for themselves’.
‫( את־הבצק‬12.39) Most Vss render as in v. 34, with even the addition of
‘their’ here in TgN,G, but TgN has ḥmyʿh, ‘leavened bread/dough’, and so makes
the statement that ‘it was not leavened’ due to a delay in the fermentation
process.
‫( ממצרים‬12.39)1o TgG adds ‘the land of’ as it did in v. 27, to match the
common formula (e.g. v. 41).
‫( עגת‬12.39) The Vss use a variety of equivalents, several of which reflect
the practice of baking bread in the ashes of a fire (LXX, Vulg, Sy: cf. 1 Kgs
19.6 and MRI [Lauterbach, p. 110]). TgG’s ‫ חלון‬is the strangest, as the word is
usually found in the context of temple ritual.
‫( גרשׁו ממצרים‬12.39) SP and 2QExa divide the words differently, ‫גרשׁום‬
‫מצרים‬, making the Egyptians the subject and the verb Piel instead of MT’s
Pual. LXX, TgG (cf. TgNmg), Vulg (cf. OL) and the earliest mss of Sy render
with active verbs and point to the same reading of the Heb. But the reading of
Kennicott ms. 129 cited in DJD III, p. 51, ‫גרשׁו מצרים‬, without the first mem,
is probably a haplographic corruption of MT. MT is supported by TgO,J,N: no
other Qumran mss preserve this phrase. Both readings make sense and both
can be explained as secondary, MT to keep the same subject throughout and
SP etc. to conform to the use of the active earlier in 6.1; 11.1, so it is difficult
to decide which is the more original. The Piel reading is early and more
widely attested and picks up v. 33 well; on the other hand the use of the rare
Pual of ‫( גרשׁ‬only elsewhere in Job 30.5) makes MT arguably the difficilior
12.28-42, 50-51 133
lectio, which should then be preferred. Most commentators tend to follow

MT, but Propp (p. 365) notes ‫ ממצרים‬earlier in the verse as an argument that it
might be secondary here.
‫( וגם צדה לא עשׂו להם‬12.39) LXX adds, reasonably enough, εἰς τὴν ὁδόν
(for the fuller expression cf. Gen. 42.25; 45.21). TgJ goes further, picking
up the inference of MRI (Lauterbach, p. 110) that the unleavened cakes met
the Israelites’ needs for thirty days, until the provision of manna began (cf.
Exod. 16.1-4). After v. 39 4QExc has an open section which is not paralleled
elsewhere.
‫( ומושׁב‬12.40) TgJ ‘And the days (which)…’, with ‘the Israelites’ in the
relative clause (cf. Sy), gets round the unique temporal use of ‫מושׁב‬.
‫( בני ישׂראל‬12.40) SP adds ‫ואבותם‬, referring to (some of) the patriarchal
period in line with the addition discussed in the next note. Many LXX mss
have αὐτοὶ καὶ οἱ πατέρες αὐτῶν for the same reason, the main O-group
and papyrus 866 (cf. Desilva, ‘Five Papyrus Fragments’, pp. 13-15) after
κατῴκησαν = ‫ ישׁבו‬and the uncials AFM and the Catena group at the end of the
same clause, but LXXB does not have it and both Rahlfs and Wevers exclude it
from their text of the original LXX. There is no trace of a Heb. equivalent in
2QExa and 4QExc or in the renderings of the other Vss. The varied locations
and wording of the plus where it does occur suggest that it belongs to a
secondary stage of the amplification of the text.
‫( במצרים‬12.40) 4QExc, TgG (as in v. 39) and later mss of Sy have ‘in the
land of Egypt’, a common expansion. SP and LXX (except for the O-text and
the longer retroversion in MRI [Lauterbach, p. 112]) also have this expan-
sion but as part of a more substantial variant which either adds ‫בארץ כנען ו‬
before the expression (SP) or καὶ ἐν γῇ Χανάαν after it (LXX). In this case
it does seem to be part of the original LXX translation (cf. Rahlfs, Wevers),
and Paul’s reference to ‘430 years’ from Abraham to Moses (Gal. 3.17)
evidently presupposes it. Although 2QExa as preserved does not exhibit the
variant, a reconstruction based on MT would be about 12 letters/spaces too
short (cf. DJD III, p. 51) and it is very likely that it originally had both ‫במצרים‬
and ‫בארץ כנען‬, though whether it followed the order of SP or LXX cannot
be determined. The reasoning behind the longer text is that 430 years was
too long for the four-generation sojourn in Egypt foretold in Gen. 15.16 (cf.
Exod. 6.16-20 and 13.18LXX: Wevers, Notes, p. 190). The problem is raised
in MRI (Lauterbach, p. 111) but solved in a different way (for other rabbinic
treatments see the next note). The ‘solution’ adopted in SP and LXX must
have been evolved in the course of discussions about biblical chronology in
the Hellenistic period, on which see B.Z. Wacholder, ‘Biblical Chronology
in the Hellenistic World Chronicles’, HTR 61 (1968), pp. 451-81. But there
was only a problem because of some late additions to the biblical narrative
(Exod. 6.13-30 and Gen. 15.16, which is dependent upon it) and MT here,
which shows no awareness of the ‘problem’ and is followed in TgO,N,G, Vulg
134 EXODUS 1–18
and Sy, undoubtedly preserves the most ancient version of the text (see further
Dillmann, Ex.-Lev., pp. 120-22; P. Grelot, ‘Quatre cents trente ans [Exode
12.34 (sic!)]: À propos de la chronologie du Pentateuque’, in G. Braulik [ed.],
Studien zum Pentateuch [FS W. Kornfeld; Vienna, 1977], pp. 91-98 [for early
evidence in Hebrew and Aramaic texts]; S. Kreuzer, ‘Zur Priorität und Ausle-
gungsgeschichte von Exodus 12,40MT: Die chronologische Interpretation des
Ägyptenaufenthalts in der judäischen, samaritanischen und alexandrinischen
Exegese’, ZAW 103 [1991], pp. 252-58 [252-55]).
‫( שׁלשׁים שׁנה וארבע מאות שׁנה‬12.40) For the time spent ‘in Egypt’ TgJ has
‘(were) thirty sabbatical years, whose sum is 210 years, and the counting of
430 years is from when the Lord spoke to Abraham, from the time when he
spoke to him on the fifteenth of Nisan between “the pieces” until the day when
they came out of Egypt’: in other words TgJ adopts the chronology followed
in SP and LXX (as does e.g. Gen.R. 44.18) and accommodates MT to fit it by
changing the numeral (cf. Gen.R. 91.2 [citing R. Abba b. Kahana, third–fourth
cent.]; PRE 48 [tr. Friedlander, pp. 374-76]; Exod.R. 18.11). The figure of
‘210 years’ had become traditional by the time of Rashi, Rashbam and Ibn
Ezra and was correlated with the numerical value of ‫( ודר‬200+4+6) in Gen.
42.2. The date of the revelation to Abraham was deduced from ‘on this very
day’ in v. 41 (see further below).
‫( מקץ‬12.40) LXX renders freely with μετά (cf. Sy). Vulg quibus expletis
avoids the repetition of the numeral, but its choice of verb does show some
trace of MT’s ‫קץ‬.
‫( שׁלשׁים שׁנה וארבע מאות שׁנה‬12.41) TgJ has a midrashic expansion, dividing
the two parts of the numeral in the order of MT at the birth of Isaac, in a
refinement of the SP/LXX chronology which is found also in MRI (Lauter-
bach, p. 111). TgG (as in v. 40) also follows the MT wording of the numeral,
against normal Aram. practice, which is to put the larger numeral first, perhaps
with this midrash in mind. TgJ has no equivalent to the second ‫שׁנה‬, perhaps
assuming that it could be understood from the context.
‫( בעצם היום הזה‬12.41) LXX did not render this phrase which was so
important to rabbinic exegesis here (e.g. MRI [Lauterbach, pp. 112-13]: see
the notes on v. 42): compare its weak equivalents for the same phrase in vv.
17 and 51. Vulg’s eadem die perhaps shows knowledge of the significance
that was seen in it.
‫( כל־צבאות יהוה‬12.41) LXX is again free, as elsewhere using a sing. form;
likewise Vulg, although the Three had corrected to the pl.27 TgJ,N,G add ‫פריקין‬,
‘redeemed’, as often elsewhere after the verb ‘go/bring out’ (see Text and
Versions on 3.10 and 6.26).
27
Wevers (Notes, p. 190 n. 55) oddly credits Theod with the rendering πάντες
οἱ ἄνδρες, apparently misreading the Latin equivalent to Syh’s reading in his own
apparatus (vires) as the pl. of vir: in fact it must represent αἱ δυνάμεις.
12.28-42, 50-51 135
‫( ליל‬12.42) SP has ‫לילה‬, the more common form and the only one used
elsewhere in biblical prose—a clear case of SP’s tendency to normalise
unusual forms. 4QExc, the only Qumran ms. to preserve the word, tantalis-
ingly reads ]‫ ליל‬before a lacuna, so its original reading is uncertain. LXX
νυκτός took the word adverbially with v. 41, ‘…by night’, recalling vv. 29-31
and perhaps Deut. 16.1, 6, but the nominative forms in Aq, Symm and Vulg
agree with the MT verse-division and syntax, like the other Vss. Some SP
mss also begin the new verse after ‫( לילה‬cf. von Gall, Tal), but many do not
(Sadaqa, Crown, Camb. 1846; cf. SamTg and perhaps the SamGk, which had
φυλαξεώς for the next word). All that can be said is that the longer spelling
permitted the interpretation found in LXX. The latter was scarcely the original
intention, as is generally recognised: both the form ‫ ליל‬and the neat balance
of v. 42 in MT are against it. TgO,N,G(AA,FF) render with ‫לילי‬, intending the sing.,
probably constr. st. (cf. TgJ,F(V) ‫)ליל‬, but it is a form that looks like and often
is the pl. constr. (Jastrow, p. 707: cf. TgJ in its expansion [below]) and it is
possible that this ambiguity generated the ‘Poem of the Four Nights’ which
was added in the middle of v. 42 in the Pal. Tg. and in an abbreviated form
in TgJ (see below).
‫שׁמרים‬1o (12.42) LXX προφυλακή is generally taken to mean ‘vigil’ here
(cf. LSJ, NETS), but the great majority of its occurrences in LXX and more
widely are in military contexts and mean ‘advance guard’ or more generally
‘defence, protection’, a sense which is found for the second occurrence of
‫ שׁמרים‬in this verse (Theod, TgJ). Here and there the most popular view was
that it meant ‘observance’ (Aq, Symm, Tgg, Sy; Vulg), whether in the sense
of (God) ‘watching over’ or of religious observance. Curiously both LXX
and Vulg supply the unexpressed verb in the present tense here, as if the
verse were part of a speech rather than of the narrative. TgF(V),G(AA) add ‘and
appointed for redemption’: TgJ has just the latter expression.28
‫( ליהוה‬12.42)1o TgO,J,F(V),G(AA,FF) have ‘before the Lord’, TgN ‘for the name
of the Lord’.
‫( להוציאם‬12.42) TgN,F(V),G(AA,FF) have ‘(at the time) when he brought the
Israelites out redeemed’ (cf. Vulg quando eduxit eos); TgJ ‘to bring out the
name of the Israelites’, but ‫ שׁמא‬is probably a scribal error for ‫עמא‬, ‘the people’.
‫( מארץ מצרים‬12.42) After their equivalents to this phrase all the Pal.
Tg. mss except for the oldest one (TgG(AA)) and TgF(P) (on which see below)
insert in slightly divergent forms ‘The Poem of the Four Nights’. In TgF(P) the
poem appears before Exod. 15.18; TgJ has an abbreviated version of it before
12.42. The poem is also included in the Mahzor Vitry (cf. Chester, p. 196
n. 45), whose origins go back to the eleventh century (but it has many later
additions), and it was very likely designed for liturgical use in celebrations of
the Passover. According to the poem the creation of the world, the covenant
28
TgG(FF) has ‘and trustworthy for redemption’, but this is probably due to
scribal error (cf. Klein 2, p. 61): this text has ‘appointed’ in v. 42b.
136 EXODUS 1–18
with Abraham and/or the birth of Isaac and his (near-)sacrifice in the Aqedah
and the deliverance from Egypt all took place on the same night and the
coming of the Messiah (omitted in TgN,G(FF)) and the final redemption of Israel
would do so too. The same dating for the covenant with Abraham was widely
current and the view that it would also be the day of future deliverance is
reported, but contested, in MRI (Lauterbach, pp. 115-16). The date of creation
was also disputed, but one view was that it occurred in the month of Nisan
(Gen.R. 22.4, cited in J.W. Bowker, The Targums and Rabbinic Literature
[Cambridge, 1969], p. 166). The first detailed modern study of the Poem was
R. Le Déaut, La Nuit Pascale (AnBib 22; Rome, 1963); see also G. Vermes,
Scripture and Tradition (Leiden, 2nd ed., 1973), pp. 193-227, and Chester,
Divine Revelation, pp. 192-99, who provides an English translation (with
notes of variant readings, but not from the more recently published Genizah
mss) on pp. 192-94. The poem is widely thought to be ‘early’, possibly even
pre-Christian (cf. Le Déaut, pp. 339-71). The use of Aramaic would support
this up to a point, but the absence of the poem from the earliest relevant ms.
of the Palestinian Targum (AA: ninth to mid-eleventh cent.) and the lack of
early citations or references should counsel caution in regard to some of the
claims that have been made.29
‫( הוא הלילה הזה‬12.42) The Tgg generally follow the wording of MT
closely here, but TgN,G(FF),F(V) have ‘it [or ‘that] (is) the night of Passover’,
giving what is surely the sense of MT without its awkward demonstrative.
LXX ἐκείνη ἡ νὺξ αὕτη most likely means ‘that night, it is…’ (BAlex): a poor
analysis of the Heb. even though the words are in the right order. Vulg hanc
observare debent… glosses over the difficulty.
‫( ליהוה‬12.42)2o See on the occurrence earlier in the verse: here TgJ and
Vulg have no equivalent.
‫( שׁמרים‬12.42)2o Again the Vss generally follow their renderings earlier
in the verse. TgJ adds ‘from the destroying angel’: here, as probably in LXX
and perhaps elsewhere, the sense of ‘protection’ by Yahweh enduring into the
future has displaced the original reference to a continuing obligation on Israel
to observe the Passover ‘for Yahweh’. This is taken furthest in the rewriting
of v. 42b in TgG(FF), which equates Passover with the fourth night as well as
the third in the poem.
‫( לכל־בני ישׂראל‬12.42) TgJ also introduces an ongoing hope by inserting
here ‘who were in Egypt and also for their redemption from their exiles’.
‫( לדרתם‬12.42) 4QpalExl certainly and 4QExc possibly (cf. DJD XII,
p. 116) had an interval after v. 42, corresponding to those in MT and SP.
29
Towards the end of the poem, which it knew from a ms. of TgF, in the section
about the coming of the Messiah the 1517 Rabbinic Bible introduced a reference
to the accompanying ‘clouds’, erroneously reading ‫ עננא‬for ‫ענא‬, ‘the flock’, which
is the reading of all the Pal.Tg. witnesses now known (see further Michael Klein
on the Targums, pp. 235-37).
12.28-42, 50-51 137
Verses 50-51 are preserved, along with vv. 43-49, in several Qumran
phylactery mss (XQ1, 8Q3, 4Q128, 136, 140) and in 4QDeutj: see Lange,
Handbuch, pp. 116-22, which draws attention to the (at least later) practice
of writing out such copies from memory (p. 121). Variants are therefore less
significant for textual criticism, but will nevertheless be recorded.
‫( ויעשׂו‬12.50) 8Q3 has ‫יעשׂו‬, which might be seen as a deliberate jussive to
conclude the preceding laws. But there is a careless error in the next line, so
this may be another. TgN has no equivalent, probably again due to carelessness.
‫( כל‬12.50) 4Q128, a Genizah ms. and LXX omit, probably to conform to
v. 28. The O-text supplies πάντες.
‫( כאשׁר‬12.50) XQ1 has instead the stronger ‫ככל־אשׁר‬, which occurs in Gen.
6.22; 7.5 and several more times in the Pentateuch.
‫( ואת־אהרן‬12.50) A Genizah ms. omits ‫ את‬according to BHS, which
also unjustifiably presumes that the absence of τῷ in LXX reflects a similar
omission in its Vorlage (cf. 5.20; 6.20; 10.8). The omission in the Genizah ms.
is probably due to assimilation to v. 28. LXX pedantically adds πρὸς αὐτούς,
‘for them’ (sc. the Israelites), as MT, SP and LXX have at Deut. 1.3 (cf. Exod.
6.13; 25.22; Lev. 27.34). Its omission here by mss AFM and the O-text (but
not by papyrus 866) is probably a correction towards MT, not merely a matter
of style (against Wevers, Notes, p. 194).
‫( כן עשׂו‬12.50) Vulg has no equivalent, probably regarding the repetition
as otiose, as in v. 28. 8Q3 has the past tense like the other witnesses here.
DJD XIV, p. 89, thinks it possible that 4QDeutj had an interval after v. 50,
as in MT.
‫( בעצם היום הזה‬12.51) The Vss mostly render as in v. 41. LXX, which had
no equivalent there, has ἐν τῇ ἡμέρᾳ ἐκείνῃ, ignoring ‫ עצם‬as it did in v. 17
and rendering ‘historically’ here.
‫( הוציא‬12.51) 8Q3 has ‫הוציאך‬, a form which occurs in 13.9 in a partly
similar expression. But it does not fit the narrative context here.
‫( את־בני ישׂראל‬12.51) TgN adds ‫פריקין‬, as the Pal. Tg. often does elsewhere
(see the note on v. 41).
‫( על־צבאתם‬12.51) Most Vss render ‫ צבאת‬as in v. 41, but Vulg reverts to
its more precise equivalent turmas, as in 6.26, 40.34 and often afterwards. Sy
replaces ‫ על‬by kwl, ‘all’, as it did in 6.26 (cf. v. 41).
After v. 51 all the Qumran phylacteries and 4QDeutj have an interval,
whether before a different passage (8Q3) or, as in MT and SP, before the
continuation in 13.1 (4QDeutj, 4Q128, 136, 140).
C h ap t er 1 2 . 4 3 - 4 9
F u rther I n s tr u cti o n s fr om Y a hwe h

abou t th e P as s over M e a l
The boundaries of this section, as in the rest of the chapter, have

been defined in this commentary according to the distinction
between narrative and law. The beginning does correspond to
an interval in the major Masoretic mss, which also appeared in
4QpalExl and perhaps in 4QExc, but at the end there is, understand-
ably, no interval until after v. 50. The section has a heading in
v. 43a (see the Explanatory Note) but no formal closure: v. 49 is,
as the recurrence of the words ‘native’ and ‘resident alien’ shows,
specifically linked to v. 48. Verses 43b-49 comprise a series of short
commandments (only in vv. 48-49 is there anything like the ‘story’
structure1 that is found in vv. 1-14[20]), in which the presence of
a number of prohibitions is a notable feature (Childs, p. 202: cf.
vv. 43, 45-46, 48b), though they are not entirely missing earlier in
the chapter (vv. 9-10, 16b, 19-20, 22b). Another recurrent element
is the expression ‘may/shall (not) eat of it’ (Heb. [lōʾ] yōʾkal bô),
which draws attention to the main (but not exclusive) topic of these
verses, i.e. who may and who may not participate in the Passover
meal. Attempts have been made to identify seven laws within the
series. Cassuto (p. 150) listed them as vv. 43b, 44, 45, 46, 47, 48a
and 48b, with v. 49 as ‘a general law in regard to the sojourner’,
which looks like eight: moreover the combination of two laws in
one in v. 46 and the subdivision of v. 48 are artificial (although
they cancel one another out numerically). Propp (p. 375) notes the
distinction in form between vv. 43b-47, 48b (apodictic) and 48a
(casuistic: it is not clear what he thinks about v. 49) and detects in
the former sequence an older ‘Pesaḥ Rule’, in which positive and
negative prescriptions alternate. This was then expanded with the
casuistic law in v. 48a. The supposed pattern does break down in
v. 46a, where an additional (supplementary) negative command
1
I owe this expression to Dr N.A. Wormell.
12.43-49 139
appears, and again the subdivision of v. 48 is artificial. If there is

an alternating pattern, it is easier to see it in the subject matter: who
may/may not eat (vv. 43b-45), how the meal shall be eaten (v. 46),
who may/may not eat (vv. 47-49). Within the two outer sub-units
there is more of a linear progression, with in each case a statement
of principle (vv. 43b, 47) followed by qualification(s) introduced
by the conjunction (Heb. we), translated here as ‘But’ (vv. 44, 48).
It is also notable that two different Hebrew verbs, rendered here
‘eat’ and ‘celebrate’, are central to vv. 43b-45 on the one hand and
vv. 47-49 on the other (for the likely significance of this change see
the Explanatory Note on v. 47).
The relationship of this section to other passages dealing with
Passover has been clarified by Houtman’s table (pp. 152-53),
although its positive relations with Num. 9.1-14, 2 Chr. 30.1-27
and perhaps Ezra 6.19-22 are underplayed through the lack of a
column for references to non-Israelites. In general it is striking
how little it shares with other passages: there is nothing in it about
the blood-rite, the final plague (or the Exodus more generally), the
date, the eating of unleavened bread or the period for which that
is prescribed. As is frequently observed, the regulations here seem
completely detached from the Exodus situation and to relate to a
later situation after the settlement in Canaan (though nothing would
actually exclude their application to Diaspora Jews). Given what
they omit, and the very limited topics with which they do deal, they
cannot have been intended to stand alone (which only makes the
heading in v. 43 more anomalous) and they must have been designed
to be a supplement to the regulations earlier in ch. 12, particularly to
vv. 1-20, with which they share a number of features (the introduc-
tory address to Moses and Aaron, the terms ‘statute’ [Heb. ḥuqqāh]
and ‘congregation’, the references to a ‘house’, the rules about the
preparation and eating of the animal, the contrast between ‘resident
alien’ and ‘native’ [cf. v. 20]). The emphasis on circumcision and
the language of v. 44 connect them with Genesis 17, but at the
same time the terminology and point of view that are shared with
passages later in Leviticus and Numbers must also be noted (see the
Explanatory Notes on vv. 45 and 48-49). The contacts with Num.
9.1-14, which adds further new provisions to the Passover law, are
particularly close: just as this section presupposes and goes beyond
Exod. 12.1-14(20), so Num. 9.1-14 presupposes what is found in
Exodus 12 as a whole and extends it further.
140 EXODUS 1–18
Critical analyses have agreed since Knobel that the section is in the broad
sense from P. Its introductory formula and its language point strongly this way
and the home setting for Passover presupposed in v. 46 both fits the earlier
Priestly Passover regulations in vv. 1-14 and contradicts the Deuteronomic
requirement for a centralised celebration (Deut. 16.1-8).2 Some scholars
say no more, whether to avoid the complexities of further sub-division of
P or because they see no need to separate this passage from the rest of the
Priestly Passover material (Wellhausen, Carpenter/Harford-Battersby, Gress-
mann, McNeile, Eissfeldt, Beer-Galling, Van Seters, Dozeman). Propp is a
recent example of the latter group (pp. 374-75): although he makes a point
of observing that ‘the terse style is not at all Priestly’, he attributes this to
the use of an older ‘Rule’ by P (see above), who then ‘explicitly quotes’ it
again in Num. 9.12-14. But the majority of scholars, beginning with Knobel
and Dillmann, have seen the section as a secondary development within the
Priestly corpus (Ps), for reasons that were summed up by Noth in his Exodus
commentary (pp. 72, 78-79, ET pp. 92, 100-101): the provisions are not
connected to the historical (i.e. narrative) situation but presuppose conditions
in the land of Canaan (esp. vv. 47-49), and the repetition of v. 41b in v. 51
(Wiederaufnahme) implies that what comes in between is a later addition. It is
possible to strengthen this argument by noting that the language of vv. 48-49
(and also the expression tôšāb in v. 44) and the concern to place the ‘resident
alien’ on the same footing as the native Israelite is characteristic, not of the
main body of P but of the Holiness Code and passages in the book of Numbers
which also appear to be later. Israel Knohl (Sanctuary, pp. 21-22) has deduced
from this that the passage is the work of ‘the Holiness School’, but this is
difficult to reconcile with the domestic setting of v. 46 and, as noted, the
expressions and ideas involved occur in Numbers as well as in the Holiness
Code (for tôšāb see also Gen. 23.2).3 A later origin, after the combination
of the Priestly and non-Priestly Exodus accounts, would need to be envis-
aged if the passage, or at least its location, were seen as dependent upon the
mention of the ‘foreigners’ in v. 38 (cf. Levin, p. 339), but the placing of such
an addition immediately after the Priestly ‘finale’ in vv. 40-42 is intelligible
enough by itself, without the need to see any original connection with v. 38.
Rendtorff compared the passage (apparently without v. 48a) to the pattern
of priestly Daat-collections which he found underlying Leviticus 6–7 and
11–15 (Gesetze, p. 58: cf. pp. 33-38, 55). What he had in mind was the
2
Ahuis’s attribution of these verses to a Deuteronomistic layer (his DtrT)
therefore cannot be sustained (Exodus 11,1–13,16, pp. 112-13, cf. p. 71), and the
argument based on a connection with vv. 3aα, 6b-8 is in any case worthless.
3
Smend (Erzählung, p. 137) found an additional argument for Ps in the
dependence of v. 44 on an addition to P in Gen. 17.12b-13a, but commentators
such as Gunkel and Westermann have seen no need to regard this section as
secondary.
12.43-49 141
‘allmähliche Anfügung einzelner Sätze zu einem bestimmten Thema, wobei

z.T. eine einheitliche Stilform Verwendung findet’. In this case, however, the
epithet ‘priestly’, in the sense that the knowledge was for the priests alone
rather than the laity, scarcely fits: one could perhaps better think of priests
formulating guidance for the laity mainly in the same form as they used for
their own ‘manuals’. But the third-person verbs here all relate (except for
v. 47) to foreigners and when the Israelites themselves come into play (as in
vv. 44, 46 and 48-49) they are addressed directly in the second person. This
then resembles rather what Rendtorff called ‘priestly Torah’ for the laity,
such as he found in Leviticus 11 and also in parts of Exod. 12.1-20 (Gesetze,
pp. 12, 22-23; cf. 56-58). One may also doubt whether there really was a
process of ‘gradual accretion’ in vv. 43-49: the rulings about the participation
of foreigners in the Passover provide what seems to be a planned compre-
hensive coverage of the various categories of foreigner and the transition
already noted from issues about ‘eating’ to those about ‘celebration’ provides
an adequate explanation for the separation of the gēr from the others here.4
It is easier to recognise a progression between this passage as

a whole and the other Priestly texts that deal with Passover. Since
it does not give any basic information about how Passover is to
be celebrated, it must be seen as a supplement to Exod. 12.1-14.
Equally clearly Num. 9.1-14, which adds the possibility of a later
celebration of Passover in the second month, knows the regular
practice of Passover in the modified form introduced by the present
passage, so it must be later than it. What is more difficult is to see
where Exod. 12.14b-17 and 18-20 fit into this development. None
of the ‘Passover’ passages gives any hint that Passover formed
the prelude to a seven-day festival of Unleavened Bread, so it is
probable that even vv. 18-20, which we have argued to be the earlier
of these two sections (see the introduction to 12.1-20), was added
after 12.43-49 and Num. 9.1-14.5 Exodus 12.14b-17, which intro-
duced the centralised celebration of the whole two-part festival as
prescribed in the Holiness Code (Lev. 23.4-8), will be later still.
If this sequence is correct, the admission of gērîm to the Passover
as well as slaves (provided they were circumcised) was not due to
the Holiness Code but to an innovation within the Priestly tradition
4
Linguistic reasons may, however, suggest that vv. 46b and 49 were added
secondarily to the main unit (see the Explanatory Note on v. 46). Both provisions
are present in Num. 9.1-14.
5
12.18-20 shares with these passages the inclusion of the gēr.
142 EXODUS 1–18
that preceded it. It was an innovation that found its legitimacy

within the context of the family unit which had once again become,
after Josiah’s short-lived imposition of the centralisation required
by Deuteronomy (2 Kgs 23.21-22: cf. Deut. 16.1-8), the main locus
for this ancient and inspiring memorial of Israel’s deliverance from
Egypt.6 The solidarity of the family unit (perhaps in the Diaspora
as much as, or even more than, in the homeland) allowed, and even
demanded, that a welcome to the community’s celebrations be
extended to those who were willing to share in its mark of identity,
circumcision. So, for some, might the hopes of prophets and psalm-
ists (Isa. 49.6; Ps. 98.4-6) begin to become a reality.
43 [Yahweh said to Moses and Aarona, ‘This is the statute for

the Passover meatb. No one of foreign birth shall eat any of itc.
44 But every slave who isd bought with money you shall circum-
cisee, thenf he may eat of itc. 45 A settlerg who is a hired man shall
not eat of itc. 46 In one house it shall be eaten: you shall not take
any of the meath outdoorsi from the house. You shall not break
any of its bones. 47 All the congregation of Israel shall celebratej
it. 48 But if a resident alien is dwelling with you and celebratingk
Passover to Yahweh, all his males shall be circumcisedl and then
hem may come near to celebrate it and hem shall be like a native
of the land. But no one uncircumcised may eat of itc. 49 There
shall ben one law for the native and the resident alien who resides
in your midst.’]

a. Unlike v. 1, which is otherwise similar, the preposition ‫ אל‬is not repeated
in MT (but see Text and Versions). This is most unusual: in the Pentateuch
there seem to be only two other such instances (Gen. 30.40; Num. 16.25),
compared with 45 places where ‫ אל‬is repeated (or, twice, followed by ‫)ל‬. G.A.
Rendsburg, ‘The Nature of Qumran Hebrew as Revealed Through Pesher
Habakkuk’, in E. Tigchelaar and P. Van Hecke (eds.), Hebrew of the Late
Second Temple Period (STDJ 114; Leiden, 2015), pp. 132-59 (143-44), has
6
The celebration of Passover and Unleavened Bread in connection with the
dedication of the Temple in 516 B.C., which is recounted in Ezra 6.19-22, would if
historical provide a terminus ante quem for the insertion of Exod. 12.18-20 and so
(on our reconstruction) a fortiori for vv. 43-49 and Num. 9.1-14. But the account
in Ezra may (like 2 Chr. 30) be an anachronistic retrojection of later practices.
12.43-49 143
noted a trend towards such omissions in LBH, but they are occasionally found
already in the classical language: in addition to the parallels cited above see
12.28 and Note a on the translation of 12.28-42, 50-51.
b. Heb. ‫חקת הפסח‬. On ‫ חקה‬see Note hh on the translation of 12.1-20: the
nomen rectum here provides ‘nearer definition’ (GK §128f) of a kind which is
not easily subsumed under the usual categories in the grammars but is quite
common with words for ‘law’ etc. in headings as here or in colophons: cf. Lev.
6.18; 7.1, 11; 11.46; 14.2 (cf. 54, 57); 15.32; Num. 5.29; 6.13, 21; 19.14 (all
with ‫)תורה‬. It seems to be a distinctively Priestly usage but occurs only here
and in Num. 9.12, 14 (a related passage) with ‫חקה‬.7 In view of what follows
‫ פסח‬must refer here (as in vv. 11 and 21) to (the meat of) the slaughtered
animal.
c. Heb. ‫בו‬, as again in vv. 44, 45 and 48: in v. 9 the common ‫ מן‬was used
for the same sense, but for ‫ ב‬cf. Lev. 22.11, 13; Judg. 13.16; Job 21.25. BDB,
p. 88 (s.v. I.2.b), compares the meaning ‘among’, implying a part of a whole,
as with ‫ שׁתה‬and ‫ לחם‬in Prov. 9.5.
d. Heb. ‫אישׁ‬, lit. ‘a man’, with ‫ מקנת־כסף‬in apposition to it, lit. ‘a purchase
of money’, giving a closer definition like many other words that follow ‫ אישׁ‬in
apposition (cf. BDB, p. 36; GK §131b), as in 2.14. The alternative is to see
‫ עבד‬as in the construct state, ‘a man’s slave’, i.e. ‘someone’s slave’, which the
MT accents seem to support (cf. the suffix added in ‫ כספו‬of SP: see also Text
and Versions): AV, RV and JPS render thus, while the other modern EVV. (and
Luther) seem to follow the first view.
e. For the form of the sentence, with the anteposed object resumed by the
pronoun ‫ אתו‬cf. GK §112mm, 143d; JM §156a, c.
f. Heb. ‫אז‬, which is here used in a logical sense, as again in v. 48 and
elsewhere (cf. BDB, p. 23, s.v. 1.b); probably with a degree of emphasis
(ibid.), implying ‘only then’.
g. Heb. ‫תושׁב‬. The word is evidently related to ‫ישׁב‬, ‘dwell’, but its frequent
association with Heb. ‫גר‬, ‘resident alien’ (e.g. Gen. 23.4), shows that it is by no
means a synonym of ‫ישׁב‬, ֵ ‘inhabitant’. Its genuine occurrences (in 1 Kgs 17.1
MT is clearly corrupt and should be emended with LXX: cf. BHS) are limited
in their distribution and probably time of origin: BDB, p. 44, describes it as
‘only P (H) and late’. Eight of the thirteen occurrences are in Leviticus 22 and
25 (mainly the latter): the others are Gen. 23.4; Exod. 12.45; Num. 35.15 (all
P); Ps. 39.13; 1 Chr. 29.15. The word does not occur in Deuteronomy, where
‫ גר‬is of course common. Cognates are attested in various forms of Aram., but
not elsewhere: the earliest occurrence of a cognate by some way is ‫ תותב‬in
Ahiqar, l. 112/160 (DNWSI 2, p. 1207; TAD 3, pp. 46-47), where it is said in
a fable that ‘there is nothing taken more lightly (‫ )קליל‬than a foreigner (‫’)תותב‬
(tr. Lindenberger, in Charlesworth [ed.] 2, p. 501, with a note that the literal
meaning is ‘sojourner, resident alien’). The date of the sayings of Ahiqar is
7
1 Sam. 10.25 perhaps provides an analogous (and non-Priestly) use of ‫משׁפט‬.
144 EXODUS 1–18
currently placed c. 600 B.C. and their likely origin in Syria (ibid., p. 482). The
lowly status of such persons corresponds to (most of) what is found in the OT
occurrences. A precise definition of the socio-economic position of a ‫ תושׁב‬has
proved difficult to attain (see briefly TWNT 5, p. 846 with n. 35 [ET, p. 848];
F. Horst, Gottes Recht: Gesammelte Studien zum Recht im Alten Testament
[Munich, 1961], p. 220; de Vaux, Institutions 1, pp. 116-18, ET pp. 74-76;
HAL, pp. 1578-79); the comment in DCH 8, p. 616, represents a widespread
view: ‘similar to the ‫ גֵּ ר‬but with fewer rights’. The problem may be that the
word did not have a precise meaning, but needed an associated word to give it
one. In legal texts it is almost always associated with either ‫( שׂכיר‬as here: cf.
Lev. 22.10; 25.6, 40) or ‫( גר‬Lev. 25.23, 35, 47 [2x]; Num. 35.15), and it has
been plausibly suggested that these pairings are examples of hendiadys: so K.
Elliger, Leviticus (HAT; Tübingen, 1966), p. 293 n. 32, for the latter pairing,
apparently unaware that E.Z. Melamed had much earlier argued that both
pairings should be understood in this way (‘Hendiadys [ἓν διὰ δυοῖν] in the
Bible’ [Heb.], Tarbiz 16 [1944–45], pp. 173-89 [175-76, 179]), as Houtman
was also to do, apparently independently (cf. 1, pp. 287, 305 n. 28; 2, p. 207).
Melamed is followed by Milgrom, Leviticus 17–22, pp. 1861-62 (on 22.10);
Leviticus 23–27, pp. 2187, 2206-208 (on 25.23, 35), 2221-23 (on 25.40), who
among other arguments points to the use of a sing. verb with the combined
phrase here in Exod. 12.45 and in Lev. 22.10. A telling additional argument
is that in Lev. 25.35-43 two successive stages in the destitution of an Israelite
are to be remedied by treatment first as a ‫( גר ותושׁב‬v. 35) and secondly ‫כשׂכיר‬
‫( כתושׁב‬v. 40). This surely implies that in these laws ‫ תושׁב‬is a general word for
a ‘settler’ (as ‫ גר‬had been in Deuteronomy and as ‫ תותב‬seems to be in Ahiqar),
but that a distinction is now being made between the ‫גר ותושׁב‬, who has the
more generous rights previously afforded to the ‫( גר‬cf. Exod. 12.48 here and
the notes below), and the ‫( שׂכיר ותושׁב‬or vice versa, as here), who is simply
a foreign hired labourer, without additional privileges. In Lev. 25.45 the pl.
is probably being used to cover both these categories, and the metaphorical
uses in Ps. 39.13 and 1 Chr. 29.15 (and possibly also Lev. 25.23) are likewise
imprecise.
h. Heb. ‫מן־הבשׂר‬, with partitive ‫ מן‬as in vv. 7 and 22.
i. Heb. ‫חוצה‬, ‘to outside’, more commonly has the article, but for its
absence cf. Num. 35.4; 1 Kgs 6.6; Isa. 33.7; Prov. 5.16; 2 Chr. 24.8; 29.16.
The sense is perhaps not ‘to anywhere away from the house’, but ‘outdoors’
(cf. v. 22, though a literary relationship with that verse need not be presumed).
j. Heb. ‫יעשׂו‬. ‫ עדה‬as a collective generally takes a pl. verb (cf. 16.1-2; 17.1;
35.20: GK §145b-c).
k. Heb. ‫ועשׂה‬. The perfect consecutive is presumably to be understood
in the same way as the imperfect ‫יגור‬, whether as a future or as iterative. In
either case this creates a difficulty, as the law goes on to insist that it is only
after the ‫ גר‬and his male relatives have been circumcised that he or they are
permitted to participate in the Passover meal. Some translations (e.g. NRSV:
12.43-49 145
cf. Houtman, pp. 208-209; Joosten, Verbal System, p. 300) therefore take the
perfect consecutive as desiderative (‘wants to celebrate…’), but while such a
usage is well attested for the imperfect (JM §113n), it is rare with the perfect
consecutive (JM §119w; Joosten cites Exod. 5.5 and 2 Kgs 14.10, but the
latter is not a real parallel).8 Perhaps, therefore, the iterative interpretation is
best, so that the law seeks to tighten up a hitherto lax attitude to the celebra-
tion of Passover.
l. Heb. ‫המול‬. The inf. abs. (Niphal) is here similar to a jussive, with ‫כל־זכר‬
most simply regarded as the subject (cf. GK §113cc,gg).
m. Heb. ‫יקרב לעשׂתו והיה‬. The first verb at least is a permissive imperfect
(like ‫ יאכל‬in v. 44: cf. GK §107s). The subject in both cases could theoretically
be either the ‫ גר‬himself or ‫ כל־זכר‬as a collective singular.
n. Heb. ‫יהיה‬, with gender discord even though the subject precedes: so also
in the same formula in Num. 15.29 and the similar one in Num. 9.14: more
examples in GK §145u and JM §150k.
Explanatory Notes
43. The opening address to Moses and Aaron recalls 12.1 (see
the Explanatory Note there, where numerous parallels in legal
material in Leviticus and Numbers are mentioned) and is itself
referred back to in v. 50: the section thus forms a supplement to the
divine instructions in vv. 1-20. The heading ‘This is the statute for
Passover meat’ (see Note b on the translation) is of a type which is
used to introduce both legal material (cf. 21.1) and various kinds
of list (e.g. Num. 33.1-2). Here it begins a series of prescriptions
which are all concerned with the eating of the Passover animal and
especially with who is permitted (or required) to eat it and on what
terms. It is in fact not one ‘statute’, at least in its present form,
but the common theme allows it to be referred to as such, both
in v. 49 (‘one law’, a different Heb. word) and in Num. 9.12, 14.
The first prescription (to which vv. 44-45 seem to be attached as
clarifications) is a ban on participation in the meal by any foreigner.
In a celebration which was to commemorate a decisive event in the
early history of Israel it was natural enough to restrict the meal to
those who could claim to be direct descendants of those who came
out of Egypt.
8
Num. 15.14, where NRSV also translates ‘wishes to’, is not an example: ‘and
offers’ is quite adequate there.
146 EXODUS 1–18
44-45. But such a general ban was bound to raise some questions
about its exact application. One that is not dealt with here is whether
a person with only one Israelite parent could be admitted (Deut. 23.3
may have originally ruled against such a person: in Priestly narra-
tives Gen. 28.2 and Num. 25.6-18 could form the basis for a similar
exclusion). The cases that are dealt with here (and also in vv. 48-49)
are based on the socio-economic status of the persons concerned.
The first is a slave acquired by purchase (on the precise grammat-
ical structure of the Heb. see Notes d and e on the translation), who
is thereby distinguished from a slave born in the household: cf. Gen.
17.12-13 for this distinction (only the latter was presumably subject
to the ‘eighth day’ rule). The need for circumcision implies that
the purchased slave was a foreigner, and in the Holiness Code it
is specifically laid down that Israelites may only take foreigners
(including ‘resident aliens’) as slaves (Lev. 25.39-46), a change
from the older legislation later in Exodus (21.11) and in Deuter-
onomy (15.12-18). Slaves are evidently regarded here as so much
part of the family that, as long as they were circumcised, they could
participate in the Passover meal (a similar way of thinking seems to
lie behind Lev. 22.11).
By contrast, another group of foreigners (the context implies that
they were foreigners) were not to be admitted: as the Heb. literally
has it, ‘a settler and a hired man’. The first expression is not the
more common word gēr, ‘resident alien’, which comes later in vv.
48-49 (on it see 2.22 and Note dd on the translation of 2.11-22),
but tôšāb, a derivative of the root yšb, ‘dwell’, which is applied to
what might be called ‘settlers’ as distinct from the ordinary ‘inhab-
itants’ (Heb. yōšebîm) of the land. tôšāb is a much rarer word than
gēr (only thirteen occurrences in Biblical Hebrew) and unlike gēr
it never occurs in the Book of the Covenant or in Deuteronomy. It
appears to be a later word from the exilic and post-exilic periods
and eight of its occurrences are in the Holiness Code, mostly in
Leviticus 25 (see further Note g on the translation). In meaning it
cannot be very different from gēr, as they are used together several
times in a metaphorical sense (Ps. 39.13; 2 Chr. 29.15; perhaps Lev.
25.23), but the separate, and more generous treatment of the gēr in
vv. 48-49 suggests that they are not identical.9 The ‘hired man’ (Heb.
9
So also J. Joosten, People and Land in the Holiness Code (VTSup 67;
Leiden, 1996), pp. 73-74, with the suggestion that the words belong to different
‘spheres’: gēr is a juridical term, while tôšāb describes a social condition.
12.43-49 147
śākîr) is more widely referred to (including Deut. 15.18; 24.14) and

occurrences of the related verb and nouns in narratives and poetry
show that such paid workers performed a variety of tasks, acting as
shepherds (Gen. 30.32; 31.7), nurses (Exod. 2.9), craftsmen (1 Kgs
5.15; Isa. 46.6) and soldiers (Jer. 46.21), for example. In a number
of cases they were working away from their homeland and probably
did not own any land.
Most commentaries and other studies (e.g. C.J.H. Wright, God’s
People in God’s Land [Grand Rapids and Exeter, 1990], pp. 99-103)
have treated tôšāb and śākîr here as two distinct categories (Wright,
p. 102, sees them as two different types of gēr), but the possibility
should be seriously considered that they form a hendiadys, as in our
translation, ‘a resident alien who is a hired man’. This is particularly
likely in the light of Lev. 25.35-43, where two distinct situations of
an Israelite who falls on hard times are compared respectively to the
gēr wetôšāb (v. 35), ‘the resident alien and settler’, who has better
conditions, and the śākîr wetôšāb, ‘the hired man and settler’ (v. 40),
who is little better than a slave: tôšāb, the newer and broader term,
has its precise meaning specified by gēr or śākîr in these later laws.
For further references and arguments in favour of this view (which
is already found in TgN: see Text and Versions) see Note g on the
translation. This category of person (or each of these categories,
on the more common view) is excluded from the Passover, being a
foreigner who lacks the close ties with an Israelite family which a
slave has.
46. The two prescriptions in this verse are addressed directly, as in
vv. 48-49, to the worshipping community (‘you’) and the variation
between sing. (vv. 46a, 48 – cf. v. 44) and pl. (vv. 46b, 49) forms of
the pronoun in MT may indicate a composite origin for these verses
(see further Text and Versions). The mention of the ‘house’ reaffirms
the family setting of the celebration that was already presumed
earlier (cf. vv. 3-4, 7, 13: also vv. 22-23), and the warning against
taking any of the meat outside recalls the danger on the first Passover
night to anyone who was outside the protected doors of the house
(cf. vv. 7, 13: also v. 22). Even in a safer situation the celebrating
group was to remain together. The wholeness of the animal was also
to be safeguarded: just as it was to be cooked whole, ‘roasted by
fire’ (vv. 8-9), none of its bones must be broken (cf. Num. 9.12).
This may be a specific prohibition related to the process of cook-
ing, as preparation for cooking by boiling in a pot would be likely
to lead to the breaking of some of the animal’s bones (cf. Mic. 3.3:
148 EXODUS 1–18
Propp, p. 418). Or it may refer to the search for additional nourish-

ment after the meal from the marrow in the bones (so TgJ: see Text
and Versions): for further suggestions see Houtman, pp. 207-208;
Propp, pp. 418-19. An early interpretation saw v. 46b as related to
Israel’s own promised preservation (Jub. 49.13-15).
47. The final verses of the section turn to a concern with those
who may (or must) not merely eat the Passover meal, but celebrate
it (Heb. ʿāśāh, lit. ‘do’, in vv. 47, 48 [2x]: the verb is not used in
this sense in vv. 1-20, but it occurs repeatedly in Num. 9.1-14),
that is take charge of the preparation and the eating of the meal (cf.
Houtman, p. 134). First it is prescribed that the whole ‘congrega-
tion’ of the people is to take part, as already stated in vv. 3 and
6 (see the notes there): it is not just for selected representatives,
it has ‘become a test case of membership in the ʿēdāh’ (Blenkin-
sopp, Pentateuch, p. 156). This requirement is presumably what lies
behind the provision in Num. 9.1-14 for a second Passover in the
following month for anyone who with good reason was unable to
participate at the normal time. There, though not here, punishment
(of an unspecified nature) is prescribed for others who fail to take
part in this important observance (Num. 9.13).
48-49. The final provisions of this section permit the ‘resident
alien’ (Heb. gēr) too, if he and his household are circumcised, to
celebrate the Passover, apparently as an independent family head.
With Joosten (People and Land, pp. 54-73) and some others I do not
regard gēr as having a meaning in Priestly texts which differs from
that which it has elsewhere in BH.
Excursus on Laws about ‘Resident Aliens’

Already in the Book of the Covenant the gēr was protected against ill-treat-
ment and expected to share in the Sabbath rest (Exod. 22.20; 23.9, 12: cf.
20.10). In Deuteronomy the Israelites’ duty to care for the gēr was extended
(or at least spelt out more clearly) to include just treatment in the law courts
(1.16; 24.17; 27.19), provision of food (with other personae miserabiles)
from the tithe of each third year (14.29; 26.12-13: cf. 10.18-19), payment of
wages on time (24.14) and a welcome into the celebration of the festivals of
Weeks and Booths (16.11, 14: cf. 26.11; 31.12). As non-Israelites they were
not bound by the food-laws (14.21); among the tasks assigned to them might
be wood-cutting and fetching water (29.11).
The Priestly source in its original form seems to ignore the existence of
gērîm in Israelite society, just as it does most other features of life outside
12.43-49 149
the ambience of the tabernacle/temple cult (Exod. 12.19 and Lev. 16.29 are
both parts of secondary additions to P). It does, however, extend the status
of gērîm to the patriarchs in Genesis (23.4; 37.1: cf. Exod. 6.4), and regula-
tions about the gēr are frequent in the later parts of the Priestly corpus. Some
of the Deuteronomic and older provisions are repeated (Lev. 16.29; 19.10
[cf. 23.22], 33-34; 25.6) and the Deuteronomic motivation based on Israel’s
time as gērîm in Egypt is taken up (Lev. 19.34). A number of laws are said to
apply equally to ‘native’ Israelites and to gērîm: about sexual relations (Lev.
18.26), about Molech worship (20.2), about blasphemy (24.16), about equal
retribution (24.22) and about accidental homicide (Num. 35.15). Of particular
relevance to the present passage are those laws about worship and ritual
which are said to apply to gērîm as well, because they imply that the latter
are permitted to participate in the practices of the cult on an equal footing,
which is quite remarkable (Lev. 17.8, 10, 12, 13, 15; 22.18; Num. 15.14, 21,
29, 30; 19.10). It is revealing that Lev. 25.47-55 envisages the possibility that
‘resident aliens’ (the Heb. is gēr wetôšāb) may become so prosperous that
they can acquire slaves of their own, and this is presumably only mentioned
because it was a reality in the time when the chapter was compiled. Wealthy
gērîm may have been particularly ready and able to participate in the worship
of the Jerusalem temple, and it seems that their participation was welcomed by
the circles who were responsible for elaborating the Priestly legislation, just as
their poorer predecessors had been welcomed by the authors of Deuteronomy.
It is not so different after all from the spirit of the prophecy in Isa. 56.3, 6-7.
If we are right in our interpretation of v. 45, the gēr here corre-

sponds to the gēr wetôšāb, whose status is made the model, in Lev.
25.35-38, for the first stage of dependency of an Israelite who falls
on hard times. Such foreigners too will have been permitted to ‘live
with’ a family, not necessarily in the same house but as part of the
same community and sharing in the protection and kindness which
that brought (for the strong sense of ‘with’ as carrying such implica-
tions see Wright, op. cit., pp. 64, 102, and TWAT 1, 486 = TDOT 1,
p. 450; the same may be true of ‘in your midst’ in v. 49). As part of
such a close association the gēr was permitted to participate in the
sacrificial worship of the temple (see the Excursus) and these verses
also specifically allow the possibility of full participation, even as
a family head, in the Passover celebration.10 This is only paralleled
10
On ‘he may come near’ McNeile commented: ‘The priestly writer here
betrays himself. The expression must mean that the worshipper is to come near
to the Temple at Jerusalem, where the lambs were killed and offered…’ (p. 77).
Possibly: but 36.2 would support a translation like ‘join in’ (sc. with the Israelite
family heads).
150 EXODUS 1–18
in Num. 9.14, 2 Chr. 30.25 and possibly (the text is uncertain) Ezra
6.21, and it looks like an advance beyond Deuteronomy, which has
nothing in 16.1-8 (the Passover legislation) corresponding to the
provisions for the participation by the gēr in the feasts of Weeks and
Booths in vv. 11 and 14: perhaps the Passover was too closely tied
up with Israel’s history and divine election for the Deuteronomists
to be able to envisage the participation of foreigners in it. Even
here a strict condition is attached: both the gēr himself (presum-
ably the family head) and all male members of his household must
be circumcised. If they were to participate in a festival so closely
bound up with Israel’s origins and identity, they must also accept
the sign of membership in the people as it had been strictly laid
down in the covenant with Abraham in Gen. 17.9-14 (cf. Lev. 12.3;
and also Josh. 5.2-10).11
Such an extension of circumcision and all that it implied to
‘resident aliens’ was a major step beyond what had been said in
Deuteronomy and the original Priestly Writing (even if the circum-
cision of Ishmael in Gen. 17.25 might be seen as a partial precedent),
and it was taken further in the later regulations for the admission of
proselytes to the Jewish people (cf. the ‘minor tractate’ Gerim, and
M. Ohana, ‘Prosélytisme et Targum palestinien: Données nouvelles
pour la datation de Néofiti 1’, Bib 55 [1974], pp. 317-32 [322-23];
J. Jeremias, Jerusalem zur Zeit Jesu: kulturgeschichtliche Untersu-
chung zur neutestamentlichen Zeitgeschichte [Göttingen, 3rd ed.,
1962], pp. 354-70, ET, pp. 320-34; E. Schürer, The History of the
Jewish People in the Age of Jesus Christ (175 B.C. – A.D. 135), rev.
ed. by G. Vermes, F. Millar and M. Goodman, III/1 [Edinburgh,
1986], pp. 159-76 [esp. 169-76]).
The full acceptance of the circumcised gēr is confirmed by the
words ‘he shall be like a native of the land’, a statement paralleled
in Lev. 24.16; Num. 15.15; Ezek. 47.22 (where it even applies to
the allotment of land), as well as in a larger number of places which
make laws applicable to both ‘gēr and native’ (Exod. 12.19; Lev.
16.29; 17.10, 12, 13; 18.26; 20.2; 24.22; 25.6; Num. 19.10; 35.15;
11
This law may have been designed to regulate more strictly a situation in
which some gērîm were already celebrating Passover: see Note k on the transla-
tion (on ‘[are] celebrating’). Another possibility (cf. Milgrom, Leviticus 1–16,
p. 1055) is that the law thereby spares gērîm who are not celebrating Passover
from the need for circumcision.
12.43-49 151
Josh. 8.33; 20.9). The equality is specifically brought out in v. 49,

‘There shall be one law for the native and the resident alien…’, and
similar expressions in Lev. 24.22; Num. 9.14; 15.15-16, 29. Both
rights and duties are covered in these laws. The full expression ‘the
resident alien who resides in your midst’ (with minor variants) is
also widespread in the later strata of the Priestly corpus (Lev. 16.29;
17.10, 12, 13; 18.26; 19.34; 20.2; 25.6, 45; Num. 15.15-16, 26, 29;
19.10; Josh. 20.9), as well as occurring in Ezek. 47.22-23. It is hard
to see how the words ‘who resides’ (haggār) would add anything
to the meaning of gēr itself: their purpose is presumably to attach
the prepositional phrase ‘in your midst’ (etc.), which underlines the
close spatial association of the gēr with the Israelite community and
the corresponding duties and privileges which go with it.
Text and Versions

Before v. 43 4QpalExl and perhaps 4QExc had an interval; so also possibly
4QDeutj and 4Q136, but in them v. 43 follows a different passage altogether
(Deut. 11.21). On these two mss and the other phylactery mss which preserve
part or all of vv. 43-49 see Text and Versions on 12.28-42, 50-51, before the
notes on v. 50.
‫( ויאמר‬12.43) Almost all witnesses agree with MT, but the erratic phylac-
tery text 8Q3 apparently read ‫)[וי]דב[ר]( וידבר‬.
֯
‫( ואהרן‬12.43) SP has the more regular ‫( ואל אהרן‬cf. Note a on the transla-
tion), which is taken in BHS to be supported by the wlʾhr(w)n of TgO,J and Sy
(also TgN). But this may simply reflect the requirements of the target language,
not a divergent Vorlage. LXX and Vulg seem to presuppose MT, which is
supported by XQ1: as the difficilior lectio this is to be preferred. LXX adds
λέγων as in 3.12; 6.6; 7.1; 9.8; 10.1.
‫( חקת‬12.43) On the spelling ‫ חקות‬in some SP mss see Text and Versions on
v. 17. For LXX (cf. Sy) νόμος cf. 13.10 (and v. 49); for Vulg religio (here in
the sense ‘taboo, rule’?) 29.9; Num. 19.2 (and above in v. 26, where it renders
‫)עבדה‬. Tgg have ‫גזירת‬, having previously used ‫ קים‬in vv. 14, 17, 24; TgN adds
‫ אורייתא‬from v. 49.
‫( בן־נכר‬12.43) The expected translation appears in LXX, Vulg, TgN and
(in pl.) Sy, but TgJ adds ‘or an Israelite who has apostasised and not returned’
and TgO has the first part of this in place of MT. MRI suggests that comparison
with Ezek. 44.9 may have been the basis for this looser interpretation (Lauter-
bach, p. 118). M. Ohana regards TgN’s literal translation as very early (before
the persecution of Antiochus), as in v. 48 (‘Agneau pascal et circoncision:
le problème de la halakha prémishnaique dans le targum palestinien’, VT 23
[1973], pp. 385-99).
152 EXODUS 1–18
‫( בו‬12.43) Only TgO,J render precisely, but the ‘from it’ of the other Vss
gives the correct sense.
‫( וכל־עבד‬12.44) 8Q3 seems to have omitted the first half of the verse,
which had the effect of making the second half apply to the foreigner in
v. 43. XQ1 and possibly 4Q140 omit the initial waw. TgJ makes explicit that
the slave is a foreigner (and that the purchaser is an Israelite), in line with
Lev. 25.39-46.
‫( אישׁ‬12.44) The connection to ‫( עבד‬see Note d on the translation) is
explicit in LXX (τινός), TgJ,N and Sy, but not in TgO (cf. AramB, ‘male’, of
the slave) and Vulg, which has no equivalent. Some variants in the Heb. of the
next phrase also imply that the ‫ אישׁ‬is the owner (see the note).
‫( מקנת־כסף‬12.44) LXX ἢ (or καὶ) ἀργυρώνητον implies a Vorlage
‫ומקנת־כסף‬, a reading which appears in XQ1 and also in the early SP ms. Camb.
1846 (‫)ומקנות‬, so including both homeborn slaves and those purchased, as in
the similar prescription in Gen. 17.13 (cf. Wevers, Notes, p. 191). The effect
is also to rule out any connection with ‫ אישׁ‬and to indicate that it was being
taken closely with ‫עבד‬. The same is implied by the standard SP reading ‫כספו‬
(the antecedent of the suffix must be ‫ אישׁ‬understood as the owner), which
also appears (spelt ‫ )כשׂפו‬in 4Q136, perhaps due to independent occurrences
of dittography.12 Vulg and TgNmg, like LXX, render by a single word, the
other Vss more literally; in 4Q128 the whole phrase may have been omitted.
According to DJD XII, p. 116, the traces that follow ‫ [אי]ש‬in 4QExc do not fit
‫מקנת‬. MT should probably be retained.
‫( ומלתה אתו‬12.44) SP, 4QpalExl (though faint), XQ1, LXX and TgO agree
with MT. The omission of the clause in 4Q136 (and possibly 4Q128) is clearly
an error, as it is presupposed by what follows. The subordinating translations
of Sy and TgNmg, the passive formulation in Vulg (cf. v. 48) and the pl. verb
in TgN (cf. v. 46) will be due to stylistic factors. TgJ has the interesting but
certainly secondary addition ‘and bath him’, which MRI (Lauterbach, p. 127)
suggests has the effect of manumission (cf. B.Yeb. 47b): it became a regular
part of the admission of proselytes. In place of ‫ אתו‬8Q3 seems to have had
‫[את ער]לתו‬, probably another sign of the influence of Genesis 17 (cf. vv. 11, 14,
23-25), although the precise phrase does not occur there (or anywhere else).
‫( אז‬12.44) So also SP, 4QDeutj, XQ1, 8Q3, 4Q136, Vulg, Tgg, Sy: LXX
καὶ τότε could be due to dittography in the Vorlage or the translator’s desire
to avoid asyndesis.
‫( תושׁב‬12.45) The Vss mostly use words, here and elsewhere, which
were, or had come to be, general expressions for foreign settlers in a place:
LXX πάροικος, Vulg advena, TgO ‫( תותבא‬cf. Sy). A distinction is generally
preserved between ‫ תושׁב‬and ‫ גר‬in vv. 48-49, as elsewhere, but Sy has twtbʾ
again in v. 48 (cf. 2.22; 18.3; 20.10; on the whole, however, as in v. 49, it
12
As Propp has seen (p. 366), this reading could also be based on Gen. 17.23.
12.43-49 153
uses other expressions for ‫)גר‬. The other Tgg make additions to clarify the
legal situation (as they do with ‫שׂכיר‬: see below). TgJ prefixes ‫דייור‬, which
means ‘lodger, traveller’, but was used in v. 19 to render ‫( גר‬so also in Gen.
23.4; Exod. 2.22; 18.3; 22.20; 23.9b; Deut. 10.19b; 27.19), so that TgJ prob-
ably follows MRI (Lauterbach, p. 121, with n. 5) in interpreting as ‫תושׁב גר‬, an
expression which was taken by the rabbis to mean a ‘semi-proselyte’, a person
who had renounced idolatry but remained uncircumcised (cf. B.Abodah Zarah
64b and further G.F. Moore, Judaism in the First Centuries of the Christian
Era in the Age of the Tannaim (Cambridge, 1927–30) 1, pp. 338-41 and 3,
p. 112; also JE 10, pp. 220-22). TgF (cf. TgNmg) has ‫תותבא דגבר‬, indicating a
status of possession, and may be based on the rabbinic interpretation of ‫ תושׁב‬as
a life-long slave (see refs. in Milgrom, Leviticus 17–22, p. 1861). Finally TgN,
anticipating a modern interpretation (cf. Note g on the translation), combines
‫ תושׁב‬with ‫שׂכיר‬: ‫תותב בר עממין לאגירא‬, a ‘gentile sojourner for wages’ (AramB
2, p. 36).
‫( ושׂכיר‬12.45) Apart from TgN (see the previous note) the Vss either
render straightforwardly (LXX, Vulg, TgO, Sy) or add the specification
‘foreign’ (TgJ,F,Nmg).
‫( בבית אחד‬12.46) TgO,J read ‘in one company’ and TgN ‘by companies’
in accordance with post-biblical practice (cf. M.Pes. 9.10; MRI [Lauterbach,
pp. 122-23]; Jos., AJ 2.14.6[312]; 3.10.5[248]; BJ 6.9.3[423]), and corre-
spondingly later in the verse.
‫( לא תוציא‬12.46) Sy makes a connection by prefixing ‘and’, as do many
LXX mss (cf. Rahlfs). SP reads the pl., as do XQ1 and perhaps 8Q3; 4QExc
most likely had the sing. (DJD XII, p. 116). No other Heb. evidence survives,
but the Vss all have the pl. Since assimilation to ‫ תשׁברו‬later in the verse may
well be the cause of all the pl. readings, and the second person sing. was used
in v. 44 (cf. v. 48), MT should probably be retained (so also Propp, p. 366):
the variation could well be due to the process of composition.
‫( מן־הבית‬12.46) XQ1 reads the longer form ‫(מן) הביתה‬, with the ending
‫ה‬- no longer indicating direction towards (cf. Qimron, p. 69). Vulg ignores the
phrase, probably seeing it as redundant, while Sy places its equivalent after
‘outside’ and has an anticipatory ‘from it’ here. TgN expands with ‘from one
house to another or even from one company to another’ (the tr. in the editio
princeps is preferable to that in AramB 2, p. 53), probably recognising that
two companies might meet in a single room, as in M.Pes. 7.13.
‫( מן־הבשׂר‬12.46) XQ1 omits the ‫ ;מן‬in LXX it is reflected in the (partitive)
genitive τῶν κρεῶν.
‫( חוצה‬12.46) XQ1 agrees with MT, but SP has ‫ החוצה‬and 4QDeutj ‫לחוץ‬.
The def. form is more common, especially in the Pentateuch (9x vs. 2x), and
SP probably modified the form here accordingly, as it did in Num. 35.4. ‫לחוץ‬
is found in MT only at Ezek. 41.17; 42.7; Ps. 41.7; 2 Chr. 32.5, but at Qumran
it seems to displace ‫ חוצה‬except in combination with other prepositions (cf.
11QT 41.12; 46.5, 14) and this may account for its appearance in 4QDeutj.
154 EXODUS 1–18
After this word TgJ has a further addition about not sending presents to
friends, which according to AramB 2, p. 195 n. 78, has no parallel elsewhere.
‫( ועצם לא תשׁברו־בו‬12.46) 8Q3 cannot have had these words at this point
but it may perhaps have interchanged them with the preceding prohibition.
TgJ adds ‘in order to eat what is within it [i.e. the marrow]’, an explanation
which alludes to the discussion in MRI (Lauterbach, p. 124); TgNmg has a more
general warning about not annulling a commandment.
‫( כל‬12.47) 4Q128 has ]‫ [ישראל יעשו‬and there is room for ‫ עדת‬before it, but
DJD VI, p. 51, says that there are no traces of ‫כ(ו)ל‬. If it was omitted (and it
is just possible that what has been read as [ ]‫ בו‬from the end of the previous
verse preserves it), it can only be an error, as the verse has no point without it.
‫( ישׂראל‬12.47) LXX, OL and Vulg prefix ‘the sons of’ as they do in vv. 3
and 6, to conform to the widespread formula. Only later mss of Sy have
this addition. TgJ adds ‘shall mix together, one with another, one family with
another (to celebrate it)’, deducing with MRI (Lauterbach, p. 127) that after
the first Passover the celebrations need not be restricted to single families.
‫( יעשׂו‬12.47) LXX and Vulg have sing. verbs, reflecting as elsewhere a
stricter view of grammatical concord: pl. verbs with ‫ עדה‬are common in MT
(see Note j on the translation), but appear in the LXX Pentateuch only at 16.1;
Lev. 24.14; Num. 14.2; 15.36 (2x); 35.25 and apparently never in Vulg.
‫( אתו‬12.47) TgN has ‘the Passover’ for complete clarity.
‫( וכי־יגור‬12.48) 4Q136 omits the waw, but it is present or represented in
all other witnesses (including XQ1) and should be retained. 8Q3 has a lacuna
between ‫ יגור‬and near the end of the verse and the space available indicates
omission of much that is in MT (cf. v. 44 and perhaps v. 46 for other careless
omissions in 8Q3). TgN and Sy render ‫ יגור‬according to its regular meaning:
LXX προσέλθῃ is a development in sense from προσήλυτος, its rendering
for ‫( גר‬see below), which appears also in v. 49. It is unclear whether TgO ‫יתגייר‬
and TgJ ‫ איתגייר‬make reference to conversion,13 as the Ithpaal is used both for
this and for the residence of a foreigner. On Vulg see the next note.
‫( אתך‬12.48) All the extant witnesses except MT have a second person
pl. suffix/pron. here. SP, 4QExc, XQ1 and 4Q136 all read ‫( אתכם‬cf. ‫ אתכםה‬in
4QDeutj, with the ‘Qumran ending’: Qimron, p. 64) and the Vss follow suit.
This could of course be due to assimilation to ‫ תשׁברו‬in v. 46 (and/or ‫ בתוככם‬in
v. 49), and MT may well be original.14 For the whole clause Vulg has quod si
quis peregrinorum in vestram voluerit transire coloniam, ‘But if any of the
foreigners wishes to transfer to your (pl.) colony’, a vivid rendering that draws
13
So M. Ohana, ‘Agneau pascal’, pp. 392-93, deducing that the full ritual
process is implied, whereas TgN reflects the earlier less developed view which
required only circumcision (see further id., ‘Prosélytisme et targum’, pp. 324-29).
14
In Num. 9.14 the same expression as here recurs with ‫אתכם‬: there the second
person pl. is used throughout vv. 1-14 (cf. vv. 3, 8, 10), in MT at least.
12.43-49 155
on the occasional equation of ‫ גר‬with a colonist in Vulg (cf. v. 49 and Lev.

18.26). There is nothing religious in the language it uses.
‫( גר‬12.48) The renderings in the Vss maintain at least the possibility of
the sociological sense of the Heb. word, even LXX προσήλυτος (cf. NETS,
‘guest’), on which see Text and Versions on 12.19. A distinction from ‫ תושׁב‬in
v. 45 is also mainly carried over by the use of different equivalents, but Sy
uses twtbʾ for both (despite the availability of ʿmwrʾ, which it has in v. 49).
‫( ועשׂה‬12.48) LXX ποιῆσαι (aor. inf.) oddly makes the celebration of
Passover the purpose of the outsider’s approach.
‫( ליהוה‬12.48) Tgg as often render ‘before the Lord’. LXX and Sy follow
MT, but Vulg’s ‘(the Passover) of the Lord’ may have a Christian liturgical
background.
‫( המול‬12.48) The sense of the Heb. Niphal inf. is only precisely rendered
by Vulg’s circumcidetur. TgO reproduces the Heb. construction in Aram.,
presumably with the same jussive sense (see Note l on the translation) but in
an active form, and TgJ represents this in a more straightforward way; likewise
TgN and Sy, with the clause subordinated to the next one. LXX περιτεμεῖς,
with its surprising return to a second person sing., perhaps read ‫ המול‬as an
imper. sing. Hiphil (which would really be ‫)ה ֵמל‬ ָ or even, like the original
writing of 4Q136, the imper. sing. Qal ‫מול‬.
‫( ואז‬12.48) TgN ignores the waw.
‫( יקרב לעשׂתו‬12.48) TgJ,N ‘shall be fit to celebrate it’, which is probably
behind Vulg’s rite celebrabit, another example of Jerome drawing on Jewish
sources.
‫( כאזרח‬12.48) 4Q128 (mis)spells ‫כיא[זר]ח‬, LXX ὥσπερ καὶ ὁ αὐτόχθων
perhaps underlines the identity of treatment by adding καί, and TgO,N and Sy
render by a pl. noun.
‫( וכל‬12.48) XQ1 and LXX ignore the waw; Vulg si quis autem matches
the opening words of the verse in a quasi-legal contrast.
‫( ערל‬12.48) TgJ and TgN limit the scope of the prohibition by contradic-
tory additions, the former asserting that uncircumcised Israelites are meant
here (with MRI [Lauterbach, pp. 127-28] and B.Shabb. 87a: cf. also Gen.
17.14), while the latter gives the contextually more likely application to
Gentiles.15 Verse 49 suggests that these are false alternatives.
‫( תורה‬12.49) TgN prefixes ‫גזירת‬, a common equivalent for ‫( חקה‬cf. v. 43
and Text and Versions there): the combined expression makes a harmonious
frame for the whole unit. TgNmg preserves a reading which uses the rare word
‫ חוזיה‬on its own for ‫( תורה‬cf. TgF at Lev. 7.7).
15
M. Ohana, ‘Agneau pascal’, pp. 395-96, argues that the application to Israel-
ites was a reaction to the persecution of Antiochus and that TgN’s more natural
interpretation is earlier than this.
156 EXODUS 1–18
‫( יהיה‬12.49) 4Q140 ‫ תהיה‬tidies up the grammar (see Note n on the transla-

tion), as the Vss naturally do also. TgJ generalises the statement to include all
the commandments, with MRI (Lauterbach, p. 128).
‫( לאזרח‬12.49) TgO,N and Sy have pl. equivalents for both this and the next
word. LXX uses its rarer equivalent ἐγχώριος here, probably for variety after
αὐτόχθων in v. 48. 8Q3 appears to have read ‫ לנכרי‬against all other witnesses,
perhaps assuming the ‘Gentile’ interpretation of v. 48b and following the
occasional alignment of ‫ גר‬and ‫ נכרי‬elsewhere (Lev. 25.44-45; Deut. 14.21).
‫( ולגר‬12.49) 4Q140 ‫( להגר‬and probably ]‫)לה[אזרח‬ ֯ unusually does not
assimilate the article after the preposition: for examples in MT (‘almost exclu-
sively in the later books’) see GK §35n.
‫( הגר‬12.49) LXX προσελθόντι is a unique equivalent to the participle,
which is clearly based on the rare use of the verb for ‫ יגור‬in v. 48. The same
uncertainty about the understanding of these expressions exists as in v. 48
(see the notes there): Vulg peregrinatur and Sy dʿmryn at least take them in a
sociological sense here.
‫( בתוככם‬12.49) LXX ἐν ὑμῖν follows a common tendency not to repre-
sent ‫ תוך‬explicitly (cf. 2.5; 3.2, 4; 9.24; 12.31) and provides no basis for
assuming a different Vorlage. The locational renderings of the phrase in all
the Vss reinforce the case for a non-religious understanding of the preceding
words.
C h ap t er 1 3 . 1 - 1 6
Laws ab ou t t h e C on s ecr ation

of th e F i r s t b or n an d t h e Festiva l
of U n leav en ed B r e a d
This further section of law is separated from the preceding one by

the concluding formulae of 12.50-51 and by a change of subject:
instead of Passover the topic is now the treatment of Israel’s
firstborn and the festival of Unleavened Bread. Unlike ch. 12
(vv. 1, 28, 43, 50), Aaron has no place here in the reception or
transmission of the law: Moses alone is involved (vv. 1, 3).1 The
beginning and end of the passage are marked by Masoretic divi-
sions, and 13.1 begins a new synagogue reading in the triennial
cycle. In the annual lectionary cycle the passage concludes the
reading that began at 10.1, thus emphasising the transition at its
end from legal to narrative text. No evidence survives for these
boundaries in the main biblical manuscripts from Qumran, but they
are recognised in the selection of passages in the phylacteries and
probably in the related mss 4QExodd (for the end) and 4QDeutj (for
the beginning). In cases where both 12.43-51 and 13.1-16 occur in
the same phylactery, there is a vacat between them in 4Q128, 136,
140 and XQ1, while in 8Q3 the two passages are on separate frag-
ments and were clearly not adjacent to one another.
Divine instructions to Moses (vv. 1-2) are followed by words of
Moses addressed to ‘the people’, first about the festival of Unleav-
ened Bread (vv. 3-10) and then about the treatment of animal and
human firstborn (vv. 11-16). Thus in chs. 12–13 as a whole Passover,
Unleavened Bread and the firstborn are each treated in both formats,
divine instructions (12.1-14[, 28, 43-51], 15-20; 13.1-2) and words
of Moses (12.21-27; 13.3-10, 11-16), with the order being the same
in each sequence and the divine instructions always preceding. Here
1
Aaron takes no further part in the departure from Egypt and is next men-
tioned in 15.20 and 16.2.
158 EXODUS 1–18
again, as we noted for the Passover in ch. 12, what is said in the two
formats is by no means the same, so that it is likely that two parallel
versions of this legal material have been combined by the compiler.
The curious separation of the two sections about the firstborn
here is probably due to a wish not to break up the closely parallel
paragraphs of Moses’ words in vv. 3-16 (cf. Propp, p. 381). It is the
more surprising that there is a Masoretic division between these two
paragraphs but not after v. 2. But vv. 1-10 and 11-16 were treated as
separate sections for inclusion in phylacteries, and a division at this
point is widely attested in the Judaean desert phylacteries, either by
a vacat or by inclusion on separate pieces of leather (4Q129, 130,
132+133, 134+136, 140, 144, 145, 155; 8Q3; XQ1+3; 34SeyPhyl;
Mur4).2 In these texts there are isolated instances of an interval after
v. 2 (34SeyPhyl) and v. 4 (4QDeutj, inferred).
The ‘words of Moses’ begin with a preamble (vv. 3-4) which
is, apart from the end of v. 3, a general exhortation to remember
the ‘day’ of the Exodus that is relevant to the whole of vv. 5-16.
These verses are distinguished from those that follow by the use of
second person plural forms rather than the sing. Formally, where the
content allows, vv. 5-10 and vv. 11-16 are very similar: each section
begins with (i) a temporal clause that defines the land of Canaan as
the place where the law is to apply (vv. 5a, 11), followed by (ii) the
central requirement (vv. 5b-6, 12), (iii) additional provisions (vv. 7,
13), (iv) a command to instruct the children about the meaning of
the ritual (vv. 8, 14-15) and (v) a concluding association of the ritual
with the Exodus by means of a comparison to visual ‘reminders’
worn on the body (vv. 9, 16). Verse 10 extends the conclusion of the
first part to emphasise that the festival is a regular annual occasion,
which the birth of a firstborn might well not be.
In contrast to the elaborate, almost sermonic, character of vv. 3-
16, the divine instructions in vv. 1-2 are a succinct command which
takes no account of complications. In this it resembles v. 12 if it is
seen in isolation and verses later in Exodus which also prescribe
the dedication of the firstborn (22.28b-29). Parts of vv. 3-16 are
also paralleled later in Exodus: for vv. 6-7 compare 23.15 and
34.18 (and Lev. 23.6-8; Deut. 16.1-8), and for vv. 12-13 compare
2
The only case where there was clearly no interval (in some cases evidence
is lacking) is XHev/Se5.
13.1-16 159
34.19-20 (and Num. 18.15-18; Deut. 15.19-23). In addition to the

central legal requirements, the ‘words of Moses’ have formulaic
and vocabulary parallels both elsewhere in Exodus and in Deuter-
onomy (see the Explanatory Notes, especially on vv. 3, 5, 8-9, 11,
14-16).
These features of the passage have been used by critical scholars to deter-
mine the process by which it was composed. Knobel already made the basic
distinction between vv. 1-2 and vv. 3-16, ascribing the former to P (his E),
because of references back to it in Leviticus and Numbers and aspects of
the language, and the latter to JE (pp. 127-28), later more specifically to his
Rechtsbuch (our E: cf. Num.-Jos., p. 532). Dillmann took a similar view, except
that he attributed vv. 3-16 to J (his C). But Wellhausen had already arrived at
the more complex view of these verses which was to be developed further by
subsequent scholars (Composition, p. 74): ‘Der Verfasser…ist, wenn nicht der
Jehovist [i.e. RJE] selber, ein deuteronomistischer Bearbeiter desselben’. He
stressed especially the author’s debt to ‘the sources’ of JE, including Exodus
34, which he regarded as distinct from J and E (pp. 95-96). But the end of
v. 7 and of v. 8 and the words ‘you shall hand over’ (Heb. wehaʿabartā) in
v. 12 came not from these sources but from Deuteronomy or ‘the writers of
the seventh century and the exile’. Several scholars of the following decades
took a similar composite view of these verses, while generally maintaining the
attribution of vv. 1-2 to P (Carpenter/Harford-Battersby, Holzinger, McNeile).
But an increasing emphasis on their Deuteronomistic affinities can be seen in
the works of Baentsch, Smend (Erzählung, pp. 132-35), Gressmann, Eissfeldt,
Rudolph, Beer, Noth, Fohrer and Hyatt. Alongside this the attribution of
vv. 1-2 to P received an initial (and isolated) challenge from Holzinger and
then, from Rudolph onwards, was generally rejected in German scholarship in
favour of either a Deuteronomistic or late Priestly origin. Outside Germany,
however, its place in the main Priestly document (or layer) has continued to
be upheld, with only a little hesitation (cf. Hyatt, Childs, Houtman [p. 148, in
effect], Van Seters, Propp, Dozeman).
The tendency to play up the Deuteronomistic affinities of vv. 3-16 was
itself subjected to a challenge similar to that discussed earlier in relation to
12.24-27 (see the introduction to 12.24-27), which has remained an important
element in the scholarly debate. First Lohfink, in his study of Deuteronomy
6, questioned whether the vocabulary and style of Exod. 13.3-16 were in
fact fully representative of the classic Deuteronomic and Deuteronomistic
conventions, although he acknowledged that they were related to them. The
presence of features that are not found in Deuteronomy and of others which
are similar but not identical to the standard formulations led him to propose
that Exod. 13.3-16 be described as ‘proto-Deuteronomic’, that is a stage on
the way to the production of Deuteronomy rather than a later text exhibiting
160 EXODUS 1–18
influence from Deuteronomy (Das Hauptgebot, pp. 122-24).3 A later study by

M. Caloz examined the linguistic evidence in painstaking detail and added
some observations on structural and legal comparisons between this passage
and the Deuteronomic and other laws that deal with the same subjects.4 Caloz
was able to give an even clearer perspective on the numerous elements in the
passage which were alien to Deuteronomy, showing that it occupied an inter-
mediate position between the older Pentateuchal sources (especially J) and
Deuteronomy itself. A particularly valuable distinction which he drew was
between the source-material and the editorial sections of the Deuteronomistic
History, since it was more often in the former that parallels to the language
of Exod. 13.3-16 could be found. All in all, Caloz concluded, the evidence
pointed rather to a pre-Deuteronomic origin than to a post-Deuteronomic one.
The ‘words of Moses’ have played a major part in the influential works
of Blum and Konrad Schmid. For Blum they are a key part of the network
of passages which he comes to define as the ‘Deuteronomistic composition
layer’ (Kd: Studien, pp. 16, 19, 35-36, 103, 106, 167-68, 172-76). The passage
is ‘closely [bündig] attached to 12.39 and makes explicit the aetiology of
Unleavened Bread’ which is implied there and in 12.33-34 (pp. 37-38). The
divine oath to the ‘fathers’ is an important link to other parts of the ‘network’
in Exodus (33.1-13) and elsewhere (pp. 103, 106) and is seen as speaking of
a future for the early post-exilic community which was both already being
realised and yet to be fully realised (p. 191). As for the dating of the passage,
Blum sought to address the arguments brought forward by Lohfink, Caloz and
others and argued that they were insufficient to undermine the presumption of
a post-Deuteronomic origin (pp. 167-68, 172-76).5
For Konrad Schmid it is really only the ‘oath to the fathers’ in vv. 5 and 11
that concerns him, since it forms a key feature of the ‘chain’ of passages which
with Rendtorff he regards as a later redactional layer that helped to make the
Pentateuch (Hexateuch?) into a theological unity (Erzväter, pp. 13, 239).
Unlike T. Römer, who thought that the reference was to Exod. 3.17 (p. 74),
and Blum, who had seen Gen. 15 as being meant, Schmid concluded that no
particular passage need be in mind.
Other recent scholars have dealt with the passage more in passing, and in
different ways. Graupner, whose interests lie elsewhere, followed the now
traditional arguments for seeing vv. 3-16 as Deuteronomistic, adding only that
the plural forms in vv. 3-4 point to a two-stage origin (Elohist, p. 68 n. 216).
3
Cf. more generally C.H.W. Brekelmans, ‘Die sogenannten deuterono
mistischen Elemente in Genesis bis Numeri. Ein Beitrag zur Vorgeschichte des
Deuteronomiums’, in Volume du Congrès Genève (VTSup 15; Leiden, 1966),
pp. 243-68, which proposed criteria to identify pre-Deuteronomic material.
4
M. Caloz, ‘Exode, XIII, 3-16 et son rapport au Deutéronome’, RB 75 (1968),
pp. 5-62.
5
See also Blum’s Komposition, pp. 374-75.
13.1-16 161
Dozeman and Propp acknowledge that the passage is ‘D-like’ (Propp, pp. 377-
78), but this does not prevent them from associating it with, respectively, the
‘non-P work’ (Dozeman, pp. 291-98) or the E source (Propp, ibid., but with
variant introductions in vv. 3-4 and later additions in v. 10 and either v. 6 or
v. 7). Van Seters too rejects the idea of a separate Deuteronomistic redaction
and attributes vv. 3-16 to J. He is thus able to retain its links to the wider nar-
rative context and explain the differences from Deuteronomy, but his exilic
dating of J means that the passage is still detached from older developments
in the tradition.
In Germany there has been a growing tendency, which picks up brief
comments made by Kohata (Jahwist, p. 275), to make much of what are seen
as connections with P as well as Deuteronomy and so to regard the passage as
very late. The ‘today’ of v. 3 is taken to connect to 12.41(P) rather than to 12.39.
Levin (pp. 339-40) found three stages of growth in (i) vv. 3a, 4, (ii) vv. 3b,
5-6 (with vv. 7-10 as a succession of additions) and (iii) vv. 1-2, 11-16, all
post-Priestly. Ahuis attributed the whole section, including vv. 1-2, to his
DtrT, who combined the older source-material with P (pp. 71-73). Gertz seeks
to strengthen the case by finding what he sees as Priestly vocabulary in the
expressions ‘remember’ (v. 3), ‘service/worship’ (v. 5), ‘reminder’ (v. 9) and
‘statute’ (v. 10) and making the separation of Passover and Unleavened Bread
and the inclusion of animals in Yahweh’s judgement (v. 15) dependent upon
the Priestly rather than the older traditions (pp. 63, 67, 68-69). Most recently
Albertz has attributed the whole of vv. 1-16 to his D redaction (although he
repeatedly mentions the possibility that vv. 3-4 may contain the continuation
of the older Exodus narrative: pp. 198 n. 1, 201, 220, 221), but he places this
after the first two Priestly Bearbeitungen, because the language of the passage
is a Mischsprache containing Priestly as well as Deuteronomic features (pp.
201-202; cf. Schmidt, pp. 557-58). This is of course largely due to the inclu-
sion of vv. 1-2, which is true of all the scholars in this group.
The older view that vv. 1-2 come from the Priestly tradition
is surely to be preferred. The only parallel with Deuteronomistic
language is the use of ‘consecrate’ in v. 2 and in view of its use
so widely in P and specifically in passages about the firstborn
(Num. 3.13; 8.17: cf. 18.17) there is no reason to speak of ‘mixed’
language here. The same passages also clearly refer back to this
one. It is more difficult to be sure whether vv. 1-2 are from the
original Priestly document or from a later redaction of it. Against
the former view it has been argued that the verses are isolated in the
context and that the fuller treatment of the subject in Num. 18.15-18
is more likely to be the original one. It is also suggested that the
requirement to consecrate the firstborn to Yahweh is unlikely to
have been made at a time when it could not yet be fulfilled. None
162 EXODUS 1–18
of these arguments is conclusive: the later elaboration of the law is

in a passage that is addressed specifically to priests (cf. Num. 18.8)
and could well presuppose a basic principle intended for the whole
people and uttered in the narrative context where the firstborn of
Egypt were certainly a central concern of P (12.12). Nevertheless
one cannot be confident about the stage in the Priestly tradition to
which it belongs and it may have been added in one of the later
strata of the tradition.
Verses 1-2 are so different from the treatment of the same topic
in vv. 11-16 that it is most improbable that they were originally
composed to be the introduction to them, as some recent commenta-
tors have supposed. The order of the paragraphs in 13.1-16 is also
difficult to square with such a view and makes it much more likely
that vv. 1-2 were attached to 12.42 or 12.51. Verses 3-16 therefore
need to be treated separately. Two preliminary remarks may be
made. One is that affinities to the Deuteronomic tradition are much
more evident here than in 12.25-27, so that even among those who
deny dependence upon Deuteronomy the description of the passage
as ‘proto-Deuteronomic’ is common. The second is that, while there
is a very close parallelism of structure and wording between the
two main parts of this section (vv. 3-10 and 11-16), it is not total –
vv. 3-4 lie outside the parallel structure and v. 10 has no corre-
sponding verse in the second part. This second observation may well
be relevant to the choice that has to be made between three main
kinds of explanation of the section: (a) a single author composed it
all, drawing on material from a variety of sources; (b) an underlying
parenesis was expanded by a redactor; (c) a combination of both
these conceptions.
To begin with literary-critical analysis, the formal anomalies
noted above combine with linguistic data to suggest that vv. 3-4
and 10 are not part of the main composition of the passage. Verses
3-4 are very likely the original continuation of the old J narrative
that ends in 12.39 (see the Explanatory Note and Albertz above).6
As they stand they read somewhat strangely, with v. 4 not being a
natural continuation of v. 3. Propp’s suggestion that two alternative
openings for the passage have been preserved is hard to parallel
6
A possible indication of Deuteronomic provenance for these verses is the
phrase ‘from the house of bondage’ in v. 3, but the surrounding phraseology is in
several ways distinct from Deuteronomy (see the Explanatory Note).
13.1-16 163
and does not work if, as seems likely, these verses originally stood
alone, without any continuation. It is more probable that at this
stage v. 4 preceded (most of) v. 3: when the law about Unleavened
Bread was expanded (and the law about firstborn was added) it was
moved to its present position to correlate more closely the (from a
narrative point of view) present experience of deliverance and its
timing with the future commemoration of it. This also enabled the
keynote of ‘remembering’ (cf. v. 9) to appear at the very beginning
of the now extended parenesis. Verse 10 includes the Priestly use of
‘statute’ (Heb. ḥuqqāh) in the singular with older vocabulary later
in the verse. Propp attributes the whole verse to the (final) redactor.
If ḥuqqāh is not simply an alternative or updating for the masculine
form ḥōq (which is attested in the sing. in some older texts: 5.14;
15.25; Pss. 2.7; 81.5-6), this may well be true for where it now
stands: and there would be a close parallel to the addition in 12.24.
But the older wording could have been taken from the end of v. 5,
where it fits well.
The main parenesis is characterised by the address of Moses to
the people in the second person singular, a form of address which
emphasises their collective unity. It is typical of passages in Exodus
in which Moses or more often God is instructing the people and
especially giving them laws (cf. 10.2; 12.44, 46, 48; 15.26; 19.23;
20.2-17, 24-26; 21.2, 14, 23; 22.17-29; 23.1-33; 34.12-26). It is
also very frequent in Deuteronomy. Second person plural forms are
found sporadically in such texts, but they are more characteristic
of dialogue in the narrative (e.g. Exod. 1.22; 3.12, 14-15, 21-22;
5.11, 18, 19; 13.19; 14.2, 13-14; 16.6, 23, 25-26; 17.2; 19.5-6, 15;
20.20, 22; 32.30). The Priestly source seems to prefer the plural
even in parenesis and law (Exod. 6.6-8; 12.1-20 passim; 25.9, 19
etc. [the many cases of the second person singular in chs. 25–31 are
addressed to Moses]).
The instruction that is given here has as its specific concern to
bring regularly to mind the mighty power which Yahweh had shown
in bringing his people out of Egypt (cf. vv. 8-9, 14-16). As such it
expands what had probably already been said in the older narratives
about the slaying of the Egyptian firstborn (11.5; 12.29) and the
festival of Unleavened Bread (12.33-34, 39; 13.3-4). If, as we have
argued, these sections originally belonged to two parallel accounts
of the Exodus, this passage presupposes their combination together.
It is most closely related to the short instructions about Unleavened
164 EXODUS 1–18
Bread which precede it (from which it draws its major theme: v. 3),
but it introduced a further ritual which it makes into a reminder of
the Exodus, the special treatment of the firstborn of humans and
animals. In both cases it is concerned to incorporate, and perhaps
modify, legal provisions which were already in circulation.
In the case of Unleavened Bread a seven-day festival in the
month of Abib is already prescribed as a memorial of the Exodus
in the Book of the Covenant (23.15a) and in Exod. 34.18. The key
words (which are the same in both passages) are reproduced exactly
in v. 6a: ‘For seven days you shall eat unleavened bread’. The same
formula appears in Deut. 16.3, where the second occurrence of
‘with it’ (Heb. ʿālāyw), referring to the Passover lamb, must be a
scribal error, as Deut. 16.4b would prohibit the eating of the meat
after the first day/night.7 The Deuteronomic law is, however, much
fuller, partly because of the integration of Unleavened Bread with
Passover and the insistence that the latter at least be celebrated
at ‘the place which the Lord will choose’ (vv. 2, 6, 7). But some
of the additional material is independent of that new setting and
may well go back to an older form of the law. This includes vv. 4a
and 8b, which are closely paralleled in Exod. 13.6b and 7bβ: it
is quite possible that a closer parallel to v. 7bα also appeared in
the pre-Deuteronomic form of the law and was adapted in Deut.
16.3aα.8 If so, then this is presumably the text on which the author
of Exod. 13.6-7 also drew. He shows no sign of knowing the law as
it was expanded and rewritten in Deut. 16.1-8.
A similar conclusion seems to follow for the law about the
firstborn in Exod. 13.12-13. In this case the Book of the Covenant
(22.28b-29) and Exod. 34.19-20 differ considerably, because the
former has no provision for the redemption of the firstborn (human
or animal). In fact the intentions of the two forms of the law may not
have been as different as they seem, as Exod. 22.28b-29 has nothing
to say about any animal except for cattle, sheep and goats, which
could not be redeemed, and it is hardly likely that it envisaged the
actual sacrifice of human firstborn, even if some ‘extremists’ under-
stood it in that way (cf. Ezek. 20.26). Most likely Exod. 34.19-20
7
So e.g. A.D.H. Mayes, Deuteronomy (NCB; London, 1979), p. 258; B.M.
Levinson, Deuteronomy and the Hermeneutics of Legal Innovation (New York and
Oxford, 1997), pp. 84-85, 88.
8
This is broadly in line with the argument of Levinson, ibid., pp. 81-89.
13.1-16 165
represents a development of the original practice, which added the

requirement for a substitute in the case of firstborn, animal and
human, which could not be sacrificed, and it is the law in this form
that the author of Exod. 13.12-13 adopted, in this case bringing it
into connection with the Exodus for the first time. The Deutero-
nomic law about firstborn (15.19-23) is quite different. It says
nothing about redemption, and follows the basic pattern of Exod.
22.28b-29, with additional provisions about the eating of the slaugh-
tered animal, either at the central sanctuary or ‘within your towns’
(vv. 20-23), which are related to Deuteronomy’s specific concerns.
The expression used for the firstborn in Exod. 13.12-13 and
34.19-20 is literally ‘the opener of the womb’ (Heb. peṭer reḥem:
see Note b on the translation). Outside these two passages it appears
only in Priestly texts (Exod. 13.1-2; Num. 3.12; 8.16; 18.15) and
in Ezek. 20.26, but it is unlikely that it was derived by our author
from the Priestly tradition itself, as his treatment of the subject is
so different (and almost certainly older). The expression may well
derive from older cultic terminology (which would explain why
Ezekiel, a priest, used it), whereas the Book of the Covenant and
Deuteronomy use the more widespread bekōr for ‘firstborn’, as does
v. 15 here as well as the Exodus narrative in 11.5; 12.12, 29.
In terms of legal history, then, it seems that Exod. 13.5-16 is
not based on the corresponding legal texts in Deuteronomy, but
on laws that are older than it but later than those in the Book of
the Covenant. In utilising these legal texts for his own purpose the
author has used language and formulae which help to determine
further his setting and method of working. To some extent he draws
on material already present in the text of Exodus itself. The names
in v. 5 have all occurred, with the epithet ‘flowing with milk and
honey’, in 3.8, 17 and recur in 33.2-3 (cf. 23.23). The divine oath
to ‘your fathers’ appears just before those last verses in 33.1, as
well as in some passages in Genesis (Exod. 32.13 may be of later
origin). The instruction of children could be modelled on 12.26-
27, and the introductory words of each paragraph, ‘When Yahweh
brings you…’, are only a theological interpretation of ‘When you
come…’ in 12.25, where the future gift of the land is also mentioned
(cf. v. 11). The specific sense of ‘service/worship’ in v. 5 occurs
elsewhere only in 12.25-26 and may well come from there. 13.3
provided the unusual phrase ‘by strength of hand’ (vv. 14, 16) and
the keyword ‘Remember’, which is recalled in v. 9 before the visual
166 EXODUS 1–18
comparison is extended in v. 16 with the mention of ‘frontlets’.

Possibly ‘the house of bondage’ in v. 14 is drawn from the same
verse, if it is original there.
But other vocabulary is new. Some of it is well known from older
narrative works like those used as sources in the Deuteronomistic
History, such as ‘Because of’ in v. 8 and ‘from year to year’ in
v. 10 (see the Explanatory Notes). Elsewhere words that otherwise
occur only in Deuteronomy are used, as with ‘young’ in v. 12 and
‘frontlets’ in v. 16, while in some other cases specific uses of words
find their closest parallels in Deuteronomy or in later laws: ‘hand
over’ (v. 12), ‘make difficult’ (v. 15) and ‘sacrifice’ of firstborn
(v. 15). None of this need imply dependence on Deuteronomy or
the later laws, since the words were presumably in use in the spoken
language before and after Deuteronomy (if not, Deuteronomy would
not have been understood), but they suggest like other aspects of
the passage that its authors may have been related in some way to
the circles that produced Deuteronomy. On the other hand, some
features of its language, like the consistent use of ‘Egypt’ without a
preceding ‘the land of’ and the absence of ‘your God’ as an epithet
of Yahweh, which is widespread in Deuteronomy, and the lack of
any trace of Deuteronomy’s concerns with a conditional covenant
and the centralisation of worship warn against a hasty categorisa-
tion of the passage as ‘Deuteronomistic’. The recent tendency to
date it to a very late stage of the composition of the Pentateuch has
no foundation at all.
‘So that the teaching of Yahweh may be in your mouth’ (v. 9)
is the reason given for remembering the Exodus at the festival of
Unleavened Bread, and the didactic purpose of the whole passage
is very clear. Like the addition in 10.1b-2, it is concerned to ensure
that the story of the Exodus is handed on from generation to genera-
tion, so that Israel will continue to be confident in the mighty power
of their God to protect and provide for them and ‘know that he is
Yahweh’. Similarly both hymns of praise and other kinds of psalm
show how the memory of the Exodus as a demonstration of Yahweh’s
‘strength of hand’ did indeed become a popular theme of worship.
Some of the evidence may be very early (see below on 15.1-21).
While this theme appears to have been especially prominent in the
northern kingdom (Pss. 77, 80 and 81), it is also attested in Judah.
It is presumably instruction at the cultic centres that the authors
of Exod. 13.3-16 particularly have in mind, whereas the Passover
13.1-16 167
provided an opportunity for it in the home (12.21-23, 25-27). No

doubt this is why both Amos and Hosea felt a need to undercut
the dangerous over-confidence which it could inspire (Amos 9.7;
Hos. 9.3). But it continued to be sung of even in late psalms (e.g.
Pss. 135; 136) and such memories and worship provided a basis for
prophetic hope after the fall of Jerusalem to the Babylonians (Isa.
42.10-13; 43.16-19; 63.11-13).
1 Yahweh spoke to Moses as followsa: 2 ‘Consecrate to me every

firstborn, what opens every wombb among the Israelites, of both
humans and animalsc: it is mine’. 3 [Moses said to the people,
‘Rememberd this day on whiche you came out of Egypt, from the
house of bondage, for by strength of hand Yahweh has brought
you out from here, and no leavened bread shall be eatenf. 4 Today
you are coming out in the month of Abib:] 5 [when Yahweh
brings youg into the land of the Canaanites, the Hittites, the
Amorites, the Hivites and the Jebusites, which he swore to your
ancestors to give you, a land flowing with milk and honey, you
shall perform this service/worshiph in this month. 6 For seven
days you shall eat unleavened bread, and on the seventh day
there shall be a festival for Yahweh. 7 Unleavened Bread shall
be eateni for the seven daysj, and no leavened bread shall be seen
with youk and no leaven shall be seen with youk, throughout
your territory. 8 You shall tell your son on that day as follows:
“It is becausel Yahweh acted in my favourm when I came out of
Egypt”. 9 It shall be a signn for you on your hand and a reminder
on your foreheado, so that the teaching of Yahweh may be in
your mouth, for by a mighty hand Yahweh brought you out of
Egypt.] 10 [You shall observe this statute at its regular time from
year to yearp.] 11 [When Yahweh brings you into the land of the
Canaanites, as he swore to you and your ancestors, and gives
it to you, 12 you shall hand overq to Yahweh all that opens the
womb, and every firstborn (lit. opener) of the youngr of animals
which shall be yours, the males that is, shall belong to Yahweh.
13 But every firstborn (lit. opener) of a donkey you shall redeem
with an animal from the flock – if you do not redeem it you shall
break its neck – and every human firstborn among your children/
sons you shall redeem. 14 When your son asks you in futures
as follows, “What does this mean?”t, you shall say to him: “By
strength of hand Yahweh brought us out of Egypt, from the house
of bondage, 15 and because Pharaoh made letting us go a difficult
thingu, Yahweh killed every firstborn in the land of Egypt, from
168 EXODUS 1–18
human firstborn to the firstborn of animals. Therefore I sacrifice

to Yahweh all that opens the womb, the males that is, and every
firstborn of my children I redeemv.” 16 It shall be a sign on your
handw and frontlets on your foreheadx, that/because by strength
of hand Yahweh brought us outy of Egypt.]’

a. Heb. ‫לּאמר‬: see Note b on the translation of 6.10–7.5. The majority of
cases in Exodus are where ‫ לאמר‬is the final word of a verse.
b. Heb. ‫פטר כל־רחם‬. The phrase, which recurs in v. 12a, is here in explana-
tory apposition (or permutation: GK §131k) to ‫בכור‬, which is now defined for
the first time as a woman’s or female animal’s first offspring. ‫ פטר‬is sometimes
used alone with the same meaning, but where the fuller phrase has already
appeared (below, vv. 12b, 13 and in 34.19b-20): the reference to a womb can
therefore always be presumed. Here and in Num. 3.12; 8.16 (‫ ;)פטרה‬18.15
and Ezek. 20.26 the phrase applies to human firstborn as well as animals,
whereas in vv. 11-16 below and in 34.19-20 this is not so clear, and HAL,
p. 874, and Ges18, p. 1049, claim that there an older usage, restricted to
animals, appears. But in the first occurrences in v. 12 and in 34.19 humans
are likely to be at least included, otherwise the verses would be unduly rep-
etitious. The exact meaning of ‫ ֶפּ ֶטר‬is disputed: according to BDB, p. 809, it
is that which ‘separates’ and then ‘first opens’ the womb (cf. the meanings
given for the verb ‫ פטר‬and its cognates, ‘separate, remove, set free’), whereas
HAL and Ges18 have ‘was durchbricht’, i.e. ‘breaks through’ or ‘breaches’
the womb, which is more closely related to the Ar. cognate, ‘split’, and an
occurrence in Ugaritic which may mean ‘breach’ (KTU 1.16.6.8; cf. DULAT,
p. 687). Neither explanation fits the occurrences of the Heb. verb very well:
there may be more to be said for an association with the sense ‘open’ which is
likely in Ps. 22.9 (of lips, in Hiphil) and 1 Kgs 6.18-35 (of flowers, in a sculp-
tured decoration). Despite the etymological difficulties, the context leaves
little doubt about the meaning (see also Text and Versions).
c. Heb. ‫באדם ובבהמה‬. The phrase also occurs in 8.13-14, but there ‫ ב‬has the
sense ‘on’ rather than ‘among’, as here and in a similar inclusive phrase, ‫בגר‬
‫ובאזרח הארץ‬, in 12.19.
d. On the use of the inf. abs. as a substitute for an imperative (here pl.) see
GK §113bb.
e. Heb. ‫אשׁר‬. For the sense ‘on which, when’ especially after ‫ יום‬cf. BDB,
p. 82.
f. Heb. ‫ולא יאכל‬. The translation as a prohibition assumes that these words
are the continuation of ‫זכור‬. They could, however, be understood (against
the Masoretic accents) as the continuation of ‫הוציא‬, indicating the ‘attendant
circumstances’ of that past event (cf. 12.34, 39): ‘with nothing leavened being
13.1-16 169
eaten’ (GK §156f). This makes the structure of the passage simpler: v. 4
continues to speak in the (narrative) present and only v. 5 begins the instruc-
tion about what is and is not to be done in the future.
g. Heb. ‫כי־יביאך‬. From here onwards the second person sing. is used, not
the second person pl. as in vv. 3-4, and this continues throughout the next
sub-section in vv. 11-16, which begins in a very similar way to v. 5.
h. Heb. ‫ועבדת את־העבדה הזאת‬. The waw marks the beginning of the main
clause of the sentence after the temporal clause introduced by ‫( והיה‬cf. ‫והעברת‬
in v. 12): GK §112ff, oo. For the sense of ‫ עבדה‬see Note p on the translation
of 12.21-27. Here the cognate verb ‫ עבד‬is naturally used with it, instead of
‫ שׁמר‬in 12.25.
i. Heb. ‫מצות יאכל‬. ‫מצות‬, repeated from v. 6, is placed before the verb for
additional emphasis, but here the verb is passive and strangely in the third
person masc. sing. The passive form may prepare for the use of the passive
later in the verse (and pick up the same passive form in the negative in v. 3).
Despite the lack of concord, ‫ מצות‬could be meant as the subject of ‫יאכל‬, as
such anomalies sometimes occur even when the subject precedes the verb
(GK §145u). Alternatively ‫ מצות‬may be the object (unmarked because it is not
definite), retained even with a passive verb, as earlier in 10.8 (so GK §121a-b).
j. Heb. ‫את שׁבעת הימים‬. The ‘accusative of time’, already used in v. 6, is here
marked as such by the object-prefix ‫ את‬because of its (retrospective) determi-
nate form (cf. Deut. 9.25 and GK §118k).
k. Heb. ‫לך‬, lit. ‘belonging to you’ and serving as a virtual possessive adjec-
tive as in many occurrences cited in BDB, p. 512 (s.v. 5a: the examples with
‫ מצא‬in Deut. 22.14; 1 Sam. 13.22 are especially close).
l. Heb. ‫בעבור זה‬. On a straightforward reading this should mean ‘because of
this’ (so LXX, TgO, Sy). The question is: what does ‘this’ refer to and how is it
connected with what follows? A traditional Jewish interpretation (see Text and
Versions), endorsed by Rashi and Ibn Ezra, took ‘this’ to mean the celebration
of Passover and Unleavened Bread (according to one view representative of
obedience to all the commandments) and held that this was being presented as
the purpose of the Exodus deliverance (cf. Cassuto). This does, however, seem
to be at variance with the commemorative intention for future celebrations
implied here (v. 9) and elsewhere, and already Rashbam and Nachmanides
were suggesting that this intention could be found here too if it were supposed
that a relative pronoun was understood after ‫זה‬, as seems necessary in some
other passages (Rashbam cited Ps. 118.24)9. This is the view that found its
way into the German and English versions of the Reformation period (e.g.
Luther, Tyndale, AV) and it has remained standard in EVV. ever since: cf.
9
Even earlier Ibn Janaḥ had proposed that the same result could be achieved
by inverting the words ‫ בעבור‬and ‫זה‬, a view that was strongly contested by Ibn
Ezra, who disputed the alleged parallels and added: ‘How can we invert (i.e. the
sequence of) the words of the living God?’ (Rottzoll [ed.], p. 360).
170 EXODUS 1–18
GK §155n and the similar construction in 4.13 and 1 Chr 15.13. A (perhaps
preferable) variation on it would be to see ‫ עשׂה יהוה לי‬as in explicative apposi-
tion to ‫זה‬: so apparently BDB, p. 260, comparing Gen. 42.18; 43.11, with ‫זאת‬
(cf. also Exod. 9.16); but ‫ זה‬is sometimes used in the same way. In this case
‫ עשׂה…לי‬would have to be translated ‘acted for me, dealt with me’ (for which
cf. 1 Kgs 8.32, 39; Ezek. 31.11).10 Another possibility is to see ‫ זה‬as not a
demonstrative pronoun but a relative particle (or even an enclitic), so that
‫ בעבור זה‬becomes equivalent to ‫ בעבור אשׁר‬in Gen. 27.10 and ‫ בעבור‬alone
elsewhere (so GK §138h; Gibson, Syntax, p. 7; Houtman, p. 213): but ‫ זה‬as a
relative is only attested in poetry and it seems that )‫ בעבור (אשׁר‬as a conjunc-
tion always means ‘in order that’, which would not fit here. Both GK and
Houtman therefore recognise that the text may be corrupt.
m. Heb. ‫לי‬. For ‫ ל‬indicating favour towards someone see BDB, p. 515, s.v.
5h (b) (α): with ‫ עשׂה‬used absolutely as here in 1 Sam. 14.6; Isa. 64.3 (cf. Ps.
68.29).
n. Heb. ‫לאות‬. Gen. 4.15 refers to a literal ‫אות‬, ‘mark’, on Cain (its location is
not specified): for such practices Gunkel compared Lev. 19.28; Deut. 14.1-2;
Isa. 44.5; Ezek. 9.4; Gal. 6.17; Rev. 13.16-17; 14.9 (Genesis, 3rd ed., p. 46),
where other words are used and the significance varies. On the significance
here see the Explanatory Note.
o. Heb. ‫בין עיניך‬, lit. ‘between your eyes’, is apparently a way of referring
to the forehead (elsewhere ‫מצח‬: cf. Ezek. 9.4) or the front of the scalp (Deut.
14.1; Dan. 8.5, 21), as in the passages similar to this one (v. 16; Deut. 6.8;
11.18).
p. Heb. ‫מימים ימימה‬. The pl. of ‫( יום‬here on the second occasion with the
‘directional he’ unusually in a temporal sense: cf. GK §90h) is used in a
number of ways without further definition. It can mean simply ‘some days’
(e.g. Gen. 40.4); in Num. 9.22 the progression suggests that it is a period
longer than a month (cf. Judg. 19.2). But the specific sense ‘a year’ is required
in Lev. 25.29; Judg. 17.10; 1 Sam. 1.21; 2.19b; 20.6; 27.7 and apparently
always in the expression found here: Judg. 11.40; 21.19; 1 Sam. 1.3; 2.19a;
cf. 2 Sam. 14.26.
q. Heb. ‫והעברת‬. The Hiph. of ‫ עבר‬has a secular legal background in Num.
27.7-8, where it is used of transferring an inheritance (cf. MRI ad loc.; Rashi),
and it appears in a religious context in Ezek. 48.14Q. A similar metaphorical
use is probably also involved here and in the more numerous occurrences
where children were ‘delivered over’ to a god such as Molech (Jer. 32.35 etc.):
so BDB, p. 718, ‘devote’. Against the common literal interpretation ‘make…
pass through (fire)’ see J. Day, Molech: A God of Human Sacrifice in the Old
Testament (UCOP 41; Cambridge, 1989), pp. 15-20. The same expression is
used with reference to child sacrifice to Yahweh in Ezek. 20.31 according
10
A similar interpretation is suggested for LXX by Wevers, Notes, p. 198: ‘(It
is) because of this, (namely that) the Lord God did to me’.
13.1-16 171
to Day (p. 19), but this is not completely certain: there is, however, some
connection with the present passage in the reference to firstborn (‫)כל־פטר רחם‬,
although here (cf. vv. 13b, 15b) human firstborn are explicitly excluded from
the sacrificial slaughter which Day rightly argues is intended for the firstborn
of animals (pp. 19-20).
r. Heb. ‫שׁגר‬. The word occurs only here and in Deut. 7.13; 28.4, 18, 51 in
BH (also in Sir. 40.19). The meaning ‘young’ (of animals, specifically oxen)’
is clear in Deuteronomy and Sirach (cf. the parallel ‫ )ילד‬and must be correct
here. It might seem to be equivalent to ‫ רחם‬in v. 12a here (cf. Cassuto), but
the collocation with ‫ פטר‬is not decisive as the latter could be used alone for a
firstborn (cf. v. 13; also 34.19b-20). ‫ פטר שׁגר‬therefore means ‘firstborn among
the young of…’: for the partitive use of the construct relationship see GK
§128i and e.g. Gen. 22.2. Like ‫עשׁתרות‬, with which it is associated in Deuter-
onomy, ‫ שׁגר‬is now known to have also been the name of a deity in non-biblical
texts (cf. DDD, 1437-40; the presence of this deity in DAPT I 7-8, 12, 16, a
text from the Jordan valley, is plausible but not certain), but there is no trace
of this in the OT passages.
s. Heb. ‫מחר‬, which was used earlier in its narrower sense ‘tomorrow’ (8.6,
19, 25; 9.5, 18; 10.4; cf. 16.23; 17.9; 19.10; 32.5), also has an extended sense
in which it means the indefinite future (Gen. 30.33; Deut. 6.20; Josh. 4.6, 21;
22.24, 27, 28).
t. Heb. ‫מה זאת‬: see Note q on the translation of 12.21-27.
u. Heb. ‫הקשׁה‬. The object is ‫( לשׁלחנו‬cf. BDB, p. 904; GK §53d-f), not an
understood ‫לב‬, which would make the idiom the same as in 7.3 (although there
the subject is Yahweh, not Pharaoh): if it were, the continuation would more
likely be ‫ ִמ ְלּשׁלחנו‬or ‫לבלתי־שׁלחנו‬. A very close parallel (of substance as well as
language) is the use of the Qal in Deut. 15.18 to refer to reluctance to release
a slave. The point here is Pharaoh’s unwillingness to release Israel, not his
(self-imposed) stubbornness.
v. For the iterative imperfect see GK §107g. The use of the part. ‫ זבח‬in the
same sense earlier in the verse is rare.
w. Heb. ‫לאות על־ידכה‬. For the rare plene spelling of the second person sing.
m. suffix in MT cf. 7.29 and GK §91d, 103g. On ‫ אות‬see above Note n.
x. Heb. ‫ולטוטפת בין עיניך‬. On ‫ בין עיניך‬see Note o above. ‫ טוטפת‬occurs in BH
only here and in Deut. 6.8; 11.18, always vocalised as a plural. In Deuter-
onomy the word is applied to the commandments, perhaps originally as a
metaphor but certainly eventually understood as referring to the tefillin or
phylacteries, small boxes in which portions of the Torah were enclosed (see
Text and Versions). In post-biblical Heb. and JAram. this is generally the
meaning (cf. Jastrow, p. 523; CAL), but not always: Tg at 2 Sam. 1.10 has
‫( טוטפתא‬sing.) for MT ‫אצעדה‬, ‘armlet’, and M.Shabb. 6.1 uses the word of an
item of women’s finery (Jastrow, ibid.; Weinfeld, Deuteronomy 1–11, p. 334).
The word may also have occurred in such a sense in Sir. 36.3 (ms. B), but
too little of the context survives for this to be certain. A similar Mandaean
word ṭuṭipta (HAL; Ges18 s.v.) is said to mean ‘amulet’ (Weinfeld, ibid.), and
172 EXODUS 1–18
etymologies based on Heb. ‫נטף‬, ‘drip’ (hence of drop-shaped beads: cf. ‫)נטיפה‬,
or Ar. ṭāfa, ‘go round’ (Akkadian ṭaṭāpu, cited in BDB, p. 377, is not in AHw
and should be discounted), have been suggested: cf. also the very specula-
tive proposal of E.A. Speiser, ‘ṬWṬPT’, JQR 48 (1957–58), pp. 208-17. The
present passage gives no hint of a reference to phylacteries and a metaphor
based on some kind of head decoration is presumably intended here: for
iconographical references see TWAT 3, 341-43 = TDOT 5, pp. 319-21, where
other relevant bibliography is cited, to which should be added the learned
studies of O. Keel, ‘Zeichen der Verbundenheit: Zur Vorgeschichte und
Bedeutung der Forderungen von Deuteronomium 6,8f. und Par.’, in P. Casetti
(ed.), Mélanges Dominique Barthélemy (OBO 38; Fribourg/Göttingen, 1981),
pp. 159-240, and J.H. Tigay, ‘On the Meaning of ṭ(w)ṭpt’, JBL 101 (1982),
pp. 321-31, which cite further iconographical evidence from the ancient
Levant and maintain that the form was originally singular and meant ‘head-
band’, perhaps with an inscribed plate (cf. Text and Versions).
y. Heb. ‫הוציאנו‬. The change from second person m.s. suffixes earlier in the
verse is at first sight surprising (and see Text and Versions) – in v. 14 ‫הוציאנו‬
is within the words which the father is to speak to his child, whereas here
Moses again addresses the people as a whole – but Moses could be envisaged
as including himself with the people in the ‘us’ (cf. 3.18; 33.16 and often in
Deut. 1–3).
Explanatory Notes
1-2. The introduction to Yahweh’s new instructions about the
consecration of the Israelite firstborn differs from the introductions
in ch. 12 (vv. 1, 43) in two ways: now Moses alone is addressed,
without Aaron, and the verb is the more formal ‘spoke’ (Heb. dibber)
rather than ‘said’ (ʾāmar).11 Probably neither variation is of great
significance. Aaron is also included with Moses at the beginning of
some of the Priestly plague-stories (7.8; 9.8: cf. 6.13), but generally
in Exodus Moses is addressed alone in both Priestly (e.g. 6.2; 11.9;
14.1; 25.1) and non-Priestly (e.g. 7.14; 10.1; 11.1; 14.18) contexts;
and ‘spoke’ appears several times in introductions to divine speech
(e.g. 6.2; 14.1; 16.11; 25.1).
Laws about the dedication of the firstborn occur in most of the
Pentateuchal legal collections: Exod. 22.28-29; 34.19-20; Lev.
27.26-27; Num. 18.15-18; Deut. 15.19-23, as well as vv. 11-16
11
The common adjunct to ‘spoke’, ‘as follows’ (lit. ‘saying’, Heb. lēʾmōr),
appears afterwards in MT, which may account for NRSV’s ‘said’ here.
13.1-16 173
below here. It was apparently an ancient and persistent custom in

Israel, comparable to the dedication of the firstfruits of the harvest
to Yahweh (cf. 22.28) and expressing both gratitude to him and the
desire for his continued blessing, and similar practices are known
all over the world (cf. J. Henninger, DBS 8, pp. 461-82; more briefly
de Vaux, Institutions 1, p. 73; 2, pp. 329-31, 390, ET pp. 42, 443-44,
489). A distinctive turn is given to the custom in two passages in
Numbers (3.11-13, cf. 40-51; 8.16-18), where the Levites are taken
by Yahweh as a substitute for all the Israelite firstborn. On this and
other details of the legislation see the notes below on vv. 11-16. It
is fact only in this chapter and in the two passages just mentioned
in Numbers that the practice is specifically connected with the
Exodus.12 Even in v. 2, where there is no explicit association with
the death of the Egyptian firstborn and the sparing of Israel, the
location of this law must imply some such connection and it is
reinforced by the use of the word ‘firstborn’ (Heb. bekôr) as well as
the legal expression ‘what opens every womb’.13 The latter phrase,
which recurs in vv. 12 and 15 (and abbreviated in v. 13), has been
variously explained (see Note b on the translation), but it clearly
defines those to whom the law applies as the firstborn of the female
rather than the male parent. Later in the chapter (vv. 12, 15) only
male offspring are covered by the law (cf. 22.28; Deut. 15.19; in
Exod. 34.19 the text is uncertain) and this is probably assumed
here. The law applies to humans as well as to animals: at this stage
no indication of how it was to be applied is given. The word used
here is ‘consecrate’ (Heb. qaddeš), which is a regular expression of
Priestly cultic vocabulary (e.g. 28.3; 29.36; 40.9), and the related
noun qōdeš (‘sacred’, lit. ‘sacredness’) is used of the firstborn in
Num. 18.17(P). In two other Priestly passages in Numbers which
refer back to this passage Yahweh himself is said to have ‘conse-
crated’ the Israelite firstborn (3.13; 8.17), using a different form
12
It is possible that the compiler of Exod. 34.11-27 also had this in mind, as he
attached the law about the firstborn to the feast of Unleavened Bread (vv. 18-20):
in the Book of the Covenant the two laws are still separate (Exod. 22.28-29;
23.15).
13
Earlier in Exodus a different kind of connection is made with the narrative
(cf. 4.22-23): there it is the familial relationship between Israel as son and Yahweh
as Father that is in view, not the duty of human fathers to present their firstborn
to Yahweh, and there is no justification for seeing the former (which is a later
addition) as the reason for the latter (cf. Propp, pp. 457-58).
174 EXODUS 1–18
of the same Heb. verb (the Hiphil), and this form is also used in
the corresponding law in Deuteronomy (15.19).14 Since the verb is
often used of the consecration of priests (e.g. 28.3), it clearly need
not imply that human firstborn were to be sacrificed. The form of
the verb here is generally analysed as a singular masc. imperative
and in the context the subject or agent is most naturally taken to be
Moses. He may be thought of as acting in a representative capacity
for the people as a whole, perhaps by ‘declaring holy’ the firstborn
of all the people (cf. ‘among the Israelites’). Another possibility
is that the form in the text is an infinitive absolute (cf. GK §52o),
which could then stand for a pl. imperative as well as a sing. Then
all the people would be addressed. It is, however, much more usual
for this to be made explicit (cf. 12.3ff.).
3-4. Moses does not immediately speak to the people about the
dedication of the firstborn – this is deferred until vv. 11-16 – but
instead takes up the topic of abstinence from leavened bread, which
was introduced as a divine instruction in 12.15-20. Two features
of vv. 3-4 distinguish them from the fuller prescriptions on the
same topic in vv. 5-10: here second person pl. forms are used, there
second person sing.; and vv. 5-10 have an almost identical struc-
ture to vv. 11-16 (see the introduction to the section), while vv. 3-4
remain outside this pattern. As they stand they can be seen as a
scene-setting introduction to what follows, but they could well have
been originally (part of) a brief older law about abstinence from
leavened bread when the Exodus was commemorated (on the last
words of v. 3 see Note f on the translation). As Albertz has seen
(pp. 220-21), the verses share some features with the last non-Priestly
narrative section in 12.29-39: the reference to Israel as ‘the people’
recalls vv. 33, 34 and 36; the avoidance of leaven repeats what the
haste of departure brought about according to vv. 34 and 39; and
the use of ‘came out/are coming out’ continues the language already
found in 12.31 (cf. 11.8). Verse 4 may have originally stood before
v. 3 – it looks more like the opening of a speech than a conclusion
(Propp, p. 378: cf. Deut. 2.18; 29.9; Josh. 3.7) – and been moved to
its present position when vv. 5-10 were added.15
14
It also appears in Lev. 27.26, but there (cf. vv. 14-19) it is used of a process
like mortgage which is not to be applied to firstlings.
15
The Samaritan text takes ‘Today’ as the final word of v. 3, but this is
unlikely: see Text and Versions.
13.1-16 175
The phrase ‘house of bondage’ (Heb. bêt ʿabādîm, lit. ‘house of

slaves’) occurs for the first time here: cf. v. 14 and 20.2, both like
this verse in discourse rather than narrative. Most famously in the
Decalogue (see also Deut. 5.6), it is almost always an evocative
rhetorical expansion of the phrase ‘the land of Egypt’ and associated
with either the verb ‘bring out’ (yāṣāʾ Hiphil) or a close synonym.
It thus adds a description of the trouble from which Yahweh had
delivered his people to the purely geographical formula found else-
where. The distribution of the fuller expression is most concentrated
in Deuteronomy (5.6; 6.12; 7.8; 8.14; 13.6, 11) and Deuterono-
mistic literature (Josh. 24.17; Judg. 6.8; Jer. 34.13): only the two
occurrences in Exodus 13 and Mic. 6.4 (possibly) stand outside
that corpus. Here two variations of expression may be significant: a
different form of the verb yāṣāʾ is used (Qal rather than Hiphil; but
the Hiphil follows in the next clause) and ‘Egypt’ rather than ‘the
land of Egypt’ (as again in v. 14) precedes.
‘By strength of hand’ (as in vv. 14 and 16) is another small varia-
tion from a much more common phrase, ‘by a strong hand’ (beḥōzeq
yād instead of beyād ḥazāqāh). The latter is again most common in
Deuteronomy and related literature (9x; 5x in Exodus; 2x in Ezekiel,
4x elsewhere), but the wording used here occurs nowhere else.
Again the variation may be a reason for not too quickly assuming
dependence on Deuteronomy.
As the name of a month, Abib is always associated with the
Israelites’ departure from Egypt and the celebration of Passover
(Deut. 16.1) or Unleavened Bread (Exod. 13.4; 23.15; 34.18). In
Hebrew it has the definite article, hāʾābîb, ‘the Abib’, which reflects
the use of the word as a common noun for ‘ears’ of grain (Exod.
9.31; Lev. 2.14). The first of these verses dates the plague of hail to
the time when the barley crop was already ‘in the ear’ (ʾābîb) and
so shares the same chronological view of the Exodus as is presup-
posed here. In Lev. 2.14 fresh ears of grain (ʾābîb) are roasted and
crushed to form an offering of firstfruits, again indicating a connec-
tion with the beginnings of harvest.16 Since Abib is not one of the
Babylonian month-names, it is generally presumed to be a name
of Canaanite origin, but unlike other ‘older’ names for months it
16
Roasted grain (Heb. qālûy, qālî) was a common kind of food (1 Sam. 17.17;
25.18; 2 Sam. 17.28), apparently especially in the spring (cf. Lev. 23.14; Josh.
5.11; Ruth 2.14).
176 EXODUS 1–18
is not so far attested outside Hebrew (cf. de Vaux, Institutions 1, p.

279, ET, p. 183; the possible occurrence in one of the Proto-Sinaitic
inscriptions is very doubtful [see J. Naveh, The Early History of
the Alphabet (Jerusalem, 1982), p. 33, for a different reading]) and
it has been suggested that it was not a name at all, but simply a
description like those in the Gezer calendar (Propp, p. 422). But the
formula ‘in the month of X’ implies that it had become a name, and
it could easily be one of the several month-names that are not so far
known from Canaanite sources.17
5. The words ‘in this month’ at the end of the verse make a
formal connection with ‘in the month of Abib’ in v. 4, but from
this verse onwards (to v. 16) the pronouns are second person sing.
instead of second person plural. This could well mean that vv. 5-16
are a later amplification of the brief instruction in vv. 3-4. The
structure of vv. 5-10 and 11-16 is very similar (see the introduc-
tion to this section): both paragraphs begin with a long temporal
clause which states that the practice concerned is to be observed
when the Israelites reach the land of Canaan (vv. 5a, 11: cf.
12.25a), which is natural for agricultural rituals. In v. 5 the land is
described by three phrases which refer to its inhabitants, its guar-
antee by a divine oath and some of its produce. In different ways
these all go beyond the briefer description of the land in 12.25
and take up expressions which are found elsewhere, both in the
book of Exodus and outside it. The list of inhabitants corresponds
almost exactly to that given in 3.8 and 17: only the Perizzites
are omitted (in MT: see Text and Versions for variations in other
witnesses). Lists with six peoples (including the Perizzites) but
with some difference in their order also appear in 23.23, 33.2 and
34.11. On the various versions of the list elsewhere in the OT see
the Explanatory Note on 3.7-8. The Perizzites are omitted only
here and, since there is no obvious reason for this being deliberate,
it may be due to scribal error. The divine oath to the patriarchs is
a frequent theme in Genesis, Exodus (13.5, 11; 32.13; 33.1) and
17
Even the large corpus of Ugaritic texts has not so far permitted the recon-
struction of a full list of month-names: those that are known are different from the
extant Hebrew and Phoenician names (see provisionally C.H. Gordon, Ugaritic
Textbook [AnOr 38; Rome, 1965], pp. 414-15; P. Xella, I Testi rituali di Ugarit
I [Rome, 1981], p. 370 etc.: Xella’s Parte II on the calendar seems not to have
appeared).
13.1-16 177
especially Deuteronomy, as well as some later books, and repre-

sents a strengthening of other passages (including Exod. 12.25)
which refer simply to Yahweh’s ‘word’ of promise. A variety of
views have been taken about the authorship of the passages in
which it occurs: it is probable that some of them (including Gen.
50.24; Exod. 33.1; Num. 11.12) belong to a pre-exilic narrative
work and that the theme was then developed further in Deuter-
onomy (with the addition of a strong conditional qualification)
and later literature (for a review of the occurrences in the Penta-
teuch and a discussion of recent scholarship see my ‘Covenant,
Oath and the Composition of the Pentateuch’, in A.D.H. Mayes
and R.B. Salters [eds.], Covenant as Context [FS E.W. Nichol-
son; Oxford, 2003], pp. 71-90 [esp. 73-82]). The produce of the
land is summarised in the phrase ‘flowing with milk and honey’,
which like the list of its inhabitants has already appeared twice in
Exod. 3.8, 17 (see further the Explanatory Note on 3.8; for a full
list of occurrences and discussion of their origin see my ‘Kd in
Exodus: An Assessment of E. Blum’s Proposal’, in M. Vervenne
and J. Lust [eds.], Deuteronomy and Deuteronomic Literature [FS
C.H.W. Brekelmans; Leuven, 1997], pp. 407-20 [esp. 410-11]).
The Israelites’ arrival in this land is said not only to be promised
by Yahweh but brought about by him: ‘When Yahweh brings you
into the land…’ (cf. v. 11), so that it is in response to this further
action on their behalf that the Israelites are to celebrate the festival
of Unleavened Bread as a memorial of the Exodus (and also to
offer their firstborn to him). Almost identical introductory formulae
appear three times in Deuteronomy (6.10; 7.1; 11.29; with different
verbal forms in 11.31; 12.20; 12.29; 17.14; 18.19; 19.1; 26.1;
31.20) in other contexts, and many scholars see this as particularly
clear evidence of a Deuteronomistic hand at work in this section. A
perhaps significant difference here is the absence of the character-
istic Deuteronomic addition of the epithet ‘your God’ to the divine
name (which is emphasised especially by Caloz, ‘Exode, XIII,3-16’
[above n. 4], p. 46).
The demand for the annual abstinence from leavened bread is
presented first in general terms (v. 5b) and then in detail (vv. 6-9)
and finally in a generalising conclusion (v. 10). The wording here,
‘this service/worship’, is identical to that used of the Passover in
12.25-26 (on the otherwise unparalleled use of Heb. ʿabōdāh see
the Explanatory Notes there). ‘This’ may either refer backwards to
178 EXODUS 1–18
the basic demand in v. 3b (as ‘this month’ refers back to ‘the month
of Abib’ in v. 4) or forwards to the specific prescriptions in vv. 6-9
(which are then recapitulated by the phrase ‘this statute’ in v. 10).
6-7. The central requirement of a seven-day period of eating
unleavened bread occurs twice, but the repetition is not without
point: the order of words and the continuation in each case show
that in v. 6 (where the wording is identical to the ‘cultic calendars’
in Exod. 23.15; 34.18, and [except for the curious, and possibly
secondary, addition of ‘with it’] Deut. 16.3) the primary concern is
with temporal aspects of the festival, while in v. 7 it is with what
is or is not to be done. The continuation in each case (the holding
of a ‘feast’ [Heb. ḥāg] on the final day and the ban on even the
possession of any trace of leaven) corresponds to a requirement
in Deuteronomy (16.8; 16.4), with small variations of wording:
in neither case does the addition appear in Exodus 23 or 34. The
removal of all leaven is also required in Exod. 12.15, 19 (where
the basic requirement is again similar, though not identical), and
in Exod. 12.16, Lev. 23.7-8 and Num. 28.18, 25 ‘holy occasions’,
with no work to be done, are prescribed for both the first and the last
day of the festival. There is clearly evidence here for a development
in the celebration of this festival (cf. de Vaux, Institutions 2, pp.
383-94, ET, pp. 484-93, though he curiously omits this passage).
While there are significant similarities to Deuteronomy 16, in two
important respects the rulings here are different and closer to the
lists of festivals in Exodus 23 and 34: the festival of Unleavened
Bread is not integrated with Passover and there is no trace of the
demand for celebration at a single sanctuary. On these grounds
one might well see Exod. 13.3-10, or at least its legal provisions,
as representing a stage of development between the older ‘cultic
calendars’ and Deuteronomy (cf. Levinson, Deuteronomy [above,
n. 7], pp. 68-69).18
8. The ‘remembering’ (v. 3: cf. v. 9) of the Exodus at the festival
of Unleavened Bread is not left to depend solely on the action (or
abstinence) that is prescribed – the more so as other associations
18
As Levinson points out (p. 67), the words ‘as I commanded you’ in Exod.
23.15 (and 34.18 is the same) must refer back to this passage, but that need not
mean that it (or even its nucleus) is older than the Book of the Covenant: the
cross-reference may be an addition from the compiler of the non-Priestly Exodus
narrative (cf. the similar, but probably later, additions in Deut. 5.12, 16).
13.1-16 179
may well have been present in the worshippers’ minds, such as the
beginning of the new harvest – but is to be formulated in words,
very like those provided for the explanation of the Passover in
12.27, addressed by parents to their children. In this case it is not
assumed (as it is below in relation to the treatment of the firstborn)
that the children will ask for an explanation of the festival, although
the wording of the explanation is an incomplete sentence which
seems to presuppose the Kinderfrage again (for discussion of the
Kinderfrage passages in general see the Excursus in the introduc-
tion to 12.21-27). Perhaps there is a reflection here of the formalised
liturgy of the festival (on the final day?), which may also lie behind
the rather strange wording of the explanation which has been so
much discussed (see Note l on the translation). In any case the
handing on of the Exodus tradition to the next generation is here too
at the centre of the remembering, just as it was in an earlier episode
of the story (10.1-2).
9-10. The paragraph concludes with an indication of the purpose
of the festival (this time for ‘you’, that is for the whole people,
not just the children), which is closely paralleled at the end of the
law about the firstborn, and (in v. 10) a further command about its
regular observance. Its purpose is explained by reference, probably,
to marks or jewellery on the hand and forehead (see Notes n and
o) which must have had some symbolic significance, perhaps as
family or tribal mementoes. The festival is said, metaphorically, to
have a similar intention, which is to keep Yahweh’s mighty deliver-
ance of Israel in the people’s minds, and indeed on their lips. ‘The
teaching’ (Heb. tôrāh) of Yahweh elsewhere usually refers to his
commandments (so in all its other occurrences in Exodus: 16.4,
28; 18.16, 20; 24.12), and it is commonly so understood here (e.g.
BDB, p. 436), to refer to the law about the festival. But it would be
odd for the festival to be designed to make people talk about the law
that prescribed it. Some verses in Psalm 119 (13, 43) use different
words of speaking about the law in general, but the final clause of
this verse points rather to talk of Yahweh’s past action than his law.19
A.B. Ehrlich therefore suggested that the genitive ‘of Yahweh’ is
objective, not subjective, meaning ‘teaching about Yahweh’ (Rand-
glossen 1, p. 314; cf. Houtman, p. 214; Propp, p. 425). This would
19
In this clause ‘by a mighty hand’ is the common beyād ḥazāqāh (see the
Note on v. 3).
180 EXODUS 1–18
be an unparalleled sense for the phrase, but it is grammatically

possible and it fits the context here much better.20 The reference
could be to the psalms of praise sung at the festival (for ‘mouth’ in
the context of praise cf. Ps. 34.1 [where ‘his praise’ is an objective
genitive] etc.) or to the kind of ‘instruction’ (Heb. tôrāh) about the
Exodus introduced by Ps. 78.1 (cf. v. 4).
‘Statute’ (Heb. ḥuqqāh) in v. 10 occurs elsewhere in the sing.
only in Priestly texts (e.g. 12.43). On the other hand ‘at its regular
time’ (Heb. lemôʿadāh) seems to be based on the festival calendars
in Exod. 23.15; 34.18, and the expression ‘from year to year’ is
characteristic of older narrative style (see Note p on the translation).
Perhaps just the words ‘You shall observe this statute’ are a late
addition to the passage (compare the Explanatory Note on 12.24)
and the rest of the verse originally stood at the end of v. 5.
11-12. The beginning of the law about the male firstborn corre-
sponds closely to v. 5, but the description of the land is briefer (with
only ‘the Canaanites’ being mentioned among its inhabitants, as
several times in Genesis) and differently worded at the end. The
initial demand, on the other hand, is more specific and longer, in
fact this law completely lacks the ‘framing’ clauses found in vv. 5b
and 10 and proceeds straight to the central requirements. Whereas
in v. 2 a single verb covers both human and animal firstborn, the
repetition of ‘to Yahweh’ here probably points (as in the transla-
tion) to two separate clauses, one using a rare verb which is not
exclusively cultic in its meaning (see Note q on the translation),
and the other being a nominal clause.21 The second clause clearly
refers to animals; the first clause may either apply generally to all
firstborn, human and animal, or specifically to humans (so e.g.
Holzinger, Cassuto). In v. 12 the terminology for the firstborn (liter-
ally ‘opener’) uses the language that has already appeared in v. 2
(see the notes there) and will occur again in 34.19-20 (but not in
what is probably the oldest law on the subject in 22.28b-29 or in
20
If it were secure, the understanding of the unique phrase tôrat hāʾādām in
2 Sam. 7.19 as ‘the manner of men’ (so Gesenius, Thesaurus, p. 628; BDB, p.
436) could provide an alternative basis for the likely meaning here. But both S.R.
Driver and H.P. Smith rejected this interpretation of 2 Sam. 7.19 and the phrase
is now generally emended or understood differently (cf. e.g. BH3, NEB, NRSV).
21
Most of the ancient translations did not recognise this and added a verb (see
Text and Versions).
13.1-16 181
Deut. 15.19-23). The word ‘young’ (Heb. šeger) occurs elsewhere

in Biblical Hebrew only in Deuteronomy (see Note r on the transla-
tion). In Deut. 15.20 the owner is permitted to eat the meat of the
animal at the central sanctuary (cf. 12.6-7, 17-18; 14.23). Nothing is
said of this here, and in Num. 18.18 it is explicitly ruled out.
13. Here and in 34.19-20 (but again 22.28b-29 and Deut.
15.19-23 are different) provision is made for ‘redemption’ in the
case of a donkey and a human child.22 ‘Redemption’ (Heb. pādāh)
here, as in the case of a slave who is to be set free (Exod. 21.8-11),
means ‘liberation by means of a ransom’ (A.R. Gray, SAHD, ‘‫’ ָפּ ָדה‬,
Conclusion), i.e. a compensatory payment (Heb. pedûyim, pidyôn).
The donkey, unlike most other domestic animals, could not be
sacrificed because it was unclean (it has undivided hooves and does
not chew the cud), and so a sheep or a goat may be sacrificed in its
place (on ‘animal from the flock’ see 12.3 and Note e on the transla-
tion of 12.1-20). But if he wishes its owner may simply kill it, in a
way that was designed to prevent any of its blood being shed. It is
not stated here or in Exod. 34.19-20 how a human firstborn was to
be redeemed (most likely it was originally in the same way), but the
option of killing it is not offered. In Priestly legislation the provi-
sion for the donkey is extended to all unclean animals and the price
is set, as also for human firstborn, at five shekels of silver (Num.
18.15-16; cf. Lev. 27.27). Whether (and if so when) there was a
time when human firstborn in Israel were regularly killed depends
on the interpretation of Exod. 22.28 (see the commentaries on that
verse) and some other passages: see J.D. Levenson, The Death and
Resurrection of the Beloved Son (New Haven, 1993), pp. 1-52.
14-15. Unlike v. 8, but like 12.26, the instruction of the children
is introduced by a question about the meaning of the rite, which is
notable for its brevity. What is being explained is indicated with
a corresponding fullness in v. 15b (the special case of the donkey
is ignored), and the explanation itself extends beyond the general
reference to the Exodus (v. 14b) to the specific connection with
Pharaoh’s recalcitrance and the final plague (v. 15a: for the latter cf.
12.27). The scope of the plague is defined in terms which are closer
to the Priestly expression in 12.12, without the elaboration about the
22
In Exod. 22.28b-29 and Deut. 15.19-23 the law about firstborn animals only
applies to cattle, sheep and goats, which suggests that donkeys (and other animals)
were not affected by it.
182 EXODUS 1–18
range of human victims that is found in 11.5 and 12.29. But such
abbreviation is natural enough in a legal text. The verb translated
‘made…a difficult thing’ earlier in v. 15 (Heb. qāšāh Hiphil) also
occurs in a Priestly passage (7.3), but the idiom is different there
and closer parallels are found elsewhere.
16. The concluding declaration of the purpose of the ritual (which
is probably not its original meaning: see the notes on vv. 1-2) is very
similar to v. 8, though somewhat shorter: see the notes there. The
main difference is the replacement of ‘reminder’ by the rare word
ṭôṭāpōt (cf. Deut. 6.18; 11.18), which probably originally meant a
kind of ornament worn, in the custom referred to in the biblical
passages at least, on the forehead (see Note x on the translation).
The application to phylacteries like those which have been found at
Qumran (small boxes containing portions of the Law) is later and
unlikely to be referred to here, whatever conclusion is reached about
the occurrences in Deuteronomy. The expression is a metaphor for
something which is prominent and attracts attention, and serves to
identify the ritual as another way in which the Exodus and the role
of firstborn in it is meant to be brought to mind. Although other
passages connect the institution of the practice with the time of the
Exodus, this is the only place where it is seen as having a similar
function to Passover and the festival of Unleavened Bread (contrast
especially Exod. 22.28b-29).
Text and Versions

The passage is preserved, in part at least, in some twenty mss from phylac-
teries and mezuzot found at Qumran or nearby (see further Lange, Handbuch,
pp. 116-22) as well as the related 4QDeutj.23
‫( וידבר‬13.1) LXX εἶπεν δέ is its standard rendering for ‫( ויאמר‬likewise
Sy wʾmr): see Wevers, Notes, p. 49 (on 4.15) for some statistics. It is not clear
whether occasional variations like this reflect the translator’s initiative or a
different Vorlage: either way ‫ ויאמר‬in 12.43 may have contributed in this case.
SP, Qumran mss and the other Vss all support MT.
‫( יהוה‬13.1) TgNmg as often adds ‘the Memra of’.
23
These include XQ4, but its readings remain in some doubt (see, in addition
to Yadin’s editio princeps, M. Baillet, ‘Nouveaux Phylactères de Qumran [XQPhyl
1-4] à propos d’une edition récente’, RdeQ 7 [1970], pp. 403-15 [412-14]) and will
not be cited here.
13.1-16 183
‫( קדשׁ לי‬13.2) TgO,J preserve the divine distance with ‘before me’ (cf.
Tg ); TgN has ‘to my name’, which AramB 2, p. 54 n. 1, suggests may reflect
Nmg
the original priestly role of the firstborn according to rabbinic tradition.

‫( כל־בכור‬13.2) TgJ adds ‘the males’ in line with v. 12 (cf. also MRI
[Lauterbach, p. 129]), while retaining MT’s sing. with ‫בוכרא‬. TgN has the pl.
as it does in subsequent phrases (likewise the mg). LXX adds πρωτογενές to
its usual equivalent for ‫( בכור‬πρωτότοκον), perhaps to clarify the next word.
‫( פטר כל־רחם‬13.2) XHev/Se 5 has the phrase after v. 4 in the next line:
there is another example of its carelessness later in this v. Space considera-
tions suggest that 4Q128 had an omission here, but since it has ‫ רחם‬this was
more likely through homoeoarkton arising from the repeated ‫( כל‬cf. DJD
VI, p. 51). All the Vss have ‘opening/that opens’ for ‫פטר‬. TgO,J,Nmg unusually
have ‫( ולדא‬normally ‘child’) for ‫( רחם‬but cf. v. 12 and 34.19); it is probably
an abbreviation for the fuller equivalent ‫בית ולדא‬. TgN adds another common
equivalent, ‫מעייה‬, to its ‫( וולד‬sic).
‫( באדם ובבהמה‬13.2) XHev/Se 5 writes the second word ‫ובה בהם‬. 4QDeutj
seems to have had a slightly longer text between ‫ לי‬and ‫( באדם‬cf. DJD XIV,
p. 89): perhaps the most likely explanation is that it inverted this pair of
words. LXX ἀπὸ ἀνθρώπου ἕως κτήνους renders freely after the equivalent
phrase in 9.25; 11.7; 12.12 (cf. also v. 15 below).
‫( לי הוא‬13.2) TgN again has ‘(belong) to my name’ for ‫ ;לי‬the mg reading
adds at the end ‘Thus says/said the Lord’, which is TgN’s rendering for ‫אני הוא‬
at the end of a verse twelve times in Leviticus 18–23 (so also TgG(F) at Lev.
23.22).
‫( ויאמר‬13.3) Vulg as often (e.g. 12.31) uses the vivid historic present ait
as a variant for dixit.
‫( זכור‬13.3) SP typically substitutes the straightforward imperative ‫זכרו‬
for the idiomatic inf. abs. of MT, which several Qumran mss attest: there is
no trace there of this SP variant. The Vss naturally use imperatives in their
renderings but this has no bearing on the readings of their Vorlagen. Sy ʾtdkrw
may be used to give the sense ‘commemorate’.
‫( אשׁר יצאתם‬13.3) SP adds ‫ בו‬in line with the requirements of strict
grammar, and two Qumran phylacteries (4Q136 and 4Q140) follow suit: so
also 4QDeutj originally acc. to DJD XIV, p. 89. LXX ἐν ᾗ need not imply such
a Vorlage, especially as LXX in Exodus several times reproduces the retro-
spective pronoun separately when it is present: likewise Vulg in qua and even
Sy d…bh, which may simply follow Syriac idiom. In any case ‫ בו‬is more likely
to have been added (it is by no means essential: GK §138b-c) than omitted,
and 4QExe, 4QDeutj, XHev/Se 5, Mur 4, 8Q3, 4Q145, 34Sey and XQ1 all
agree with MT. TgJ,N add ‫ פריקין‬as they often do after forms of ‫ יצא‬referring to
the Exodus (see e.g. later in this verse).
‫( ממצרים‬13.3) The reading ‫ מארץ מצרים‬appears in SP, 4QExe and some
Qumran phylacteries (4Q128, 136, 140), and it is reconstructed in a lacuna in
4QDeutj. Some LXX mss, including B, have ἐκ γῆς Αἰγύπτου: this reading
184 EXODUS 1–18
was followed by Rahlfs but not Wevers (see THGE, pp. 238-39, for his
reasons). Only later mss of Sy have the addition here. ‫ מצרים‬alone seems to be
characteristic of such expressions in 13.3-16 (so also SP in vv. 8, 9, 13, 16)
and ‘the land of’ here is probably due to influence from 12.51.
‫( מבית עבדים‬13.3) The phrase is omitted, no doubt accidentally (homoe-
oteleuton? – cf. Propp, p. 367), in 4QExe. LXX ἐξ οἴκου δουλείας, replacing
the personal expression with an abstract one, occurs here for the first time and
became standard: similar expressions are adopted in most other translations,
ancient (TgJ combines one with ‘of slaves’) and modern.
‫( בחזק יד‬13.3) Again most of the Vss render freely, ‘with a strong hand’,
assimilating to the more common expression in v. 9: TgO,J follow MT’s
wording, TgJ with a double rendering of ‫( חזק‬as in vv. 14, 16), perhaps to
highlight its strangeness (it does not occur outside this passage).
‫( יהוה‬13.3) TgNmg as usual adds ‘the Memra of’.
‫( אתכם‬13.3) TgN again adds ‫ פריקין‬following a form of ‫יצא‬.
‫( ולא‬13.3) Vulg ut non (comedatis), which seems for Jerome to be a free
variant for ne (cf. Gen. 3.1; 4.15 etc.), makes abstinence the purpose, perhaps
of the remembering rather than the deliverance (though see the note below on
‫ בעבור זה‬in v. 8).
‫( היום אתם‬13.4) SP ‫היום ואתם‬, with the clear implication (exhibited in the
punctuation of many SP mss) that ‫ היום‬was to be read with v. 3, thus requiring
immediate abstinence from leavened bread as in 12.8. But this seems not to
be the concern of this passage, and the non-Priestly narrative attributes such
abstinence at the Exodus to the haste of the departure rather than a prior
command (cf. 12.39).
‫( יצאים‬13.4) The spelling with a yodh after the ṣade in 4QpalExm and
XHev/Se 5 is probably not an error but an orthographical variant, in view of
the forms ‫ אנישי‬and ‫ משיחה‬cited by Qimron, p. 66 (§330.1d).
‫( בחדשׁ האביב‬13.4) The def. art. with ‫ אביב‬prompted LXX (τῶν νέων:
cf. Aq. τῶν νεάρων), Vulg (novarum frugum) and Sy (dhbbʾ, ‘of flowers,
blossoms’) to identify the meaning (cf. the Explanatory Note) rather than
simply transliterate. TgJ prefixed ‘on the fifteenth of Nisan, that is’, adding
the exact date and (as in ch. 12 several times) the later name of the month.
TgN adds ‫זמן‬, ‘the time of’, which it used in 12.17, 41 where MT has ‫עצם‬, and
may have intended a similar emphasis here and in 23.15; 34.18; Deut. 16.1,
where the same addition is made. But it may just be its way of recognising the
obsoleteness of the name Abib.
‫( והיה‬13.5) Vulg, as in v. 11, recognises the ancillary role of the formula
and renders simply by (cum)que.
‫( יבאיך‬13.5) Here and throughout the verse, and indeed throughout the
rest of the passage to v. 16, TgN uses second person pl. forms as in MT of vv.
3-4; likewise Sy to v. 8.
13.1-16 185
‫( יהוה‬13.5) SP, LXX, 4QExe, probably 4QDeutj, two Qumran phylacteries

(4Q140, 145) and TgJ add ‘your God’, probably by assimilation to the similar
Deut. 7.1 (cf. v. 11).
‫( אל־ארץ‬13.5) 4QDeutj and 4Q130 have the inappropriate ‫( אל־הארץ‬the
latter also in v. 11), probably through recollection of Deut. 6.10 or a similar
passage.
‫( הכנעני‬13.5) Apart from Vulg all the Vss render this and the following
gentilics in the pl., but the Heb. texts all support MT’s sing. In LXX the def.
art. is used only on this first occasion, but its force can extend to the coordi-
nated nouns that follow (cf. BDF §276[1]).
‫( והחתי‬13.5) SP omits the waw as in 3.17, making the first three names
into a group (cf. 4Q136). In 4QExe none of the names has waw, and omissions
also occur in 4QDeutj, 4Q140, 144 and 145. LXX and the other Vss agree with
MT, as does 4QpalExm (for once not supporting SP): this reading is probably
original. After ‫ והחתי‬4QExe and LXX have ‘(and) the Hivite(s)’, moved
forward from later in the list, as in LXXF,M in 3.17, where Wevers thinks it
is the original LXX reading. Whatever the reason for this (cf. perhaps Josh.
3.10), it is unlikely to be the original Heb. order here.
‫( והאמרי‬13.5) After this SP has ‫והפרזי הגרגשׁי‬, making the full list of seven
peoples in exactly the same way as it did in Exodus 3 (see Text and Versions
on 3.8, 17): there would be room for the same addition here in 4Q145, but the
actual words do not survive. 4QDeutj and some other phylactery texts (4Q128,
140; 4Q144 is not so clear) had only ‫ הפרזי‬at this point, as in MT in Exodus 3,
and added ‫ הגרגשׁי‬at the end to make up the seven names. In LXX mss various
orderings are attested, but in that which Rahlfs and Wevers consider most
ancient (cf. THGE, pp. 157-58) the Perizzites also occur after the Amorites,
and the Girgashites before them (as in Deut. 7.1). Sy just adds wdprzyʾ at the
end of the list. While omission by parablepsis is always a possibility in lists
such as this, the varied locations of the names which are not in MT here and
the probability of influence from other passages makes it most likely that the
‘short’ text of MT is the oldest that is extant. But its unique omission of the
Perizzites may be due to a scribal slip.
‫( אשׁר‬13.5) 4QExe, 4QDeutj, 4Q140 and a Genizah ms. (acc. BHS) have
‫ כאשׁר‬as in v. 11 (cf. Sy), a clear case of secondary assimilation.
‫( נשׁבע‬13.5) TgN repeats the subject ‘the Lord’ and TgJ adds ‘by his
Memra’.
‫( לך‬13.5) 4QDeutj seems to have read ‫ לה[ם‬and DJD XIV, pp. 89-90,
reconstructs [‫ ]ולזרעם‬in the lacuna at the end of a line, as in Deut. 11.9. But
this probably does not leave sufficient space for the following phrase at the
beginning of the next line, and it is more likely that ‫ להם‬was immediately
followed by ‫ארץ‬. For ‫ להם לתת‬cf. Deut. 11.21, which probably appeared in the
previous column of 4QDeutj (DJD XIV, p. 88). In any case the variant is due
to contaminatio from a similar passage.
186 EXODUS 1–18
‫( זבת‬13.5) For Tgg’s ‘producing’ see Text and Versions on 3.17.

‫( חלב ודבשׁ‬13.5) For TgN’s generalisation of the blessings of the land see
Text and Versions on 3.8, 17: here TgNmg records the variant ‘tasty’ for TgN’s
‘sweet’, a word which TgN and TgF use elsewhere in the same expression (e.g.
Deut. 6.3): for Talmudic parallels to it see AramB 2, p. 55 n. 4.
‫( ועבדת‬13.5) The ‘redundant’ waw is reflected in LXX (καὶ ποιήσεις)
as well as Tgg, but not in Vulg or Sy. The sacral nature of the ‘service’ is
conveyed by the use of the root plḥ in Tgg (cf. LXX λατρείαν; Vulg morem
sacrorum).
‫( שׁבעת ימים‬13.6) SP, LXX and several Qumran phylacteries (4Q132, 136,
140, 144, 145) read the numeral as ‘six’, but 4QpalExm does not anticipate
the SP reading in this case. The variant makes no substantial difference, but
it does bring the wording into conformity with Deut. 16.8. One might at least
consider whether ‘six’ was the original reading here and was changed to
conform to Exod. 12.15 (cf. the next note): so Propp, p. 368.
‫( תאכל‬13.6) LXX ἔδεσθε contrasts with its second person sing. forms
in v. 5 and v. 7b (on v. 7a see below): the pl. forms in Sy and TgN are less
surprising, as they fit the practice of these Vss throughout vv. 5-8. LXX or
its Vorlage must be harmonising with the pl. form in 12.15: the same has
happened in two phylactery texts (4Q140 and 34Seyphyl).
‫( ליהוה‬13.6) Tgg have ‘before the Lord’, as often elsewhere.
‫( יאכל‬13.7) There is no Heb. evidence for a variant and Tgg follow MT.
But LXX and Sy ease the difficult grammar by repeating their second person
pl. forms from the previous verse, and Vulg takes the same route – oddly after
its vesceris before, but in line with its earlier rendering of ‫ יאכל‬in v. 3. Propp
reads ‫ תאכל‬to avoid the grammatical difficulty (p. 369: cf. Note i on the trans-
lation), but the argument is weak.
‫( את שׁבעת הימים‬13.7) Sy renders as in v. 6, ignoring the def. art. here.
‫( ולא‬13.7) SP24 and 4Q144 omit the waw and LXX and Vulg also have
no ‘and’: since the tendency of scribes was to add the conjunction and the
asyndeton is rhetorically effective, this could well be the original reading.
‫( לך‬13.7) Whereas most Vss render literally but clumsily ‘to you’, Vulg
apud te and Sy bkwn give the likely sense. Both these versions also abbreviate
the end of the verse, ignoring the distinction between ‫ חמץ‬and ‫( שׂאר‬cf. Text and
Versions on 12.19). The second occurrence of ‫ לך‬is omitted by XHev/Se 5, no
doubt through carelessness (cf. its spelling ‫ ירארה‬earlier).
‫( גבלך‬13.7) LXX τοῖς ὁρίοις σου and Vulg finibus tuis use the pl. as
Greek and Latin idiom prefer: TgN ‫ תחומיכון‬is more surprising.
‫( והגדת‬13.8) TgJ,N ‘and you shall teach’ gives a more specific equivalent.
24
Von Gall gives the MT reading here, with no mention of a variant: the error
was pointed out by Baillet, ‘Corrections’, p. 29. Sadaqa, Tal and Crown all read
‫ולא‬, as does Camb. 1846.
13.1-16 187
‫( לאמר‬13.8) XHev/Se 5 omits; in Sy the d with the next word may serve
as its equivalent.
‫( בעבור זה‬13.8) LXX διὰ τοῦτο, TgO ‫ בדיל דא‬and Sy mtl hnʾ render
‫ בעבור‬by ‘because of’. Wevers suggests that LXX means ‘(It is) because of
this (namely that) the Lord God did to me…’ (Notes, p. 198: the OL propter
takes it this way) and he thinks that Vulg hoc est quod (fecit…) implied this
too. But Vulg probably just ignored ‫בעבור‬. The meaning of LXX, TgO and Sy
could alternatively be related to the interpretation that is more fully spelt out
in TgJ,N (see Note l on the translation: also MRI [Lauterbach, p. 149]). They
have ‘because of this commandment’: TgN adds ‘of the unleavened bread’ and
TgNmg also has ‘and of the bitter herbs and the Passover meat’. That is, the
ritual is the reason for the Exodus. AramB 2, p. 196 n. 4, notes the similarity
of ‫מצוותא‬, ‘the commandment’, and ‫ מצות‬in v. 7 and plausibly suggests that
this could have given rise to the interpretation (cf. MRI on 12.17 [Lauterbach,
p. 74]).
‫( יהוה‬13.8) LXX κύριος ὁ θεός is a rare expansion of the divine name in
Exodus, only certainly attested elsewhere in 34.14 (Wevers, THGE, p. 241,
argues against Rahlfs that in 13.9 it is secondary). In MT ‫ יהוה אלהים‬appears
in 9.30, but it is probably a conflation of two older readings (see Text and
Versions there). Sy ʾlhy, ‘my God’ (5b1 ʾlhʾ), and the addition of ‘the Memra
of’ (TgJ,Nmg) are further variations.25
‫( לי‬13.8) After ‫ לי‬TgJ ‫ניסין ופרישׂן‬, ‘signs and wonders’ (cf. 15.11), and TgN
‫נצחני קרבינן‬, ‘victories in our battles’, supply what they saw to be the missing
object(s) of ‫( עשׂה‬but see Note m on the translation).
‫( בצאתי‬13.8) TgN ‫ באפקותן‬looks like an Aphel inf. (cf. Stevenson §20),
which would mean ‘when (he) brought us out’: cf. the causative forms in vv.
3 and 9. TgN also adds ‫ פריקין‬as in vv. 3-4 and elsewhere.
‫( ממצרים‬13.8) Some LXX mss insert ‘the land of’ as elsewhere and the
phylactery texts 8Q3 and 4Q145 may also have done so.
‫( והיה‬13.9) SP ‫והיו‬, with the pl. perhaps referring to the ‫( מצות‬or [cf.
Houtman, p. 213 n. 140] ‫ )?הימים‬in v. 7. Phylacteries are another possibility
(Propp, p. 370), but there is no specific word for them in the context and ‘there
is no evidence that they were used at any time’ by the Samaritans (Crown et
al., A Companion, p. 185).26
‫( לך‬13.9) Vulg has no equivalent.
‫( לאות‬13.9) Vulg quasi signum clearly identifies the sense as metaphor-
ical (as again below). By contrast TgN’s pl. ‘signs’ fits a reference to the use
of phylacteries (see the next note); TgJ ‘(you shall have) this miracle (clearly
inscribed)’ has the same interpretation in mind and renders freely here.
25
The simple substitution of ὁ θεός for ‫ יהוה‬in LXX is much more common:
see 5.21; 6.26; 8.25-26; 9.5; 10.11, 18; 13.21; 14.31.
26
The SamTg has no reference to phylacteries here.
188 EXODUS 1–18
‫( על־ידך‬13.9) Mur4, LXX, Vulg, TgO and Sy agree with MT, but TgJ has
‘on the phylactery of the hand, at the upper part of the left (hand)’, in exact
accord with the rulings given in MRI (Lauterbach, pp. 150-52); likewise more
briefly TgN, ‘upon your arms’. The pl. ‫ ידיך‬is attested here in Heb. by SP
(where the sense must be metaphorical) and in several of the Qumran phylac-
tery texts (XQHev/Se5; 4Q130, 132, 136, 140, 145; XQ1): so also in v. 16
and in Deut. 6.8 and 11.18, presumably with reference to all the individuals
in the community.27
‫( ולזכרון‬13.9) Vulg again prefixes quasi; TgN adds ‫ טב‬again, as in 12.14
(see Text and Versions there). XHev/Se5 has the variant spelling ‫ולאזכרון‬, on
which see Qimron, p. 39.
‫( בין עיניך‬13.9) LXX πρὸ ὀφθαλμῶν and Vulg ante oculos are free
renderings that imply a metaphorical understanding here too (cf. Wevers,
Notes, p. 198; BAlex, p. 157). By contrast TgJ gives a very full practical guide:
‘clearly inscribed on the phylactery of the head, fixed before your eyes, at the
upper part of your head’ (cf. MRI [Lauterbach, pp. 152-53]).
‫( למען‬13.9) Vulg et ut might simply intend a further purpose for the
celebration (so the Douay tr.), but the omission of ‘and’ at the beginning of
v. 10 suggests that Vulg may have taken the purpose clause with what follows
it, with v. 9b understood as a parenthesis. This is easier in Latin than in Heb.
‫( יהוה‬13.9)1o TgN adds ‘constantly’ for emphasis, as does Vulg (semper).
‫( ביד חזקה‬13.9) Camb. 1846 and two SP mss cited in von Gall read ‫ביד‬
‫החזקה‬, presumably influenced by the definite phrase that occurs four times in
Deuteronomy. 4Q130 has ‫בחוזק יד‬, assimilating to the unusual phrase in vv. 3,
14 and 16, while TgJ (as in v. 3) uses its composite equivalent ‘by the strength
of a mighty hand’.
‫( יהוה‬13.9)2o TgN adds (as elsewhere after a form of ‫ יצא‬referring to the
Exodus) ‫פריקין‬.
‫( ושׁמרת‬13.10) LXX surprisingly has the pl. form φυλάξεσθε after the
sing. in v. 9: either the translator or his Vorlage presumably had 12.17 in mind.
On Vulg custodies (without et) see the note on ‫ למען‬in v. 9: the alternative to
the explanation given there would be that v. 10 is seen as simply an elabora-
tion of v. 9aβb (cf. the addition of semper there).
‫( את־החקה הזאת‬13.10) The rendering of ‫ החקה‬in LXX (τὸν ν́ομον) seems
to equate it with the ‫ תורה‬in v. 9 (for the equivalence see 12.43 and Wevers,
Notes, p. 191), while those of Aq, Symm, Vulg (cultum), TgN (‘the statute
of this law’: cf. 12.49) and Sy (pwqdnʾ hnʾ wnmwsʾ [Sy’s word for ‫ תורת‬in
v. 9] hnʾ) distinguish them. TgJ’s addition ‘of the phylacteries’ very clearly
takes v. 10 as further guidance about their use (see the detailed discussion in
27
GSH §55bγ (pp. 244-45) lists a number of other places where MT and SP
differ similarly over the second person m.s. suffix, enough perhaps to suggest that
‫יך‬- might be a phonetic/orthographical variant rather than a pl. form.
13.1-16 189
MRI [Lauterbach, pp. 154-57], which takes the same view but in the process
exposes its difficulties).
‫( למועדה‬13.10) Some Qumran phylacteries read ‫ למועדו‬or ‫( למועדוה‬4Q129,
140, 145), but this is only a phonetic variation: see Qimron, pp. 39-40 and
also the next note. LXX κατὰ καιροὺς ὡρῶν combines two of its renderings
for ‫( מועד‬cf. the variation in Num. 9.2-3), perhaps in an attempt to specify
‘the right times’. TgNmg has ‘to be doing it’ (sc. the statute), freely following
Deuteronomic usage (e.g. Deut. 5.1).
‫( מימים ימימה‬13.10) 4Q129, 136, 140, 145 have the same unusual substi-
tution of ‘o’ for ‘a’: see the previous note. Most Vss render literally, without
attempting to interpret the idiom, but TgO ‘from (right) time to (right) time’
does so. TgJ ‘on workdays but not on sabbaths or festivals, by day and not
by night’ finds here the rulings specified in MRI (see above). TgF,Nmg ‘from
those days to those months’ mix possible senses of ‫ ימים‬in a way that remains
obscure.
‫( יהוה‬13.11) LXX and SP add ‘your God’ as they do in v. 5 (see the note
there). TgNmg prefixes ‘the Memra of’. Vulg has no equivalent to the divine
name here, no doubt presuming that it can be supplied from the end of v. 9.
‫( אל־ארץ הכנעני‬13.11) 4Q130 again has ‫( הארץ‬see the note on v. 5) and also
replaces ‫ אל‬with ‫על‬, perhaps a slip due to a common phrase. As in v. 5 most of
the Vss render the gentilic in the pl., but Vulg has MT’s sing.
‫( לך ולאבתיך‬13.11) LXX has simply τοῖς πατράσιν σου, matching the
most common formula elsewhere (esp. in Deuteronomy but also in v. 5). ‫לך‬
is of course justified by verses like 3.8 and 6.8 (the latter is cited by MRI here
[Lauterbach, p. 158]). The O-text (cf. Syhex) adds σοι καί. Propp (p. 370)
adopts the shorter LXX reading, but it is probably a secondary assimilation
to v. 5.
‫( והעברת‬13.12) LXX* (again with an ‘unnecessary’ καί as in v. 5b) and
TgO rendered with words meaning ‘remove’, a possible sense of ‫העביר‬, but a
well attested LXX variant and the other Vss recognised the sense ‘set apart’
(see Note q on the translation and MRI [Lauterbach, p. 159]).
‫( כל פטר רחם‬13.12) The Vss generally render as expected from v. 2 but
TgN, its mg (with which cf. TgF(VN) at 34.19) and Sy expand with words corre-
sponding to the prefixed ‫ בכור‬there. LXX adds τὰ ἀρσενικά, ‘the males’, here
to match the second half of the verse or, as Propp suggests (p. 370), v. 15.
‫( ליהוה‬13.12)1o TgO,J as often have ‘before the Lord’ and TgN repeats its
‘to the name of the Lord’ from v. 2 (see the note there), both here and at the
end of the verse.
‫( שׁגר בהמה‬13.12) For ‫ שׁגר‬LXX has its usual ἐκ τῶν βουκολίων, ‘from the
herds’, which clearly does not fit here. Vulg ignores ‫שׁגר‬, as it does in Deut.
7.13 (but not in Deut. 28), while TgO,N and Sy treat it as an equivalent to ‫רחם‬.
TgJ, ‘which its mother bears (prematurely?)’, explains from a sense which the
verb ‫ ְשׁגַ ר‬can have in Aram.: so also MRI (Lauterbach, p. 160), though appar-
ently of an animal which is not subject to the law. For ‫ בהמה‬some Qumran
190 EXODUS 1–18
texts (4Q130, 155; possibly 4Q134) read ‫( בבהמה‬cf. v. 2), and 8Q3 may have
had ‫בבהמתך‬, though only the last letter survives and the text seems to be
abbreviated (DJD III, pp. 150-51). LXX ἢ ἐν τοῖς κτήνεσιν σου may well
therefore be based on a Vorlage different from MT (but scarcely preferable to
it): in v. 2 it has ἕως κτήνους.
‫( יהיה לך‬13.12) SP has ‫יהיו‬, an inferior reading that is influenced by the
context (perhaps esp. the pl. ‫ הזכרים‬that follows). The omission of ‫ לך‬in XHev/
Se5 is a simple error of copying.
‫( הזכרים‬13.12) Sy has dkrʾ, maintaining the earlier collective sing.
wording. All the other Vss except TgN add ‘you shall consecrate’ on the basis
of v. 2, but this is unnecessary: either the force of ‫ והעברת‬is carried over or, as
in our translation, v. 12b is a nominal clause.
‫( וכל־פטר חמר‬13.13) LXX and Vulg do not represent the initial waw
and Vulg does not render ‫ כל‬either (cf. v. 12b). LXX and TgJ,N repeat their
formulae for ‘opening the womb’ (as in v. 12b), but TgO, Vulg and Sy substi-
tute the simpler ‘firstborn’, with ‘male’ added by Sy according to 7a1. For ‫חמר‬
Sy has dbʿyrʾ, ‘of cattle’, thus allowing redemption of all animals with a lamb
and ignoring the special provision for the (unclean) donkey. Weitzman (Syriac
Version, p. 155) thinks a copying error (cf. v. 12) is a more likely explanation
than the influence of non-rabbinic interpretation; similarly Propp, p. 371.
‫( תפדה‬13.13)1o LXX ἀλλάξεις, ‘you shall exchange’, followed via the
OL by Vulg, is a rare rendering of ‫פדה‬, found only in this verse (twice) and in
Lev. 27.27, all cases where substitution of one animal for another is involved.
Aq, Symm and Theod replaced it with the normal equivalent λυτρόω.
‫( בשׂה‬13.13) Most of the Vss render with words for ‘lamb’, but the Three
seem, in slightly different ways, to have allowed for the wider meaning
implied in 12.3-5 (cf. O’Connell, Theodotionic Revision, p. 154, and for paral-
lels to Theod’s ἐκ ποιμνίου Lev. 27.26 and Isa. 66.3).
‫( ואם לא‬13.13) One phylactery ms. from Qumran (4Q130) omits the waw.
‫( תפדה‬13.13)2o SP reads ‫תפדנו‬, with the suffix pedantically added to
specify the object: the same reading appears already in two Qumran texts
(4Q130, 155). In BH the suffix is often dispensed with in such cases (cf.
GK §117f), but SP adds it in a number of other places (cf. GSH §55bγ [pp.
245-46]). Traces of the same addition appear in some LXX witnesses, Sy,
TgN and more surprisingly in Symm and Theod (cf. O’Connell, Theodotionic
Revision, pp. 117, 120-21), but it is not certain that they are based on a
different Vorlage from MT. LXX again uses ἀλλάσσω, but Vulg has redemeris
this time.
‫( וערפתו‬13.13) SP and all the other Heb. evidence (XHev/Se5, Mur4,
4Q130, 133 (damaged), 155, 34SeyPhyl, XQ3) agree with MT. LXX,
however, has λυτρώσῃ αὐτό, ‘you shall redeem it’, which is quite contrary
to the Heb. and will reflect a less drastic (monetary) alternative to killing
the donkey (cf. LXX on 34.20). This presumably represents the contem-
porary practice of Alexandrian Jews: cf. Z. Frankel, Über den Einfluss der
13.1-16 191
palästinischen Exegese auf die alexandrinischen Hermeneutik (Leipzig, 1851,

repr. 1972), pp. 98-99, with a reference to Philo, Spec. Leg. 1.135, where
the same provision is applied to horses and camels. Propp (p. 372) suggests
that LXX read the Heb. as ‫וערבתו‬, but λυτρόω is never used for ‫ערב‬. The
Three correct this in line with the different interpretations given elsewhere
(cf. Salvesen, Symmachus, pp. 87-88): Symm ἀποκτενεῖς, ‘you shall kill’,
corresponds to TgN,F, Vulg and Sy, and more distantly to TgO,J, ‘strike’,
while Theod νωτοκοπήσεις (cf. Isa. 66.3) and Aq τενοντώσεις (in 34.20
τενοντοκοπήσεις, where Symm has τραχηλοκοπήσεις) perhaps both mean
‘cut the neck’ (for νωτοκοπήσεις O’Connell, Theodotionic Revision, p. 289,
usefully compares the use of νῶτος for ‫ ע ֶֹרף‬in the καιγε sections of Samuel–
Kings [2 Sam. 22.41; 2 Kgs 17.14]), as later in Rashi and Ibn Ezra (cf. M.Bek.
1.7, TgNmg and TgJ on 34.20, all mentioning a ‫קופיץ‬, ‘axe’).
‫( וכל־בכור אדם‬13.13) LXX again does not represent the waw.
‫( בבניך‬13.13) TgJ adds ‘but not among your slaves’: according to E.
Levine, ‘A Study of Targum Pseudo-Jonathan to Exodus’, Sefarad 31 (1971),
pp. 27-48 (44-45; cited in AramB 2, p. 197 n. 12), the extension of this
requirement to slaves by the early Karaites is here rejected.
‫( תפדה‬13.13)3o Both TgJ and Vulg (pretio) specify that a monetary
payment was to be made, in line with Num. 18.16.
‫( והיה‬13.14) Vulg again (see the note on v. 5) dispenses with an equiva-
lent to ‫ היה‬and here LXX and Sy (and a Genizah ms. acc. BHS) also do so, as
they do with ‫ (ו)יהי‬in v. 15: perhaps in mid-paragraph such elaboration seemed
especially unnecessary.
‫( ישׁאלך בנך‬13.14) TgN renders in the pl., ‘your sons…’, just as it has used
a pl. ‘you’ ever since v. 3 (but see below on ‫)ואמרת אליו‬.
‫( מחר‬13.14) Most Vss stay woodenly with their words for ‘tomorrow’,
but here LXX, which has αὔριον elsewhere, uses μετὰ ταῦτα. This renders
the similar word ‫ אחר‬in 5.1 (cf. 3.20; 11.1, 8; 34.32), so LXX might have
found this in its Vorlage here. However, it is also possible that the Exodus
translator recognised the extended sense of ‫( מחר‬see Note s on the translation):
‘μετὰ ταῦτα carries this sense much better than αὔριον would have done’
(Wevers, Notes, p. 201).
‫( לאמר מה זאת‬13.14) Sy unusually turns ‫ לאמר‬into a finite verb, wnʾmr
lk; TgJ specifies that ‘the commandment of the firstborn’ is meant (cf. its less
happy expansion in v. 10). The misspelling ‫ מא זת‬in XHev/Se5 is particularly
slipshod.
‫( ואמרת אליו‬13.14) Here uncharacteristically (see the note above on
‫ )ישׁאלך בנך‬TgN has sing. forms, presumably as part of its patchy correction to
MT (the mg records the more usual pl. here). Early variants read ‫ לו‬for ‫אליו‬
(4Q130: LXX αὐτῷ and Vulg ei could be representing this or ‫ )אליו‬and add ‫כי‬
afterwards (4Q130, 135, possibly 133: again LXX [ὅτι] and Sy [d] could, but
need not, reflect knowledge of such a reading). MT probably has the original
text in each case.
192 EXODUS 1–18
‫( בחזק יד‬13.14) On the slight variations in some of the Vss see Text and
Versions on v. 3.
‫( ממצרים‬13.14) TgJ,N add as often ‫פריקין‬, but TgN accidentally omits ‫ממצרים‬
(its mg corrects this). LXX ἐκ γῆς Αἰγύπτου is a rare departure from MT
in the original translation of this expression: Wevers, THGE, pp. 238-39,
considers a different Vorlage (for which there is some evidence in v. 3) the
likely cause.
‫( מבית עבדים‬13.14) XHev/Se5 carelessly omits, perhaps by homoe-
oteleuton. On TgJ’s double rendering see Text and Versions on v. 3.
‫( ויהי כי־הקשׁה‬13.15) In addition to ignoring ‫ (ו)יהי‬Vulg has nam cum for
‫כי‬, adding a causal connection which could be derived from ‫כי‬. For ‫ הקשׁה‬TgO,N
have exactly corresponding renderings which can bear what is probably the
sense of MT, ‘refused’ (cf. Note u on the translation). LXX’s ἐσκλήρυνεν
(followed by an inf.) was presumably intended in the same sense, as Wevers
proposes (Notes, pp. 201-202), but the intrans. use of the active is most
unusual: only 2 Kgdms 2.10 (also with inf.) is at all similar. The other Vss
variously work from 7.3, the only other occurrence of ‫ קשׁה‬Hiphil in Exod.,
and either supply ‘the Memra of the Lord’ as the subj. (TgJ: cf. MRI [Lauter-
bach, p. 167]) or render ‘was hardened’ in the passive (Sy and Vulg, the latter
with the addition of et nollet after TgO,N).
‫( יהוה‬13.15) LXX has no equivalent, regarding the subj. as clear enough
from v. 14b: it would be even more so if LXX, like TgJ, had meant Yahweh to
be cause of the hardening.
‫( כל־בכור‬13.15) TgN and Sy have the pl., the former as throughout the
remainder of the verse and the latter only as in the final occurrence.
‫( בארץ מצרים‬13.15) Sy prefixes d, ‘which (were)’.
‫( מבכר אדם‬13.15) LXX has the pl. here and in the next phrase (for both
nouns).
‫( ועד‬13.15) SP omits the waw, as it did in similar phrases in 9.25 and
12.12: so here also XHev/Se5 and 4Q130 (whereas Mur 4, 34SeyPhyl, XQ3
and 4Q129 agree with MT). The omission of ‘and’ in LXX (and Vulg) is not
significant: see Wevers, THGE, pp. 164-65.
‫( בהמה‬13.15) Vulg oddly has pl. iumentorum after hominis in the previous
phrase.
‫( אני זבח‬13.15) 1Q13 ‫( ואני‬if this placement of fr. 44 is correct) and 4Q135
‫ [זבח] אני‬must be careless errors, whereas the substitution of ‫ אנכי‬for ‫ אני‬in
4QExd, 4Q129, 130 suggests a minority local variation, with its origin perhaps
in copying from memory. TgN characteristically has the pl. ‘we sacrifice’ (as
later ‘we redeem’, for which the mg has ‘I…’).
‫( ליהוה‬13.15) TgO,J,N as usual have ‘before the Lord’: TgNmg ‘to the name
of the Lord’, reverting to the main text’s formula in v. 2.
13.1-16 193
‫( כל־פטר חמר‬13.15) Another recurrent Qumran variant, this time an

impossible reading, is ‫ וכול‬here in 4Q129, 130 and 135 (this time 4QExd
agrees with MT). TgNmg inserts ‫קדמי‬, its word for ‘firstborn’ (pl.) in v. 2, and
then has the sing. like MT instead of TgN’s pl.
‫( וכל־בכור בני‬13.15) SP, assimilating to v. 13, reads ‫וכל־בכור אדם בבני‬:
so also 4Q129, 130 and very likely 4QExd (on the basis of the length of
the preceding line: DJD XII, p. 128). The shorter (and surely original) MT
reading appears in XHev/Se5, Mur 4, 8Q3, 4Q135, 155 and XQ3 and is
reflected in all the Vss. Vulg again has the pl. primogenita here.
‫( אפדה‬13.15) The (presumably) iterative sense of the imperfect is
accurately represented by Vulg’s present redimo and TgN’s part., but missed
in LXX (λυτρώσομαι!), TgO,J,Nmg and Sy.28 TgJ, as in v. 13, adds ‘with money’.
‫( והיה‬13.16) As in v. 9 SP has ‫( והיו‬so also perhaps 4Q133, 135), with the
pl. here referring to the firstborn. SP also has ‫ לך‬afterwards (likewise 4Q129,
130 and Sy [but not 5b1]), no doubt an addition to match v. 9.
‫( לאות‬13.16) TgJ amplified, as in v. 9, with ‘clearly inscribed’.
‫( ידכה‬13.16) The unusual plene spelling of MT is reproduced in a number
of phylacteries from Qumran and its surroundings (XHev/Se5, Mur 4, 8Q3,
XQ3 and apparently 34SeyPhyl), where it is more normal (cf. Qimron, pp.
58-59). SP and other phylacteries (4Q129, 130, 135) have ‫( ידיך‬or ‫)ידיכה‬,
apparently a pl. form (cf. TgN ‘your arms’), but see the note on v. 9. LXX,
Vulg, TgJ (with the addition of ‘left’ as in v. 9) and Sy all have the sing. like
MT, which is probably original.
‫( ולטוטפת‬13.16) TgO,J,Nmg explicitly equate the ‫ טוטפת‬with phylacteries, as
they do in the other two occurrences of the word (Deut. 6.8; 11.18). The other
Vss all provide an explanation or a substitute for it. TgN (with the addition of
‘good’ as in v. 9) and Sy use the term that is parallel to ‫ אות‬in v. 9, ‘reminder’.
Vulg adspensum quid ob recordationem, ‘something hung up as a reminder’,
is partly related to this, but seems to draw on another line of interpretation
(possibly based on an etymology) which appears in Philo, Spec. Leg. 4.137-
39, the OL in Deut. 11.18 (mobilia), Vulg at Deut. 6.8 (et movebuntur) and in a
Greek rendering σαλευτόν, ‘shaken’, which appears in Syhex and in two Greek
mss (cf. BAlex, pp. 52-55). The latter equivalent is also given (in the pl.) for
Theod at Deut. 6.8. Symm had the obscure διεσταλμένοι there, while Aq
both there and here had νακτά, ‘solid things’, which Salvesen following Field
plausibly connects with the phylacteries (Symmachus, pp. 150-51, in a full
review of the ancient evidence for the Deuteronomic occurrences). The word
was evidently a mystery to the translators, and LXX’s ἀσάλευτον here (and in
the pl. in Deuteronomy), ‘immoveable, unchangeable’, a word that is applied
to laws (Philo, Vita Mosis 2.14) and legal documents and decisions, could
28
LXX shows some awareness of the iterative imperfect in 1.12; 5.4-15;
33.7-11. Evans (Verbal Syntax) does not discuss this issue.
194 EXODUS 1–18
be a guess from the context. BAlex, pp. 54-55, alternatively suggests that it
is a deliberate inversion of σαλευτόν, which had become unacceptable as a
description for the law, by an exegetical process known from the Targumim
and elsewhere in the LXX. On this view σαλευτόν would be the original
LXX reading, largely suppressed in the tradition. The use of sing. equivalents
in some Vss adds weight to Propp’s argument from Qumran orthography
(p. 373) for seeing ‫ טוטפת‬as sing. (see also Note x on the translation).
‫( בין עיניך‬13.16) TgJ here has ‘between your eye-lids’ (or ‘eye-brows’?).
‫( בחזק יד‬13.16) Once again some of the Vss diverge from the unusual
expression in MT: see Text and Versions on v. 3.
‫( הוציאנו‬13.16) SP reads ‫הוציאך‬, ‘brought you out’, which matches the
pronouns earlier in the verse and also the end of v. 9 (similarly LXX, Sy).
No Qumran ms. has this variant, while XHev/Se 5, Mur 4, 4Q130, 133,
XQ3, Vulg and Tgg agree with MT, which as the difficilior lectio should be
preferred (see also Note y on the translation). TgN adds ‫פריקין‬: its mg prefixes
the verb ‫ פרק‬instead, with the same theological effect.
C h ap t er 1 3 . 1 7 - 2 2
Aspec t s of I s r ael’ s D epa rtur e
Between the legal provisions of 12.43–13.16 and the dramatic

narrative of events at the ‘sea’ in ch. 14 a short narrative section
deals (mainly: v. 20 is an exception) with general aspects of the
Israelites’ onward journey. In this it is similar to 12.34-42 (the itin-
erary-notes in 12.37 and 13.20 make an explicit link between the
two passages) and there are links to topics that were central to the
earlier narrative: Pharaoh’s eventual agreement to let the Israelites
depart (v. 17a) and the departure from Egypt itself (v. 18b, with
the verb ‘went up’ [Heb. ʿālāh] as in 12.38: cf. also ‘from here’
[sc. Egypt], again with ʿālāh, in v. 19). But, while this is still a
transitional passage, the journey ahead through ‘the wilderness’ is
now very much in view, both in the description of the route that
was taken (v. 18a) and in the location of Etham ‘on the edge of the
wilderness’ (v. 20b). The presence of Yahweh, in the pillar of cloud
and fire, as his people’s guide (vv. 21-22) introduces a new theme
which will recur in different forms in the following narrative (see
the Explanatory Note).
It is therefore not surprising that the possibility has been raised,
and vigorously debated, that this section and, more significantly, the
sea narrative which follows it, belong not to the Exodus story itself
but to the wilderness narrative. G.W. Coats especially contested
Noth’s version of the common view that the sea-narrative had
once been the heart of the Exodus story and in the written narrative
remained its finale or ‘postlude’ and even its ‘goal and climax’ (cf.
Noth, p. 82, ET, pp. 104-105; Coats, ‘Traditio-Historical Character’:
so still in Exodus 1–18, pp. 99-122). Childs agreed with Coats as far
as the J narrative was concerned, but held that in P the sea-narrative
was tightly bound into the Exodus story by a series of theological as
well as thematic motifs (‘A Traditio-Historical Study of the Reed Sea
Traditions’, VT 20 [1970], pp. 406-18; cf. his Exodus, pp. 221-24).1
1
Coats’s view was rejected outright e.g. by D. Patrick, ‘Traditio-History of
the Reed Sea Account’, VT 26 (1976), pp. 248-49; Kohata, Jahwist, p. 296 n. 113;
Houtman 2, pp. 233-34.
196 EXODUS 1–18
Much of the debate naturally centres on the analysis and contents of

the following chapter and can better be dealt with in the next section
of the commentary: here we shall be concerned only with arguments
based on 13.17-22, to which Coats devoted a separate study (‘An
Exposition for the Wilderness Tradition’, VT 22 [1972], pp. 288-95;
more briefly in Exodus 1–18, pp. 107-108).
Coats argues that 13.17-18 relate to the whole wilderness journey,
citing the expressions ‘lead’, ‘way’ and ‘armed for battle’, which
recur in the later narrative, as well as ‘wilderness’ itself. But this is
not conclusive and the mention of Yam Suf in v. 18 in fact suggests
that the route of the departure from Egypt and the journey to the next
significant point is meant, no more. Coats also takes the itinerary-
note in v. 20 to support his argument, because it forms part of the
‘wilderness chain’, but since its range only extends to ‘the edge
of the wilderness’ it seems that entry into the wilderness still lies
ahead. His best argument comes from the reference to the pillar of
cloud and fire in vv. 21-22, which is certainly often associated with
the journey through the wilderness. But, as with the itinerary, it is
perhaps with the people’s movement as such that this motif is to be
associated and so it is not necessarily decisive for where the wilder-
ness theme begins.2 It is in any case doubtful whether the varied
components of the present passage add up to anything that could
be called an ‘Exposition’ in the way that Gen. 12.1-3 and Exodus
1.1-14 can perhaps be so described (on them see Coats, ‘Structural
Transition’, pp. 129-38). It may therefore be better to recognise,
as Coats does elsewhere (e.g. ‘Traditio-Historical Character’, pp.
254, 265; Exodus 1–18, p. 103), at least as far as this passage is
concerned, that the wilderness narrative is not sharply demarcated
from the Exodus story: one may do better to speak of a gradual
transition which begins already in 12.37-42 and continues here.3
The beginning of the section is marked by a Masoretic division
and the opening of a new reading in the one-year lectionary, and
the end by a Masoretic division. A vacat in 4QExc marks the end of
the passage: at the beginning the only evidence from Qumran is in
2
One might even argue that the effect of the itinerary-note, covering a specific
stage of the journey, is to tie down any hints of a longer-term reference in vv. 17-19
and 21-22 to a limited section of the route.
3
Cf. Coats’s more recent description of the passage as ‘A Transition to the
Wilderness Traditions’ (Exodus 1–18, p. 107).
13.17-22 197
those phylactery manuscripts which include 13.11-16 as an extract

from the biblical text (on them see the introduction to 13.1-16).
Sub-division of the passage can proceed on the basis of both form
and content: (i) a theological explanation for the route of the Israel-
ites’ departure from Egypt (‘God’ as subject: vv. 17-18a); (ii) a
further description of the manner of the Israelites’ departure (v. 18b:
cf. 12.37b-38); (iii) a report of Moses’ fulfilling of Joseph’s adjura-
tion of the Israelites to take his ‘bones’ with them when they leave
Egypt (v. 19); (iv) an itinerary-note describing the next stage of the
Israelites’ journey (v. 20); (v) a report of Yahweh’s guiding presence
in the pillar of cloud and fire (vv. 21-22). The grammatical subject
changes repeatedly from sub-section to sub-section, but in such a
way as to form a chiastic pattern: God (vv. 17-18a); the Israelites
(v. 18b); Moses (v. 19); the Israelites (‘they’: v. 20); Yahweh/the
pillar of cloud and fire (vv. 21-22). Its coherence is complicated by
the use of different expressions for God and also for the Israelites
(‘the people’ in vv. 17-18a; ‘the children of Israel’ in vv. 18b-19;
‘they’ in vv. 20-22). It is not surprising that critical scholars have
generally regarded the passage as composite in origin.
Knobel distinguished between the Grundschrift’s itinerary-note in v. 20

and the Jehovist’s addition of vv. 17-19, 21-22, all of which he later attributed
to his Rechtsbuch (or E: Exod.-Lev., p. 111; Num.-Jos., p. 532). Wellhausen’s
detachment of vv. 21-22 for J (cf. the different expressions for God) estab-
lished a pattern of analysis (17-19 = E; 20 = P; 21-22 = J) which was widely
followed subsequently and is in essence the view of Graupner and Baden still
today (Elohist, pp. 71-77; Composition, pp. 124-25, 199, 207).4 In their quest
for evidence of a ‘third early source’ Smend (pp. 137-40) and his followers
attributed v. 20 to J1 (L, N). Apart from this the first main divergence from
Wellhausen’s analysis came, not surprisingly from Rudolph, who argued
(pp. 27-28) that vv. 17-19 could not be E because they were isolated from
the rest of the Exodus narrative: he saw them as comprising a redactional
‘commentary’ on the change of route described (according to him by J) in
14.1-2 and a late addition (v. 19) based on Gen. 50.24-25 (as Gressmann,
Anfänge, p. 55 n. 3, had suggested earlier). He assigned v. 20 (as well as vv.
21-22) to J, as he had done with the earlier itinerary-note in 12.37. In this
(but not in his elimination of E) he was followed by Noth, Hyatt and initially
4
Early on (cf. Holzinger) it was suggested that the temporal clause at the
beginning of v. 17 was an addition of RJE because it used the terminology of J,
and a minority of scholars have continued to take this view (e.g. Noth, Kohata,
Gertz, Graupner).
198 EXODUS 1–18
Coats (‘Traditio-Historical Character’, p. 255), but the idea then lost favour
until its recent revival in a different form (see below). Coats had abandoned it
by 1972, as part of a more radical fresh analysis of the passage (‘Wilderness
Itinerary’, pp. 145-47; cf. ‘Structural Transition’, pp. 138-39; still maintained
in Exodus 1–18, pp. 102-108), which also denied that vv. 17-19 were from E.
In the case of vv. 17-18 this was because E had no wilderness narrative (which
may seem irrelevant, but for Coats the whole of vv. 17-22 was an ‘exposition
of the wilderness theme’ [see above]).5 Coats attributed these verses to P or
PS, noting that Gen. 50.22-26 (to which v. 19 refers) was also ‘secondary P’
and so anticipating what has become recently a much more common view
(see further below). By contrast Kohata (rather like Dozeman more recently)
saw vv. 21-22 as a late redactional addition related to Deuteronomistic and
secondary Priestly passages elsewhere (pp. 292-93).
These views form a transitional stage to the most recent developments,
which are characterised by diversity and, as elsewhere, a growing tendency
to date components of the passage to a late period. Particularly influential
has been Rendtorff’s inclusion of v. 19 among texts deriving from a very late
theological redaction of the Pentateuch (Problem, pp. 75-77), which Blum
described more precisely already in 1984 and now identifies as part of a ‘Josua
24 Bearbeitung’ (see also Studien, pp. 363-65), to which Gen. 33.19; 50.25;
Josh. 24.32 also belong. Both Blum himself and others who have followed
him (K. Schmid, Gertz, Albertz) find here evidence of a ‘Hexateuchal’ dimen-
sion to the editing of Israel’s traditions at a very late stage. Levin, Gertz and
Albertz also regard vv. 17-18 as very late in origin (it is not clear where Blum
and Schmid would place them – but for Blum see now his ‘Die Feuersäule in
Ex 13-14 – eine Spur der “Endredaktion”?’, in his Textgestalt und Komposi-
tion. Exegetische Beiträge zu Tora und Vordere Propheten [ed. W. Oswald,
FAT 69; Tübingen, 2010, pp. 137-56), where he argues that they must belong
to a pre-Priestly version of the route-description, because they exclude the
very route by Lake Sirbonis which [he believes] is implied by 14.1-2). For
these scholars the oldest elements of the passage are the itinerary-note in
v. 20 and the beginning of v. 21 (all of 21-22 according to Albertz), some of
which may be pre-exilic. Equally different from the older consensus, but in
the opposite direction, is the unitary treatment of the older components of the
passage (vv. 17-18[a], 21-22) by Vervenne (‘The “P” Tradition’, pp. 77-79:
5
Coats could have simply been following Rudolph in this, but he was in
Heidelberg in 1970–71 for research leave (cf. the beginning of ‘The Wilderness
Itinerary’) and it is more likely that he was influenced by ideas that were already
beginning to take shape there at this time: R. Kessler submitted his Querverweise
as a dissertation (in which he concluded that v. 19 was a late redactional addition:
his view of vv. 17-18 is not so clear) in 1972 and has confirmed privately that
he and Coats were in close contact during its preparation. See also below on
Rendtorff.
13.17-22 199
JE), Van Seters (Life, pp. 128-31: J) and Propp (p. 476: E, probably). Both
Vervenne (pp. 85-86: P) and Propp (ibid.: R) regard v. 20 as later. Dozeman
assigns vv. 17-18a, 19-20 (but not vv. 18b, 21-22) to ‘Non-P’.
It is somewhat artificial to separate the analysis of these verses

from that of ch. 14, to which they are connected in various ways.
But a striking feature of them is the reappearance of ‘God’ (Heb.
ʾelōhîm) as a divine title four times in vv. 17-19 alongside ‘Yahweh’
in v. 21, which recalls the situation in chs. 3–4 especially. There the
variation corresponded in our view to a distinction between two
accounts, traditionally called J and E, extracts from which had been
intertwined by a redactor. Subsequently there has continued to be
evidence to support such a distinction, but without this particular
variation being present. Scholars of an earlier generation, and
some still today (e.g. Graupner, Baden), have had no hesitation in
assigning vv. 17-19 to E and vv. 21-22 to J. It could be said that no
better explanation of the variation here has been forthcoming. Both
sections deal with the same topic, the journey of the Israelites, albeit
in different ways, and if they had a common origin one might have
expected them to be adjacent to one another. Their separation could
be due to v. 20 being a later insertion, and below that will be seen to
be likely. However, the ‘them’ of v. 21 has no obvious antecedent in
v. 19, so once again there is a difficulty. This too might be overcome
if v. 19 were also a later insertion, as a significant group of scholars
now suppose. We will return to this issue later too. For the moment
it is sufficient to note that there is a rather similar situation in 14.19-
20. Here the pillar of cloud and fire is mentioned again, moving
from in front of the Israelites to behind them, but in v. 19a ‘the angel
of God’ (Heb. malʾak ʾelōhîm) is said to do exactly the same. Such an
angel appears as an alternative mode of guidance for the Israelites
in 23.20 and 33.2. The duplication in 14.19-20 is remarkable and
probably due to the combination of different accounts, one of which
uses ‘God’, ʾelōhîm, while the other speaks of the pillar of cloud and
fire, exactly as here. Indeed the second account seems to make a
specific reference back (‘from before them’) to v. 21 here, implying
that it comes from the same account. With some confidence, there-
fore, we might conclude that vv. 17-19 are from the same parallel
account which uses ‘God’ in 14.19a. The objections to deriving
these verses from a continuous source made by Rudolph and Coats,
for example, are not strong. The opening words of v. 17 in fact
200 EXODUS 1–18
make an explicit link to the earlier narrative in 12.31-32, which we

have argued is, like the main non-Priestly account of the plagues,
from E (see the introduction to 12.28-42, 50-51).6 Verses 17-19 also
connect well to the following episode of the deliverance at the sea in
ch. 14. It is true that the expression Yam Suf in v. 18 does not occur
again in the narrative until 15.22, but after the identity of the sea
concerned had been stated once ‘the sea’ would be quite sufficient
to refer back to it (in the non-Priestly narrative [see the introduction
to 14.1-31] in 14.9aα, 21aβ, 30).
In v. 19 there is also a connection (and an explicit one) to the
earlier non-Priestly narrative, in this case to Gen. 50.25, where
Joseph makes his brothers swear to bring his bones with them when
they leave Egypt. The connection is confirmed by the use of Heb.
ʾelōhîm, ‘God’, there too as well as here: and consequently in the
older source-analysis Gen. 50.25 was also generally attributed to
E. Since the 1970s (with isolated antecedents in Gressmann and
Rudolph), as we have seen, this very connection between Gen.
50.25 and Exod. 13.19 (as well as Gen. 33.19 and Josh. 24.32) has
come to be regarded as an argument for assigning both verses to
a late redactional layer which presupposes the join between the
Pentateuch (or Tetrateuch) and the Deuteronomistic History. The
fact that these verses also link the patriarchal traditions and the
Exodus traditions is naturally a further argument for a late date in
the eyes of those who regard the combination of these traditions as
exilic or later. In themselves the closing verses of Genesis (50.22-
26) are also now often regarded as a string of late additions to the
main narratives in this part of Genesis (the Joseph-story and P),
though not necessarily for compelling reasons.7
It is clear that Josh. 24.32 is dependent upon Gen. 33.19 as well
as Gen. 50.25 and/or Exod. 13.19 and probably Gen. 48.22. One
possible interpretation of this is certainly that all these verses were
added as a single process of ‘Hexateuchal’ redaction. Another
possibility, which should not be dismissed too quickly, is that all
6
This is another place (cf. also 12.21-27) where our view actually makes for
a more straightforward analysis of the Exodus-story than the usual older view that
the main narrative was from J.
7
See e.g. C. Westermann, Genesis 12–50 (BKAT; Neukirchen, 1982), pp. 234-
42; Gertz, Tradition, pp. 358-65. Against such views see briefly my remarks in
‘Transition’, pp. 73-75.
13.17-22 201
these verses belonged to a continuous old narrative extending

from Genesis to Joshua. Since the work of Martin Noth the idea
of ‘Pentateuchal’ sources continuing into Joshua has lost its former
popularity. But there are some signs of a renewal of interest in it,
and with good reason.8 Along with the itinerary-link between Num.
25.1 and Josh. 2.1; 3.1, Josh. 24.32 has such strong connections
with older ‘Pentateuchal’ tradition that it could well have formed
the conclusion to a narrative account which included the conquest
of Canaan. Graupner, who attributes v. 19 to E, has a different
approach to the issue, which also deserves further consideration. In
the first place, he objects to a redactional origin because in (what
are to him) clearly redactional passages the divine name is used,
not ʾelōhîm, ‘God’.9 Secondly, he observes, following Noth and
Fritz, that the links in the chain are not all strong. Genesis 50.25
and Exod. 13.19 show no clear sign of a reference to Gen. 33.19 and
Josh. 24.32 and it could very well be the case that Josh. 24.32 is an
isolated and late combination of data found in the older narratives in
the Pentateuch.10 In that case the sequence of verses in Genesis and
Exodus need be neither ‘Hexateuchal’ nor redactional.
As for v. 20, its links to 12.37a require that it be understood as
part of the same itinerary-chain, which we have argued is most
likely to derive from a Deuteronomistic redactor (see the introduc-
tion to 12.29-42, 50-51). There is no more reason here than before
to consider a Priestly origin for it. It has no narrative attached to
it and intrudes into the more general statements about the Israel-
ites’ journey in vv. 17-19 and 21-22. Verses 21-22 are somewhat
verbose, but there is no justification for regarding any part of
them as secondary. If (as seems likely) they are not the original
continuation of vv. 17-18/19, they have no direct connection to the
immediate context and serve mainly to prepare for what is said in
14.19b-20, 24. If we are right that 12.33-34 (37b-38?), 39 are from
J, then 13.21-22 may have followed them (or rather 13.3-4: see
8
Cf. Kratz, Komposition, pp. 286-304.
9
Graupner, Elohist, p. 71, with n. 271 (cf. p. 74 n. 242). This observation does
seem to be generally true, but Exod. 20.1 (which Graupner and others regard as a
secondary addition to the Sinai narrative: p. 126) looks like an exception, which
is a problem, unless the reasons for discounting the E origin of Exod. 20.1-17 are
weaker than Graupner thinks.
10
Graupner, Elohist, p. 75 n. 249; so now Schmidt, p. 578.
202 EXODUS 1–18
the introduction to 13.1-16), either immediately or closely, in the

original sequence. It is understandable that, with their concern with
the continuing journey rather than the departure itself, they should
have been placed after vv. 17-19 in the combined non-Priestly
narrative.
It is appropriate that as the Israelites’ journey gets under way
the theme of divine guidance should be prominent in both the older
narrative accounts, though with what is perhaps a characteristic
difference of emphasis. In vv. 17-19 it is the providential, unseen
direction of events by God that is prominent, an insight that could
be generated by nothing more than informed reflection on the tradi-
tion that Israel’s journey into the wilderness was to have as its
preliminary goal ‘the mountain of God’ (cf. 3.1, 12) in (probably the
southern part of) the Sinai peninsula. They therefore did not take the
coast road and even a slight familiarity with Israel’s early history
would indicate that this would avoid the need to pass through ‘the
land of the Philistines’, where Israel’s traditional enemies would
have been a serious obstacle to progress and a cause of discour-
agement and even rebellion against God. A deeper reflection on
God’s care for his people would see the detour as part of his plan
to spare them from such problems. In vv. 21-22 the guidance is
direct and visible, and also unpredictable: it derives from the actual
presence of Yahweh himself at the head of his people, just he had
earlier ‘come down’ to deliver them from Egypt (3.8), and as they
knew him in later times to be present in their midst, especially in
the Jerusalem temple (e.g. Ps. 46.6). These different theological
perspectives are combined together here without difficulty, just
as they are also combined with the more mundane itinerary-note
which traces another stage in the people’s progress away from the
authority of Egypt – or so it seems.
17 When Pharaoh let the people go, God did not lead thema by
the wayb to the land of the Philistines, althoughc it was near; for
God thought, ‘In cased the people change their minde when they
see war and return to Egypt’. 18 But God made the people go
roundf by (the?) wayg of the wilderness to the Yam Suf h. Armed
for battlei the Israelites went up from the land of Egypt. 19 Moses
took the bonesj of Joseph with him, for he had made the sons of
Israel swear, as follows: ‘God will indeed show concernk for you
13.17-22 203
and you shall bring my bonesj up from here with you’. 20 [The
Israelites departedl from Succoth and camped at Etham on the
edgem of the wilderness.] 21 [Yahweh was goingn before them
by day in a pillar of cloud to guide themo on the wayp, and by
night in a pillar of fireq to give them light, to travelr by day and
by night. 22 The pillar of cloud would not departs by day nor the
pillar of fire by night before the people.]

a. Heb. ‫ולא נחם‬. The Qal of ‫ נחה‬is less than half as common as the Hiphil,
which has the same meaning (cf. v. 21), but it occurs twice more in Exodus,
at 15.13 and 32.34. Both forms are more frequent in poetry (esp. Psalms) than
in prose.
b. Heb. ‫דרך ארץ פלשׁתים‬. ‫ דרך‬can be used as a preposition to mean ‘towards’
(cf. 1 Kgs 8.44, 48), but the pronoun ‫ הוא‬in the next clause confirms that it
is here a noun, which is used adverbially, as often with the names of routes.
c. Heb. ‫כי‬. ‘Although’ is the only sense of ‫ כי‬that fits here (cf. Childs,
p. 217; Houtman, p. 249): for a fuller list of exx. than BDB’s otherwise purely
poetic occurrences (p. 473) see DCH 4, p. 387 (but AHI 1.004.4 is probably
not an example). Genesis 8.21 is surely another case. A. Aejmaleus sought to
discount this meaning of ‫ כי‬altogether (‘Function and Interpretation of ‫ כי‬in
Biblical Hebrew’, JBL 105 [1986], pp. 193-209 [205-207]), but her view is
only partly supported in JM §171a-b.
d. Heb. ‫ פן‬usually follows a main clause, but this is one of several passages
where it occurs idiomatically at the beginning of an utterance to indicate that
what has been done is a precaution against some undesired outcome (BDB,
p. 814): cf. Gen. 26.9; 31.31; 38.11; 42.4. In all these cases, as here, the clause
is introduced by a form of ‫( אמר‬it is omitted in Gen. 26.7), which can best be
translated ‘thought’ (BDB, p. 56).
e. Heb. ‫ינחם‬. The sense ‘repent’ which has become traditional for ‫ נחם‬Niphal
is only rarely applicable, only in Jer. 8.7 and perhaps Job 42.6 (cf. TWAT 5,
368-78 = TDOT 9, pp. 342-50). In many cases an ‘affective dissociation’
from a previous attitude, leading to a change of purpose, is involved, hence
the senses ‘be sorry, regret’ and ‘be consoled’. ‘Change of mind’ alone is most
often found in theological contexts, whether affirmed or (in 1 Sam. 15.29a,
like the Hithpael in Num. 23.19) denied in relation to God. 1 Samuel 15.29b
regards it as a weakness of human beings and that comes closest to the usage
here. The paronomasia with ‫ נחם‬earlier in the verse is perhaps deliberate (cf.
Cassuto, p. 156; Houtman, p. 250).
f. Heb. ‫ויסב‬. The Hiphil of ‫ סבב‬has causative meanings related to the senses
‘turn, rotate’, ‘go around (in an area)’, ‘go round (a city or an area)’ and
204 EXODUS 1–18
‘surround’. Only the second and third of these come into consideration here,
and the addition of ‫ דרך המדבר ים־סוף‬makes the third most probable: 2 Kgs
3.9 is another example where the area is not explicitly defined, although the
context indicates the route and destination. The ‘going round’ may involve
a complete circuit (Josh. 6.11: cf. 6.3 etc.; Ps. 48.13), but not necessarily: a
part of a circuit (Gen. 2.11, 13; 2 Sam. 5.23 par.; Ezek. 47.2; 2 Chr. 13.13) or
the avoidance of the most direct route (Num. 21.4; Deut. 2.1, 3; Judg. 11.18;
2 Kgs 3.9) may be meant. The latter is evidently intended here.
g. ‫ דרך‬could, as in v. 17, mean ‘way, road’ here, but ‘the way of (to?
through?) the wilderness’ is too vague to be a likely name of a route, so a
prepositional use, either ‘through’ (other, at least possible, cases are 2 Sam.
4.7; 18.23; Jer. 39.4a) or ‘towards’, is probable.
h. ‫ ים־סוף‬is an adverbial ‘accusative’ of direction (GK §118f) indicating the
next important destination. It can scarcely modify ‫ המדבר‬in view of its def.
art.: alleged exceptions (cf. GK §127f) are dubious. On the sense of ‫ ים־סוף‬see
the Explanatory Notes on 10.19, 13.17-18 and 15.4.
i. Heb. ‫חמשׁים‬. A military sense for this word (with its three other occur-
rences in MT: Josh. 1.14; 4.12; Judg. 7.11) has long been recognised (see
Text and Versions; cf. Luther ‘gerüstet’, Tyndale and AV ‘harnessed’, RV
‘armed’ [avoiding what is now an archaic sense of ‘harness’]). BHS recog-
nises two further instances by emendation in Num. 32.17 (cf. LXX, Vulg)
and Neh. 4.11 (cf. LXX). The use of ‫ חלוצים‬in Num. 32.20 (cf. also v. 17)
reinforces the contextual support for this meaning. Ar. ḫamīsun, ‘army’, and
ESA ḫms, ‘bearers of arms’, show a similar meaning in related languages;
according to some also ‫ חמשם‬in Moabite (KAI 181.28: see DNWSI, p. 386),
but most tr. ‘fifty’ there. On the basis of Ar. some see the form as derived
from a denominative of ‫חמשׁ‬, ‘five’: cf. BDB, p. 332, ‘in battle array’; HAL,
pp. 317-18; Ges18, pp. 368-69, where evidence of such denominatives in
Ug., Ar. and Eth. as well as Heb. is cited. The sense might then be ‘in five
divisions’, ‘in ranks of five’ or ‘in groups of fifty’ (DCH 3, p. 259), but the Vss
give no support to such refinements and for Heb. at least the sense ‘armed’
or ‘in marching order’ is preferable. Propp (p. 488) favours ‘resolute’, but
the argument is weak. See further Ska, Le passage, pp. 14-17. Syntactically
‫ חמשׁים‬is a ‘predicative accusative of state’ (JM §126a-b: cf. GK §118-p),
placed before the verb for emphasis.
j. Heb. ‫את־עצמות‬, perhaps better ‘body’ (cf. BDB, p. 782), as Joseph had
been embalmed (Gen. 50.26).
k. Heb. ‫פקד יפקד‬, as in Gen. 50.25 (cf. v. 24): for the meaning see Note b
on the translation of 3.16-22.
l. Heb. ‫ויסעו‬. In the full itinerary formula here ‫ נסע‬has its standard meaning
‘departed’ (contrast 12.37).
m. Heb. ‫קצה‬, ‘end’, is often used of the ‘edge, extremity’, for example, of
a region: Num. 20.16 and 22.36 show, respectively, how ‘on the edge’ may
mean ‘just outside’ or ‘just within’ the region (BDB, p. 892). Here the former
is perhaps more likely.
13.17-22 205
n. Heb. ‫ויהוה הלך‬. The word-order is regular for a nominal subject followed
by a participle (JM §154fc). The clause is probably independent rather than
circumstantial in the narrow sense (against Childs, p. 234): its length and the
close connection with v. 22 would suggest this, and JM §159f provides some
parallels. Of course the wider context determines the tense as past durative
(examples in JM §121f and Joosten, Verbal System, p. 247).
o. Heb. ‫לנחתם‬,ַ understood in the MT vocalisation as the Hiphil inf. cons.
with the prefix unusually elided (cf. GK §53q). In this case (but not in all) the
Qal inf. could have been read without change to the consonantal text or the
sense, and the Qal appears in v. 17 (see Note a). Possibly the Masoretes chose
to read the Hiphil here because it was the more common form (and cf. Neh.
9.19: Propp, p. 465); alternatively they may, like Rashi, have understood it in
a causative sense so as to identify the pillar of cloud as the intermediary (first
object) used by God to guide the Israelites.
p. Heb. ‫ הדרך‬cannot be a direct obj. of ‫( נחה‬Qal or Hiphil) and must be
understood adverbially, as is more common with names of routes (cf. v. 17
and Note b).
q. Heb. ‫ולילה בעמוד אשׁ‬. The participle ‫ הלך‬is understood with this clause
too.
r. Heb. ‫ללכת‬. Presumably this is not parallel to the two preceding infinitives
but gives the reason for the dual provision of cloud and fire.
s. Heb. ‫ימישׁ‬. The Hiphil of ‫ מושׁ‬can be intransitive (cf. 33.11) like the Qal.
The imperfect is clearly iterative as, e.g., in 13.15.
Explanatory Notes
17-18. The narrative resumes with the first of two observations
about God’s leading or guidance of ‘the people’: a second, with a
different focus, follows in vv. 21-22. Here, as again in 14.1-2, where
God (Yahweh) is said to give Moses instructions for the people about
their next camping-place, the direction in which God led his people
is at the centre. If Canaan was the eventual destination, there were
several possible routes to there from eastern Egypt. The shortest
and most favoured way was the coastal road, which is described
here as ‘the way to the land of the Philistines’, since as both biblical
evidence (e.g. Josh. 13.2-3) and archaeology (e.g. T. Dothan, The
Philistines and their Material Culture [New Haven and London,
1982]) indicate the Philistines with their five famous cities occupied
the coastal plain in the south-west of Palestine.11 In biblical narrative,
11
Recent discoveries and studies of Hieroglyphic Luwian inscriptions from
Aleppo and nearby have indicated the existence of a ‘land of Palistin/Walistin’ in
206 EXODUS 1–18
especially in the books of Samuel, they are often at war with Israel
and Judah. So the prospect that encounter with them might deter
the escaping Israelites from proceeding further would have been a
plausible reason, in the minds of the passage’s original audience,
for God to have chosen a different route out of Egypt for his people.
In fact such considerations would have been anachronistic, because
at the likely time of the Exodus events, in the thirteenth century
B.C., the Philistines were not yet settled in the region: their arrival
occurred, as Egyptian evidence shows (especially Papyrus Harris I
and inscriptions and reliefs from Medinet Habu: Dothan, pp. 1-13;
ANET, pp. 262-63), early in the twelfth century. But already before
this the coast road, known as ‘the Way(s) of Horus’ had often been
used by Egyptian armies and was heavily fortified (see the classic
study by A.H. Gardiner, ‘The Ancient Military Road between
Egypt and Palestine’, JEA 6 [1920], pp. 99-116, supplemented from
more recent exploration in Hoffmeier, Israel in Egypt, pp. 164-75,
182-98, and further in Ancient Israel in Sinai, Chapters 4 and 5).
Still there must be some doubt whether an Israelite writer, even of
the monarchy period, would have been as well informed about the
presence of Egyptian forces along this route as modern archaeolo-
gists are now.12
The route which the Israelites were thought to have taken is
described here in terms that are either vague in the extreme (‘the
wilderness’) or highly contested in their interpretation (‘Yam Suf’).
It is at least clear that it was seen as a detour (see Note f on the
translation), in contrast to the way that was ‘near’, which suggests a
route either east or south-east into the Sinai peninsula. It is curious
that nothing is said here about the need to visit Mount Horeb/Sinai
(cf. 3.1, 12) or ‘the mountain of God’ (cf. 18.5), but perhaps Yam
this region of northern Syria in the eleventh and tenth centuries: for a convenient
summary with references see T.P. Harrison, ‘Recent Discoveries at Tayinat
(Ancient Kunulua/Calno) and their Biblical Implications’, in C.M. Maier (ed.),
Congress Volume: Munich 2013 (VTSup 163; Leiden, 2014), pp. 396-425 (402-
405), where mention is also made of occurrences of ‘the land of the Peleset’ in
inscriptions of Rameses III (p. 405). On the latter see D. Kahn, ‘The Campaign
of Rameses III against Philistia’, JAEI (online) 3:4 (2011), pp. 1-11 (to which
Harrison refers), where this ‘land’ is identified with the region in Syria, not with
south-west Canaan (as most scholars have thought hitherto).
12
There is evidence for contact with the Delta region, presumably along the
coast road, in Isa. 30.1-5; 31.1-3.
13.17-22 207
Suf is mentioned as the next significant place on the journey. In

10.19 (see the Explanatory Note) it is where the swarms of locusts
are driven from the Nile Delta by a ‘west’ (lit. ‘sea’) wind: either
the Gulf of Suez or one of the lakes east of the Delta is probably
meant there, and the same possibilities exist here. Recently renewed
credence has been given to the suggestion that in antiquity the Gulf
of Suez may have extended further north than today, perhaps as
far as the Bitter Lakes and Lake Timsah (Van Seters, ‘Geography’,
pp. 258, 272-75; Albertz, pp. 27-29, 184; cf. earlier my Way of the
Wilderness, pp. 73-74). More detailed discussion must await the
Explanatory Notes on 14.1-4 and 15.4.
A different kind of question is raised by the use of the title ‘God’
(Heb. ʾelōhîm) in place of the divine name Yahweh three times in
these verses (and again in v. 19, where it is related to Gen. 50.24-25:
see further below on this). Apart from what may be appellative uses
(‘a god’) in 8.15 and 9.28, this is the first time that ʾelōhîm has been
used in this way in Exodus since a series of occurrences in chs. 1–6
(1.17, 20, 21; 2.23, 24, 25; 3.1, 4, 6, 11, 12, 13, 14, 15; 4.20, 27; 6.2).
There it has traditionally been seen as a marker of either the Priestly
source (in 2.23-25; 6.2) or the Elohist source (the other passages)
of the Pentateuch (so still, in essence, Graupner, Der Elohist). More
occurrences of this feature appear in 14.19; 17.9; 18.1, 5, 11, 12, 15,
19, 21, 23; 19.3, 17, 19; 20.1, 19, 20, 21; 24.11, 13 and occasionally
in the Book of the Covenant and later. The contrast in this regard
with 13.21 and 14.1 is particularly notable and is discussed further
in the introduction to this section.
The statement that the Israelites left Egypt ‘armed for battle’
(Heb. ḥamušîm: the other occurrences and most of the renderings in
the ancient Versions firmly support this interpretation [see Note i on
the translation]) is at first sight surprising, especially so soon after
God has been said to wish to prevent his people from facing a battle.
But a battle is described in 17.8-13 and more follow in Numbers
(14.40-45; 21.1-3, 21-35; 31. 1-12) and of course in Joshua. The
formal itinerary notes scattered through the narrative (e.g. 12.37;
13.20) belong to a genre which was particularly favoured in descrip-
tions of military campaigns (see the Excursus in the introduction
to 12.28-42, 50-51). The same conception has even influenced the
Priestly tradition, both in the expression ‘in their tribal divisions’
(Exod. 6.26; 7.4; 12.17, 41, 51) and in the organisation of the camp
and the order of march described in Num. 2.1-34; 10.14-28.
208 EXODUS 1–18
19. The plural subjects (vv. 18b, 20; cf. ‘the people’ in vv.
17-18a) are interrupted by a statement about Moses in the singular
which adds a further detail about the Israelites’ departure: he takes
Joseph’s ‘bones’ (ʿaṣmôt: the broader sense ‘body’ is perhaps more
appropriate, as in Note j on the translation) with him. Joseph, who
is clearly the ‘he’ of the explanatory clause (some versions of the
text add his name), had made ‘the sons of Israel’ swear that they
would do this (Gen. 50.25): the phrase is conveniently ambigu-
ous, as it could mean Joseph’s brothers (cf. Gen. 50.24), who were
like him literally sons of Israel (Jacob’s other name), or the whole
family, which had already begun to increase in numbers before
Joseph died (cf. Gen. 50.23). This important motif of continuity
reaches its goal in Josh. 24.32, where the Israelites (Moses having
already died) bury Joseph’s ‘bones’ at Shechem in land which his
father Jacob was said to have bought long before from the Canaan-
ite inhabitants of the place (cf. Gen. 33.19; also 48.21-22, where
the means of acquiring the land is different). There is thus a genu-
inely ‘Hexateuchal’ character to this series of passages, whether (as
long supposed) as part of the Elohist source-document or (as has
recently become a widely held view) as a later redactional layer
added to the older narrative (on the debate see the introduction
to this section). The verse also repeats Joseph’s assurance to his
family that ‘God’ would ‘show concern’ for them and bring them
back to Canaan from the land of Egypt where they had first come in
a time of famine (Gen. 50.24-25), an assurance which Moses had
also given, in the same words, as part of his original message from
Yahweh in Exod. 3.16. As the Israelites leave Egypt, the providen-
tial care of their God for them over the generations is once more
underlined.
20. Now another stage in their itinerary (cf. 12.37 and the notes
there) takes the Israelites on to ‘the edge of the wilderness’, in par-
allel to the larger-scale description of their route in v. 18. Again
there is a close equivalent in the continuous itinerary in Num. 33.6.
On Succoth see the note on 12.37: Etham has not been convincingly
identified with a toponym known from ancient Egyptian (or other
texts). Geographically the Egyptian word ḫtm, ‘fort’, is an attractive
possibility, as it not surprisingly occurs in a number of references
to the eastern border area, including the famous ‘Report of a Fron-
tier Official’ in Papyrus Anastasi VI and another model letter (in
13.17-22 209
Papyrus Anastasi V) reporting the escape of two runaway slaves:

in both it occurs with the likely equivalent of Succoth.13 But the
correspondence of Heb. aleph (as in Etham) and Egyptian ḫ would
be very irregular. A more recent suggestion is that Etham represents
the name of the god Atum and is a (probably late) shortened from
of the name Pithom (on which see the Explanatory Note on 1.11),
without the Egyptian prefix ‘house/temple of’ (so M. Görg, ‘Etham
und Pitom’, BN 51 [1990], pp. 9-10; conveniently summarised in
ABD 2, p. 644). But neither of the places commonly identified with
Pithom (see again the note on 1.11) is likely to represent a camping-
place beyond Succoth ‘on the edge of the wilderness’. Even though
a convincing identification remains elusive, the description of this
part of the route as a whole and especially the words ‘on the edge of
the wilderness’ strongly suggest that the itinerary envisages a place
at or beyond the eastern end of Wadi Tumilat in the vicinity of Lake
Timsah.14
21-22. The closing verses of the section return to the subject of
divine guidance of the Israelites, here throughout their journey (cf.
v. 22) and with the emphasis now on the way in which this divine
guidance was made known. Later, after the departure from Mount
Sinai, this ‘pillar of cloud and fire’ is associated respectively with the
Priestly desert shrine or ‘tabernacle’ (40.36-38; Num. 10.11-12) or
with the ark of the covenant (Num. 10.33-36; cf. 14.44; Josh. 3.2-6,
14-17). At this stage only the pillar of cloud and fire is mentioned,
with Yahweh ‘in’ it (cf. 14.24) at the front of the people as they
travel, as their ‘leader’ in a very precise sense. The separation of
these verses from vv. 17-18 and, probably, the use of the name of
God itself rather than the title ‘God’ suggest that they derive origi-
nally from a different strand in the composition in the narrative, one
which reappears in the account of the sea-crossing in 14.19b-20,
24 as well as in Num. 14.14, Deut. 1.33 and the summaries of the
Exodus story in Pss. 78.14; 105.39; Neh. 9.12, 19. A related but
distinct tradition associated the pillar of cloud (but not fire) with
‘the tent of meeting’ in non-Priestly passages (33.9-10; Num. 12.5;
13
For further details and discussion of this possibility see Cazelles, ‘Les
localisations de l’Exode’, pp. 358-60; Hoffmeier, Israel, p. 182.
14
Cf. Hoffmeier, Israel, p. 182.
210 EXODUS 1–18
Deut. 31.15: cf. Ps. 99.7).15 The Priestly version of the ‘guidance’
theme (cf. above) evidently combined the two traditions and applied
them to its much larger tent-shrine. The specific concept of a ‘pillar’
of cloud or fire draws on the more widely attested ideas of divine
appearance or presence in cloud and fire (cf. 3.2-6; 16.10; 19.9,
16-18; 24.15-18: see the fuller discussion in T.W. Mann, ‘The Pillar
of Cloud in the Reed Sea Narrative’, JBL 90 [1971], pp. 15-30,
and Houtman 2, pp. 254-56). A variety of suggestions have been
made about the origins of this motif (ibid.): most likely it arose from
the combination of common practices of travel in desert areas with
the theological ideas already mentioned, which find close paral-
lels in portrayals of the storm-god in particular (cf. Clements, God
and Temple, p. 22; J. Day, Yahweh and the Gods and Goddesses of
Canaan [JSOTSup 265; Sheffield, 2000], pp. 91-98).
The final words of v. 21 imply that the Israelites travelled by
night as well as by day, which goes beyond the statement that the
pillar of fire was ‘to give them light by night’. The same implication
is evident in Deut. 1.33, Ps. 78.14 and Neh. 9.12, 19, but there is no
indication in the main wilderness narrative that this was generally
the case, only in the immediately following instance of the episode
at the Red Sea, where the pillar of cloud is in fact behind the Israel-
ites rather than in front of them and Yahweh protects and fights for
them rather than showing them the way (14.19b-20, 24). Of course
travelling by night in a desert region would have avoided the heat
of the day (cf. perhaps 1 Kgs 19.8) and the narrator may have had
such a practice in mind here.
Text and Versions

‫( ויהי‬13.17) As often LXX, Sy and Vulg have no equivalent, for stylistic
reasons. Vulg begins the verse with igitur, scarcely to make a connection with
what has just preceded, but resumptively (cf. OLD, p. 822, s.v. 5) to pick up
the earlier narrative in 12.31-32 which was interrupted by the legal texts in
12.43–13.16: ‘So, when…’
‫( אלהים‬13.17) Both here and later in the verse SP, LXX and Sy agree
with MT, but the Tgg all use forms of the divine name (TgNmg even has ‘the
Memra of…’) as they regularly do throughout chs. 1–3 where the absolute
form occurs. Vulg follows suit here, though not always elsewhere (e.g. v. 19).
15
Blum, Studien, pp. 140-41, tentatively suggests that this may be a ‘transfor-
mation’ of the ‘pillar of cloud’ tradition here.
13.17-22 211
‫( דרך‬13.17) The Vss all render with a noun (on the problem see Note b
on the translation).
‫( כי‬13.17)1o Equally the Vss all have words for ‘because’, except for
Vulg which renders freely with the rel. pron. quae and so avoids the problem
discussed in Note c on the translation.
‫( קרוב הוא‬13.17) TgF(P),G(X)16 add here their version of the long haggadic
expansion which TgJ (cf. MRI) has after ‫( בראותם מלחמה‬see the note there),
where it fits better. Placed here the explanation follows the sequence of events
more closely but its relevance is obscured, the more so as TgF(P),G provide no
lead-in to the narrative of an earlier war (see below).
‫( כי אמר אלהים‬13.17) TgNmg makes the reference to God’s thoughts
explicit by adding ‫במחשבת דליה‬.
‫( ינחם‬13.17) LXX, Vulg and TgJ preserve the sense of MT: TgO (which
is followed exactly here, as throughout the translation of the verse itself, by
TgF(P),G) and Sy in different ways give the contextually apt sense ‘be afraid,
tremble’, while TgN has ‘their heart be broken’, presumably in disappointment
or grief (for the form cf. Jastrow, p. 1645).
‫( בראותם מלחמה‬13.17) TgJ inserts here an interpretation which understands
the war to be not a Philistine attack on the main Exodus group but an earlier
event which affected only a large group of Ephraimites. This legend is found
already in MRI (Lauterbach, pp. 170, 172-73: after an alternative explanation
based on Num. 14.45), where its origin in a combination of 1 Chr. 7.21b-23
and Ps. 78.9-10 is clearer. It is from the latter passage that the idea of a trans-
gression of a divine command is taken, but it is made to refer to a refusal to
wait for the time of the Exodus prescribed by God (cf. TgJ on Gen. 50.25). In
TgJ a connection is also made with Ezekiel 37, and this was developed further
in the generally very similar forms of the legend which appear in TgG (Klein
1, pp. 222-23; cf. 2, pp. 61-62) and TgF(P) (see more fully J. Heinemann, ‘The
Messiah of Ephraim and the Premature Exodus of the Tribe of Ephraim’, HTR
8 [1975], pp. 1-15 [esp. 10-15]; M.J. Mulder, ‘1 Chronik 7.21B-23 und die
rabbinische Tradition’, JSJ 6 [1975], pp. 141-66 [esp. 149-54; on pp. 160-63
Mulder even supposes that the origin of the legend is not exegetical but based
on old oral tradition]).
‫( ושׁבו מצרימה‬13.17) TgJ inserts ‘and when they see this they fear’ between
the end of its addition and these words of MT, recalling and virtually repeating
its translation of the previous clause of MT. TgF(P),G have the same insertion
here, but because they place the major addition earlier in the verse (cf. above)
it immediately follows the words that it was designed to recall, which is
another sign that the placement of the addition in TgF(P),G is not original.
‫( ויסב‬13.18) TgF and Sy wdbr lacks the element of ‘around’ and makes the
contrast with ‫ ולא נחם‬in v. 17 more direct.
16
Klein 1, p. xxxvii, accepts a date for ms. X in the ‘Middle Period’
(mid-eleventh to late fourteenth cent.).
212 EXODUS 1–18
‫( אלהים‬13.18) TgNmg again has ‘the Memra of the Lord’, as does TgF(VN)
here.
‫( המדבר‬13.18) So also 4QExc, but SP reads ‫מדבר‬, i.e. constr. st. to be taken
with the following ‫ים־סוף‬: ‘the wilderness of the Yam Suf’. Sy and TgJ have
dymʾ, implying the same understanding and possibly the same text as SP; the
other Vss in various ways avoid such a connection and presumably reflect
MT’s wording. The choice between the two readings is difficult: both make
sense and neither is clearly derivative from the other. ‘The wilderness of the
Yam Suf’ occurs nowhere else, but such naming of desert areas is widespread
(e.g. 15.22; 16.1). Some weight might be given to the fact that 14.3 speaks of
‘the wilderness’, but it is scarcely decisive. It is, however, possible to argue
that ‫ ים־סוף‬was misunderstood because it lacks an explicit indication of direc-
tion (contrast 10.19) and this led to its being taken as a defining genitive,
hence the change to ‫ ;מדבר‬and the combined evidence of MT, LXX and
4QExc, along with most of the Tgg, would also favour, on balance, seeing
‫ המדבר‬as the original reading.
‫( ים־סוף‬13.18) As elsewhere (cf. 10.19) LXX and Vulg equate with ‘the
Red Sea’, while the other Vss simply reproduce the Heb. expression.
‫( חמשׁים‬13.18) Von Gall and Sadaqa give the same reading for SP, but most
mss (including the older ones in von Gall, Crown, Tal and Camb. 1846) have
‫( חמישׁים‬likewise the older SamTg text [J]) and this should be adopted as the
SP text. What it was taken to mean remains uncertain, though the later SamTg
text (A) has ‘armed’. This is also the interpretation of TgO, Aq, Symm, Sy
and Vulg here (cf. MRI [Lauterbach, pp. 174-75]), and the ‘armed with good
work(s)’ of TgN,F is a moralising elaboration of it. LXX has military interpre-
tations in Josh. 1.14 and 4.12 (as well as Num. 32.17 and Neh. 4.11, where
‫ חמשׁים‬may well be the original reading), but its πέμπτῃ δὲ γενεᾷ here relates
the word to the numeral ‘5’, as do Theod’s πεμπταίζοντες (apparently ‘on
the fifth day’) and TgJ’s ‘each with five children’. The reason for LXX’s
decision here is presumably the ‘four generations’ of Gen. 15.16, which also
affected LXX’s (and SP’s) interpretation of 12.40: perhaps, therefore, the SP
variant was originally also intended to bring out a connection with ‘5’ and the
correct interpretation ‘armed’ was only adopted later. On the opposed views
in early Jewish interpretation see S.E. Loewenstamm, The Evolution of the
Exodus Tradition (Jerusalem, 1992), pp. 228-32.
‫( בני ישׂראל‬13.18) TgN,F add ‫פריקין‬, as the Pal.Tg. often does after ‫יצא‬
when it refers to the Exodus. Sy has dbyt ʾysr(ʾ)yl as it does occasionally for
‫( בני ישׂראל‬again in v. 19).
‫( מארץ מצרים‬13.18) Propp reads ‫( ממצרים‬p. 464), but with very little
support.
‫( ויקח‬13.19) TgO,J,F have ‫ואסיק‬, ‘and…brought up’, making Moses’ action
correspond exactly to what is prescribed later in the verse. The other witnesses
agree with MT: Vulg’s quoque for waw is a frequent stylistic feature and
rarely represents the presence of ‫גם‬.
13.17-22 213
‫( את עצמות יוסף‬13.19) TgJ has ‘the coffin within which were Joseph’s
bones from the Nile, and he was taking (it) with him’, so reading more into
‘brought up’ and alluding to a legend that Joseph had been buried in the Nile
(TgJ on Gen. 50.26). MRI (Lauterbach, pp. 176-77) mentions this tradition
along with a different one which is also found in T.Sim. 8.3-4.
‫( השׁבע השׁביע‬13.19) SP, LXX and TgF(P) add ‘Joseph’, the latter with ‘in
his lifetime’. There is a lacuna at this point in 4QExc but probably no room for
the addition (DJD XII, p. 116). Its purpose will have been to make absolutely
clear that there has been a change of subject.
‫( פקד יפקד‬13.19) Tgg and Sy have ‘will surely remember’, with the addi-
tion of ‘in his goodly compassion’ in TgN,F; LXX and Vulg have ‘will surely
visit’ which has influenced many EVV. (cf. Text and Versions on 3.16 and
4.31: also J.W. Wevers, ‘An Early Revision of the Septuagint of Numbers’, in
Harry M. Orlinsky Volume, EI 16 [1982], pp. 235*-39* [238*]).
‫( אלהים‬13.19) So also SP, Vulg and Sy (4QExc does not survive), but
this time LXX joins the Tgg in reading ‘the Lord’. This particular variation
between LXX and MT is very rare and elsewhere it only occurs where ‫יהוה‬
closely precedes or follows (3.4; 9.30; 18.1; 20.1). In this case (as Propp
[p. 465] points out) it involves a divergence from LXX’s own precise render-
ing of Gen. 50.25MT: and it is not due to the influence of v. 21 below, since
there LXX has ὁ θεός for MT’s ‫יהוה‬. Perhaps the translator (or his Vorlage) is
recalling (3.16 and) 4.31, where ‫ יהוה‬is used in connection with ‫פקד‬.
‫( והעליתם‬13.19) Most mss of Sy have, a little freely, ʾsqw (imper. without
waw), but 5b1 agrees with MT, as it tends to do.
‫( ויסעו‬13.20) LXX ἐξάραντες δέ, with the frequent participium coni-
unctum for the first of two finite verbs in the Heb. (see further Aejmaleus,
‘Participium Coniunctum as a Criterion of Translation Technique’, VT 32
[1982], pp. 385-93 = On the Trail, pp. 7-16). ἐξαίρω is used for ‘depart’ again
in 14.19 and 19.2, and often in Numbers 1–12, but this sense appears to be
post-classical (LSJ, p. 582, cites only Polybius 2.23.4 outside LXX): ἀπαίρω,
which is more common for ‫( נסע‬cf. 12.37 and repeatedly in Num. 33.1-49),
is the classical expression (LSJ, p. 175). LXX adds οἱ υἱοὶ Ἰσραήλ as the
subject because there is no direct connection with v. 19.
‫( מסכת‬13.20) The second corrector of LXXF (Fb) records συσκιασμοί/
ῶν, ‘shady places’ and σκηναί, ‘tents, booths’, as alternative renderings of
‫סכת‬: the latter will be, as in 12.37, from Symm, the former probably from Aq,
to whom it is attributed in Pss. 27.5; 31.21; 60.8; Amos 5.26. TgJ (like MRI
here) recalls the legend of the ‘clouds of glory’ which it introduced at greater
length at 12.37 (see Text and Versions there) and will repeat later (AramB 2,
p. 194 n. 62).
‫( באתם‬13.20) LXX ἐν Ὀθόμ is far from the vowels of MT, but these are
attested in the spellings of Aq, Symm, Theod and Vulg (as well as TgO). An
interpretation in LXXF (ἐν σημείῳ), which was known to Origen (cf. BAlex,
p. 161), was based on a reading with at least the first of LXX’s vowels.
214 EXODUS 1–18
‫( בקצה המדבר‬13.20) SP prefixes ‫ אשר‬to make the connection explicit (cf.

MT and SP in Num. 33.6), as do TgO,J,F,Nmg with )‫( ד(אית‬cf. Sy).17 Most of the
Vss assume that the near ‘edge’ of the desert is meant, but Aq, Symm, Theod
τὴν ἐρημοτάτην (followed by Vulg in extremis finibus solitudinis) oddly
indicate its farthest limit.
‫( ויהוה‬13.21) SP and most of the Vss agree with MT (with the addition
of ‘the Memra of’ in TgN,F), but LXX has ὁ δὲ θεός, a variation that is quite
frequent (see Text and Versions on 13.8): here it could be due to the use of
‘God’ several times in vv. 17-18. TgJ amplifies as in 3.6 (cf. 12.12) with ‘the
glory of the Shekinah of’).
‫( הלך לפניהם‬13.21) LXX and Tgg aptly paraphrase with ‘led (before)
them’.
‫( לנחתם‬13.21) 4QExc agrees with the consonants of MT, but some impor-
tant SP mss (including Camb. 1846 and those used by Sadaqa, Tal and Crown
[except for his Chester Beatty ms]) have ‫להנחתם‬, the full spelling of the Hiphil
form indicated by MT’s vowels. The renderings of TgO,J,N derive it from ‫נחה‬,
but TgF and Sy saw a form of ‫נוח‬, ‘(give) rest’, here. LXX δεῖξαι αὐτοῖς, ‘to
show them’, is a free rendering, perhaps designed to enable ‫ הדרך‬to be taken
as the object: Vulg followed this and is probably the source of ‫ למחווי להון‬in
TgNmg.
‫( הדרך‬13.21) SP omits the def. art. (cf. SamTgJ), against all the other
witnesses (inc. 4QExc ]‫)הדר[ך‬, for no obvious reason: maybe a mechanical
(but inappropriate) assimilation to vv. 17-18 is involved (cf. below on ‫ ענן‬and
‫ אשׁ‬in v. 22).
‫( ולילה‬13.21) TgF ‘and also (‫ )ואוף‬in the night’ need not imply a different
Vorlage.
‫( בעמוד אשׁ‬13.21) TgN has ‫ דעננא דאישׁתא‬‫א‬‫בעמוד‬, ‘in a pillar of fiery cloud’,
implying that there was a single pillar of cloud which became fiery in the
night, as 14.20 seems also to suggest. TgJ elaborates in a different way on the
basis of 14.19-20.
‫( להאיר להם ללכת יומם ולילה‬13.21) LXX has nothing for this: its Vorlage
may have suffered omission by homoeoarkton (‫להאיר‬‫)לא‬, as Wevers suggests
(Notes, p. 207), or alternatively the translator may have made the same
error. There is no reason to suppose that the shorter text is more original, as
M. Vervenne has proposed (‘The Question of “Deuteronomic” Elements in
Genesis to Numbers’, in A. Hilhorst et al. [eds.], Studies in Deuteronomy [FS
C.J. Labuschagne; VTSup 53; Leiden, 1994], pp. 243-68 [267]). Symm and
Theod supplied the missing text and it was taken up into the O-text. Vulg ut
dux esset itineris utroque tempore is probably a free rendering of MT, avoiding
unnecessary repetition. TgNmg cites an abbreviated version of the final phrase
which seems to continue the theme of light on the journey (see AramB 2, p. 57).
17
4QExc preserves only the first word of the verse, but Sanderson (DJD XII,
p. 116) considers that ‘spacial reconstruction favours the shorter reading’.
13.17-22 215
‫( לא ימישׁ‬13.22) The Hiphil is probably also the reading of 4QExc (cf. DJD
XII, p. 116), but SP has the Qal imperf. ‫ימושׁ‬: the same divergence from MT
occurs at 33.11. The sense is intransitive in either case (see Note s on the trans-
lation). The variant could have arisen from the similarity of yodh and waw
in early forms of the square script (in Prov. 17.13 the Qere and Kethiv vary
in the same way), but it is more likely that SP made a deliberate change (or
choice) to ensure the consistent use of the Qal throughout the Pentateuch (cf.
‫ משׁו‬in Num. 14.44) and avoid any ambiguity which the Hiphil might cause.
The Vss divide over the sense which they give to the verb here, between ‘fail,
cease’ (LXX, Vulg, TgJ,N,F: cf. Jer. 17.8) and ‘depart’ (TgO, Sy), which is more
regular in BH.18
‫( עמוד הענן‬13.22) 4QExc omits the article with ‫ענן‬, no doubt through
miscopying from the previous verse (see also the next note). Von Gall prints
‫ עמוד ענן‬as the SP reading, but it appears only in a few of his mss, all of
the fifteenth cent., and the majority (inc. Camb. 1846 and all those used by
Sadaqa, Tal and Crown) like LXX agree with MT. The error evidently took
place within the Samaritan tradition and not at its source. TgF prefixes ‫מאור‬,
‘the light of’.
‫( ועמוד האשׁ‬13.22) For the waw TgO,F have ‫וא(ו)ף לא‬, ‘and also not’. 4QExc
again omits the article and this time all SP mss do so too. As before the reason
will lie in miscopying from the previous verse (for SP cf. the note on ‫הדרך‬
in v. 21).19 LXX again agrees with MT. TgF prefixes ‫נהור‬, ‘the brightness of’.
‫( לילה‬13.22) TgN adds ‘leading and making ready and shining’ to specify
what the cloud and fire do not ‘cease’ to do.
‫( לפני העם‬13.22) LXX adds παντός, ‘all’, as it does frequently in Exod.
(e.g. 7.5; 9.23; 10.4-5, 15, 23; 11.3; 12.30; 14.4, 6, 17-18). TgO and Sy have
‘from before the people’ in line with their interpretation of ‫ ימישׁ‬as ‘depart’.
TgNmg adds ‫בית ישׂראל‬, specifying which people is meant, an addition which
(in Exodus at least) is most frequently found in TgJ and TgF, though not here.
18
In early post-biblical Heb. the sense ‘cease, be lacking’ is better attested
(DCH 5, p. 189).
19
Or do the indeterminate forms in 14.24 play a part in the variants here and
in the preceding phrase?
T H E D E L IV E R A N C E AT T H E SEA
(14.1–15.21)
C h ap t er 1 4 . 1 - 3 1
T he C ros s i n g of th e S ea a nd t he
D est r u cti on of th e E g y pt i a ns
The boundaries of this chapter are marked by intervals in MT

(which also has one after vv. 14 and 25) and SP (with further
divisions after vv. 10a, 14, 18 and 25). At Qumran the beginning
of the new chapter is marked in 4QExc, but there is no evidence
in regular biblical mss for or against a division at the end of ch.
14. All we have is the curious manuscript 4QExd, which appears
to have continued directly (after an interval) from 13.15-16 to
15.1, thus omitting 13.17–14.31 altogether, at least at this point.
While it is conceivable that the omission was accidental, there is
no obvious explanation for it (by parablepsis, for example) and it is
more probable that 4QExd only contained excerpts from the biblical
text, probably made for liturgical purposes (cf. Lange, Handbuch,
p. 61). It would then be more informative about the boundaries of
13.1(10)-16 and 15.1-18(?) than about those of 14.1-31. A division
after v. 14 is likely in 4Q365 and one possibly occurred there in
4QpalGen-Exl, which probably also (like SP) had one after v. 18.
4QpalExm had a division after v. 25. 4Q365 additionally had one
between vv. 12 and 13, before the words ‘and stood behind them’
(wyʿmwd mʾḥryhmh) at the end of v. 19 and perhaps after ‘Israel’
in v. 20. 4QpalGen-Exl, which survives for vv. 19-20, did not have
divisions at these latter points.1
1
4QExc probably did not have a division where SP has one in v. 10. According
to DJD XII, p. 118, 4QExc had divisions after vv. 14, 18, 25 and 31, but they are
reconstructed in a long section for which no text survives and the suggestion is
unduly speculative.
14.1-31 217
The divisions of the chapter which are best attested in the

manuscript tradition (before v. 1 and after vv. 14 and 25) all coincide
with the beginning of a divine speech which initiates a new stage
in the action: the encampment of the Israelites by the sea and the
Egyptians’ pursuit (vv. 1-14); the dividing of the waters and the
entry of the Israelites and the Egyptians into the sea (vv. 15-25);
and the overwhelming of the Egyptians by the returning waters (vv.
26-31).2 Each of these sections could be further divided between the
divine speeches (vv. 1-4, 15-18, 26) and what follows from them in
a combination of narrative and human speeches (vv. 5-14, 19-25,
27-31). An alternative analysis might seek to do more justice to the
theological elements of the narrative in its canonical form, which are
not limited to the ‘instruction and obedience’ pattern of the explicit
divine speeches: vv. 30-31 in particular deserve greater attention,
but so does the succession of divine actions in vv. 21, 24, 25 and
27. H.C. Schmitt therefore proposed the following sub-division of
the chapter (‘ “Priesterliches” und “prophetisches” Geschichtsver-
ständnis’, pp. 149-52):
1-4 (with 13.17-22): Exposition (divine guidance)

5-29: Corpus/main narrative
5-9 Egyptian pursuit
10-12 Fear and lament of Israel
13-18 Promise of Yahweh’s intervention
19-29 Yahweh’s intervention, first for Israel (19-23), then
against Egypt (24-29)
30-31: Conclusion (deliverance and faith)
For Schmitt, the ‘main narrative’ can then be seen to exhibit ‘a linear
narrative development’, from crisis (Not) to promise (Verheissung)
and deliverance (Rettung). Here too of course smaller units in
each section could be distinguished, as in the sub-divisions of our
Explanatory Notes. When this is done, it becomes more obvious
that the chapter has been put together by the combination of two,
and possibly three, different versions of the story.
2
The division of the chapter into two main parts (vv. 1-14 and 15-29, with
vv. 30-31 as an ‘Epilogue’) by Gertz (pp. 193-95) in effect combines the second
and third Masoretic sections into one.
218 EXODUS 1–18
We can again begin our survey of critical analysis of the text with Knobel,
who had already separated out a Grundschrift in a form that was very close to
later definitions of the Priestly account (similarly Nöldeke in 1869) and (by
1861) had distinguished (in a way that did not stand the test of time) sections
of the ‘JE-Bearbeitung’ which derived respectively from his Rechtsbuch and
his Kriegsbuch (cf. Num.-Josh., pp. 532, 548). A variety of attempts were
made over subsequent decades to identify J and E components through most
of the narrative, but with only limited agreement: only vv. 16aα (Moses’
staff) and v. 19a (the angel of God) were regularly assigned to E. Holzinger
acknowledged that in vv. 5b-7 and 20 the assignment of the material between
the sources J and E was impossible (p. 44). Wellhausen had even doubted
whether P was preserved in the second half of the chapter: he attributed most
of vv. 15-18, 21-23 and 26-29 to E (pp. 76-77), because he considered the
usual characteristics of P to be missing (p. 75). The same argument was taken
further by Smend, who denied the presence of P in the chapter altogether and
so (!) found room for the contributions of three old narrative sources here, as
well as a late theological redaction in vv. 4, 9 and 17-18 (Erzählung, pp. 138,
143). Fohrer was to follow his analysis closely (pp. 97-110), but Eissfeldt
suspected that the redactional layer might in fact be elements of a P narrative
(Hexateuch, p. 37) and Beer took this view for vv. 4aβb and 17-18 (pp. 73-74).
The great majority of scholars, here as elsewhere, have found quite sufficient
evidence to identify a complete Priestly strand in the chapter which does have
significant similarities to Priestly texts elsewhere (see the Explanatory Notes
for the main examples).
Within the non-Priestly text an early casualty was v. 31. Carpenter/
Harford-Battersby attributed it to RJE, especially because of the unusual
designation of Moses as Yahweh’s ‘servant’, and Baentsch (p. 127) noted the
parallels to this in redactional parts of Joshua (1.7, 13, 15 etc.) alongside Num.
12.7-8E. He also considered the phrase ‘the great act of power’ an echo of
Deut. 34.12 and the repetition of ‘and Israel saw’ suspicious. The verse might
therefore, at least in its present form, be from a Deuteronomistic redactor.
Gressmann (Mose, p. 108 n. 1; Anfänge, p. 55 n. 2) and McNeile were more
confident about a redactional origin, and Rudolph did not include it in the
verses which he attributed to J. As elsewhere Rudolph denied the presence
of E altogether (pp. 28-31), finding no need for source-division in vv. 5-7,
assigning v. 16aα to a redactor and simply deleting the reference to ‘the angel
of God’ in v. 19a! In v. 15 he emended the singular ‘do you cry out’ to the
plural to avoid the discrepancy with v. 10. Rudolph’s analysis was recognised
to be faulty but, apart from the special case of Fohrer (pp. 97-110), it had
the effect of reducing E’s contribution to the narrative to fragments (at most
[Noth, p. 84, ET, p. 106] vv. 5a, 6 or 7, 11-12, 19a, 25a: cf. Hyatt, Childs). By
the 1980s E was credited only with vv. 5 and 19a (W.H. Schmidt, Kohata), and
the new developments in Pentateuchal criticism (see below) seemed to have
finally eliminated it altogether. But around 2000 two powerful voices (Propp
14.1-31 219
and Graupner) reasserted the need for the hypothesis of a third, Elohistic,
narrative strand in the chapter.3 Propp (pp. 476-81) is more sure that there
is an extensive E component than about where exactly it is to be found (cf.
Holzinger above): he seems most confident about vv. 14, 19a, 21aβ, 24-25,
28b and 31b. Graupner’s analysis is based on the subtraction of the gener-
ally agreed P component and the detection of duplications or other kinds of
difficulty in what remains (pp. 77-89). Alongside a well preserved J narrative
there are therefore other elements, some of which he treats as redactional (vv.
11-12, 16aα, 31), but there is a residue which actually contradicts the J version
and this must comprise extracts from a parallel old account (vv. 5a, 7aαb, 19a,
25a), which can be assigned to E on the basis of v. 19a. In the last fascicule of
his commentary (2019) W.H. Schmidt accordingly adds v. 25a and possibly
some of vv. 6-7 to his list of E fragments (pp. 619-26).
The ‘new developments’ are often seen to begin with H.-C. Schmitt’s article
published in 1979, but it turns out on closer examination not to be as radical
a departure from older views as it at first appears. It is true that what Schmitt
compared was the viewpoints (in his wording the ‘Geschichtsverständnis[se]’)
of P and the final redactor, but this was clearly at the time a choice made to get
round the already uncertain scholarly opinion about the older components of
the narrative, and the only verse whose composition Schmitt firmly assigned
then to the final redactor is v. 31. Otherwise it is the redactor’s arrangement of
the existing Priestly and non-Priestly material which indicates his viewpoint.4
Shortly afterwards Kohata’s dissertation divided most of the chapter between
J, E and P, but argued that the J account had undergone a redaction (especially
in vv. 7, 11-12[, 14?], 19b, 24-25, 31) which, with the inversion of a traditio-
historical distinction introduced by Noth (pp. 91-95, ET, pp. 115-20), brought
‘Yahweh war’ ideology into a narrative which had originally presented an
account of the episode at the sea which differed markedly from the familiar
picture (Kohata, pp. 281-89). Blum’s account of the chapter is again transi-
tional rather than a thoroughgoing move to an explanation by successive
Bearbeitungen (see esp. Studien, pp. 17-18, 30-32, 39-40, 256-62). The older
3
T. Yoreh, The First Book of God (BZAW 402; Berlin, 2010), makes the E
strand the basic one, supplemented by a Yahwistic revision, but his analytical
procedure lacks rigour, as his reconstruction of E shows (pp. 192-201): vv. 5a, 6-7,
9aα, 19a, 20*, 21aα1βb, 22-23a, 27aα, 28aαb. He uses some ‘Priestly’ elements to
provide the narrative continuity that is obviously desirable in a Grundschrift, very
like Van Seters’s ‘expansion’ of J at the expense of P.
4
This understanding of Schmitt’s position finds some confirmation in his
Arbeitsbuch zum Alten Testament (UTB 2146; Göttingen, 2005), where in ch. 14
only v. 31 is included in his ‘spätdeuteronomistische Pentateuchredaktion’ (p. 243;
cf. pp. 245-46; on p. 244 vv. 13-14 are included as well). Schmitt’s most recent
analysis of the chapter, however, in a Workshop at the 2013 Congress of IOSOT
in Munich, attributed much more to a post-Priestly redactional layer or layers.
220 EXODUS 1–18
substratum of the chapter is attributed as usual to his D-Komposition, but

it involves a ‘partielle Bearbeitung’ of older source-material which is only
clearly visible in vv. 13-14, 30-31 (pp. 39-40: for the supporting ‘parallels’
from elsewhere in the D-Komposition and other Deuteronomistic texts see
pp. 30-32, 47, 104-105). In relation to the Priestly component of the chapter,
Blum acknowledges that it presents an ‘almost complete’ narrative (p. 257)
which because of the ‘gap’ before v. 15 (Moses’ ‘cry’ has not been previ-
ously mentioned and it is implausible to see v. 15a as referring back to the
people’s ‘cry’ in v. 10bβ) must have been taken from ‘a pre-existing, already
formulated text’ by the author of the P-Komposition (p. 260). The latter was
in effect combining two ‘sources’ here: all that makes Blum attribute this to
a compositional layer rather than an old-fashioned redactor is his conviction
that elsewhere the evidence points to this kind of relationship between the
Priestly material and the older substratum.5
Some other recent scholars have adopted a wholly ‘supplementary’
approach to the chapter. This is true in the main of M. Vervenne’s study,
which like Blum’s appeared in 1990, even though it is cautiously expressed
and allows that ‘the P redactors might also have disposed of existing Priestly
material’ (‘The “P” Tradition’, p. 88).6 Vervenne concludes that ‘a profound
study of the composition of the sea-narrative reveals that the Priestly redac-
tion has worked over an existing narrative into a new composition… As a
result, there seems to be no strong evidence that the P edition of at least this
narrative contains an independent and complete Priestly narrative’ (p. 87; cf.
p. 85). The ‘existing narrative’, with ‘the label “JE” ’, is ‘the result of a proto-
deuteronomic redaction of existing materials’, which are not investigated
further (pp. 78-79). Van Seters takes a similar view, except that the ‘existing
narrative’ is his late Yahwist, which exhibits (post-)Deuteronomistic features
in vv. 10, 11, 13-14, 21, 28, 31 (Life, pp. 134-36, 143, 145, 148-49). In his
analysis the P material is much more ‘incomplete’ as an account of the episode
because he argues that several elements that are usually attributed to P must in
fact belong to J (vv. 10bβ, 22a, 23, 28: cf. pp. 131[with n. 8]-34). This analysis
also provides the J narrative with an explicit account of the sea-crossing, the
absence of which is often commented upon (see further below).7
5
Not quite everywhere else, for Blum acknowledges that the situation in
Num. 16 is very similar (pp. 261, 263-71). Whether Blum’s view of the rest of the
Priestly narrative is really tenable is of course open to question: see on Exodus my
‘Composition of the Book of Exodus’.
6
Vervenne seems to be referring here to places other than Exod. 14: cf. his
critique of Weimar’s view that Pg reworked an older narrative of the sea-crossing
on pp. 78-85.
7
Van Seters’s analysis has found little acceptance (but see Dozeman below),
probably because the language of vv. 22a, 23, 28 points much more obviously to
the usual attribution of them to the P strand of the chapter (cf. esp. yabbāšāh for
‘dry land’ in v. 22a as in v. 16 vs. ḥārābāh in v. 21J).
14.1-31 221
After some years during which it seemed to have been abandoned (see
below),8 this approach to the chapter has recently received renewed support.
Dozeman’s commentary (cf. pp. 300, 303) follows Van Seters’s reallocation
of some material from P to J and, as if in compensation, assigns the references
to the pillar of cloud and fire (vv. 19b, 20aβ, 24aβ) to P (presumably because
of 40.36-38, though the expressions there are different). It is hard to see how
the P component, thus defined, could ever have existed separately (and note
‘adds’ twice at the top of p. 317; on pp. 41-42 Dozeman is more cautious
about the nature of P). Berner, who is in no doubt that source criticism is to
be abandoned (Exoduserzählung, pp. 7, 449), has reconstructed the earliest
layer of the text as comprising vv. 5a, 6, 10bα, 13a, 14, 19b, 20aαγb, 21aα2β,
24aαb, 25b, 27aα2βb, 30, which was successively amplified in no fewer than
eleven distinguishable stages of redaction. Several of these made very minor
contributions: the most significant are the main Priestly redaction in vv. 1-2a,
4, 8a, 10abβ, 15, 15aβb, 17abα, 18a, 21aα1b, 22, 23aαb, 26abα, 27aα1, 28-29
and a ‘theological’ (Deuteronomistic) redaction in vv. 11, 13b, 31aβb (see
pp. 400-403, with the chart on pp. 403-405). Berner here, like Blum, regards
the existence of an older basis for the Priestly Bearbeitung of this chapter as
probable (p. 437).9 According to Albertz too (most clearly on pp. 224-27) the
Priestly sections of the chapter are from a (single) Bearbeitung.
The view that Exodus 14 is based on two (but only two) originally separate
accounts of the episode, which were drawn from longer ‘sources’, is also
strongly maintained at the present time, whether the non-Priestly component
is attributed to a ‘Yahwist’ who included the story of the patriarchs as well
(Levin, Baden) or to an ‘old Exodus story’ (K. Schmid, Gertz).10 The main
difference between Levin and Baden is that the former (pp. 341-47) assigns
considerable sections of the chapter to redactional expansions of the original
accounts or of the combined narrative (so vv. 5b, 7, 19a, 20aβγ, 25a are Js,
parts of vv. 3-4, 8, 17-18 and 29 are Ps, and vv. 10a, 11-12, 15aβ and 31
are from the final redactor or later), while Baden maintains that everything
can be assigned to J or P (pp. 196-213). Schmid does not provide a detailed
analysis of the chapter as a whole, but to judge from the passages which he
8
Both Gertz (pp. 203-206) and Graupner (p. 80 n. 260) were strongly critical
of it in the form proposed by Blum and Vervenne.
9
In the opening paper to a workshop at the 2013 Congress of IOSOT in
Munich Berner proposed that all the non-Priestly material might be post-Priestly,
apparently because of Deuteronomistic features (Yahweh-war) which he was
inclined to regard as late. The assumption then would presumably be that P pro-
vided the Grundschrift for this particular episode at least: its relative completeness
would be compatible with this. Gertz and Schmitt argued strongly against the
elimination of any pre-P version.
10
Blenkinsopp’s broad division of the chapter into two ‘narrative strands’,
elsewhere called P and D and extending back into Genesis (pp. 143-44, 158),
would belong more closely with Levin and Baden.
222 EXODUS 1–18
does discuss it is clear that for him vv. 4, 8-9, 17-18, 22, 23, 26 and 28-29
(at least) belong to P (pp. 226, 269), while an older version had the ‘Yahweh
war’ ideology of vv. 13-14 and (24-?)25 (cf. pp. 160-61). Other elements are
late, perhaps very late, such as v. 5a (with its unique reference to a ‘flight’
of the Israelites: p. 152) and vv. 30-31 (which are ‘too advanced’ to be early
and go with chs. 3–4: pp. 144, 196, 248). Gertz provides a verse-by-verse
analysis along similar lines (for a summary see p. 396), with the majority of
the redactional additions being attributed to the ‘final redactor’ who combined
the non-P and P accounts (vv. 2bβ, 5b, 8b, 9 (most), 11-12, 16aα (most), 20aγ,
24 (the addition of ʾēš, ‘fire’), 25a, 31: pp. 214-32). But a substantial nucleus
(including vv. 5a and 30, against Schmid) is retained for the old non-Priestly
account. In its main stages of composition this scheme has a certain simplicity
(too much for Berner, pp. 2-3), but its claims to isolate later additions are as
far-reaching as any.11
As will be clear, some recent analyses of the chapter have made

its composition an extremely complex process. It may be doubted
whether this is necessary. The widely accepted view that two versions
of the story have been interwoven needs only a little elaboration to
do justice to some residual unevennesses in the parallel versions.
The main arguments will be spelt out in the Explanatory Notes and
the conclusions can be just briefly summarised here. Verses 1-4
introduce the Priestly version, which is continued in v. 8. The section
between (vv. 5-7) belongs to the non-Priestly version, and so does
the beginning of v. 9 (to ‘the sea’). The remainder of v. 9 is Priestly
in character, but is most likely a secondary addition. Verses 10-15
and the first clause of v. 16 (to ‘your staff’) are non-Priestly; the
rest of v. 16 and vv. 17-18 are Priestly (cf. the connections between
vv. 17-18 and v. 4), and they lead on (with close verbal correspond-
ence) to vv. 21-23. Verses 19-20 are non-Priestly, as is the middle
of v. 21 (‘and Yahweh made…dry land’), which has no equivalent
in vv. 16-18. Verses 24-25 are also non-Priestly (cf. the temporal
expression and the pillar of cloud and fire, as in v. 20). In vv. 26-28
the two accounts are interwoven: the language of v. 26 is Priestly
(cf. vv. 16-18) and the beginning of v. 27 (to ‘the sea’) and most
of v. 28 (to ‘after them’) echo its wording closely, while most of
11
For Gertz’s more recent observations on the narrative, which include some
minor changes to his analysis and a vigorous defence of a ‘two-source’ approach
(against e.g. Berner), see his ‘The Miracle at the Sea’, in Dozeman et al., The Book
of Exodus, pp. 91-120.
14.1-31 223
v. 27 and the end of v. 28 introduce alien expressions (and duplicate

the statement about the return of the waters) and are non-Priestly.
Verses 29 and 30-31 conclude the Priestly and non-Priestly accounts
respectively, using characteristic language from earlier in the narra-
tive (cf. vv. 13 and 22 especially).
There remain some unevennesses in the non-Priestly version
which require explanation. The two halves of v. 5 seem to be of
separate origin, not so much because the idea of the Israelites
‘running away’ contradicts their being ‘let go’ as because the
singular verbs in vv. 6-7 more naturally continue v. 5a than the
plural in v. 5b.12 The beginning of v. 9 is the natural continuation of
v. 5b, while v. 7 leads well into the focus on Pharaoh in v. 10. Verse
15 follows vv. 10-14 awkwardly in two ways: it is not Moses but
the people who have ‘cried out’ (in v. 10), and after Moses has just
instructed the people to ‘keep your position’ (in v. 13) it is odd to
find Yahweh requiring them to ‘set out’. In vv. 19-20 the statements
about the angel and the pillar of cloud have long been recognised
as doublets. In v. 16 Moses is to raise his staff, presumably to make
the waters withdraw, but in v. 21 Yahweh does this by sending a
strong wind (compare the similar distinction in 10.12-13). The
jamming of the Egyptian chariots’ wheels in v. 25 is only a problem
if they are in motion, but in v. 24 they seem still to be stationary.
Verses 30-31, finally, are repetitive, with two separate references to
what the Israelites ‘saw’ (vv. 30b, 31a) and two separate statements
about how they reacted (v. 31), which are not combined together
elsewhere.13
These variations are not sufficient to reconstruct two more or less
complete earlier versions of the story from the non-Priestly account
in the way that the Priestly account and the non-Priestly account as
a whole can be disentangled. They could in theory be due to a series
of redactional additions to a single old narrative. But they look more
like variants in the way that the story was told than purposeful alter-
ations, and this suggests that they arise from what were once two
independent older accounts of the episode. The parallel non-Priestly
versions are clearest at the beginning and the end (vv. 5a, 6-7 par.
vv. 5b, 9aα; vv. 30, 31b par. v. 31a); elsewhere it seems that one
12
Cf. Gertz, p. 214.
13
For this distinction see Smend, Erzählung, p. 143, and those who followed
him.
224 EXODUS 1–18
version was mainly used, with insertions from the other version to
fill it out. Certainty is impossible, but it appears that the version
which contributed the most to the non-Priestly plague-story (‘E’) is
here used in extracts (vv. 5b, 9aα, 15, 19a, 21aβ, 25a, 31a), while the
main account is from the version that was only drawn on occasion-
ally earlier (‘J’): vv. 5a, 6-7, 10-14, 16aα, 19b-20, 24, 25b, 27b, 28b,
30, 31b. We must presume that this change was due to the redac-
tor’s preference for the themes that were highlighted in this version,
such as Yahweh’s total victory over Pharaoh’s military might, the
‘Yahweh war motifs’, the change in the Israelites’ attitude from fear
to faith and the vindication of Moses as Yahweh’s chosen leader
of the people. Nevertheless his inclusion of certain elements of the
other version must surely also be indicative of his interests. Several
of them seem at first sight merely to fill out the narrative, but there
is a recurring focus on Yahweh’s mighty action in them (especially
at the beginning of v. 31), which would have reinforced the similar
emphasis in the version from which this redactor drew most of his
material.
An important problem that arises from the critical analysis of
Exodus 14 into two or more once independent sources is whether
the non-Priestly account(s) included the crossing of the sea by the
Israelites. If, as they usually are because of their close relation to the
divine instructions in v. 16, vv. 22-23 are attributed to the Priestly
account, there remains nothing in the extant text of the chapter to be
a non-Priestly version of this part of the story. In early days scholars
found no difficulty in presuming that it had been present in the older
source(s): it was simply not taken up by the ‘final’ redactor who
combined ‘JE’ with P, no doubt because the Priestly version, with
the waters like walls on either side of the passageway, was more
dramatic (so e.g. Wellhausen, pp. 76-77; Holzinger, pp. 44-45;
Baentsch, p. 126).14
It seems to have been Eduard Meyer who first declined to make
this assumption and inferred that before P the event was presented
in a different way (Die Israeliten, pp. 22-24), and he was followed
in this successively by E. Sellin, Rudolph (pp. 28-31) and Noth
(pp. 93-95, ET, pp. 118-19): ‘J does not speak of a passage of Israel
14
Of course there was less of a problem for those who followed Smend in
denying that P had any part in the chapter or at least in vv. 21-23: they could and
did attribute the crossing narrative there to one of the older sources (usually E).
14.1-31 225
through the sea. Israel remained in their camp and according to v.

30 perhaps saw nothing at all of the actual flight and catastrophe of
the Egyptians, but merely its consequence.’ For Noth it is ‘the fear
of God which causes the Egyptians to rush into the sea near where
they and the Israelites were encamped… J incorporated in his
description elements of a variant narrative of the miracle at the sea
which…perhaps told how the Egyptians had encamped at a place
from which the sea had gone away and how they were then over-
whelmed by the “returning” sea.’ From Noth this general approach
has been taken up by W.H. Schmidt (p. 603; cf. Exodus, Sinai und
Mose, p. 63), Kohata, Levin (p. 344: J’s source had no reference to
the Israelites’ crossing, and J’s own additions [p. 341] did not intro-
duce one), Graupner (pp. 78, 84) and Baden (p. 206). Not everyone
has been convinced by it, however: Blum doubts if the evidence is
sufficient to support an ‘alternative version’ and thinks it would be
readily understandable if ‘the composition’ had omitted an older
account of the Israelites’ crossing to avoid unnecessary duplication
(Studien, p. 257 n. 96). Such caution is to be commended. Others
seek more explicit support for a crossing in an older account. Van
Seters, who is followed by Dozeman, transfers some of the Priestly
account to J, so that the latter has parts of vv. 22-23 (Life, pp. 131-
34: see further above), and P as a redactional layer amplifies what
J says about this. Yet the vocabulary is much more characteristic of
P. Propp (pp. 483-84, 537) argues that the Song of Moses, which
is included in JE, speaks of a crossing in 15.16. But even if that
is what 15.16 means (see the Note), the case for the poem being
included in JE is weak, however old the poem may be.
In Exodus 14 itself there are some possible indications of a
sea-crossing (or at least an attempted one) by the Israelites and the
Egyptians in the non-Priestly account as we have defined it. But
they are not as clear in the main strand of that account (‘J’) as in
some of what we have taken to be extracts from the parallel old
account (‘E’) that have been added to it. At best one might argue
that Yahweh’s ‘shaking’ of the Egyptians out (of their chariots) into
the sea (v. 27b) makes more sense if they had been attempting to
cross it. In the parallel account, on the other hand, the Israelites are
instructed to ‘set out’ (v. 15b) after they have been caught up while
‘encamped by the sea’ and the angel who moves behind them is
said to have been previously ‘going in front’ of them (v. 19a), just
before Yahweh makes the sea ‘flow away’ (v. 21aβ). This suggests
226 EXODUS 1–18
preparations for a crossing. Once the two old accounts had been
combined (‘JE’), the Egyptians have their chariot wheels ‘jammed’
(or ‘removed’: v. 25a), so that they are slowed up, and they decide
to ‘flee from Israel’ (v. 25b) before the waters return (vv. 27-28):
this is more readily understandable if they were pursuing the Israel-
ites through the passage opened up for them than on the alternative
reconstruction suggested. The presence of these motifs in an old
account of the episode is to be expected in view of its presence in
two psalms which are likely to be pre-exilic (both at the point in
question probably reflecting traditions with a north Israelite origin:
Pss. 77.20-21; 78.13, 33; for the northern associations cf. ‘Joseph’
in 77.16 and ‘the Ephraimites’ in 78.9). If, as some suppose, this
‘crossing’ motif was not an original part of the tradition but was
modelled on the story of the crossing of the Jordan in Joshua 3–4,
its introduction must, it would seem, have occurred by the eighth
century B.C. at the latest.15 Our ‘J’ account, on the other hand, if
its mention of a crossing was not simply omitted by the redactor
who combined it with P, might conceivably represent a form of the
‘sea’ tradition which had remained unaffected by assimilation to
the Jordan narrative, as many believe to be the case in the Song
of Moses in the next chapter (but see the Explanatory Notes on
15.8-9).
To return to the Priestly account, it turns out (when taken alone)
to exhibit a three-part structure which corresponds closely to the
Masoretic divisions before vv. 1, 15 and 26. Each section begins
with the introduction to a speech by Yahweh, in which Moses is
commanded either to instruct the Israelites to do something (vv.
1-2) or to do something himself (vv. 15aα[?], 16aβ; cf. v. 26). When
they or he obey, the announced train of events is set in motion (vv. 8,
21b-23, 28a, 29), described in much the same words as in the divine
speech. So Moses is at the centre of this version too, but every-
thing he does is dictated by a divine command, and the Egyptians
too behave exactly as Yahweh directs. The only action which is
specifically attributed to Yahweh, however, is the making stubborn
15
Against Van Seters (Life, pp. 140-45), who argues that it is post-Deuterono
mistic. Despite what he says about the lateness of any connection between the
Jordan crossing tradition and the shrine at Gilgal (pp. 142, 145), the latter remains
a very plausible locale for such a development of the tradition: cf. Kraus, Gottes-
dienst, pp. 179-89, ET pp. 152-61; Cross, Canaanite Myth, pp. 103-105, 138-42.
14.1-31 227
of Pharaoh’s or the Egyptians’ hearts (vv. 4, 8, 17); elsewhere the

impression given is that it is the obedient action of the Israelites or
Moses that makes things happen.
The final redaction, which combined the J account (itself amplified
by elements from E) with the P account, has used the divine instruc-
tions created by P to introduce each stage of the action: the Egyptian
pursuit of the Israelites, the dividing of the waters and the sea-cross-
ing, and the return of the waters and the destruction of the Egyptian
army. At the beginning of the second stage some non-Priestly mate-
rial has been inserted into the divine speech (vv. 15aβb, 16aα). The
remaining non-Priestly text has been added within or at the end of
the Priestly sequences, in such a careful way that it can often be seen
as providing the detail that is missing from the bare Priestly narra-
tive. This is especially the case with what Vervenne has described as
the ‘framing’ or ‘sandwiching’ of non-Priestly text within two parts
of the Priestly account in vv. 21 and 27-28 (cf. ‘The “P” Tradition’,
p. 87). But the arrangement is dictated by the material itself and is
just as easy to attribute to a separate redactor as to the Priestly writer
himself. In some places, however, the combination of material is
not so careful and exposes itself to the modern reader for what it
is: thus in the delay of the ‘making stubborn’ until after Pharaoh
has begun his preparations (v. 8) and in the repetition of ‘he/the
Egyptians pursued’ in vv. 8-9 and the return of the waters in vv.
27-28. There are also cases where the non-Priestly text has had the
effect of adding material which significantly modifies the simple
Priestly presentation of the episode in the terms outlined above. The
pattern of instruction and obedience is in danger of being under-
mined by the Israelites’ rebellion in vv. 11-12 and is only rescued
by Moses’ intervention in vv. 13-14. These latter verses, together
with vv. 24-25, also introduce an old theology of warfare which
was of no interest to the Priestly writer. Finally, and connected with
this, vv. 30-31 conclude the combined narrative with an expansive
attribution of the outcome to Yahweh’s power (which has also been
an aspect of the non-Priestly ‘detail’ added in vv. 21, 24, 25 and 27)
and a description of the people’s response to it as ‘fear’ and ‘faith’
(in Moses as well as Yahweh).
Theologically, therefore, the Priestly component of the narra-
tive does not have the last word by any means, even if its motif
of command and obedience remains prominent. In contrast to
much recent scholarship, we have not seen this result as being due
228 EXODUS 1–18
to multiple layers of post-Priestly redaction but as deriving from

the balanced if selective combination of material from three older
accounts.16 In most cases our reasons for this conclusion can rest
on an approach that does not adopt a complex explanation when a
simpler one is available. But the case of v. 31 needs closer exami-
nation. As already observed (p. 218), a redactional origin for it
was already suspected at the beginning of the twentieth century,
because of what was taken to be Deuteronomic language. But the
expressions concerned were either not uniquely Deuteronomic (cf.
Num. 12.7-8) or only broadly similar to a Deuteronomic expression
(Deut. 34.12). Consequently in mid-century a number of commen-
tators went back to the older view that v. 31, like v. 30, was from
J(E) (Noth, Hyatt, Childs; so more recently Propp and Baden). The
recent revival of the redactional view turns out to be due largely
to a surprising argument based on the reference to ‘faith’ and an
important and much quoted study of this concept (specifically Heb.
heʾemîn) by (the younger) R. Smend. The issues have been addressed
in the introduction to 4.1-9. Too much weight has been attached to
the famous passage in Isa. 7.1-9, which is unlikely to be the origin
of the concept of faith in God in the Old Testament, especially as
an object is lacking in v. 9: this is likely to be a secondary develop-
ment from the fuller expression ‘believe in (someone)’. Most likely
Isaiah elevated a widespread aspect of Israelite piety (cf. TWAT 1,
613-15 = TDOT 2, pp. 92-93) and perhaps ‘Yahweh War’ (cf. von
Rad, Heilige Krieg, pp. 10, 31 and passim) into a test of Ahaz’s
loyalty to Yahweh by which his and his people’s future would be
determined. The concept of faith is equivalent here in v. 31 to not
being afraid (v. 13), which prophets outside Israel also advocated in
the name of their gods (cf. ANET, pp. 449-51, 605-606). There is
every reason to associate the language used here with that in 4.1-9,
31, but as there it is also reasonable to attribute it to a pre-exilic
narrative source rather than to a late redactional layer.
The other element of Israel’s response in v. 31, the fear of
Yahweh, is characteristic of the passages traditionally attributed to
the E source in the Pentateuch (cf. in general Wolff, ‘Zur Thematik’,
16
Only a few additions of ‘new’ material have been added by the final redactor
to smooth out some of the divergences between the sources he used. Examples are
‘king of Egypt’ in v. 8 (from v. 5), the more precise detail in v. 9 (from ‘all’ to the
end: taken from vv. 2-3 and 23) and ‘horses’ in v. 23 (from 15.1, 19, 21).
14.1-31 229
62-67; Graupner, Elohist, pp. 152-53). ‘To fear Yahweh’ is certainly

a central demand of the Deuteronomic pareneses (e.g. Deut. 5.29;
6.2, 13, 24), but that does not mean that all occurrences of it are
Deuteronomistic, as it is a very widespread aspect of Israelite (and
more generally ancient) piety. Here at least it has no particular
association with the observance of the law (as also in 1.17, 21; 3.6).
As we have argued earlier, faith in Yahweh and fear of Yahweh are
indicators of two different strands of the narrative of Exodus and
the Pentateuch more generally and are part of the reason for seeing
v. 31 as combining elements of the two non-Priestly accounts that
lie behind the present text, as Rudolf Smend (the elder) and his
followers saw. An association of them with a Deuteronomistic layer
of redaction is quite unnecessary.
These responses of the people do, however, contribute powerfully
to the impact of the canonical form of the sea narrative and raise it
above the level of a mere rejoicing at the destruction of the pursuing
army of the Egyptians. They ensure that the event is recalled as a
mighty act of Yahweh himself, just as the ritual celebrations of the
release of the Israelites itself were designed to do (12.27; 13.8-9,
14-16). Moses too, both as the prophetic representative of Yahweh
who assures the people that ‘Yahweh will fight’ for them (v. 14) and
as the one who seems to have divine power in his hand (vv. 16, 21,
26, 27), gains yet more authority over them – or so it seems. In fact,
as the journey onwards will show, this is by no means the end of
Israel’s rebellions against their divinely appointed leader and guide.
[From here to 17.7 the ‘J’ strand is regarded as the main non-Priestly account
and additions to it from the ‘E’ strand (as well as those from elsewhere) are
enclosed in square brackets.]
1 (Then) Yahweh spoke to Moses as followsa: 2 ‘Speak to the

Israelites, that they should turn back and campb in frontc of
Pi-hahiroth, between Migdol and the sea, in frontc of Baal-zephon
– you shall campd opposite it by the sea. 3 Pharaoh will say/
think, “They – meaning the Israelitese – are in confusionf in the
landg: the wilderness has shut them inh”. 4 I will make Pharaoh’s
heart stubborn and he will pursue them, so that I may win
gloryi over Pharaohj and all his army and so that the Egyptians
may know that I am Yahweh.’ They [sc. the Israelites] did so.
5 It was reportedk to the king of Egypt that the people had run
awayl [and Pharaoh and his servants changed their attitudem to
230 EXODUS 1–18
the people and said: ‘Whatn have we done in that we have let
Israel go free from serving us?o’] 6 He harnessed his chariot(s)p
and took his people with himq: 7 he took six hundred choicer
chariots (and) all the Egyptian chariotrys, with officerst over
them allu. 8 Yahweh made the heart of Pharaoh [king of Egypt]
stubborn and he pursued the Israelites, as the Israelites were
leaving defiantlyv. 9 [The Egyptians pursued them and overtook
them as they were encamped by the sea] [– all the horses (of/
and) the chariotry of Pharaohw and his horsemenx and his army –
by Pi-hahiroth in front of Baal-zephon.] 10 yWhen Pharaoh
approachedz, the Israelites looked upy and saw to their surpriseaa
the Egyptians advancingbb behind them and they were very
afraid. The Israelites cried out to Yahweh and 11 said to Moses,
‘Was it because there were nocc graves in Egypt that you took us
to die in the wilderness? What have you done to usdd in bringing
us out of Egypt?ee 12 Is this not what we said to you in Egypt,
“Leave us alone, and let us serve the Egyptians”, for serving
the Egyptians was better for us than dying in the wilderness.’
13 Moses said to the people, ‘Do not be afraid: keep your
position and (you will?) seeff the victorygg of Yahweh which he
will bring about for you today, for the Egyptians whomhh you
have seenii today you will never see again: 14 Yahweh will fight
for you and you shall remain silentjj’. 15 [Yahweh said to Moses,
‘Whykk do you cry out to me? Speak to the Israelites that they
should set out,ll] 16 and you, lift up your staff and stretch out
your hand over the sea and divide it, so that the Israelites may
entermm into the midst of the sea on dry ground, 17 while I for
my partnn am making the hearts of the Egyptians stubborn, so
that they go in after them. In this way I intend to win gloryoo
over Pharaoh and all his army, his chariots and his horses
18 and the Egyptians will know that I am Yahweh when I win
glory over Pharaoh, his chariots and his horses.’ 19 [Then the
angel of God who was going in front of the armypp of Israel
moved and went behind them,] and the pillar of cloud moved
from in front of them and stood behind them: 20 it came between
the armypp of Egypt and the armypp of Israel, and there was the
cloud with (?) the darknessqq and it lit up the night, and neither
(army) came near the other all through the night. 21 Moses
stretched out his hand over the sea, [and Yahweh made the sea
flow awayrr by a strong east wind all night and made the sea dry
land.ss] The waters were divided 22 and the Israelites entered into
the midst of the sea on dry ground and the waters were a wall for
themtt on their right and on their left. 23 The Egyptians pursued
and went in after them, [all the horses of] Pharaoh, his chariotsuu
14.1-31 231
and his horses/horsemen, into the midst of the sea. 24 At the

morning watch Yahweh looked down on the Egyptian armypp in a
pillar of cloud and firevv and set the Egyptian armypp in confusion.
25 [He jammedww the wheels of their chariots and made them
drive with difficulty,] and the Egyptians said, ‘Let us [lit. me]
flee from Israel’, for Yahweh was fighting for them against the
Egyptians. 26 Yahweh said to Moses, ‘Stretch out your hand over
the sea, so thatxx the waters may come back upon the Egyptians,
upon their chariots and their horses’. 27 So Moses stretched
out his hand over the sea, and as morning approachedyy the sea
came back to its usual place,zz and as the Egyptians were fleeing
towards itaaa Yahweh shook the Egyptians out into the midst of
the sea. 28 The waters came back and covered the chariots and
horses of all the Egyptian army who came into the sea after
them: not even one of them was left. 29 But the Israelites had
gonebbb on dry ground through the midst of the sea and the waters
were a wall for them on their right and on their left. 30 Yahweh
delivered Israel on that day from the power of the Egyptians and
Israel saw the Egyptians deadccc on the seashore. 31 [Israel saw
the great act of powerddd which Yahweh inflicted on the Egyptians
and the people were in awe of Yahweh] and they believed in
Yahweh and in Moses his servant.

a. See Note b on the translation of 6.10–7.5.
b. Heb. ‫וְ ישׁבו וְ יחנו‬. Simple waw with the jussive often expresses purpose (cf.
GK §165a), but in cases like this after a verb of saying an indirect command
is likely to be intended (cf. 6.11; 11.2; 14.15; 25.2), equivalent to a direct
command as in 16.12 and 31.13 (cf. JM §177j,k where 8.4 is cited as an
example).
c. Heb. ‫לפני‬. The use of ‫ לפני‬with places is ‘not very common’ (BDB,
p. 817), and most of the occurrences relate to a sacred building or part of it
(cf. 16.34; 30.6; 40.5-6). With place-names it appears elsewhere only in the
parallel verse in Num. 33.7 (but there with ‫ )מגדל‬and in Gen. 23.17, where it
used (with ‫ )ממרא‬instead of (and equivalent to?) the more common ‫( על־פני‬cf.
Gen. 23.19; 25.9; 49.30; 50.13). ‫ על־פני‬often but not always means ‘east of’
(BDB, pp. 818-19), and that might be the sense of ‫ לפני‬here. But in its second
occurrence it is equivalent to )‫נכח(ו‬, which does not seem to have such a sense
(for its use with place-names see Judg. 19.10; 20.43; Ezek. 47.20 [all preceded
by ‫)]עד‬. Whether the meaning is in effect ‘outside’ the places mentioned or
‘opposite’ them (i.e. across a stretch of water) remains uncertain: on possible
identifications of the places see the Excursus in the Explanatory Notes.
232 EXODUS 1–18
d. Heb. ‫תחנו‬. For the transition from third person pl. to second person pl.
cf. 12.4.
e. Heb. ‫לבני־ישׂראל‬, with ‫ ל‬here in the sense of ‘about, of’ (BDB, p. 514):
the translation given is a paraphrase of this.
f. Heb. ‫ נבכים‬is the Niphal part. of the rare verb ‫בוך‬, which occurs only
here, in Joel 1.18 (of cattle in a time of drought) and Esth. 3.15 (of the city
of Susa after the issue of Haman’s decree), with a related noun ‫ מבוכה‬in Isa.
22.5 (of a city under attack) and Mic. 7.4 (of time of social disintegration).
These other occurrences suggest a general sense of ‘distress, anxiety’, and
Ar. bāka provides support for the specific meaning ‘be disturbed, confused’,
which is generally adopted, with some support from the Vss (Tgg generally
[cf. MRI here (Lauterbach, pp. 190-91)]; LXX and Sy in Esth. 3.15; possi-
bly LXX here and in Isa. 22.5, if πλανάομαι and πλάνησις are understood
metaphorically): on their other interpretations see Text and Versions. There
is no basis in BH for the sense ‘wander’ (so HAL, p. 111, here and in Joel
1.18; Ges18, p. 132, here; cf. NRSV ‘wandering aimlessly’). For a possible
Ar. cognate meaning ‘be pressed’, see C. Rabin, ‘Etymological Miscellanea’,
ScrH 8 (1961), pp. 384-400 (388), but there is no need to invoke this to
explain the sense here.
g. Heb. ‫ בארץ‬most obviously means ‘in the land’, but after a passive verb
‫ ב‬is often instrumental, which might justify the rendering ‘by the land’ (sc.
the terrain).
h. Heb. ‫סגר עליהם‬. ‫ סגר‬is most often followed by an object such as ‫דלת‬,
‘door’, though it can be omitted even when the sense is literal (Gen. 7.16) and
even more so when it is not (cf. Job 12.14, also with ‫)על‬.
i. Heb. ‫ואכבדה‬, with waw and the cohortative expressing purpose (JM
§116c), as again in v. 17. The sense ‘win glory’ for ‫ כבד‬Niph. is apparently
found only in exilic and later texts (cf. vv. 17-18; Lev. 10.3; Isa. 26.15; Ezek.
28.22; 39.13; Hag. 1.8; Sir. 33[36].4: in 1 Sam. 6.22 the meaning is probably
different). The underlying basis for this seems to be a reflexive rather than
a passive use of the Niphal (cf. THAT 1, 801 = TLOT, p. 595), i.e. ‘glorify
myself’. Some hold that the idiom ‘has more to do with a demonstration of
Yahweh’s power’, i.e. with his showing or revealing himself to be glorious,
in view of the ‘recognition-formula’ which follows here and in v. 18 and
also in Ezek. 28.22 (TWAT 4, 21 = TDOT 7, p. 20: cf. HAL, p. 434; Ges18,
p. 522 – all apparently based on W. Zimmerli, Ezechiel, 2 [BKAT; Neukirchen,
1969], p. 692 [on 28.22], ET, p. 98, and bibl. there). Such a sense is more
characteristic of the Hithpael than the Niphal (cf. GK §54e) and is perhaps
conveyed more by the ‘recognition-formula’ (where it is present) than by the
verbal form itself.
j. Heb. ‫בפרעה‬. The sense of ‫ ב‬here and in the next phrase (as after ‫ כבד‬Niph.
also in vv. 17-18) is presumably one of supremacy (as with ‫ )משׁל‬or hostility
(as with ‫נלחם‬: cf. BDB, p. 89, II.4): its use with ‫ נקם‬Niph. (e.g. Jer. 50.15)
may also be related. Alternatively ‫ ב‬may be instrumental, as with ‫ קדשׁ‬Niphal
14.1-31 233
in Lev. 10.3. E. Jenni (Die hebräischen Präpositionen, 1: Die Präposition

Beth [Stuttgart, 1992], pp. 100-103, 113-15) places the occurrences of ‫ ב‬with
‫ כבד‬Niphal in this chapter (and ‫ קדשׁ‬in Lev. 10.3) with those examples of the
related beth causae which accompany verbs expressing personal status and
specifically the glory or honour of God (‫כבד‬, ‫פאר‬, ‫קדשׁ‬, ּ‫גבה‬, ‫)שׂגב‬.
k. Heb. ‫ויגד‬. The impersonal passive construction is common in older narra-
tives (e.g. 1 Sam. 23.7, 13; 27.4) but is not found in P.
l. Heb. ‫ברח‬. The verb is only rarely used of flight from an attacking army
(e.g. Jer. 4.29; 39.4): most often it describes ‘evasion of and escape from
continuing, unpleasant, dangerous situations, e.g. tensions and tragedies
within the tribe’ (TWAT 1, 778-79 = TDOT 2, p. 250). Here the Israelites are
regarded as runaway slaves (cf. 1 Kgs 2.39), who have taken advantage of
the temporary release from their labours which Pharaoh had granted them (cf.
Houtman, pp. 226, 259).
m. The ‘change’ indicated by ‫ הפך‬with ‫ לב‬or ‫ לבב‬can be a change of mind
(Hos. 11.8), but it can also be an emotional change (Lam. 1.20; 5.15) or a
change of attitude, as here (cf. the following ‫ )אל־העם‬and in Ps. 105.25. For
other examples of a single nomen regens with two (or more) genitives see GK
§128a: ‫ פרעה ועבדיו‬naturally ‘form one closely connected whole’ (ibid., n. 3:
cf. 5.21; 7.10, 20, although in these and some other places the nomen regens
is repeated).
n. Heb. ‫מה־זאת‬. Here, as in v. 11, ‫ זאת‬is an enclitic (not, as in 13.14, a
demonstrative pronoun), which strengthens the interrogative ‫( מה‬GK §136c,
noting that ‫ זאת‬rather than ‫ זה‬is used [in fact only] when ‫ עשׂה‬follows): the case
of Jonah 1.10 shows that the meaning of ‫ זאת‬is not ‘this’ in such cases.
o. Heb. ‫מעבדנו‬. For ‫ מן‬with an inf. cons. expressing a negative consequence
see GK §119y; JM §169h. But here after ‫ שׁלח‬Piel the ‫ מן‬can almost have its
regular sense of ‘from’, even though the verb is generally used without such
a prepositional modifier in Exodus: only in 6.11; 7.2; 11.1, 10; 12.33 (all with
‫ ארץ‬or an equivalent).
p. Heb. ‫רכבו‬. Perhaps collective, as in v. 7 and later in the chapter, but the
sing. ‫ויאסר‬, ‘harnessed’ (the regular expression for attaching the horses to a
chariot: Gen. 46.29; 1 Kgs 18.44; 2 Kgs 9.21 – cf. 1 Sam. 6.7, 10),17 favours
a reference to Pharaoh’s own chariot here, as does the separate reference in
v. 7 to what he ‘took’ with him.
q. The inverted word-order (O-P) in the second half of the verse probably
implies no emphasis on the ‘people’ (i.e. his military personnel: Childs, p. 218)
but is simply for variety (cf. Muraoka, Emphasis, pp. 39-40; JM §155oa).
r. Heb. ‫בחור‬. The rendering ‘picked’ (NRSV etc.) represents the pass.
part. form correctly, but the word need not imply an act of selection on this
particular occasion: ‫ בחור‬is used, esp. with ‫( אישׁ‬and even alone), to mean
17
Also, in Aramaic, in the Tel Dan inscription, ll. 6-7.
234 EXODUS 1–18
‘choice’, i.e. the finest or the best, apparently always for fighting (e.g. 1 Sam.
24.3; 1 Kgs 12.21); cf. the use of ‫ מבחר‬in 15.4.
s. Heb. ‫וכל רכב מצרים‬. It is not clear grammatically whether the waw is
additive, so that ‘all the other (sc. ordinary) Egyptian chariotry’ is meant (so
NRSV), or explicative (cf. GK §154a n. 1[b]), so that the 600 would comprise
the whole chariot force available to Pharaoh.
t. Heb. ‫ושׁלשׁם‬. Apart from cases where the meaning is ‘a third’ (of a
measure: Isa. 40.12; Ps. 80.6) or ‘a triple’ (of a musical instrument: 1 Sam.
18.6) and Prov. 22.20Q, where the intended meaning is uncertain, there are
sixteen occurrences of ‫ שׁלישׁ‬in BH, five sing. and eleven pl. (to those listed
by BDB, p. 1026, add Exod. 15.4; 1 Kgs 9.22; 2 Chr. 8.9). Older EVV. (e.g.
Tyndale, AV) rendered it by ‘captain(s)’ (cf. Luther, ‘Hauptleute’). Gesenius,
Thesaurus, p. 1429, took the primary sense to be ‘chariot warrior’, and BDB
followed him closely: ‘adjutant or officer (best explained as third man [in
chariot])’. P. Haupt, ‘The Hebrew Term šālîš’, Beiträge zur Assyriologie 4
(Leipzig, Baltimore and London, 1902), pp. 583-87 (cf. JBL 21 [1902], pp. 74-
77), identified the ‘third man’ specifically with the shieldbearer, and this
view was widely, but not universally, followed (e.g. Noth, p. 89, ET, p. 112).
In 1979 B.A. Mastin, ‘Was the Šālîš the Third Man in the Chariot?’, in J.A.
Emerton (ed.), Studies in the Historical Books of the Old Testament (VTSup
30), pp. 125-54, challenged this interpretation in a detailed examination of
biblical and extra-biblical evidence and proposed instead that it meant a
kind of officer, most likely one ‘of the third rank’ (which had according to
Gesenius, ibid., been the view of Drusius [c. 1600]). HAL, pp. 1412-13, appar-
ently unaware of Mastin’s article, continued to maintain the ‘third man in a
chariot’ view, finding support in Akk. tašlīšu with the same meaning (AHw,
p. 1339); likewise Ges18, p. 1365, with additional bibl. (inc. Mastin); and
DCH 8, p. 392 (again unaware of Mastin and other recent studies), gives
‘third man (in a chariot), officer, adjutant’. But M. Vervenne, ‘Hebrew šālîš –
Ugaritic tlt’, UF 19 (1987), pp. 355-73, concludes that the meaning in both
the Exodus occurrences is ‘knights’ (cf. S.R. Driver ad loc.), in the sense
of ‘someone with a thorough military training in the use of arms and in
horsemanship’ (p. 361); and both Houtman (1, pp. 63-64) and Propp (pp. 492-
93) have also accepted Mastin’s critique of the ‘third man’ interpretation.
The strongest objection to the latter is that most of the biblical occurrences
of ‫ שׁלישׁ‬have nothing obvious to do with chariots. Hebrew epigraphy may
provide some support for ‘of the third rank’: two inscriptions from Ḥorvat
ʿUza (AHI 2, 37.004 and 005) include occurrences of šlšy, as well as other
ordinal numbers, in what may be lists of soldiers (cf. qṣyn in both inscrip-
tions) ordered according to rank (see Davies, ‘Some Uses of Writing in
Ancient Israel’, in P. Bienkowski et al. [eds.], Writing and Ancient Near
Eastern Society [FS A.R. Millard; LHBOTS 426; London, 2005], pp. 155-74
[156-57]).
14.1-31 235
u. Heb. ‫על־כלו‬. The sense to be given to ‫ על‬depends upon the view taken
about the meaning of ‫שׁלישׁ‬: if it is ‘third man in a chariot’ ‫ על‬must mean ‘on’;
if it is a kind of officer, as seems more likely, then ‘over’ (in the sense of ‘in
command of’) is also possible.
v. Heb. ‫ביד רמה‬, lit. ‘with raised hand’. In Num. 15.30 the phrase is used
of a deliberate sin in contrast to an unintentional one, an act of contempt and
defiance towards God (v. 31).18 Here and in the very similar context of Num.
33.3 it expresses the same attitude of defiance towards the Egyptians: cf. the
related expressions in Deut. 32.27 and Mic. 5.8 as gestures of victory and in
1 Kgs 11.26-27 of a rebellion. The clause is a regular circumstantial noun
clause (GK §141e).
w. Heb. ‫כל־סוס רכב פרעה‬. This could mean ‘all Pharaoh’s chariot horses’,
but the expression ‫ סוס רכב‬occurs nowhere else and elsewhere in the chapter
the chariots are always mentioned in their own right. In a very similar listing
in v. 23 there is also no waw between ‫ סוס פרעה‬and ‫( רכבו‬in accordance with
GK §154a n. 1[a]), but asyndeton here between what would be two nomina
regentia (‘the horses and chariots of Pharaoh’) scarcely conforms to a regular
grammatical pattern. The awkwardness of MT is probably due to redactional
expansion in this part of the verse, which is likely on other grounds (see the
Explanatory Note).
x. Heb. ‫ופרשׁיו‬. ‫ ָפ ָרשׁ‬can mean both ‘horseman’ and ‘horse’ (perhaps origi-
nally ‘mare’: cf. D.R. ap-Thomas, ‘All the King’s Horses?’, in J.I. Durham
and J.R. Porter [eds.], Proclamation and Presence [FS G.H. Davies; London,
1970], pp. 135-51) and Mowinckel argued from its frequent association with
chariots that it (always) has the specific meaning ‘chariot horse’ (‘Drive and/
or Ride in O.T.’, pp. 289-95). The word also appears in vv. 17, 18, 23, 28
and in 15.19. It might seem from the present text that the meaning ‘horse’ is
excluded by the occurrence of ‫ סוס‬earlier in the list here, in v. 23 and in 15.19:
hence the rendering ‘riders’ or ‘chariot-drivers’ (e.g. NRSV) has sometimes
been adopted throughout. But in all three verses there is good reason to see ‫סוס‬
as part of a redactional addition (see the Explanatory Notes), so that there is no
need to avoid the rendering ‘horse’ for ‫ פרשׁ‬in the original narrative. It is only
in the redacted text that it becomes a problem. According to the Masoretic
vocalisation the noun would be of the qattal type, which often identifies a
person’s role in society (e.g. ‫ח ָרשׁ‬,ָ ‘craftsman’: JM §88Ha), but this vocalisa�-
tion is universal in BH, even where the meaning is certainly ‘horse’, and it
may in such cases be erroneous. Definite cases of the meaning ‘horseman’
are in fact relatively few in number (cf. HAL, p. 919; Ges18, p. 1085; against
BDB, p. 832).
18
Further occurrences in legal contexts appear in 1QS 5.12; 8.17, 22; 9.1; CD
8.8 (=19.21); 10.3 and in several texts from Qumran Cave 4 (see the list in DCH
7, p. 442).
236 EXODUS 1–18
y. The word-order in the first clause marks it as a circumstantial clause

which is in effect subordinate to the second clause (so Joosten, Verbal System,
p. 361, with further examples [e.g. Gen. 4.1; 1 Sam. 19.18] on p. 132).
z. Heb. ‫הקריב‬. The Hiphil is apparently intransitive here (or ‘internally
transitive’: GK §53d-e), for which there is no exact parallel in BH. But ‫הקריב‬
‫ לבוא‬in Gen. 12.11 is at least semantically equivalent and ‫ הקריב‬here is perhaps
an abbreviation of this expression.
aa. Heb. ‫והנה‬: the translation brings out both the visual implication and the
element of surprise which are present here, as often in the use of ‫ הנה‬in narra-
tive (cf. BDB, p. 244, s.v. c; Houtman 1, p. 18).
bb. Heb. ‫נסע‬. The sing. will be due to the treatment of ‫ מצרים‬as a collective
sing. (as below in v. 25: cf. GK §145f).
cc. Heb. ‫המבלי אין‬. The ‫ אין‬is unnecessary, but also occurs in 2 Kgs 1.3, 6,
16, again in lively polemical discourse, where the pleonasm no doubt has a
rhetorical purpose.
dd. See Note n above: ‫ ָלּנוּ‬has dagesh forte conjunctivum according to GK
§20f.
ee. Here the action called in question is expressed, not by a ‫כי‬-clause as in
v. 5, but by a gerundival inf. cons. (GK §114o). For the unusual form of the
suff. ‫אנו‬-
ָ see GK §61c, h.
ff. Heb. ‫וראו‬. The second imperative may simply represent a further
command, but frequently such pairings are to be understood as a condition
and the consequence of its fulfilment (GK §110f).
gg. Heb. ‫את־ישׁועת יהוה‬. The immediate expression here corresponds to a
widespread characteristic of the ‫ הושׁיע‬word-group: that the focus is ‘on the
agent or the action rather than the beneficiary or the danger/need’ (SAHD
summary, p. 7), so that ‘help’ or better ‘victory’ is preferable as a translation
to ‘deliverance’. But the clauses which follow introduce references to both
the beneficiary and the danger, and so does the occurrence of the verb ‫ הושׁיע‬in
v. 30 below, so these features, while not central, do place Yahweh’s action here
very much in the context of inter-human relationships and his own interaction
with human communities in the narrative setting (cf. ‫ ויושׁע‬in v. 30).
hh. Heb. ‫אשׁר‬. Since there is no preceding antecedent, ‫ אשׁר‬seems here
to introduce an ‘independent relative clause’ (GK §138e) and would then
be translated ‘those whom’. However, the sentence is complicated by the
occurrence of what would be the natural antecedent, ‫את־מצרים‬, within the
‫אשׁר‬-clause and this has led to the alternative suggestion that ‫ אשׁר‬here means
‘as’ (GK §161b: see also Text and Versions), as it may do in a few other
passages (BDB, p. 83 [cf. 1121] lists Deut. 15.14; 1 Sam. 16.7 [in both
of which it prefers emendation]; Isa. 54.9; Jer. 33.22; 48.8; Ps. 106.34 as
places where this meaning has been proposed). But this produces an unlikely
weakening of the sense, even if v. 30b might seem to support it, and perhaps
‫ את־מצרים‬is a later explanatory addition, like ‫ את מלך אשׁור‬in Isa. 7.17. Another
possibility arises from the fact that the antecedent of a relative clause is
14.1-31 237
occasionally repeated within the clause (Gen. 13.16; 49.30; 50.13; Jer. 31.32:
cf. JM §158h n. 1): perhaps here it has simply been drawn forward into the
clause, which may then be understood as relative (cf. Vulg).
ii. Heb. ‫ראיתם‬. A perfective understanding (‘you have seen’) is not impos-
sible, but JM §112a includes this and some other instances of ‫( ראה‬and ‫)שׁמע‬
among stative verbs, where the perfect has a present tense meaning. Joosten,
Verbal System, pp. 197-202, does not include them in his treatment of stative
verbs, and nor does GK §106g. In the other examples cited in JM for ‫ ראה‬and
‫( שׁמע‬1 Kgs 20.13; Jer. 7.11; Pss. 35.22; 74.9; Lam. 3.59; Jer. 4.31; Job 3.18;
Ruth 2.8) a present tense translation is either not compelling or (in poetry) not
due to the use of these verbs.
jj. Heb. ‫תחרישׁון‬. For the sense ‘silent’ rather than ‘still’ (NRSV) see the
analysis in BDB, p. 361. The ‘paragogic’ nun, which is an ancient feature
of the language, is particularly common in pausal forms (JM §44e) and here
brings Moses’ short speech to a forceful conclusion.
kk. For ‫‘ = מה‬why?’ in place of ‫למה‬, cf. 17.2 and BDB, p. 553 s.v. 2a(b).
ll. Heb. ‫ויסעו‬. On the indirect command construction see Note b above.
mm. Heb. ‫ויבאו‬. Simple waw with the imperfect/jussive expressing purpose
according to GK §165a.
nn. Heb. ‫ואני הנני‬. The combination of the independent personal pronoun
and following ‫ הנה‬serves both to underline the contrast between what Moses
and Yahweh are to do and, probably, to give special emphasis to Yahweh’s
role: cf. JM §146a,d, 156m, and Muraoka, Emphasis, pp. 62, 140.
oo. Heb. ‫ואכבדה‬, with simple waw and the cohortative adding the larger
purpose of Yahweh’s action (JM §116c): the translation adds ‘In this way’ to
break up what would otherwise be a long and complex sentence. For the use
of ‫ כבד‬Niphal see Note i above.
pp. Heb. ‫מחנה‬. From its association with the verb ‫ חנה‬the noun has a primary
sense of ‘encampment’, both of travellers in general (e.g. Gen. 32.22) and
of an army ready for battle (e.g. 1 Sam. 4.5-7). The secondary meanings,
‘those who encamp’ and specifically ‘an army, force’ (e.g. Josh. 8.13), will be
derived from this. While the use of ‫ חנה‬in vv. 2, 9 might seem to favour the
former sense in the occurrences in vv. 19-20, the immediate context favours
the latter: nothing has been said about the Egyptians encamping, in v. 24 the
reference is clearly to the Egyptian force, and the two instances of ‫ זה‬in v. 20b
most naturally refer back to the repeated ‫ מחנה‬in v. 20a.
qq. Heb. ‫ויהי הענן והחשׁך‬. MT is puzzling (on the different LXX reading see
Text and Versions) but since ‫הענן‬, viewed as fiery in the night (as in v. 24),
must be the subject of ‫ ויאר‬it seems that ‫ והחשׁך‬does not add a second subj.
for ‫ ויהי‬and the waw is to be regarded as waw concomitantiae (cf. 12.8; 21.4:
BDB, p. 253; GK §154a n. 1[b] end; JM §151a), so ‘with the darkness’ (so
NRSV: as proposed already by Ehrlich, Randglossen 1, p. 318). The expres-
sion remains unusual and the text may be corrupt or the result of careless
editing. M. Vervenne regards both this clause and the next as glosses: he sees
238 EXODUS 1–18
‫ והחשׁך‬as an apodosis(!) and the subject of ‫ ויאר‬as ‫ עמוד האשׁ‬in 13.21 (‘Exodus
14,20MT-LXX: Textual or Literary Variation?’, in J.M. Auwers and A. Wenin
[eds.], Lectures et relectures de la Bible [FS P. Bogaert; BETL 144; Leuven,
1999], pp. 3-25 [21-24]). See also G. Steins, ‘Exodus 14,20 – ein neuer Blick
auf ein altes Problem’, ZAW (2009), pp. 273-76, who sees the angel/Yahweh
as the subj. of ‫( ויאר‬with an allusion to Gen. 1). For other views see the
summary by Childs, p. 218.
rr. Heb. ‫ויולך‬. For ‫ הלך‬Hiphil of waters being made to ‘flow’ see Ezek.
32.14: the Qal is also used with waters as the subj. (e.g. Isa. 8.6 and AHI
4.116.4-5). ‫ הלך‬frequently has the sense ‘(cause to) go away’ (e.g. Gen. 18.33;
Exod. 3.19; Job 12.17).
ss. Heb. ‫לחרבה‬, a different word from that used in vv. 16, 22, 29 (‫)יבשׁה‬: for
‫ חרבה‬cf. Gen. 7.22; Josh. 3.17; 4.18; 2 Kgs 2.8; Ezek. 30.12; Hag. 2.6.
tt. Heb. ‫והמים להם חמה‬. A standard circumstantial noun-clause (GK §141e,
156a,c).
uu. Heb. ‫רכבו‬, without waw, as often occurs in the middle of a list (GK
§154a n. 1[a]).
vv. Heb. ‫בעמוד אשׁ וענן‬. Since ‘fire’ and ‘cloud’ are closely connected a
single nomen regens can govern them both (see Note m above). The absence
of the def. art. with ‫ אשׁ‬and ‫ ענן‬is more surprising (contrast ‫ עמוד הענן‬in v. 19
referring back to 13.21 [also 13.22]): it is almost as if the ‘pillar’ were being
mentioned for the first time here. This might suggest that 13.21-22 and perhaps
(part of) 14.20 too belonged to a very late redactional layer, though one might
expect such a redactor to be well aware of the earlier mentions of the ‘pillar’.
An ingenious solution has been proposed by E. Blum, ‘Die Feuersäule in Ex
13-14 – eine Spur der “Endredaktion”?’, in id., Textgestalt und Komposition
(ed. W. Oswald, FAT 69; Tübingen, 2010), pp. 137-56: there were not two
pillars but one, with a changing appearance. Here, as night turned to day, the
column was half-fire and half-cloud, a phenomenon not hitherto mentioned in
the text and therefore fittingly expressed without the def. art.
ww. Heb. ‫ויסר‬. The simplest interpretation is to see this as from ‫ סור‬Hiphil
in its common sense, ‘removed’, which is appropriate enough in the context.
But it is preferable to adopt the variant reading ‫ויאסר‬, from ‫אסר‬, ‘bind’ (see
Text and Versions).
xx. Heb. ‫וישׁבו‬. See Note mm.
yy. Heb. ‫לפנות בקר‬. In Judg. 19.26 this point in time clearly precedes
morning (v. 27), just as in Gen. 24.63 and Deut. 23.12 ‫ לפנות ערב‬precedes
evening, so that ‫ פנה‬will here have the sense ‘turn towards, approach’.
zz. Heb. ‫לאיתנו‬. ‫ איתן‬is normally an adj., ‘continuous, enduring’ (of water
in Deut. 21.4; Amos 5.24), but it is used as a noun in Ps. 74.15 and in the pl.
as the name of a month in 1 Kgs 8.2, where BDB, p. 450, renders ‘steady
flow(ings)’.
14.1-31 239
aaa. Heb. ‫לקראתו‬. NRSV renders ‘(were fleeing) before it’, but this cannot
be right: all the parallels and the etymology indicate that the sense is ‘towards
it’: the waters were closing in ahead of the Egyptians and so they faced an
insuperable obstacle.
bbb. Heb. ‫ובני ישׂראל הלכו‬. The waw is adversative here and the verb is to
be understood in a pluperfect sense, as the inversion of subject and predicate
show (cf. JM §155nb; GK §106f). Both features made the use of the regular
waw consecutive inappropriate.
ccc. Heb. ‫מת‬. The sing. is at first surprising with ‫מצרים‬, but is readily expli-
cable since ‫ מצרים‬is treated as a collective sing. elsewhere in the chapter (cf.
the suffixes in v. 25 and the sing. part. in v. 10).
ddd. Heb. ‫את־היד הגדלה‬. Since ‫ יד‬is often used metaphorically for ‘power’
(cf. ‫ מיד־מצרים‬in v. 30), it is a small further step for it to mean a specific ‘act
of power’ (‫ גבורה‬is used similarly in the plural): cf. Deut. 34.12; Ps. 78.42;
Job 27.11.
Explanatory Notes
1-4. The new chapter begins (after a petuchah in MT which is
anticipated in 4QExc) with the same words as were used in 13.1
to introduce legal material (see the Note there). Here Yahweh’s
words contain an unexpected instruction for the Israelites to ‘turn
back’ from the journey of departure which they have begun (so
already 12.37, 41, 51) and encamp at a place ‘by the sea’ which
is defined with unusual precision (v. 2). In the present form of the
text it is natural to take the sea to be the Yam Suf mentioned in
13.18 and the starting-point for this manoeuvre to be Etham ‘on the
edge of the wilderness’ (13.20). It is, however, exceptional for any
movement on the journey to be the subject of such a divine commu-
nication (33.1, of the departure from Sinai, is the closest parallel in
the Pentateuch), and 13.21 has just indicated a different means by
which Yahweh is said generally to have guided his people. Here it is
no doubt the need to explain the purpose of this deviation from the
direct route (vv. 3-4) that is responsible for the form of guidance that
is used. The immediate purpose is that, when Pharaoh hears of it (it
is assumed that he will) he will (mis)understand it as a failure of the
Israelites to continue their journey into the wilderness which has left
them confused about what to do next. But what might be seen as the
natural reaction of the Egyptian king – to pursue the Israelites and
recover his work-force – is traced instead to a fresh development of
240 EXODUS 1–18
Yahweh’s own plan, as stated already in 7.3-5(P), which introduces

the defeat of the Egyptian army as both a new result of Pharaoh’s
stubbornness and the means by which the Egyptians will be brought
to recognise Yahweh for who he is.
This theological interpretation of not only the destruction of the
Egyptian chariot-force but also of Pharaoh’s decision to pursue
the Israelites is a recurring feature of the chapter as a whole (cf.
vv. 8, 17-18). But it is very different from the divine purpose in
bringing the Israelites to Yam Suf according to 13.17-18, and it
is curious to say the least that in vv. 5-7 Pharaoh is described as
making his preparations for a pursuit of the Israelites before his
heart is made stubborn in v. 8. Already here it is apparent that two
different accounts of the deliverance of Israel at the sea have been
interwoven in the present text of Exodus (see further the Note on
vv. 5-7 and the introduction to this section).
The very precise location given in v. 2 (and v. 9) for the encamp-
ment ‘by the sea’ may well have originally pointed clearly, for its
authors and some at least of its readers, to a quite specific place
on the eastern border of Egypt. That it no longer does so (as the
unresolved scholarly debate shows) is due to the fact that modern
research, using archaeological and textual evidence from ancient
Egypt, has identified possible equivalents to the names at several
places in the border region. My earlier brief review (Way of the
Wilderness, pp. 80-82) distinguished four main alternatives: (1) the
region of Lake Sirbonis on the Mediterranean coast east of Port
Said; (2) a location close to the north end of the modern Suez Canal;
(3) a place further south in the Isthmus of Suez (Lake Timsah or the
Bitter Lakes); and (4) the traditional site at the head of the Gulf of
Suez. It was concluded then that (3) was the most likely. This is still
possible, but the case for (2) is perhaps the strongest now (see the
discussion in the Excursus below).
Excursus on the Place-Names in Exodus 14.2 and 9

Full discussions of the growing Egyptian evidence have been provided by
O. Eissfeldt, Baal-zephon, Zeus Casios und der Durchzug der Israeliten
durchs Meer (Halle, 1932), H. Cazelles, ‘Localisations’, both of whom
favoured identifications in the vicinity of Lake Sirbonis on the Mediterranean
coast, and Hoffmeier, Israel in Egypt, pp. 169-71, 182-91 (where he locates
the names ‘somewhere around the Lake Timsah and the Bitter Lakes region’)
14.1-31 241
and Ancient Israel in Sinai, pp. 71-73, 94-109 (now proposing a more north-
erly location near el-Qantara on the line of the modern Suez Canal). Since
Eissfeldt and Cazelles wrote, important new archaeological and geological
work has been done in north-eastern Egypt, and this provides the context
within which the relevant Egyptian textual evidence needs now to be assessed
(see M. Bietak, ‘Avaris and Piramesse: Archaeological Exploration in the
Eastern Nile Delta’, PBA 65 [1979], pp. 225-89 [274-81]; Hoffmeier, Israel
in Egypt, Chapter 7; Ancient Israel in Sinai, Chapter 5; also my ‘The Wilder-
ness Itineraries and Recent Archaeological Research’, in J.A. Emerton [ed.],
Studies in the Pentateuch [VTSup 41; Leiden, 1990)], pp. 161-75 [161-67],
on E. Oren’s work in the vicinity of Lake Sirbonis).19
Pi-hahiroth
As transmitted by the tradition the name has the appearance of a Hebrew or
at least Semitic expression: ‘the mouth of the ḥîrōt’ (in Num. 33.8 MT and
LXX read simply ‘the ḥîrōt’, but it is not certain whether this is a textual
error – the Samaritan Pentateuch has ‘Pi-hahiroth’ as in v. 7 – or a real alter-
native form of the name). ḥîrōt could be the plural of a noun ḥîrāh, but no
such noun is known in Hebrew, although it bears some similarity to words for
‘hole, cave’ and (in post-biblical Hebrew) ‘liberty, license’ (cf. the explana-
tions in MRI [Lauterbach, p. 188] and Text and Versions on these verses).20
But, as the case of Succoth shows (see the Explanatory Note on 12.37-38),
an apparently Hebrew name may conceal an Egyptian toponym and modern
scholars have generally explained Pi-hahiroth in this way. Cazelles, Autour
de l’Exode, pp. 216-20, after discounting some less likely suggestions, offers
three possibilities. The third-century B.C. Demotic papyrus Cairo 31169,
3.18, includes the name ḥnyt-ta-ḥerty, where ḥnyt is a word for ‘canal,
stretch of water’. Cazelles considered that the context implied a location near
Pelusium, but the preceding names in the list point rather to the Wadi Tumilat
area (see below), as D.B. Redford (who transcribed the name ḥnt-t3-Ḥ3-r-ti)
19
For the issues discussed here, as distinct from the location of Yam Suf,
H. Lamberty-Zielinski, Das “Schilfmeer” (BBB 78; Frankfurt, 1993]) is chiefly
of value for its survey of the impact of geographical studies on biblical research
(pp. 5-16).
20
DCH 3, p. 217, records an occurrence in 3Q15 (the ‘Copper Scroll’) 8.4
with the possible sense ‘rock’, but this reading of that text now seems to have been
abandoned. DCH also cites (p. 391) D. Winton Thomas for an unpublished sugges-
tion that the word here might mean ‘court’ (with an Ar. cognate), which would fit
the rendering of LXX. Recently A. Wilson-Wright has pointed to words in Syriac,
Sabaic and Arabic meaning ‘camp’ as possible cognates for ḥîrōt, also comparing
LXX, and seen this as appropriate to a location on the ancient Egyptian military
road or route along the Mediterranean coast (‘Camping along the Ways of Horus:
A Central Semitic Etymology for pî ha-ḥîrot’, ZAW 129 [2017], pp. 261-64).
242 EXODUS 1–18
reaffirmed in locating it near to Lake Timsah (‘An Egyptological Perspec-

tive on the Exodus Narrative’, in A.F. Rainey [ed.], Egypt, Israel, Sinai
[Tel Aviv, 1987], pp. 137-61 [142-43]; followed by Van Seters, ‘Geography
of the Exodus’, pp. 274-75).21 Alternatively the name might be related to
pa-ḥwyr in the thirteenth-century description of Pi-Ramesse on P.Anastasi
III, 2.9 and 3.4 (ANET, p. 471, tentatively renders ‘the Her canal’), where it
is said to produce natron and flax for the royal palace. From the proximity
to the ‘Shi-Hor’ or ‘waters of Horus’ (also mentioned in 2.9: see ABD 5, p.
1212, for its probable location in the north-east part of the Nile Delta) and
the occurrence of minerals near the Mediterranean coast Cazelles deduced
that pa-ḥwyr was also in this region. Hoffmeier, who had earlier argued for
a Semitic meaning, ‘canal’, for the name Pi-hahiroth based on Akkadian
(Egypt, pp. 169-71), has more recently brought that view into conjunction
with this toponym and given it a specific identification with the ‘Eastern
Frontier Canal’ (Sinai, pp. 105-108). He proposes that like the other names
in Exod. 14.2 Pi-hahiroth was in ‘the area between the north side of the
el-Ballah Lake system and the southern tip of the eastern lagoon’, the latter
being not Lake Sirbonis/Bardawil (on which see below on Baal-zephon) but
an inlet on the ancient Mediterranean coast south of where Pelusium was
later to be built. If this general approach is followed, the ‘Pi’ could be a
transcription of the Egyptian definite article pa.22
Cazelles’ third suggestion, which is firmly linked to the Wadi Tumilat in
the south, is to see the name Pi-hahiroth reflected in p(a)-ḥrw in the stele of
Ptolemy II Philadelphus which was found by E. Naville at Tell el-Maskhuta
(ll. 11, 13: cf. Naville, The Store-City of Pithom [London, 1885], pp. 17-18 and
pls. VIII-X). Naville transcribed the name as Kharma, so similar to pa-ḥrm,
the name given to a canal of the eighth nome in the Edfu geographical list,
but the m(a) can be disregarded.23 According to Cazelles the name is attested
in the Ramesside era (P.Anastasi V.11.4) and the Persian period (Shalluf stele
of Darius, l. 13) as well as later.
21
A.A. Burke, ‘Magdalūma, Migdālîm, Magdoloi, and Majādil: The Histor-
ical Geography and Archaeology of the Magdalu (Migdāl)’, BASOR 346 (2007),
pp. 29-57 (35), states that Cairo 31169 may contain a text which is much older
than this copy, perhaps from the Late Bronze (or Iron) Age. The papyrus (found
at Saqqara) contains lists of several kinds and is clearly a school-book: the
geographical list at the beginning will be an extract from a larger compilation (cf.
W. Spiegelberg, Die Demotischen Denkmäler, II.2 [Leipzig, 1908], pp. 266, 270,
277-78).
22
Hoffmeier (Sinai, p. 107) therefore abandons his earlier view that it was the
Hebrew for ‘mouth of’.
23
Cazelles, Autour de l’Exode, p. 219.
14.1-31 243
Migdol
Migdol is easily recognisable as a variant of the common Hebrew word for
‘tower’, migdāl, and several such names are attested in the border region of
north-east Egypt. A place of this name was located on the main road out of
Egypt along the Mediterranean coast in several periods: reliefs of Seti I in
the Nineteenth Dynasty portray a ‘Migdol of Menmaʿrēʿ’ as the third station
on ‘The Ways of Horus’ (the classic study is A.H. Gardiner, ‘The Ancient
Military Road between Egypt and Palestine’, JEA 6 [1920], pp. 99-116; but
see also Aharoni, Land of the Bible, pp. 46-48 [with a list of the first places
on the route], and the important new studies of Hoffmeier, Egypt, pp. 183-89;
Sinai, pp. 89-105);24 the name Mag[da]li has been restored in Esarhaddon’s
account of his invasion of Egypt on his tenth campaign in 671 (ANET, p. 292);
Herodotus reports (2.159) that Necho (II) ‘attacked the Syrians by land and
defeated them at Magdolus’ (according to many a mistake for ‘Megiddo’, as
in 2 Kgs 23.29-30); a fifth-century Aramaic letter found at Elephantine (Padua
1: TAD 1/A3.3) had been written at Migdol; and a clear indication of its
location between Pelusium and Sile, twelve Roman miles from each, is given
in the sixth-century A.D. Itinerarium Antonini (O. Cuntz, Itineraria Romana,
1 [Leipzig, 1929], p. 23). Recent archaeological investigation of the area has
been unable to identify a single site which was occupied at all these periods.25
Tell el-Herr, much favoured as the site (e.g. Cazelles, Autour de l’Exode,
pp. 213, 216), has been shown by the excavations of M. Abd el-Maksoud
and D. Valbelle to have been occupied only from Persian to Greco-Roman
times and so cannot even be the site of the Migdol mentioned in Jer. 44.1;
46.14 and Ezek. 29.10; 30.6. A good candidate for the latter is Tell Qedua
(T. 21), excavated by E. Oren and D.B. Redford, which has remains from
the seventh and sixth centuries (only) (Hoffmeier, Sinai, pp. 95-96). Tell
Hebua (I), sometimes identified with Migdol in earlier times under its former
name Tell Samout,26 has extensive New Kingdom structures, but epigraphic
evidence found in 1999 and 2005 makes it virtually certain that it was the site
of the key border fortress of Tjaru in the Ramesside period (Hoffmeier, Egypt,
pp. 183-87; Sinai, pp. 92-94; Abd el-Maksoud and Valbelle, ‘Tell Héboua–Tjaru.
24
Very likely this is the place later called ‘Migdol of Ramesses prince of
Heliopolis’ in the records of Year 8 of Ramesses III (cf. Cazelles, Autour de
l’Exode, p. 211).
25
See the review of the evidence, textual and archaeological, by B.E. Scolnic,
‘A New Working Hypothesis for the Identification of Migdol’, in J.K. Hoffmeier
and A.R. Millard (eds.), The Future of Biblical Archaeology (Grand Rapids,
2004), pp. 91-120; also Hoffmeier, Sinai, pp. 94-105; and Burke, ‘The Historical
Geography…’.
26
H. Brugsch, L’Exode et les monuments égyptiens (Leipzig, 1875), p. 20; cf.
Cazelles, p. 210.
244 EXODUS 1–18
L’apport de l’épigraphie’, Revue d’égyptologie 56 [2005], pp. 1-44 [esp. 7-8,

20-22]). Sile (its Roman name) may have been at the formerly preferred
location of Tell Abu Sefe four miles to the south. Hoffmeier has excavated
another New Kingdom site, Tell el-Borg, which lies about three miles east
of Tell Hebua (Hoffmeier and Abd el-Maksoud, ‘A New Military Site on the
“Ways of Horus” – Tell el-Borg 1999–2001: A Preliminary Report’, JEA 89
[2003], pp. 169-97; Hoffmeier, Sinai, pp. 97-105). The structures, including
two successive forts, date to the Eighteenth and Nineteenth Dynasties and
much about them would favour an identification with New Kingdom Migdol.27
But Hoffmeier reasonably prefers to identify Tell el-Borg with ‘The Dwelling
of the Lion’, the station between Tjaru and Migdol in the Seti I reliefs – the
short distance between Tell el-Borg and Tell Hebua (Tjaru) favours this – and
to look for Migdol further to the south or south-east, where Oren’s survey
located some New Kingdom sites.
There clearly was, in view of the Egyptian textual evidence, a New
Kingdom Migdol on the coast road out of Egypt, even if its precise location
is currently unknown. In later times too a place of this name existed in the
region, both in the seventh and sixth centuries (Tell Qedua) and from the
Persian period to Roman times (Tell el-Herr). It would probably now be the
majority view that the Migdol of Exodus 14 and Numbers 33 is to be found
here, whether as a New Kingdom site or as a toponym from a later period,
and Hoffmeier has in his more recent book taken this view, having earlier
preferred a location east of Wadi Tumilat, to the south (Sinai, p. 108; contrast
Egypt, p. 189). Hoffmeier’s earlier view was based on the mention of Succoth
(in the Wadi Tumilat) shortly before (Exod. 12.37; 13.20) and the ‘Migdol
of Seti-Merneptah’ in P.Anastasi V, 20.2-3 (cf. ANET, p. 259), which was
apparently in the same region. He now believes, like others (e.g. Cazelles,
p. 211), that this Migdol is identical to the one near the Mediterranean
portrayed on the reliefs of Seti I. But his earlier view, which was once
widely shared (cf. Cazelles, pp. 194-97) and appears to be presupposed
in a recent article of Van Seters (‘Geography of the Exodus’, pp. 274-75),
still has something to be said for it. Kakemwere, pursuing two slaves in
the border region, reports that at ‘the keep of Tjeku’ he was told that ‘They
say in the south that they passed in the third month of Shōmu, day 10’. It
is only logical to suppose that the ‘fortress’ to which he proceeded and ‘the
stronghold’ where the fugitives were seen were in this direction: the view that
they were to the north takes for granted an identification with places known
from other texts, not this one, and one which is intrinsically improbable (cf.
R.A. Caminos, Late Egyptian Miscellanies [London, 1954], p. 257, whose
translation I quote here). Some support for the existence of not just one but
four places called Migdol/Migdal in the area east of Wadi Tumilat can also
27
Cf. Scolnic, art. cit., pp. 118-19 (but not on p. 120); Hoffmeier, Sinai, p. 99.
14.1-31 245
be found in Cairo Papyrus 31169.28 Again there is a widespread view that

the places referred to lay near the Mediterranean coast (cf. Cazelles, pp.
214-15). But this may be to over-simplify the distribution of such names
and the arrangement of the names on the papyrus seems to be against it.
Column 1 lists names in the west of the Nile Delta, column 2 includes first
names in the northern Delta and then those in the north-east, including Zeus
Casios, followed by Daphnae in l. 18. The end of column 2 comes south to
the region west of Wadi Tumilat and early in column 3 Bubastis and Ṯkw
occur. It is reasonable to see the names at the end of column 3, including the
four with the element mgdl (Eg. mktr) as lying further to the east, rather than
in the north. Cazelles invokes the names including the element grr (3.12-
13) in favour of his interpretation because they ‘naturally evoke’ the Greek
name Gerra in the vicinity of Lake Bardawil/Sirbonis. But they may well
be distinct places in a more southerly region: so at least the careful arrange-
ment of the names in this text would suggest.29 The papyrus is admittedly a
late one, from the early third century B.C., but for that period at least it is
most ‘naturally’ understood to indicate the existence of four places known as
Migdol in the region east of Wadi Tumilat, even if specific identifications for
them cannot be confidently suggested.
Baal-zephon
The name is evidently a shortened (Aramaic?) spelling of Baal-zaphon, a
deity widely attested in Ugaritic, Phoenician, Akkadian and Egyptian texts
(for details see DDD, 289-93), who in the Hellenistic period became known
as Zeus K/Casios. Although his origin lay in Syria, he was worshipped (as
the protector of Phoenician sailors) all around the Mediterranean, including
in Egypt, where the evidence reaches back into the second millennium
B.C. There he is associated with Memphis (ANET, pp. 259-60), Daphnae/
Tahpanhes (KAI 80) and Tell el- Dabaʿ (a seal of the eighteenth century B.C.)
and under his Greek name with Mount Casios on the Mediterranean coast
(Herodotus 2.6, 158; 3.5), by Lake Bardawil/Sirbonis. Early authors placed
Baal-zephon near the north end of the Gulf of Suez (Jos., AJ 2.315; Etheria,
Per. 7.4), in line with the traditional equation of the Gulf of Suez with the
sea crossed by the Israelites. Many maps followed suit.30 But some of the
28
Cf. Spiegelberg, Demotische Denkmäler, II.2, pp. 270-80, and above n. 21
on this text; also G. Daressy, ‘La liste géographique du pap. 31.169 du Caire’,
Sphinx 14 (1910–11), pp. 155-71.
29
So Daressy, pp. 166-67, with detailed textual and geographical correspond-
ences, and Burke, ‘The Historical Geography’, pp. 33, 35 cites the order as a
reason for locating the Migdal sites in ‘the eastern delta’.
30
See, e.g., J.R. Bartlett, ‘The “Way of the Wilderness” on Sixteenth-Century
Maps’, in J.K. Aitken, K.J. Dell and B.A. Mastin (eds.), On Stone and Scroll (FS
G.I. Davies; BZAW 420; Berlin and Boston, 2011]), pp. 169-91.
246 EXODUS 1–18
Egyptian and Greek evidence led scholars already in the nineteenth century to
locate Baal-zephon at Ras Qasrun on the spit which separates Lake Bardawil/
Sirbonis from the sea and to associate the events described in Exodus 14 with
similar catastrophes that later befell armies invading Egypt (esp. Brugsch,
L’exode). Excavations in the early twentieth century added further evidence
of the cult of Zeus Casios in the region and with the first publication of
texts from Ugarit, which included references to both Baal-zaphon the deity
and the sacred Mount Zaphon (Jebel el-Aqra), Otto Eissfeldt was able to
produce a synthetic study (Baal Zaphon: above p. 240) which has remained
the standard work on the subject. Eissfeldt, who was followed in this by
Noth (‘Der Schauplatz des Meereswunders’, in J. Fueck [ed.], Festschrift
O. Eissfeldt zum 60. Geburtstage [Halle, 1947], pp. 181-90 [181]), located
Baal-zephon at Maḥammadiye, at the western end of Lake Sirbonis, because
the (late) archaeological evidence was most concentrated there, but the
literary evidence favours Ras Qasrun, where the height of the eminence (‘plus
de 100 mètres’ according to Cazelles) fits Mount Casios better (Eissfeldt,
Baal Zaphon, pp. 39-43; Cazelles, Autour de l’Exode, pp. 200-203; on later
discussion see Lamberty-Zielinski, “Schilfmeer”, pp. 10-16). It is possible,
even likely, that the main cultic sites were located at places some distance
from the mountain (as at Ugarit), such as Maḥammadiye and even Pelusium,
but they did not bear the name Casios (or, presumably, Baal-zephon).
The evidence for the cult of Zeus Casios in Egypt is all from the Hellen-
istic and Roman periods, although presumably Herodotus’s use of the name
already presupposes it in the fifth century B.C. How much further back
it can be traced is a matter of doubt. Cazelles points to the name Ḫṯyn in
P.Anastasi I. 27.4, the name of a place on the coast east of the Nile Delta, as
possible evidence of it in the Ramesside period (pp. 202-203).31 But exten-
sive archaeological survey work by E. Oren in the 1970s established that
there was no occupation at Ras Qasrun or anywhere else on the spit before
the Persian period (including Maḥammadiye). This could well be because
the spit (and Lake Sirbonis behind it) did not exist much before the Persian
period: geological investigation by D. Niv concluded that it was of ‘relatively
recent’ origin.32 This does not exclude an identification of Baal-zephon with
Ras Qasrun altogether, but it does mean that it would only be possible at a
very late stage of the composition of the Pentateuch. Given the difficulties of
31
Also perhaps P.Cairo 31169 2.14-17 (cf. Daressy, ‘La liste géographique’,
161).
32
See Oren, ‘The Survey of Northern Sinai 1972–8’, in Z. Meshel and I.
Finkelstein (eds.), Qadmoniot Sinai (Tel Aviv, 1980), pp. 101-58 (Heb.), esp.
pp. 114, 122, 124; Davies, ‘Recent archaeological research’, pp. 163-66, with
English translations of key extracts.
14.1-31 247
fitting such an identification into other aspects of the biblical evidence (even
elsewhere in Exod. 14.2),33 alternative possibilities need to be considered.
A sixth-century B.C. papyrus letter from Saqqara (KAI 50.2-3) pronounces
a blessing on the recipient ‘by Baal-zaphon and all the gods of Tahpanhes’,
Tahpanhes being the later Daphnae in the north-east Delta: the text suggests
that Baal-zaphon was a (or the) major god of the city and on this basis N.
Aimé-Giron, followed by W.F. Albright and more recently M. Bietak, proposed
locating Baal-zephon there.34 It has been objected that Baal-zephon was not
the name of Tahpanhes, but it could perhaps have been part of it or nearby.
The theory of Hoffmeier that the proximity of Baal-zephon is reflected in ‘the
waters of Baal’ which are mentioned in P.Anastasi III.8 just before pa-ḥwyr
and the Shi-hor could gain some support from this later evidence (Sinai,
p. 107). Since the evidence from Memphis and Tell el- Dabaʿ noted above
need not relate to a place named Baal-zephon in their immediate vicinity,
the only other possible pointers to its location are the meagre and doubtful
evidence from the vicinity of Clysma (cf. Cazelles, pp. 204-206, 212-14)
and the restored reference to a Migdol of Baal-zaphon in P.Cairo 31169 3.22
(according to Daressy, ‘La liste géographique’, p. 169; the reading is accepted
by Redford, in Rainey [eds.], Egypt, Israel, Sinai…, p. 144), which in view of
the order of the list is more likely to be in the vicinity of Wadi Tumilat than
on the Mediterranean coast. A place of that name is not inconceivable in the
area as a shrine used by sailors on the Red Sea, especially when the latter was
connected, in whatever way, with at least the east end of the Wadi Tumilat.
As currently understood, the available data for locating the place-

names offer several possibilities of identification. Pi-hahiroth could
be either in the middle of the Suez isthmus or further north near
el-Qantara. Migdol could also be near el-Qantara or, perhaps, at the
eastern end of Wadi Tumilat. Baal-zephon could (if a late toponym)
be at Ras Qasrun on the Sirbonis spit, or near Tahpanhes or (though
the evidence is weak) at the eastern end of Wadi Tumilat. If the
evidence for the three names is taken together, Baal-zephon is the
only name with support for a location near Lake Sirbonis, which
33
See J. Simons, The Geographical and Topographical Texts of the Old Testa-
ment (Leiden, 1959), pp. 234-41, 248-50; Davies, Way of the Wilderness, pp. 81-82.
34
Cf. N. Aimé-Giron, ‘Baʿal Ṣaphon et les dieux de Taḥpanḥès dans un
nouveau papyrus phénicien’, ASAE 40 (1941), pp. 433-60 (447-60); W.F. Albright,
‘Baal-Zephon’, in W. Baumgartner et al. (eds.) Festschrift Alfred Bertholet zum
80. Geburtstag (Tübingen, 1950), pp. 1-14 (13); M. Bietak, ‘Comments on the
“Exodus” ’, in Rainey (ed.), Egypt, Israel, Sinai, pp. 163-71 (167); cf. ÄuL 10
(2000), p. 186.
248 EXODUS 1–18
makes it unlikely that the encampment was intended to be there; all

three names could be located either near el-Qantara or near Wadi
Tumilat, but the evidence for the former is much stronger in the
cases of Migdol and Baal-zephon. In both cases a plausible ‘sea’
can be found in the vicinity, Lake Ballah in the first case and Lake
Timsah or the Bitter Lakes (perhaps linked at the time to the Gulf of
Suez) in the second. If continuity with the name Succoth is sought,
the Wadi Tumilat option is much more attractive, as the journey
north to Lake Ballah is (as Redford put it) ‘most unlikely’. But these
verses could originally have belonged to a different version of the
Exodus story, so that argument need not be decisive.35 Proximity to
Succoth/Ṯkw is, however, more clearly implied in the connected
itinerary in Num. 33.1-49 (cf. vv. 6-8) and we have previously taken
the view that this was the source for the detailed itinerary-material
in Exodus and the earlier chapters of Numbers (see the Excursus
on Itineraries in the introduction to 12.28-42, 50-51 and more fully
‘…The Composition of the Pentateuch’, pp. 5-8). If one of the
more northerly locations for the toponyms in Exod. 14.2 were to
be preferred, there would be a case for adding Num. 33.7b-8a to
the sections of that chapter which we have earlier recognised as
being due to post-Priestly revision of it (on this see briefly ‘Recent
Archaeological Research’, pp. 173-74; more fully in a part of my
doctoral dissertation, published as The Wilderness Itineraries in
the Old Testament [Cambridge, 1975], which was not among the
sections of it which have been published before [pp. 57-58])36.
5-7. The account of how Pharaoh learned of the Israelites’
departure and decided to pursue them could scarcely be more
different from what vv. 3-4 would lead us to expect (their natu-
ral sequel appears only in v. 8). Pharaoh, who is at first called
simply ‘the king of Egypt’ (v. 5a: cf. 1.15, 17, 18; 2.23; 3.18-19;
5.4), is told simply that the Israelites (referred to as ‘the people’
[v. 5: cf. 13.17-18, 22 and elsewhere]) have ‘run away’, with no
hint of their curious manoeuvre; and his decision to go after them
is based not on their vulnerability but on second thoughts about
35
Van Seters, ‘Geography’, pp. 270-76, thinks it is decisive, because although
13.20 is J and 14.1-4 is P he believes that the P narrative was from the beginning
a supplement to J and so naturally would represent a consistent route.
36
For a similar view of Num. 33.6-11 and a consideration of some of its impli-
cations see Propp 2, pp. 749-50.
14.1-31 249
his release of them (v. 5b). A contradiction is sometimes seen

within v. 5 between the idea that Israel had ‘run away’ and the
statement that the Egyptians had ‘let (them) go’, as if the former
presupposed a different version of the Exodus story in which the
Israelites had escaped without being given formal permission to
leave (e.g. Noth, pp. 87-88, ET, pp. 111-12; Graupner, pp. 81-82).
But both make sense in the light of the actual request of Moses for
‘three days’ of leave (5.3 etc.): ‘run away’ is a natural expression
of the Egyptians’ realisation that the Israelites did not intend to
return (cf. Houtman, p. 259; Propp, p. 492). A better case for the
division of v. 5 can be made from the singular verbs in vv. 6-7 with
no explicit subject, which connect much better with v. 5a than
with v. 5b (Gertz, pp. 214-15). The plural subjects in v. 5b lead on
most naturally to the beginning of v. 9. Pharaoh’s change of mind
recalls his earlier withdrawals of permission for Israel to leave in
the non-Priestly version of the plague-story (8.11, 28; 9.34), so
that vv. 5-7 will be the continuation of that (and of 12.29-39 etc.).
The differences from the opening of the Priestly strand of this
chapter in vv. 1-4 are then not surprising.
Pharaoh leads his army in his chariot according to this version of
the story, as in the Egyptians’ own written and pictorial records of
their campaigns (e.g. ANET, pp. 234-38; ANEP, nos. 314-16, 322,
345; cf. Keel, Die Welt, pl. 405a = ET, ibid.). Pharaoh’s ‘people’
here will mean his army (for this sense of Heb. ʿām cf. Num. 20.20;
1 Kgs 20.10: also TWAT 6, 192 = TDOT 11, p. 176) and, in the light
of what follows, specifically the infantry. But it is the chariot force
on which attention is focused: this was the strike force of an ancient
army, not so much for its mass and speed but because it provided a
mobile detachment of archers, who stood on the chariot’s platform
beside the driver (for details see ABD 1, pp. 888-92, and more fully
M.A. Littauer and J.H. Crouwel, Wheeled Vehicles and Ridden
Animals in the Ancient Near East [Leiden, 1979]). The figure ‘six
hundred’ seems to be a common one in biblical battle narratives
(Judg. 18.11, 16, 17; 20.47; 1 Sam. 13.15; 14.2; 23.13; 27.2; 30.9;
2 Sam. 15.18) and probably bears no intentional relation to the ‘six
hundred thousand’ of 12.37: it is more significant that they are all
‘on foot’. While the Hebrew could be taken in either way (see Note s
on the translation), ancient parallels make it more likely that the ‘six
hundred’ are an elite division of Pharaoh’s chariotry (for ‘choice’ –
also of soldiers in general – see the Annals of Ramesses III, ANET,
250 EXODUS 1–18
p. 262) than that they comprise the whole of his chariot force: the
numbers recorded as captured by some Pharaohs or arrayed against
Shalmaneser III of Assyria are much larger (ANET, pp. 237, 247,
278-79), and 1 Kgs 10.26 credits Solomon with 1400.37 On the
‘officers’ in charge of the force see Note t on the translation and
15.4: the view that Heb. šālîš meant ‘the third man in a chariot’ is
less likely.
8-10. Only now is the reason for Pharaoh’s action that was
announced in v. 4 stated in the narrative (with precise verbal corre-
spondence), which confirms that the different version in vv. 5-7 has
been taken from another account, while v. 8 again comes from P.
‘Defiantly’ is literally ‘with raised hand’ (Heb. beyād rāmāh), as
in the corresponding place in Num. 33.3. The phrase is used in
legal contexts to distinguish deliberate from unintentional wrong-
doing, but the sense is the same, since deliberate defiance of God is
involved there (cf. Num. 15.30-31). The boldness of Israel (which
contrasts with what vv. 10-12 will say) will have been inspired, or
at least reinforced, by the knowledge that Yahweh was intent on the
final destruction of the Egyptians (v. 4b). The beginning of v. 9 (to
‘by the sea’), which partly duplicates what has already been said in
v. 8, must be from the non-Priestly account. The rest of the verse
corresponds to the wording of the Priestly account (see [a] vv. 4,
17, 18, 23, 26, 28; [b] v. 2) and so could be part of it, but in view
of the untidy combination of phrases which relate respectively to
the Egyptians and the Israelites it is more likely to be a redactional
amplification of the narrative at this point. The narrative of P would
then resume in v. 15 (but see the note there) or v. 16. The listing of
the components of Pharaoh’s force here is closest to that in v. 23
(which also follows the verb ‘pursued’), but ‘and his army’ is added
at the end, presumably to link up with the wording of v. 4 (compare
the inclusion of ‘all his army’ in v. 17, more logically at the begin-
ning of the list). ‘Horsemen’ is to be preferred for Heb. pārāš in this
redactional addition because the meaning ‘horses’ would be otiose
after the occurrence of another word with this meaning (sûs) just
before (see Note x on the translation). In v. 10 the viewpoint shifts
from the pursuing Egyptians to the Israelites’ catching sight of
37
For later biblical references to Egyptian chariots see Isa. 31.1; 36.9; Jer.
46.9; Song 1.9; also 1 Kgs 10.29 (though some see ‘Egypt’ there as an error: on
the problem see G.H. Jones, 1 and 2 Kings [NCB; London, 1984], 1, pp. 230-31).
14.1-31 251
them and their resultant fear. The end of the verse (‘they were very
afraid’) contrasts with the Israelites’ attitude in v. 8 and is answered
by Moses’ words ‘Do not be afraid’ in v. 13: there is no reason
why the whole of v. 10 should not be attributed to the non-Priestly
account.
11-14. The Israelites do not only ‘cry out to Yahweh’ (v. 10b: cf.
their earlier cry in 2.23; 3.7, 9 and that of the Egyptians in 11.6 and
12.30), they complain to Moses that they are now going to ‘die in the
wilderness’ (vv. 11-12). In comparison to this even serving, or being
slaves to, the Egyptians was preferable: any kind of life is better
than death. The fact and nature of the complaint are similar to those
uttered at later points in the narrative of Exodus and Numbers, both
in what are generally judged to be parts of the non-Priestly version
of the story (17.3; Num. 11.5-6 [cf. 18]; 14.3-4; 16.13-14; 21.5; cf.
Deut. 1.27; 9.7 [‘from the day you came out of the land of Egypt’!],
22, 23) and in Priestly texts (16.3; Num. 14.2; 20.3-5). But ‘dying
in the wilderness’ cannot be an anticipation of one or more of these
later crises, as the threat of the Egyptians is the immediate danger. So
G.W. Coats must be right to argue that these verses presume that the
Israelites are already ‘in the wilderness’ before they pass ‘the sea’
(see the discussion in the introduction to 13.17-22 and especially
Coats’s article ‘Traditio-Historical Character’ [pp. 256-58]), and
this is also consistent with one possible interpretation of ‘by the
way of the wilderness to the Yam Suf’ in 13.18 (cf. Notes g and h
on the translation of 13.17-22). Equally it is implied at the end of
v. 11 that the Exodus is past, though the fact that the Egyptians are
still, or appear to the Israelites to be, a threat means that any simple
thematic distinction between ‘the Exodus story’ and ‘the wilder-
ness story’ at this point has to be suspect. What is more important
is that already here that note of ‘rebellion in the wilderness’ (cf.
Coats’s book of that name [1968]), with its challenge to Israel’s
central confession of faith, is being sounded for the first time. While
it serves to present the people in a negative, unbelieving way, it also
highlights the dangers and deprivations of the wilderness journey
and the deliverances and provisions by which Yahweh repeatedly
responded to such complaints, and not least in the present case (see
below). The people’s claim in v. 12 that they had already objected
to the plan to leave Egypt when they were still there created a
problem which was already felt (and solutions attempted: see Text
and Versions) in early times, because nothing in the earlier narrative
252 EXODUS 1–18
corresponds precisely to what they say. It is sometimes suggested

that the Israelite supervisors’ criticism of Moses and Aaron in 5.21
may be meant, but that makes a rather different point and perhaps
the people are simply made to recall here an earlier objection which
was not included in the version(s) of the story which we possess.
There could be various reasons for this (cf. Noth, p. 89, ET, p. 113;
Propp, p. 495); in any case the effect is to extend the people’s
rebellion back even further and to confirm the apprehension which
Moses expressed in 4.1.38
Moses’ response to the people’s fear and complaint is on this
occasion assured and like that of a kind pastor (seelsorgerlich:
Albertz, p. 241): he brings a word of comfort which begins like
the salvation-oracle that answers a cultic lament, ‘Do not be afraid’
(cf. v. 10: J. Begrich, ‘Das priesterliche Heilsorakel’, ZAW N.S. 11
[1934], pp. 81-92; K. Koch, Was ist Formgeschichte? Neue Wege
der Bibelexegese [Neukirchen, 2nd ed., 1967], pp. 214-15, ET
pp. 175-76). But there is no direct divine speech here: Moses as
Yahweh’s servant (v. 31) takes the responsibility for countering the
people’s anxiety by once again becoming the mouthpiece for what
Yahweh is about to do (cf. 3.16-17; 4.30; 12.23). It is described as a
‘deliverance’ or ‘victory’ Heb. yešûāh: see Note gg on the translation)
and Yahweh ‘fights’ for his people (again v. 25), so the catastrophe
for the Egyptians is this time (unlike in ch. 12) given a military
character which foreshadows (or reflects) Yahweh’s involvement in
Israel’s later wars (so also in 15.3). Here the fact that, as earlier
in the Exodus narrative, Yahweh acts without any human involve-
ment (‘Keep your position’ [v. 13]; ‘you shall remain silent’ [v. 14])
certainly differentiates this episode from most of the narratives
which represent ‘Holy War’ or ‘Yahweh War’ (cf. von Rad, Heilige
Krieg; R. Smend, Jahwekrieg, chs. 1–2). But, as von Rad saw, the
non-Priestly narrative of Exodus 14 shares a number of features
as well as an overall ethos with some ‘post-Solomonic’ narratives
which likewise emphasise divine intervention and minimise human
military activity (Josh. 6; Judg. 7; in some respects 1 Sam. 17: von
Rad, Heilige Krieg, pp. 42-50), as well as with a strand in the teach-
ing of the prophet Isaiah which carried the transformation of the
38
M. Weiss, ‘Weiteres über die Bauformen’, pp. 203-204, saw it as a case of
the literary device of Rückwendung (‘flashback’), but the explicit reference to an
earlier time makes it less abrupt than other examples of this feature.
14.1-31 253
old ‘Holy War’ tradition even further (Isa. 7.1-9; 30.15; 31.1, 3-4:
von Rad, Heilige Krieg, pp. 56-62). So neither an unqualified asso-
ciation of what Moses says with the ‘Holy War’ tradition (Noth,
p. 89, ET, p. 113) nor a virtually complete dissociation (Albertz,
p. 242 with n. 19) is satisfactory: the truth lies somewhere in
between and the possibility should also be left open that the Exodus
narrative is in its special features not a result of the later develop-
ments in the ‘Holy War’ tradition but a contributor to them. These
motifs have a relationship not only to early Israelite warfare but to
wider mythological traditions of the ancient Near East (cf. Cross,
Canaanite Myth, pp. 112-41; Miller, The Divine Warrior).
15-18. The rebuke of Moses by Yahweh with which this section
begins is doubly surprising: nothing has been said of any ‘crying
out’ by Moses (although in subsequent narratives he is said to
have done so: 15.25; 17.4; Num. 12.13), and on the contrary he
has just spoken to the people in a way that one would presume to
be pleasing to Yahweh. It is often suggested that the ‘crying out’
that is criticised is that of the people in v. 10, with Moses being
seen by Yahweh as their representative (Noth, p. 90, ET, p. 113;
Houtman, p. 266; Dozeman, p. 314; Berner, p. 358). But in addition
to the different subjects of the verb, such an interpretation ignores
the exchange between the people and Moses in vv. 11-14, and it is
more likely that v. 15 comes from a different version of the story:
this would also explain the contrast between ‘keep your position’ in
v. 13 and ‘set out’ here.39 Verses 16-18 clearly belong to the same
(Priestly) strand of the narrative as v. 4, and v. 15 might do so too,
if it could be presumed that compiler omitted a variant Priestly
version of the complaint in v. 11 which attributed it to Moses.40
This is one possibility considered by Propp (p. 479), but as he
points out the verb ṣāʿaq (‘cry out’) is not used elsewhere by P:
the parallels noted above are all from non-Priestly passages.41 We
must therefore envisage the incorporation here of an extract from
a second non-Priestly account of the episode, of which there are
in fact further traces elsewhere in the text (see the notes below on
39
As will become clearer, this contrast probably reflects two very different
portrayals of what happened at ‘the sea’.
40
The Peshiṭta actually has words to this effect between vv. 14 and 15, but they
are almost certainly a secondary amplification (see Text and Versions).
41
In 2.23 P used the variant form zāʿaq.
254 EXODUS 1–18
vv. 19-21 and 25, and earlier on vv. 5 and 9). Most of v. 16 presents
further Priestly instructions from Yahweh, which are continued in
vv. 17-18. But the words ‘And you, lift up your staff’, unlike the
continuation in the rest of v. 16, have no equivalent in the fulfilment
of these commands in vv. 21-23, which otherwise follow them
very closely. These words too are therefore not likely to have been
an original part of P. Their origin in a non-Priestly version of the
story is confirmed, not only by earlier references to Moses’ staff
(see the notes on 4.1-5; 7.15-17; 9.23; 10.13), but by its mention in
17.5. Probably they originally led into vv. 19b-20 and the raising
of Moses’ staff brought about the movement of the pillar of cloud
there, not the withdrawal or dividing of the sea.
The Priestly instructions to Moses will then begin with ‘Stretch
out your hand’ in v. 16. Remarkably Moses himself is credited with
the ability to ‘divide the sea’, though it is clear that he does so only
by means of a divinely commanded action. The instruction for
Aaron in 8.1(P) is similar. Only now is Yahweh’s plan more fully
revealed, but at this stage it is still not apparent how the Egyptians’
pursuit of the Israelites into the dried up sea will end in Yahweh
winning glory over Pharaoh and his army. On the expressions used
in vv. 17-18 see Notes i and x on the translation.
19-20. The execution of Yahweh’s commands is delayed until
v. 21 and first the narrative continues with some themes which show
none of the characteristics of the Priestly narrative but do pick up
some features of the non-Priestly version(s) of the story. In v. 19 and
the first part of v. 20 (to ‘the army of Israel’) different means used by
Yahweh to reveal himself to Israel and guide them on their way are
(re)introduced. The second, ‘the pillar of cloud’, is already familiar
from 13.21-22; the first, ‘the angel of God’, has not appeared before
in a guiding role, but it will recur in 23.23, 33.2 and Num. 20.16 and
it recalls ‘the angel of Yahweh’ in 3.2 and similar manifestations of
God in Genesis (see the Explanatory Note on 3.2-3). The reference
to the angel here (‘the angel of God who…’) seems to presup-
pose an earlier mention of its guiding role, which must have been
included in a larger narrative of which only parts have been used by
the compiler of the present text. Given the use of ‘God’ rather than
the divine name, this probably originally followed 13.17-18(19).
The two parallel statements here are so similar (both begin with
the same verb ‘moved’ [Heb. nāsaʿ] and both use the expression
‘the army [here maḥnēh, lit. ‘camp’, not ḥêl, the word used in the
14.1-31 255
Priestly account in vv. 3, (9,) 17 and 28] of Israel’) that there must
be some connection between them: either one was modelled on the
other or both are dependent on a common source, whether written
or oral. In each case what had hitherto been a means of guidance
now becomes a source of protection, separating the Israelites from
the Egyptians behind them.
The translation and interpretation of the second half of v. 20 is
very difficult and uncertain (see Note qq and Text and Versions),
but MT probably represents the oldest recoverable form of the text
and it yields a sense that fits the context (cf. NRSV), viz. that when
darkness came the cloud was transformed into fire (cf. 13.21-22;
14.24) and lit up the night, with the result that the two ‘armies’ were
kept apart (how or why is not explained). The implication seems to
be that neither the Egyptians nor the Israelites were moving at this
point, which fits vv. 9 and 13-14 but perhaps not v. 15b.
21-23. Almost all of this sub-section conforms closely to the
wording of Yahweh’s instructions in vv. 16-18 and so will also
belong to the Priestly account. The main exception is in the words
‘and Yahweh made the sea flow away by a strong east wind and
made the sea dry land’ in v. 21. These words break the connec-
tion between Moses’ action and the ‘division’ of the sea in v. 16; in
addition ‘dry land’ (Heb. ḥārābāh: cf. Gen. 7.22; Josh. 3.17; 4.18;
2 Kgs 2.8) is a different word from ‘dry ground’ (Heb. yabbāšāh)
in vv. 16 and 22 (also 29), and the ‘flowing away’ of the sea is a
different scenario from the creation of two ‘walls’ of water in v. 22
(cf. v. 29).42 There is good reason, therefore, to see the middle of
v. 21 as an extract from the non-Priestly version of the story which
the compiler inserted at the most convenient point in the Priestly
account, as critical scholars have generally done. In addition the
words ‘all the horses [here Heb. sûs] of’ do not appear in vv. 17-18
and were probably inserted here by a redactor to match the wording
of 15.1 and 21. As a result the meaning of Heb. pārāš just after-
wards was shifted ftrom ‘horses’ to ‘horsemen’ (see Note x on the
translation).
24-25. Again there is no trace of the Priestly account until v. 26
(where the continuation joins on very well to the end of v. 23) and
the word ‘army’ (lit. ‘camp’), ‘a pillar of fire and cloud’ (v. 24) and
42
The language is obviously metaphorical, but it is worth observing that ‘wall’
(Heb. ḥōmāh) is the word for a city wall rather than the wall of a house (qîr).
256 EXODUS 1–18
Yahweh fighting for the Israelites (v. 25) reintroduce features of the
non-Priestly version (cf. vv. 14, 19-20). As Blum has well seen, the
latter theme on the lips of the Egyptians recalls (and even ‘fulfils’)
the fear expressed in 1.10 by Pharaoh (but on behalf of ‘us’) that the
Israelites will ‘fight’ against them – the only (but very significant)
difference being that they now recognise that it is Yahweh whom
they have to fear and not just the Israelites (Studien, p. 9). But in
the present text there is no non-Priestly account of the Israelites
or the Egyptians crossing the ‘dry land’ from which the sea had
withdrawn: presumably it was omitted by the compiler in favour of
the more dramatic Priestly version (for other explanations see the
introduction to this section). From here on the narrative concen-
trates mainly on the fate of the Egyptians, to such an extent that
many commentators have doubted whether the non-Priestly account
even included the Israelites’ crossing of the sea (on this problem too
see the introduction).
The non-Priestly version notes the progression of time again:
‘at the morning watch’, i.e. in the latter part of the night (cf. ‘all
the night’ in vv. 20-21 and ‘as morning approached’ in v. 27). For
the expression cf. 1 Sam. 11.11: ‘the middle watch’ in Judg. 7.19
suggests that the night was divided into three ‘watches’. The precise
connection between Yahweh’s ‘looking down’ and the ‘confusion’
of the Egyptians is not specified, and perhaps no external interven-
tion is to be envisaged at this point. The extant text finds one in the
‘jamming’ (or ‘removal’: see Note ww on the translation and Text
and Versions) of the chariots’ wheels in v. 25, but this assumes that
the chariots are already in motion and may have been drawn from a
different non-Priestly version of the story. Divinely caused confu-
sion or panic in the enemy army is a regular feature of ‘Yahweh
war’ narratives and laws (cf. 15.14-16; 23.27; Deut. 7.23; Josh.
10.10; Judg. 4.15; 7.22), and is explicitly connected with Yahweh’s
support for Israel in the final clause of v. 25 (which, in view of the
‘the Egyptians’, is better understood as the narrator’s comment than
as words of the Egyptians, who would have said ‘us’). The outlook
of this strand of the narrative thus continues what has been said in
vv. 13-14 (see the notes there).
26-29. The finale is introduced by a further extract from the
Priestly account, with the same instruction to Moses, which he
carries out, to raise his hand over the sea (vv. 26-27a) as in vv. 16 and
21. But there is a clear duplication in the repetition of the ‘return’
14.1-31 257
of the sea/waters in vv. 27 and 28, each time with the effect on the
Egyptians following, but in different terms. The use of the expres-
sions ‘horses’ (Heb. pārāš) and ‘army’ (here Heb. ḥêl: see the note
on vv. 19-20) points to v. 28 as containing the Priestly version: its
language also coincides in other respects with vv. 23 and 26.43 Most
of v. 27 (from ‘as morning approached’) will then be from the non-
Priestly account, with another precise specification of the timing (cf.
vv. 21-22, 24) and the Egyptians not simply ‘covered’ by the sea
but ‘shaken out’ (from their chariots?) into it.44 In the combined text
v. 27 must mean that the Egyptians have given up their pursuit of
the Israelites but are cut off by the returning waters before they can
reach the point from which they have entered the sea. But it is widely
held that in the original separate non-Priestly narrative the Israelites
did not cross the sea, so that the Egyptians would only have put
themselves in danger when they fled into its oncoming waters in the
‘confusion’ described in v. 24. In any case the combined narrative as
it stands ends with a recap of the Israelites’ crossing of the dried-up
sea in safety in the Priestly language of v. 22: the only variations are
‘gone (through)’, implying completion, instead of ‘entered’ and the
inversion of ‘on dry ground’ and ‘(through) the midst of the sea’ to
underline the contrast with the fate of the Egyptians.
The account of the destruction of the Egyptian force, which
perhaps had more than one version in the biblical tradition (certainly
there is a distinction in form between the poems in ch. 15 and the
components of the prose narrative in ch. 14, not to speak of other
passages in both prose and poetry elsewhere), has no close parallel
in ancient Near Eastern or classical literature. But there are some
accounts of military forces meeting a disaster in a watery context
in other ways, which at least end in ways that resemble the fate
of the Egyptian force in Exodus. Some are accompanied by vivid
illustrations.
43
But the final words ‘not even one of them was left’ might equally well
belong to the non-Priestly version. ‘After them’ (sc. the Israelites) is awkward in
v. 28, as the last reference to Israel was in v. 25. Did v. 29 originally stand earlier
in the text, before v. 28? Or are the words ‘who came into the sea after them’ a
secondary addition modelled on v. 23?
44
‘Morning’ is also elsewhere the regular time of Yahweh’s deliverance of his
people (cf. Ps. 46.6 and TWAT 1, 751-54 = TDOT 2, pp. 226-28).
258 EXODUS 1–18
From Egypt itself there is the portrayal, in both texts and reliefs,
of a sea-battle between the Egyptians and a naval force of Sea
Peoples (c. 1190) on the walls of the mortuary temple of Ramesses
III at Medinet Habu. The attacking force is attempting to make an
entry through the mouths of the Nile, but it is repelled and captured,
with at least some loss of life, by Egyptian forces on land and
water.45 In a different, inland, setting reliefs of Ashurbanipal from
the South-West Palace at Nineveh show a battle between Assyrians
and Elamites at a town by the river Ulai (c. 653), where warriors,
horses, weapons and parts of chariots can be seen floating in the
river. Some of the defeated Elamites were apparently driven into
the river and perished there. The main portrayal of the river scene
extends across the bottom register of three adjacent slabs: another
shows it from a different perspective.46 Finally, Lake Sirbonis on
the Mediterranean coast east of the Nile Delta was well known
to classical writers as a place of danger for travellers and armies.
Diodorus (Hist. 1.30), in describing the strength of Egypt’s natural
boundaries, reports that ‘many of those who were unfamiliar with
the perils of the lake shore had lost whole armies through losing
their way’. Later he instances the examples of Artaxerxes III in 342
B.C. (16.46) and Antigonus in 306 (20.73-74). Strabo also writes
of the flooding of the road (Geog. 1.58) and, in general terms, of
catastrophes suffered by armies here (16.758).47 Further references
to the dangers of the Sirbonis region appear in Herod. Hist. 2.6;
45
For the texts see RITA 5, pp. 27-30, also 32-34 (inscription of Ramesses III’s
8th year); the reliefs appear in H.H. Nelson, Medinet Habu I: Earlier Historical
Records of Ramesses III (OIP 8; Chicago, 1930), pls 36-37. Discussion and inter-
pretation in Nelson, ‘The Naval Battle Pictured at Medinet Habu’, JNES 3 (1943),
pp. 40-55; E. Noort, Die Seevölker in Palästina (Palestina Antiqua 8; Kampen,
1994), pp. 56-57, 64-72, 110, with numerous illustrations.
46
The technical publication of the reliefs, with introduction, is in R.D. Barnett
et al., Sculptures from the Southwest Palace of Sennacherib at Nineveh (London,
1998), pp. 94-100 with pls 286-320 (esp. a) pls 300-311; b) pls 296-297). An
earlier version of pl. 311 appeared in ANEP, no. 204: part of this is redrawn in
Keel, Bildsymbolik, pp. 206-207 (ET, p. 227) with fig. 310 and a caption citing
Exod. 15.1 and Ps. 106.11. The historical circumstances of the war against
Teumman are described in CAH2 III/2, pp. 52 and 147-54; some of the textual
evidence appears in ARAB, nos. 1068-72, and in COS 4, p. 187.
47
Strabo’s most detailed example is located much further north and relates to
an undated border dispute between the forces of Tyre and Ptolemais/Acco, when
14.1-31 259
3.5; Polybius Hist. 5.80, Strabo, Geog. 1.50; 16.760 and Plutarch,
Antony 3.48 Some historians have even regarded this as the location
of the events portrayed in Exodus 14–15: see the Excursus on 14.1-4
and n. 85 to the Explanatory Note on 15.8.
30-31. The surviving conclusion is from the non-Priestly version
of the story: if there was a corresponding Priestly conclusion it
would have taken up the language of Yahweh’s ‘winning glory
over the Egyptians’ and making them acknowledge him (cf. vv. 4,
17-18). Instead we read of Yahweh’s ‘deliverance’ of Israel as
in v. 13 and his ‘great act of power’ against the Egyptians. By a
subtle shift of meaning the Israelites are no longer ‘afraid’ of the
approaching Egyptians (vv. 10, 13) but ‘in awe of Yahweh’ (v. 31),
and now they ‘believe’ again in Yahweh and Moses ‘his servant’
(cf. 4.31). The latter designation of Moses is rare outside the Deu-
teronomic tradition, but it occurs in Num. 12.7. The repetitiveness
of the conclusion (especially ‘Israel saw’ in v. 30b and 31a) has led
to the suggestion that v. 31 is a redactional addition (so especially
H.-C. Schmitt, ‘ “Priesterliches” und “prophetisches” Geschichts-
verständnis’). But Israel seeing the Egyptians dead on the seashore
does not seem a very plausible ending of the original narrative
(although it too has at least a loose connection with v. 13), and the
dual responses of the people in v. 31 match well those attributed to
E (‘fear of God’) and J (‘faith in God’) in earlier criticism. There
may therefore be further evidence here of the merging of two older
versions in the non-Priestly account (cf. vv. 13-16, 19-20).49 The
inclusion of Moses with Yahweh as the object of Israel’s trust is
particularly notable (but cf. 4.1-9; 19.9). In the narrative context it
most naturally looks forward to Moses’ role as leader (not always
so positively regarded by the people) on the journey through the
wilderness.
what sounds like a tsunami surged over the coastal plain and dragged many of the
defeated army to their deaths (ibid.). He then adds that ‘similar phenomena’ have
occurred ‘near Mount Casios close to Egypt’ (in fact by Lake Sirbonis).
48
For extended discussion and translations of texts see O. Eissfeldt, Baal
Zaphon, esp. pp. 39-71; more briefly Herrmann, Israels Aufenthalt, pp. 83-92 (ET,
pp. 56-64).
49
For the attribution of vv. 30 + 31b and v. 31a to different sources see Smend,
Erzählung, p. 143 (cf. Eissfeldt and Beer ad loc.).
260 EXODUS 1–18
Text and Versions

‫( וידבר‬14.1) In 4QExc there is a lacuna, but space for the verse at the
beginning of a new line after a vacat following the end of ch. 13. The Vatican
ms. of TgO and one early printed ed. omit this introductory verse, as they often
do (cf. AramB 6, pp. 36-37).
‫( יהוה‬14.1) TgNmg, G(J) add ‘the Memra of’, while TgF(P) has ‫דיבריה דייי‬, ‘the
speech of the Lord’, a rarer periphrasis which also occurs in TgF(P) at 19.3.
‫( לאמר‬14.1) Sy reads wʾmr lh as in 6.29 (cf. MT at 6.2), except for 5b1
which has wʾmr for ‫ וידבר‬instead (cf. 13.1).
‫( דבר‬14.2) Sy has ʾmr, again showing its freedom in rendering verbs of
saying (see the previous note and 13.1).
‫( בני ישׂראל‬14.2) Sy dbyt ʾysrʾyl: see 11.7 and Text and Versions there.
‫( וישׁבו‬14.2) The majority of SP mss read the plene form ‫( וישובו‬contra von
Gall), but 4QExc agrees with MT. TgJ adds ‘backwards’, perhaps to exclude
the auxiliary use of ‫שׁוב‬.
‫( לפני פי החירת‬14.2) So also SP and apparently 4QExc; the transliterations
in Vulg and the Syh representation of Aq, Symm and Theod support the MT
vocalisation. LXX has ἀπέναντι τῆς ἐπαύλεως (so also in v. 9). ἀπέναντι
as an equivalent to ‫ לפני‬is rare (but cf. 30.6) and might render ‫( פי‬in the sense
‘entrance’?) as well here. ἔπαυλις is a favourite word of the LXX transla-
tors and is used elsewhere for ‫( חצר‬cf. 8.9; Lev. 25.31) and ‫( טירה‬Gen. 25.16;
Num. 31.10), either of which the Exodus translator may have read here. But
ἔπαυλις was also used as a place-name (cf. Preisigke 3, p. 294) and in later
times there was a place called Epauleum near to modern Suez (Etheria, Per.
7.4): possibly the translator equated ‫ פי החירת‬with this (cf. my Way of the
Wilderness, pp. 5-6). TgO,J and Sy rendered ‫ פי‬by its Aram. equivalent ‫( פום‬cf.
LXX at Num. 33.7), while TgN,F less obviously used ‫פונדקי‬, a loan-word from
Greek meaning ‘taverns’. For ‫ החירת‬MRI (Lauterbach, p. 188) already knew
two interpretations, ‘rocks’ and ‘freedom, license’: TgJ (with some elaboration
also found in MRI) and perhaps TgO and Sy followed the former and TgNmg
the latter. TgJ also equates the place with Tanis in the northern Nile Delta,
which makes some sense as a place to which the Israelites ‘turned back’ from
Pelusium, where TgJ located Ramesses (like Josephus and TgN,F) in 12.37 (cf.
Way of the Wilderness, pp. 12-13, 19, 21). This identification of Pi-hahiroth
could be due to the biblical reference to ‘Zoan’, the older name of Tanis, as the
site of the Exodus miracles (Ps. 78.12-13). But Pi-hahiroth was also identified
with Pithom (MRI, ibid.), and it had been equated with Tanis in 1.11 by TgJ,N,F.
‫( בין מגדל‬14.2) LXX ἀνὰ μέσον Μαγδώλου (cf. Vulg inter Magdolum)
‘preserves the contemporary pronunciation’ (Wevers, Notes, p. 208), i.e. of the
Greco-Roman period: it is also attested in the Itinerarium Antonini (third cent.
A.D.) as the name of a place twelve Roman miles from Pelusium (O. Cuntz,
Itineraria Romana I [Leipzig, 1929], p. 23).
14.1-31 261
‫( בעל צפן‬14.2) LXX Βεελσεπφών (with the Aram. vowels in Βεελ-)

and Vulg Beelsephon transcribe as the name of a place, likewise TgO and
Sy. TgJ,N,F(VN) substitute ‫טעות‬, ‘idol’ for ‫בעל‬, but probably still regard ‫ צפן‬as
a place-name; TgF(P), which adds ‫ טעותא‬after the expression, regards it all as
the name of a deity. TgJ has a long addition which sees Baal-zephon as the
supreme Egyptian god who survived the attack described in 12.12 (see also
the note on v. 3 below).
‫( נכחו‬14.2) LXX ἐνώπιον αὐτῶν (corrected to the sing. in Aq, Symm
and Theod: cf. Vulg in conspectu eius) probably had both the places ‘before’
which the Israelites were to encamp in mind. Some witnesses to TgO have the
same intention.
‫( תחנו‬14.2) LXX has the sing. στρατοπεδεύσεις for no obvious reason
(presumably the people are viewed collectively) and Aq is credited with the
same reading. But Symm, Theod and the other witnesses all agree with MT’s
pl.
‫( על־הים‬14.2) LXX’s ἐπὶ τῆς θαλάσσης can mean ‘by the sea’ (LSJ,
p. 621), and perhaps Vulg’s super mare can too. TgJ,F(P) have ‫על גיף ימא‬, ‘on the
sea-shore’, to remove any doubt about the sense of ‫על‬.
‫( ואמר‬14.3) Sy wʾmr missed the future sense of the perfect consecutive
(cf. Propp, p. 466).
‫( לבני ישׂראל‬14.3) 4QExc and one SP ms. read ‫ אל‬for ‫ל‬, as often after verbs
of saying in the sense ‘about’ (e.g. Gen. 20.2: cf. BDB, p. 40): as the easier
reading it is probably secondary. Sy, Vulg, LXXA and most of the Tgg give
the sense ‘about’, which more likely represents the true sense of MT here
(see Note e on the translation) than a divergent Vorlage. The sense evidently
escaped the original LXX translators, who paraphrased with τῷ λαῷ αὐτοῦ,
Οἱ υἱοὶ Ἰσραήλ (OL’s fili Israel is doubtless a partial correction of this), and
TgJ who introduced a reference to ‘Dathan and Abiram, the Israelites who
were left in Egypt’, an otherwise unknown tradition (cf. TgNmg, which has a
conflation of all the Targumic renderings of this verse).
‫( נבכים‬14.3) So also 4QExc and most SP mss (the variant ‫ נביכים‬reflects
the different Sam. pronunciation: GSH §82e). The meaning of this rare word
is variously represented here: ‘wandering’ (LXX, TgN,F), ‘confused’ (TgO,J: cf.
MRI [Lauterbach, pp. 190-91]), ‘confined’ (Vulg, probably inferred from the
next clause, like Rashi’s ‫כלואים‬: his comparisons with words for ‘depth’(?) are
far-fetched) and even ‘foreigners’ (Sy dnwkryyn). MRI (ibid.) also recalls the
derivation from ‫בכה‬, ‘weep’, which was common in other places where the
root appears.
‫( בארץ‬14.3) The renderings in TgN (‘on the way’) and TgF (‘in the wilder-
ness’) are perhaps the products of their (mis)understanding of ‫נבכים‬.
‫( סגר‬14.3) Most of the Vss correctly understand ‫ המדבר‬as the subj. (and
LXX added γάρ to establish a connection with the previous clause), but
TgJ,N,F(P) introduce ‘the [or “my”] idol/Baal of Zaphon’ from v. 2 as the subj.
262 EXODUS 1–18
(TgF(VN) have ‘Peor’), so that ‫ המדבר‬can be an explicit obj. of the verb. MRI
(Lauterbach, p. 191) is similar but makes Yahweh the subj. – an unlikely idea
for Pharaoh and perhaps an orthodox way of repudiating the elaboration in
the Tgg while still dealing with the grammatical ‘problem’. TgO’s ‫אחד‬, which
is usually transitive, was perhaps meant to be understood in the same way.
‫( המדבר‬14.3) TgN,F prefix ‫נגדוי‬, ‘the passes (or paths) of’, to clarify the
sense: probably TgJ originally had this reading too, in place of the unintel-
ligible ‫ נגחוי‬which is also cited in TgNmg (cf. AramB 2, p. 199 n. 7).
‫( וחזקתי‬14.4) Sy (except for 5b1) prefixes ‘The Lord said to Moses’, to
prevent the following words from being seen as the continuation of Pharaoh’s
declaration. Many SP mss (cf. Tal, Crown as well as von Gall’s apparatus) read
the Hiph. ‫ והחזקתי‬instead of the Piel, as they also do in 4.21 and 14.17. This
reflects a general tendency in SH to use the Hiphil more (cf. GSH §171cα), but
in BH ‫ חזק‬Hiphil is not used in this way. See also Text and Versions on 4.21.
‫( את־לב־פרעה‬14.4) TgJ adds ‫יצרא‬, ‘the inclination’, as usual (see Text
and Versions on 11.10). Vulg has simply ‘his heart’, as Pharaoh has been the
speaker in v. 3.
‫( אחריהם‬14.4) Sy (but not 5b1), Vulg and some LXX mss have ‘(after)
you’, which will have seemed more natural in words that Moses was to speak
to the Israelites. The more awkward reading of MT, SP and the other Vss will,
however, be the original one.
‫( ואכבדה‬14.4) The Vss mainly render by simple future passive verbs, but
TgNmg ‫( אייקר מימרי‬first person sing. impf. Pael) preserves something of the
reflexive sense (cf. TgF(P)).
‫( ובכל־חילו‬14.4) TgO,J use ‫משׁרית‬, ‘camp’, for ‫חיל‬, as they do in vv. 9 and
17 and also for MT’s ‫( מחנה‬more naturally) in vv. 19, 20 and 24: the variation
in wording and probably also the sense (cf. Note pp on the translation) are
thus obscured.
‫( מצרים‬14.4) LXX adds πάντες, ‘all’, as it does frequently (see Text and
Versions on 13.22).
‫( כן‬14.4) TgF(P) adds ‘according to his Memra (or better, “his word”)’.
‫( ויגד למלך מצרים‬14.5) TgJ has ‘The slave supervisors who went with
Israel related’ (cf. TgNmg) to explain how the report was made, alluding to a
fuller midrash in MRI (Lauterbach, p. 194), but oddly omits any mention of
Pharaoh at this point.50
‫( ברח‬14.5) LXX, TgJ,N,F and Vulg have the expected ‘fled’, but TgO and
Sy use ‫אזל‬, with the weaker sense ‘went (away)’, as TgO does sometimes
elsewhere (cf. B. Grossfeld, ‘The Relationship between Biblical Hebrew ‫ברח‬
and ‫ נוס‬and their Corresponding Aramaic Equivalents in the Targum – ‫ערק‬,
‫אפק‬, ‫אזל‬: A Preliminary Study in Aramaic–Hebrew Lexicography’, ZAW
50
The words ‫ דאזלו עם ישׂראל‬could in theory be translated ‘that the people of
Israel had gone’ and be viewed as an alternative translation (close to TgO and Sy)
of the next clause in MT (cf. AramB 7, p. 39 n. 2). But TgJ does not use ‫ עם‬in the
constr. st. before ‫ישׂראל‬, preferring the det. st. followed by ‫ בית‬or ‫( בני‬cf. vv. 3, 25).
14.1-31 263
91 [1979], pp. 107-23 [107-108]). The other instances are concentrated in

the stories of Jacob (e.g. Gen. 31.20-22) and Balaam (Num. 24.11) and are
designed to avoid criticism of respected figures: the same would apply here,
especially with the contrasting description in v. 8 to follow (see also MRI
[Lauterbach, p. 194]).
‫( לבב‬14.5) SP has ‫לב‬, the form generally used in Exodus and so probably
a case of it ‘correcting’ to the regular form.
‫( אל־העם‬14.5) SP reads ‫ על‬for ‫ אל‬here, as it did in 12.22; cf. LXX ἐπί and
Vulg super. The variation could be due to failure to understand the sense of
‫ אל‬here (on which see Note m on the translation). TgN ‫ לות‬corresponds well to
‫אל‬, but TgO ‫ ב‬and TgJ,F ‫( על‬cf. Sy ʿl) imply a hostility which TgJ emphasises
by adding ‘to evil’.
‫( מה־זאת עשׂינו‬14.5) Most of the Vss translate literally, with ‘this’ for ‫זאת‬,
but Vulg’s quid voluimus facere, ‘What did we mean to do?’ perhaps shows
some awareness of its rhetorical dimension (cf. its similar rendering in v. 11,
though there with hoc).
‫( ישׂראל‬14.5) LXX prefixes τοὺς υἱούς, no doubt assimilating to vv. 2-3.
‫( מעבדנו‬14.5) LXX, Vulg and TgN,F follow the Heb. construction closely,
but TgO,J and Sy employ nouns meaning ‘servitude’. The use of ‫ קדם‬after ‫פלח‬
instead of a direct object in TgN,F is more widespread with reference to the
worship of God, but is extended in the Pal. Tgg to the service of men (cf. Gen.
14.4 and v. 12 here in TgN and Gen. 27.29; 29.15 in TgF.
‫( ויאסר‬14.6) LXX ἔζευξεν and Vulg iunxit employ the idioms appro-
priate to the target languages. ἔδησεν in the mg of LXXF is probably from one
of the Three: cf. the similar ‘corrections’ in Gen. 42.16 and Isa. 61.1. LXX’s
addition of Φαραώ reintroduces him as the subject but is scarcely original.
‫( את־רכבו‬14.6) The Vss are divided between a precise sing. (TgO,J,F, Vulg),
which TgJ at least51 takes to refer to Pharaoh’s own chariot (with an allusion to
a comment found in MRI [Lauterbach, pp. 198-99]), and a pl., which under-
stands the Heb. collectively (LXX, TgN, Sy): SP supports MT’s sing. form.
‫( ואת־עמו‬14.6) SP, TgO,J and Sy follow MT’s text exactly: the remaining
Vss modify it in two opposite ways, LXX and Vulg by adding ‘all’ and TgN,F
by (pedantically) limiting the force to those able to fight.
‫( לקח‬14.6) LXX συναπήγαγεν and Vulg adsumpsit depart from a simple
equivalent to avoid repetition in the next verse, while TgJ adds ‘with persua-
sive [lit. soft] words’, alluding to a midrash in MRI (Lauterbach, p. 200).
‫( בחור‬14.7) All the Vss (except TgN, which has no equivalent at all)
render with words that can mean either ‘chosen’ or’choice’.
‫( וכל רכב מצרים‬14.7) Most of the Vss render the waw with ‘and’, but
Vulg does not, implying that this phrase explains the previous one. LXX καὶ
πᾶσαν τὴν ἵππον τῶν Αἰγυπτίων, ‘and all the cavalry of the Egyptians’ (cf.
Lee, Lexical Study, p. 35), actually alters its translation and so avoids repeti-
tion; Sy’s omission of an equivalent to the second ‫ רכב‬does not alter the sense.
51
According to the editio princeps: the London ms. omits the whole verse.
264 EXODUS 1–18
TgJ has a long addition explaining that the animals involved were those of the
Egyptians who heeded Yahweh’s word in 9.20 (cf. MRI [Lauterbach, p. 201]).
‫( ושׁלשׁים‬14.7) Von Gall printed the same reading for SP, but this is a
mistake, as nearly all the mss read ‫ושׁלישׁים‬, thus supporting MT’s vocalisation,
as do several of the renderings in the Vss (on which see further Salvesen,
Symmachus, pp. 89-90, and Mastin [in Note t on the translation], pp. 138-42).
LXX and Theod have και τριστάτας, a word used several times for ‫ שׁלישׁ‬but
of uncertain meaning, as it is not attested outside LXX. LSJ, p. 1823, gave
‘one who stands next to the king and queen, vizier’, but the Supplement,
p. 296, suggests that it may be a calque (coined by the Exodus translator,
perhaps) of Heb. ‫שׁלישׁ‬, designating a military officer of high rank or an
attendant on the king. Aq τρισσοί (prob.), ‘triples’, and Symm ἀνὰ τρεῖς,
‘in threes’, relate the sense to the numeral ‘three’ and may be alluding to the
interpretation taken by TgJ from MRI (Lauterbach, p. 202, where the Eng. tr.
wrongly applies ‘the third’ to an occupant of the chariot): ‘to each chariot he
[sc. Pharaoh] added a third mule to pull (it) and to pursue (them) in haste’
(AramB 2, p. 199). TgO,F and Sy have ‘warriors’, the first interpretation offered
in MRI, ibid. (see also the next note), which TgN ‫ורברבנין‬, TgNmg ‫( פולימרכין‬a
Greek loanword) and Vulg duces elevate into ‘commanders’ (see also the
next note). Of these only TgF incorporates an attempt to relate the word to the
numeral, with ‘triply armed’, which is again in MRI, ibid.
‫( על־כלו‬14.7) Several of the Vss clearly give ‫ על‬the sense ‘(with authority)
over’ by adding explanatory words: Vulg (duces) totius exercitus, TgO,N.F
‘(were) appointed’. LXX (ἐπὶ πάντων) and Sy could have this sense, but it
is not certain. Only TgJ certainly does not, but for it ‫ על‬means ‘(harnessed)
to’ rather than ‘upon’. Early evidence for a reference to chariot-riders seems
to be limited to the comments of Origen and Basil (cf. Mastin, pp. 139-40).
‫( יהוה‬14.8) TgF(P) adds ‘the Memra of’.
‫( לב‬14.8) TgJ prefixes ‫ יצרא‬as in v. 4 and elsewhere.
‫( בני ישׂראל‬14.8) TgF(P) prefixes ‫עמא‬, ‘the people of’.
‫( יצאים‬14.8) Von Gall printed this as the text of SP, but nearly all mss have
the plene form ‫יוצאים‬. LXX ἐξεπορεύοντο and TgJ ‫ נפקין‬render MT precisely,
but Vulg egressi erant (cf. TgO,N,F ]‫ נפקו[ן‬and Sy), while true to (some of)
the narrative context (cf. 12.41, 51: also Num. 33.3), would really require a
Vorlage ‫ יצאו‬to express the pluperfect (GK §142b). Even if such a Heb. text
had existed, however, it would have to be regarded as the easier reading (and
hence inferior) than that of MT and SP, which implies that the Exodus was still
continuing. TgN,F(PVN) as often add ‫פריקין‬, ‘redeemed’.
‫( ביד רמה‬14.8) LXX, Vulg, TgJ and Sy translate literally (with the
addition of ‘prevailing over the Egyptians’ in TgJ to explain it, as in one
interpretation in MRI [Lauterbach, p. 204]), but TgO,N,F(P) read ‫ברישׁ גלי‬, ‘with
head uncovered’ (another interpretation [of MT] given in MRI, ibid.: for other
occurrences see AramB 2, p. 59 n. 6). This perhaps carries the sense ‘defiantly’
(so AramB 7, p. 38), though Jastrow, p. 248, plausibly suggests ‘openly’.
14.1-31 265
‫( אחריהם‬14.9) Vulg, not content this time with a simple pron. obj. as in
v. 3, has vestigia praecedentium, ‘the footsteps of those who had gone before’,
in a stylistic flourish that picks up a Ciceronian phrase (cf LS, p. 1354).
‫( וישׂיגו‬14.9) LXX καὶ εὗρον and Vulg et reppererunt are imprecise if
widely used equivalents for ‫ נשׂג‬Hiph. The sense is better represented in TgO,J
and Sy: TgN,F(P) ‫וארעו‬, ‘and they joined, met, found’ is closer to LXX and Vulg.
‫( על־הים‬14.9) TgF(P) (but not this time TgJ) inserts ‘the shore of’ as in v. 2.
TgJ follows these words with a long addition about the Israelites collecting
pearls and precious stones carried by water from the garden of Eden: no
source for this is known.
‫( כל־סוס רכב פרעה‬14.9) SP has the same reading as MT, but the sudden
change back to a (fuller) description of Pharaoh’s force and the unique expres-
sion ‫ סוס רכב‬caused difficulties for some of the Vss. LXX καὶ πᾶσα ἡ ἵππος
καὶ τὰ ἅρματα Φαραώ makes the rest of the verse indicate where Pharaoh’s
army was now located (for further evidence of this interpretation see below on
‫ )על־פי החירת‬and treats the ‘cavalry’ and chariots as two separate components
of it: Vulg also adds ‘and’ before ‘the chariots’ for the same reason. Sy simply
uses the easier wording of v. 23: ‘all Pharaoh’s horses and his chariots’. TgN
‘all the horses, the chariots of Pharaoh’ achieves the same result without
adding ‘and’. The other Tgg render the Heb. in the most natural way and
take ‫ כל־סוס‬as in the construct state. This difficulty at least is probably due to
redactional activity (see the Explanatory Note), and LXX, Vulg, Sy and TgN
are secondary attempts to deal with it.
‫( ופרשׁיו‬14.9) Vulg has no equivalent here, probably because equitatus,
its rendering for ‫סוס‬, was taken to include the riders as well as the horses (in
v. 23, however, it has equites in a similar situation). All the other Vss give the
sense ‘and his horsemen’: LXX has no explicit equivalent for the suffix, as
one is not needed in Greek (one is supplied by Aq and Theod, followed by the
O-text). The Tgg and Sy use the Aram. cognate of ‫פרשׁ‬, which by their time
meant only ‘horsemen’, although it occurs in Old Aram. (as occasionally in
Heb.) as an alternative word for a (war?) horse: see Jastrow, p. 1243; DNWSI,
p. 945; CAL.
‫( וחילו‬14.9) Many SP mss read the pl. form ‫וחיליו‬, although in vv. 4 and
28 they have the sing. like MT: here it is probably a scribal error due to the
pl. suff. on ‫ופרשׁיו‬. TgN,F have the pl. too, in all three cases, perhaps influenced
by 12.41 where TgN uses the same form to render the pl. of ‫צבא‬. Vulg prefixes
universus, thus matching ‘all his army’ in v. 4. TgO,J render ‫ חיל‬as there (see
the note).
‫( על־פי החירת‬14.9) LXX represents ‫ על‬with ἀπέναντι as it had for ‫ לפני‬in
v. 2 (cf. TgF), but the other Vss have either ‫( על‬TgO,J, Sy) or a rendering which
represents it closely: Vulg in, TgN ‫סמיך ל‬. The recognition of the variation may
at least have assisted the reading of this part of the verse as the location of the
Egyptian army rather than of Israel (cf. the note above on ‫)כל־סוס רכב פרעה‬,
which is clearly adopted by Vulg and Sy, as their prefixing of erant and šryn
266 EXODUS 1–18
here shows, and could be intended in the other Vss (especially in TgJ with its
long insertion in the middle of the verse: cf. AramB 2, p. 199) and even in MT.
‫( ופרעה הקריב‬14.10) The SP mss make this the end of v. 9 (cf. Sy), clearly
treating it as an independent clause and the climax of the Egyptian approach.
This hardly does justice to the inversion of the normal word-order. Vulg
cumque adpropinquasset Pharaoh is a decisive witness to the opposite view,
with subordination to what follows. TgJ, picking up its rendering in v. 2, has
‘Pharaoh saw that the idol of Zaphon had been spared and offered offerings
before him’, giving ‫ הקריב‬its regular sacrificial interpretation, like one view
cited in MRI (Lauterbach, p. 205: cf. TgNmg). LXX’s προσῆγεν often bears
this meaning in Leviticus and Numbers (Wevers, Notes, p. 213), but it can also
mean ‘advance’ (LSJ, p. 1499).
‫( בני ישׂראל‬14.10) Sy reads dbyt ʾysrʾl the first time, as in v. 2, but reverts
to precision for the second occurrence (which Vulg simply omits).
‫( את־עיניהם‬14.10) LXX τοῖς ὀφθαλμοῖς (following ἀναβλέψαντες),
without αὐτῶν, but it was added by the Three and the O-text.
‫( והנה‬14.10) SP prefixes ‫ויראו‬, and this is an early addition, since it is also
found in 4Q365 (according to DJD XII, p. 118, it is unlikely to have been
in 4QExc). The Tgg all agree with MT, but LXX ὁρῶσιν (cf. Vulg, Sy) has
been seen as further evidence for the SP reading. It may, however, simply be
an interpretation of ‫( והנה‬cf. Gen. 37.29). In narrative ‫ והנה‬is often preceded
by a form of ‫( ראה‬e.g. Exod. 2.6; 3.2: cf. BDB, pp. 244, 907), but not always
(e.g. Gen. 15.12; 37.7, 9, 29). Most likely the reading of 4Q365 and SP is an
expansion of the original text here, to supply the fuller idiom.
‫( נסע‬14.10) SP has the pl. ‫ נסעים‬and again there is early evidence for
this reading, in 4Q365 and probably in 4QExc, where the scroll is damaged
but some of the letters survive. MRI (Lauterbach, p. 206) defends the sing.
reading in a way that suggests it is familiar with the alternative. The Vss
render with pl. forms (except for Vulg, which ignores the word), but this
need not mean that they are based on a Vorlage different from MT: they
may simply have been rendering according to the sense (and their render-
ings of ‫ מצרים‬in the pl.). LXX ἐστρατοπέδευσαν (a curious equivalent for
‫נסע‬, but cf. Gen. 12.9; Deut. 1.40) is often seen as supporting the SP reading
(e.g. BHS), but may actually count against it: the aorist tense is most easily
understood as based on ‫ נסע‬read as perfect. Most likely ‫ נסע‬is the original
reading (see Note bb on the translation for the use of collective singulars
in this passage) and the pl. form is an adaptation to the more common form
of expression. 4Q365 had an addition afterwards about the numbers of the
Egyptian forces (cf. v. 7).
‫( ויצעקו‬14.10) TgJ,N,F and Sy have ‘and…prayed’, presenting the Israel-
ites’ reaction not (yet) as a complaint but as a response worthy of God’s people
(cf. the long comment to this effect in MRI [Lauterbach, pp. 206-209]). LXX,
Vulg and TgO,Nmg follow MT and SP.
14.1-31 267
‫( ויאמרו‬14.11) TgJ limits this response to ‘the wicked of the generation’ as

distinct from those who ‘prayed’ (see the previous note and cf. 32.19).
‫( המבלי אין־קברים‬14.11) The Vss (except TgN,F which reproduce it with
an almost identical form, perhaps by secondary assimilation to MT) do not
explicitly represent the redundant ‫בלי‬, but the rhetorical and ironic paraphrases
in LXX (cf. LSJ, p. 1304, for παρά = ‘just because’) and Vulg may have been
attempts to do justice to it. TgJ,F,Nmg add ‘for us’.
‫( במצרים‬14.11) LXX has the fuller expression ἐν γῇ Αἰγύπτῳ here (but
not later in the verse or in v. 12), a variation which also occurs in 12.30, 40.
‫( למות‬14.11) LXX and TgN have ‘to kill (us)’, sharpening the accusation
against Moses: both versions have the same variant at 21.14. The evidence is
scarcely sufficient for this reading to be preferred to that of MT and SP (which
gains some support from ‫ ממתנו‬in v. 12), especially as the variant could well
be due to influence from 16.3; 17.3; Num. 16.13. TgF has ‘to bury us’, which
is further away from the other witnesses and pedantically based on ‘graves’
earlier in the verse.
‫( במדבר‬14.11) TgF adds ‘of Zaphon’ from v. 9: the combination is other-
wise unknown.
‫( מה זאת‬14.11) Compare the note on v. 5: here Vulg and Sy too render
‫ זאת‬by ‘this’ (hoc) or ‘thus’ (hknʾ) and make grammatical space for it by using
‘why’ (quid; lmnʾ) for ‫מה‬.
‫( להוציאנו‬14.11) LXX has simply ἐξαγαγών, with the obj. to be supplied
from the preceding ἡμῖν: the Three and the O-text add ἡμᾶς to conform to
MT.
‫( במצרים‬14.12) After this TgJ inserts the opening words of the Israelite
supervisors in 5.21, not in its own rendering but interestingly in that of TgN:
‘May the Lord be revealed against you and judge’. This correlation follows
MRI (Lauterbach, pp. 209-10), but SP preserves a different view: see the end
of the next note.
‫( חדל‬14.12) SP adds ‫נא‬: according to DJD XII, p. 118, it is not possible
to tell if 4QExc had this or not. There is no trace of it in the Vss. Perhaps SP
sought to make the Israelites more respectful towards Moses. Elsewhere it
has ‫ נא‬when MT does not in Gen. 47.19 and Exod. 12.3 (apart from places
where ‫ נא‬appears in expansions of the text), while the reverse is true in Exod.
32.32. Probably the shorter text is more original in all these cases. SP, perhaps
doubting whether the different wording in 5.21 could be meant, has inserted
the words from here to the end of the verse after 6.9, to ensure that the state-
ment has a basis in the earlier narrative (compare its other harmonisations in
chs. 7–11).
‫( ונעבדה‬14.12) LXX, Vulg and Sy treat this as the beginning of a purpose
clause, but in this case the focus on serving the Egyptians in v. 12b makes the
normal cohortative sense more likely.
268 EXODUS 1–18
‫( טוב לנו‬14.12) Vulg multo melius est strengthens the Israelites’ claim and
(by not rendering ‫ )לנו‬generalises it: see also the note below on ‫ממתנו‬.
‫( את־מצרים‬14.12) Vulg has eis, avoiding the repetition.
‫( ממתנו‬14.12) TgN,F slavishly render the Qal inf. by ‘to kill’, as LXX and
TgN did in v. 11: here it does not fit the context at all.
‫( במדבר‬14.12) LXX and Sy add ‘this’, perhaps from 16.3: in any case it
emphasises the contrast with being in Egypt.
‫( ויאמר משׁה אל־העם‬14.13) TgJ,N,F(VN),G(FF,J) begin the verse with a midrash
(also found in MRI [Lauterbach, p. 214]: see AramB 2, pp. 60 n. 9 and 200
n. 20, for further references) about the reactions of four groups of Israelites to
the Egyptian pursuit, which then divides up Moses’ words in vv. 13-14 into
separate responses to each of them. Earlier versions of the tradition occur in
Philo, Vita Mosis 1.172-75, and Ps-Philo, LAB 10.3. TgF(P),G(U) have the same
addition (with minor variations, as in the main group of texts) before 15.3: on
the displacement in these ‘festival-liturgical Targums’ see Klein, Fragment-
Targums, pp. 21-22. In all these texts there is naturally no equivalent to
‫אל־העם‬, since the people are divided into groups.
‫( אל־תיראו‬14.13) LXX, as in a few other places (e.g. 20.20), employs the
positive instruction θαρσεῖτε, ‘take courage’, to represent this expression.
‫( התיצבו‬14.13) The Vss render with words meaning ‘stand still, stand
firm’: this is also true of Tgg ‫אתעתדו‬, as its opposition to throwing themselves
into the sea in the midrash makes particularly clear (cf. CAL).
‫( את־ישׁועת יהוה‬14.13) The Tgg and Sy as often render ‫ ישׁועה‬by ‫פורקנא‬,
‘redemption’ (in the pl. in TgN,F(VN)).52 For the divine name LXX has τὴν παρὰ
τοῦ θεοῦ, with the preposition emphasising that the nomen rectum indicates
the source of the deliverance. The use of θεός is a common variation in LXX
(cf. 13.21 and Text and Versions on 13.8) and there is no direct evidence of
a variant Heb. text: 4Q365 appears, like MT and SP, to have had ‫ יהוה‬here.
TgG(FF) adds ‫ אלהכון‬and TgG(J) substitutes ‫אדני‬.
‫( אשׁר‬14.13) SP and 4Q365 read ‫כאשׁר‬. All the Vss except TgO and Vulg
render with words for ‘as’, but ‫ אשׁר‬itself can sometimes bear that meaning
(see Note hh on the translation) and they (or some of them) may simply be
rendering the MT reading (which appears in the citation in MRI [Lauterbach,
p. 214]) in that way. Vulg clearly takes ‫ אשׁר‬as a relative pronoun (quos) with
‘the Egyptians’ as the antecedent and TgO (d) may have done so too.53
‫( תסיפו‬14.13) Almost all SP mss read ‫תוסיפון‬, including those used in
recent editions: von Gall’s choice of ‫ תוספון‬is inexplicable. SP’s preference
52
Vulg uniquely has magnalia, ‘mighty acts’, for ‫ ישׁועה‬here, apparently
influenced by the description of the Exodus deliverance as ‫גדול‬, ‘great’, in v. 31
and Deut. 10.21; 11.2; 2 Sam. 7.23; Ps. 106.21.
53
Some early printed editions of TgO have ‫כמא‬, ‘as’, here (cf. TgF(P)), but only
one of Sperber’s mss agrees.
14.1-31 269
for the longer form of the imperfect here has parallels in 1.22 and 3.21, but
elsewhere the reverse variation is found (GSH §63b). Where the longer form
survives there is a case for regarding it as more original, though influence
from Aramaic (as indeed from an adjacent form [cf. the end of v. 14]) is also
a possibility where the evidence is divided, as here.
‫( לראתם‬14.13) 4Q365 seems to have had ‫לראתו‬, with a collective sing.
suffix which would conform to some other references to the Egyptians in this
chapter (vv. 10 and 25): but in v. 10, where 4Q365 is preserved, it has the pl.
reading of SP! In any case this one witness is scarcely sufficient to outweigh
the combined evidence of SP, MT and LXX for the pl. suffix here. TgN,P(VN)
add ‫בשׁעבוד‬, ‘in slavery’, qualifying the absolute statement of MT and perhaps
seeking to reconcile it with v. 30.
‫( עד־עולם‬14.13) LXX εἰς τὸν αἰῶνα χρόνον is a unique rendering of
‫ עד־עולם‬in the Pentateuch (contrast e.g. Gen. 13.15; Exod. 12.24), with only a
few parallels in the translated books of LXX (e.g. Isa. 9.6). It was evidently a
Greek idiom, but not surprisingly χρόνον was obelised by Origen and omitted
in some hexaplaric mss.
‫( יהוה ילחם לכם‬14.14) LXX, TgO, Vulg and Sy reproduce MT, with which
SP and 4Q365 also agree. On the expanded text of TgJ,N,F(NV),G(FF,J) in this verse
and of TgF(P),G(U) before 15.3 see the first note on v. 13. It is clear that MT
underlies them all, despite some variations between them.
‫( ואתם תחרישׁון‬14.14) Sy kd ʾntwn thwwn šlyn, ‘when you will be silent’,
makes these words into a condition for the fulfilment of the preceding
promise. After them most mss (not 5b1) add ‘and Moses prayed before the
Lord’ to provide a basis for (their interpretation of) the beginning of v. 15.
Both these variants seem to be special developments in the Syriac tradition:
SP, LXX, Vulg and TgO all agree with MT. 4Q365 does not preserve ‫תחרישׁון‬,
but there is space for it (and an interval after it) in a lacuna.
‫( יהוה‬14.15) TgF,Nmg add ‘the Memra of’.
‫( מה תצעק‬14.15) SP and 4QpalExl (]‫ )מה תצע[ק‬agree with MT, but 4Q365
‫ תזעק‬follows the spelling generally preferred at Qumran, as it does in 15.25
(similarly 1QIsa in five cases out of six acc. to TWAT 2, 630 = TDOT 4,
p. 114). This contrasts with SP’s alteration of the only two cases of ‫ זעק‬in
the Pentateuch to ‫( צעק‬see Text and Versions on 2.23). LXX and Vulg render
the Heb. as expected, but TgJ,N,F and Sy render ‫ צעק‬by ‘pray’, as they do in
v. 10 (cf. MRI [Lauterbach, p. 216]); these Tgg also prefix ‘stand and’. TgJ,N,F
back up the implied criticism of Moses by adding ‘look, the prayer of my
people preceded yours’, a consideration based on v. 10 and attributed to R.
Simon b. Judah in MRI (Lauterbach, p. 221: cf. Exod. R. 21.1). By contrast
all that TgO has is ‘I have accepted your prayer’, which follows the strongly
pro-Moses interpretation of R. Aḥa (also in MRI [Lauterbach, pp. 221-22],
along with much additional basis for the challenge to Moses in MT) and
totally displaces the question in the Heb. TgN,F have their own version of both
these amplifications.
270 EXODUS 1–18
‫( אלי‬14.15) Sy and TgJ,N,F have their usual ‘before’ after words for prayer.
‫( דבר‬14.15) Sy has ʾmr, ‘say’, which fits its rendering of ‫ ויסעו‬as direct
speech (nšqlwn, without waw) better.
‫( אל־בני־ישׂראל‬14.15) Sy kl byt ʾysrʾl employs its alternative designation
for the people as, e.g., in v. 10.
‫( ויסעו‬14.15) LXX καὶ ἀναζευξάτωσαν, using a rare equivalent for
‫( נסע‬as in Exod. 40.36-37). The Three evidently used (a compound of)
αἴρω here (Wevers cites their reading in Latin: et tollant), in line with LXX
elsewhere. Numbers 2.17 provides evidence for the Three preferring different
compounds of αἴρω and Aq’s choice there of ἀπαίρω, which in fact corre-
sponds to what the Sam. Gk. has here with a ἵνα to indicate purpose (cf. Vulg
ut). On ἀπαίρω and ἐξαίρω as equivalents for ‫ נסע‬see Text and Versions on
13.20.
‫( ואתה‬14.16) TgF(P) adds ‘Moses’, which was perhaps the more necessary
after its expansion of v. 15.
‫( נטה‬14.16) TgO as usual (cf. vv. 21, 26) but oddly has ‫ארים‬, ‘lift up’,
whereas the other Tgg and Sy use ‫ארכן‬, ‘incline’, which is closer to one of the
senses of ‫נטה‬. LXX ἔκτεινον and Vulg extende give the more usual sense of
‫נטה‬, ‘stretch out’.
‫( ידך‬14.16) TgJ adds ‘with it’, i.e. the staff, perhaps recalling the wording
of 8.1.
‫( ובקעהו‬14.16) LXX has ῥῆξον here for ‫( בקע‬contrast ἐσχίσθη in v. 21):
its usual meaning ‘break’ scarcely fits, but similar uses are found in Gen. 7.11
and Num. 16.31 and διαῤῥήγνυμι is occasionally found with such a sense in
classical Gk. Ms. Fb records the expected correction to σχίσον, and it may
well be from Aq, who is credited with such a change in Isa. 59.5; 63.12; Ezek.
13.13; Ps. 78.15.
‫( ויבאו‬14.16) LXX, Tgg and Sy render with verbs for ‘enter’, but only
Vulg ut gradiantur clearly expresses the idea of purpose that is probably
intended here.
‫( בתוך הים‬14.16) Vulg in medio mare fits its avoidance of the sense
‘enter’: LXX εἰς μέσον τῆς θαλάσσης, on the other hand, and presumably
the other Vss too take ‫ ב‬in the sense ‘into’.
‫( מחזק‬14.17) Most SP mss (including Tal, all of Crown’s and this time
Camb. 1846) spell ‫מחזיק‬, in line with the widespread Hiph. reading in v. 4, on
which see the note. LXX, Vulg and TgJ render with future forms, but this need
not presuppose a different Vorlage. In general the Vss in this verse follow their
renderings of the corresponding phrases in vv. 4 and 8 but here Sy uses ʿbʾ (as
in 7.22; 8.11; 9.35) in place of its more common equivalent qšʾ, which was
used in v. 4, apparently just for variety.
‫( את־לב מצרים‬14.17) 4Q365 and LXX assimilate to vv. 4-5 by adding the
expected reference to Pharaoh’s heart (for which a correction to 4Q365 has the
spelling ‫ לבב‬as in v. 5 according to MT). The expanded readings are inferior
to MT and SP, which fit well with the treatment of the Egyptians as a totality
14.1-31 271
in what follows (esp. v. 23).54 LXX also adds πάντων after τῶν Αἰγυπτίων,
as it did in vv. 4 (see the note there) and 6 and will do again in v. 18. Here
(and in v. 4) the addition could be based on ‫ ובכל־חילו‬elsewhere in the verse.
‫( ויבאו אחריהם‬14.17) Vulg ut persequantur vos again brings out Yahweh’s
purpose here (as does Sy dnʿlwn here): the content of its rendering follows
what it had in v. 4, where it was closer to MT.
‫( ברכבו‬14.17) LXX and Sy prefix ‘and’: the distinction of what follows
from Pharaoh’s ‘army’ may be connected in LXX with its rendering of ‫פרשׁיו‬.
LXX also (like Vulg) has no equivalent to the suffix, as in vv. 9-10: Aq, Theod
and the O-text add αὐτοῦ.
‫( ובפרשׁיו‬14.17) LXX uses ἵππος for ‫ פרשׁ‬here and in v. 18, in contrast to
its preference for ἱππεύς in v. 9, where the preceding ‫ סוס‬had already been
rendered by ἵππος. In v. 23, where ‫ סוס‬occurs again, it takes up another word
for ‘horse-rider’, ἀνάβατης, and it continues to use this for the occurrences
of ‫ פרשׁ‬in vv. 26 and 28 (and 15.19), where there is no such constraint. Vulg
has ‘horsemen’ (equites) throughout, as do Tgg and Sy. The effect of LXX’s
partial departure from this meaning is to give greater prominence to the
Egyptian horses themselves in ch. 14, which is reinforced by its mistranslation
of the second occurrence of ‫ רכב‬in v. 7.
‫( וידעו מצרים כי־אני יהוה‬14.18) No text from this verse is preserved in
4Q365, but there is not room for all of it. These words would fill the space
available exactly: the scribe (or author) may have omitted v. 18b as being
redundant after v. 17b. Accidental omission by homoeoteleuton (one sugges-
tion in DJD XIII, p. 267; Propp, p. 468) would not explain how only a part
of the verse was left out. SP and LXX have ‘all the Egyptians’ as they do in
the identical expression in 7.5: see Text and Versions there. 4QpalExl has a
lacuna at this point.
‫( בפרעה‬14.18) SP and Sy add (perhaps independently) ‘and his army’. In
this case 4QpalExl certainly did not have the extra text, which is undoubtedly
a secondary expansion based on v. 17.
‫( ברכבו ובפרשׁיו‬14.18) LXX, Vulg and Sy have ‘and’ before this phrase,
which is unnecessary in the Heb. and certainly secondary. LXX καὶ ἐν τοῖς
ἅρμασιν καὶ ἐν τοῖς ἵπποις αὐτοῦ omits the possessive pronoun with the
first noun (cf. Vulg), but the Three and the O-text restored conformity with the
Heb. On LXX’s use of ἵππος for ‫ פרשׁ‬here see the final note on v. 17.
‫( אלהים‬14.19) Tgg replace this with the divine name (see Text and
Versions on 13.17).
‫( ההלך‬14.19) TgO,J,F ‫( מדבר‬Pael ‫)דבר‬, ‘who was leading’, and TgN ‫דהוי עתיד‬
‫למהלך‬, ‘who was ready to go’, slightly modify the sense, perhaps because the
Israelites were now not moving.
54
4QpalExm does not preserve this part of the verse, but there would have been
no room in it for the longer text (DJD IX, p. 35).
272 EXODUS 1–18
‫( מחנה‬14.19) TgN has the pl. here and for most of the occurrences of ‫מחנה‬
in ch. 14 (cf. Text and Versions on ‫ וחילו‬in v. 9).
‫( ישׂראל‬14.19) LXX τῶν υἱῶν Ἰσραήλ makes the same addition as it had
in v. 5. TgN pedantically adds ‘from in front of them’ (as in the second half
of the verse).
‫( וילך‬14.19) Vulg abiit, TgO,J ‫ואתא‬, ‘and came’, and even TgN ‫( וקם‬again
matching the second half of the verse) need not presuppose a different Vorlage
from MT and SP (cf. LXX, Sy).
‫( מאחריהם‬14.19) LXX ἐκ τῶν ὄπισθεν is (apart from the lack of αὐτῶν,
which appears at the end of the verse and is added in the O-text here) a very
literal rendering which might seem likely to confuse a Greek reader. But,
though relatively rare, it was a Greek idiom (cf. ἐκ τῶν ὀπίσω later in the
verse and in 26.22 for ‘at the back’ and LSJ, p. 1239, for ἐκ τοῦ ὀπίσω, ‘on
the back’, in a first-cent. BC papyrus), that presumably developed from the
more widespread and older use of ἐκ δεξιᾶς and ἐξ ἀριστερᾶς, ‘on the right/
left’.
‫( עמוד הענן‬14.19) LXX prefixes καί (= ‘also’) and Vulg cum eo pariter, to
underline the parallel movements of the angel and the cloud. In general Vulg
has a very free and stylish rendering of this verse.
‫( מפניהם‬14.19) In place of this word 4Q365, which seems to have
followed MT up to this point in the verse, had a different text which began
‘from the army of Egypt to be in the army…’: the continuation is lost until
the last two words of the verse, which in essence follow MT. No other texts
have an expansion at this point (but note the addition in TgJ at the end of the
verse). For discussion see B. Nitzan, ‘Ancient Jewish Traditions of Biblical
Commentary in Qumran Literature’, in K.D. Dobos and M. Köszeghy (eds.),
With Wisdom as a Robe (HBM 21; Sheffield, 2009), pp. 288-300 (290-92).
‫( ויעמד מאחריהם‬14.19) TgO,J render the verb ‫ושׁרא‬, ‘and encamped, stayed’,
to bring out the specific sense of ‫ עמד‬here. TgJ adds as an explanation that the
cloud intercepted the arrows and stones which the Egyptians were firing at the
Israelites (the stones with catapults according to MRI [Lauterbach, p. 227]).
Sy varies its rendering of ‫מאחריהם‬, with bstr (a Persian loan-word with the
same meaning) instead of btr.
‫( ויבא‬14.20) Vulg has no equivalent and joins the rest of the first clause
of the verse to v. 19. This need not imply a different Vorlage from the other
witnesses, as Vulg often compresses the sense of its original for stylistic
reasons.
‫( מחנה‬14.20)bis As in v. 19 TgN has a pl. rendering for this word.
‫( ויהי הענן והחשׁך ויאר את־הלילה‬14.20) A full discussion of the textual
evidence and its interpretation was given by M. Vervenne (see Note qq on
the translation). Although von Gall prints ‫ והחשׁך‬as the SP reading, most SP
mss (inc. Sadaqa, Tal, all of Crown’s and Camb. 1846) omit the waw and
the two versions of SamTg agree, so SP at some stage diverged from MT.
14.1-31 273
The SamGk καὶ ἦν τὸ νέφος καὶ τὸ σκότος, however, seems to indicate
that the divergence was either not early or not total. Further evidence of the
omission of the waw exists in 4Q365 ‫הענן חושׁך‬, where the omission of the
def. art. makes a predicative interpretation of ‫ חושׁך‬even more straightforward
(cf. the widespread interpretation of the whole middle section of the verse
which is examined below). But LXX, Aq, TgO and Sy clearly represent the
waw. LXX καὶ ἐγένετο σκότος κὰι γνόφος at first seems to ignore ‫הענן‬
altogether (cf. its rendering of the next clause): σκότος generally represents
‫ חשׁך‬and γνόφος stands mostly for other words for ‘darkness’ or ‘gloom’.
But γνόφος does also render words for cloud, including ‫ענן‬/‫( עננה‬Deut. 4.11;
5.22; Job 3.5; Isa. 44.22), so it is possible that in its reformulation of this
clause (presumably to deal with its obscurity, but cf. also Josh. 24.7) LXX
was working from MT and reversed the order of its equivalents for literary
effect (elsewhere in LXX γνόφος generally follows σκότος). In the next
clause LXX καὶ διῆλθεν ἡ νύξ, ‘and the night passed’ (cf. LSJ, p. 426) is
even further away from MT and SP, where ‘the night’ is the object, not the
subject, and there is a quite different verb, ‘gave light’. Vervenne (pp. 23-
25) suggests that LXX had a Vorlage ‫( ויעבר‬as already proposed in BH3) and
that MT/SP ‫ ויאר את־הלילה‬is a gloss from the same redactor who, he believes,
added the idea of ‘light by night’ at the end of 13.21. More likely LXX,
having solved the problem of the previous clause by recourse to Josh. 24.7,
now had to find a passable sequel for it and came up with this rather bland
and redundant expression (cf. BAlex, p. 167; Propp, p. 469). There is no need
to suppose that it had anything different from MT/SP (cf. Aq, TgO, Sy, Vulg)
before it here (likewise Wevers, THGE, p. 147). Sy had added ‘all night’ to
the previous clause and has it again here in place of ‘the night’ as obj. of the
verb. This could be justified by taking ‫ את־הלילה‬as an ‘accusative of time’,
but the further variation from MT in Sy’s addition of ‘for the Israelites’
indicates that Sy has preserved here a little of the Targumic and rabbinic
interpretation which has transformed the middle of the verse into a contrast
between the Israelites and the Egyptians. The same is probably true of Vulg
et erat nubes tenebrosa et illuminans noctem, which seems contradictory as
it stands (on Symm see below). TgO goes a little further than Sy by inserting
‘for the Egyptians’ as well as ‘for the Israelites’ between ‘darkness’ and
‘gave light’ (cf. MRI [ed. Lauterbach, p. 226], and 10.23), and it too renders
‫ את־הלילה‬by ‘all night’. For the full picture, however, one must turn to TgJ
and the various forms of the Pal. Tg. The differences between them are in
the wording rather than the substance (TgNmg reproduces the wording of TgF),
and TgJ can serve as a representative of them all: ‘The cloud was in part light
and in part darkness; and on one side it made darkness for the Egyptians and
on the other side it made light for the Israelites all the night’ (the words of
MT are italicised in the translation to make clear how what were perceived
as gaps in it were filled in). Symm καὶ ἦν ἡ νεφέλη σκότος μὲν ἐκεῖθεν
274 EXODUS 1–18
φαῖνουσα δὲ ἐντεῦθεν plainly belongs here too (cf. Salvesen, pp. 90-92,
who adds further evidence). For rabbinic parallels to the additions and to
those later in the verse, see AramB 2, p. 61 n. 12. It seems likely that the only
clear Heb. variant in this part of the verse, the omission of waw before ‫(ה)חשׁך‬
in 4Q365 and most of the Samaritan evidence (cf. Symm), presupposes this
elaborate and forced interpretation and that the reading of MT is the oldest
extant textual form. Since sense can be made of it, even if with difficulty (cf.
Note qq on the translation), there is no necessity for an emendation (so also
Ska, Passage, pp. 17-18, who gives some examples of older emendations;
see further BH3 ad loc. and S. Goldman, From Slavery to Freedom [New
York, 1958], pp. 390-91).
‫( ולא קרב‬14.20) Vulg begins the clause with ut (here of result) to subor-
dinate it to what precedes, and both LXX and Vulg paraphrase the rest of the
verse while preserving the sense.
‫( זה אל־זה‬14.20) TgO follows the Heb. idiom precisely, but TgN,F and Sy
employ pl. pronouns to fit the context: similarly TgJ,Nmg ‘camp/army to camp/
army’. TgJ,N,F clarify further by adding ‘to (make) lines of battle’. TgNmg also
introduces a Talmudic tradition that ‘the angels of the service did not say the
service’ that night (sc. in heaven: cf. B.Meg. 10b and B.Sanh. 39b).
‫( ויט‬14.21) The Vss mostly render ‫ נטה‬as they do in v. 16 (see the note
there), but here Sy has wʾrym like TgO. Vulg cumque extendisset typically
subordinates to what follows.
‫( את־ידו‬14.21) LXX idiomatically has just τὴν χεῖρα (cf. Vulg), but the
Three and the O-text added αὐτοῦ. TgJ adds ‘with the staff’ from v. 16 and
also a description of it which in part parallels those which it gives at 2.21 and
4.20 (see Text and Versions there) but also goes beyond them to include the
inscribed names of the twelve tribes and their ancestors.
‫( קדים‬14.21) For the renderings of LXX, Vulg and Sy see Text and
Versions on 10.13.
‫( לחרבה‬14.21) The Vss do not distinguish between this word and ‫יבשׁה‬
(except for Aq χέρσον). LXX ξηράν ignores the ‫ל‬, for which Aq and Symm
have εἰς.
‫( ויבקעו המים‬14.21) LXX as often has the sing. ὕδωρ for ‫מים‬: the pl. τὰ
ὕδατα of Fb may well be from Aq. TgJ adds ‘into twelve divisions corre-
sponding to the twelve tribes of Jacob’, an elaboration found in other Jewish
and early Christian writings (see AramB 2, p. 201 n. 28). For the verb most
Sy mss have ʾtplgw like the Tgg, but 5b1 has ʾttrʿw, which has more the sense
of ‘break’ (cf. LXX in v. 16).
‫( בתוך הים ביבשׁה‬14.22) Vulg per medium maris sicci combined the two
phrases into one (as again in v. 29 with sicci maris to conform to the Heb.
order there).
‫( והמים‬14.22) Vulg erat enim aqua brings out a logical connection, as can
be appropriate even without waw (cf. GK §158a).
14.1-31 275
‫( חומה‬14.22) The Vss were concerned either to make explicit that there
were two walls (LXX by repeating τεῖχος; Tgg by using a pl. form) and/or
to avoid a literal interpretation by adding ‘congealed’ (TgJ), ‘as’ (Vulg, TgJ,
Sy: cf. MRI [Lauterbach, p. 237]) or ‘of water’ (TgN). TgF has ‘high (walls)’
and TgJ took the wonder of the change to extremes with ‘300 miles high’: no
parallel to this is cited.55
‫( מימינם ומשׂמאלם‬14.22) LXX did not render the suffixes: they are restored
in Aq and Symm and the O-text (though the evidence in the latter case is
more meagre than usual). Vulg was content to include eorum the first time and
Theod αὐτῶν the second, a good compromise between accuracy and style.
‫( וירדפו‬14.23) Vulg persequentes, neatly avoiding a succession of coordi-
nated clauses.
‫( סוס‬14.23) On LXX’s ἡ ἵππος (as in v. 9) see the note on ‫וכל רכב מצרים‬
in v. 7.
‫( רכבו‬14.23) Although the copula is not needed in the middle of a list
(see Note uu on the translation), 4QpalExl reads ‫( ורכבו‬cf. LXX, Sy). It is
undoubtedly secondary. LXX as often has no equivalent to the suffix: αὐτοῦ
was added by the Three and the O-text.
‫( ופרשׁיו‬14.23) On LXX’s οἱ ἀναβάται (again without αὐτοῦ: cf. Vulg)
see the note on ‫ ובפרשׁיו‬in v. 17.
‫( אל־תוך הים‬14.23) Vulg per medium maris is less appropriate here than
for the slightly different Heb. in v. 22, but in view of Vulg’s tendency to
translate freely it scarcely points to a different Vorlage from MT (with which
SP and 4QExg agree). The renderings reused in LXX and Sy fit both contexts
well. TgF(P) (but not TgF(V)) has ‫ ימא רבא‬for ‫הים‬, a phrase which it and TgG(U)
also use in 15.5, 8, again without any special reason. The expression was
used, both in BH (Num. 34.6-7) and in MH, for the Mediterranean, but that
hardly seems likely to be meant here. It occurs in 15.8 in TgJ,N,G(W), as well as
TgF(P),G(U), and this may therefore be where it was used first: TgF(P) and TgG(U)
will then have added it to other references to the Egyptians’ destruction. See
Text and Versions on 15.5 for its exegetical background.
‫( ויהי באשׁמרת הבקר‬14.24) Vulg iamque advenerat vigilia matutina looks
at first like the translation of a Heb. text which lacked the preposition ‫ב‬, but
more likely it is a typically vigorous paraphrase of MT. TgN,F have ‘time’
rather than ‘watch’ (cf. MRI [Lauterbach, pp. 237-39], which compares refer-
ences to ‘morning’ in general).
‫( וישׁקף‬14.24) The Vss use equivalents here which do not convey the
element of ‘down’ in the usual sense of ‫( שׁקף‬contrast, e.g., LXX’s preference
for compounds of κύπτω elsewhere): perhaps because the context implies
55
The readings of TgN and TgF are very similar and one may be a corruption
of the other.
276 EXODUS 1–18
that Yahweh in the column of cloud and fire is ‘at ground level’. Vulg et ecce
respiciens both expands (with ecce: cf. iam earlier) and adjusts (with a part.
in place of a finite verb: cf. the beginning of v. 23) the wording of the Heb.
for stylistic reasons.56
‫( יהוה‬14.24) TgNmg and TgF(VN) prefix ‘the Memra of’, as often. TgJ,N,F add
‘in anger’: as MRI (Lauterbach, p. 240) pointed out, ‫ שׁקף‬could refer to favour
as well as punishment and the Tgg make immediately plain which it is here.
‫( אל‬14.24) SP reads ‫ על‬and the renderings of LXX, Vulg and TgJ,N,F
interpret correspondingly, whereas 4QpalExl has ‫ אל‬like MT and TgO and
Sy agree. Both prepositions are found after ‫ שׁקף‬in BH, but ‫ על‬is much more
common (‫ אל‬only elsewhere in 2 Kgs 9.32) and this may have caused its
intrusion here.
‫( בעמוד אשׁ וענן‬14.24) 4QExg has ‫ם‬-‫] אשׁ וע‬. The mem is probably due to
what has been identified as a phonological convergence between m and n at
the end of words at Qumran (cf. Qimron, p. 27). The traces of the previous
letter would fit what DJD XIII, p. 146, calls a ‘thin letter (such as waw or
yodh)’: since nun was also a very ‘thin’ letter in mss of the same date (e.g.
4QGend), it must also be a possibility.57 TgN,F(VN) have instead of this phrase
‘and he cast upon them naphtha and fire and hailstones’, and TgF(P) and TgJ
respectively added similar material after it or within it.58 MRI (Lauterbach,
p. 245) cites Ezek. 38.22 as a parallel to this midrash, but Ps. 18.13-14 (with
the same verb ‫ המם‬as here following in v. 15) is closer and more likely to be
the source of it.
‫( ויהם‬14.24) Some SP mss (inc. Tal, Rylands 1 and Camb. 1846) read
‫ויחם‬, an unintelligible reading in the context which will be an example of a
‘relatively frequent’ (GSH §12h) scribal confusion of ‫ ה‬and ‫ ח‬in SP mss (cf.
the variant ‫ וינחגהו‬in v. 25): SamTgJ has ‫‘ = וארתע‬frightened’, supporting the
reading with ‫ה‬. Vulg curiously (and prematurely!) renders interfecit, ‘killed’,
a sense that it has for some, but by no means all, other occurrences of
‫המם‬/‫( הום‬cf. 23.27; Deut. 2.15; 7.23; Ps. 144.6; Est. 9.24). The basis for this
seems to lie in LXX’s (and OL’s) renderings of ‫המם‬/‫ הום‬in Deut. 2.15 (cf. TgO)
and 7.23, where the context could have suggested the sense ‘kill’.
56
Sy ʾtḥzy, ‘appeared’ (cf. 3.1; 16.10) is even further from MT. Propp (p. 469)
suggests influence from MH ‫ שׁקף‬Niphal (cf. Jastrow, p. 1625), which could gain
support from Sy at Num. 21.20; 23.28.
57
One way of making sense of what DJD suggests reading would be to see
here another occurrence of the obscure (and according to some corrupt) hapax
legomenon ‫ ֲעיֹם‬in Isa. 11.15, for which the sense ‘glow’ has been suggested on the
basis of Ar. parallels (cf. BDB, p. 744; HAL, p. 773; H. Wildberger, Jesaja 1-12
[BKAT; Neukirchen, 1972], p. 464).
58
TgF(P) has lost text by homoeoteleuton but enough remains to establish the
reading which it inherited.
14.1-31 277
‫( מחנה‬14.24) Vulg exercitum displaces its earlier use of castra for ‫מחנה‬,
whereas exercitus has stood for ‫ חיל‬in vv. 4, 9 and 17 (cf. v. 28 and 15.4). No
doubt the variation is due to Jerome’s understanding of the governing verb
rather than a different Vorlage.
‫( ויסר‬14.25) SP reads ‫ויאסר‬, from ‫אסר‬, ‘bind’, and LXX and Sy presume
this reading (cf. Ezek., Exag. 232-33). ‫ אסר‬is used of preparing a chariot for use
in v. 6: a different meaning for it would be required here, appropriate to wheels,
namely ‘jam’. TgJ ‫ונסר‬, ‘and he sawed’, will be based on the analysis of MT’s
consonants as derived from the Heb. cognate (the verb occurs in MH, but in
BH there is only a noun spelt ‫)משׂור‬: the sense is, however, most unlikely here.
The other Vss (and MRI [Lauterbach, p. 241]) seem all to be based on the MT
reading, understood as ‘removed’ in TgO,N,F and Symm and slightly differently
in Vulg subvertit, ‘overturned’(?). This makes good sense in itself, but removal
of the chariots’ wheels would surely cause more than ‘difficulty’ for the drivers:
they would not be able to drive them at all (cf. below on the additions in
TgJ,N.F)! MT’s reading could originally have been a defective writing of ‫ויאסר‬,
with the weak letter aleph suppressed (so J.H. Stek, ‘What Happened to the
Chariot Wheels of Exod 14:25?’, JBL 105 [1986], pp. 293-94, with parallels:
see also GK §23f; Qimron, p. 22). In this case the reading of SP, LXX and Sy
should be preferred (with HAL, p. 707; Ges18, p. 880). For recent defences of
MT see Ska, Passage, p. 19, and Houtman, p. 272.
‫( את אפן מרכבתיו‬14.25) LXX τοὺς ἄξονας, ‘the axles’, for ‫אפן‬, ‘wheels’,
is a unique rendering, perhaps accommodating MT more precisely to experi-
ence of wheeled vehicles in Egypt: corrections to the regular equivalent
appear in Aq (rotam, pres. τὸν τροχόν) and Fb (τοὺς τροχούς). The use of
the pl. is natural. The sing. suffix of ‫מרכבתיו‬, like other pronominal references
to the Egyptians in this verse, is also generally rendered in the pl. in the Vss.
Perhaps because this is a different word for ‘chariots’ (instead of ‫)רכב‬, TgJ,N.F
employ here ‫רידווה‬, whose precise significance is uncertain (contrast Jastrow,
p. 1473 with AramB 2, p. 61 n. h).
‫( וינהגהו בכבדת‬14.25) The Vss do not recognise the likely causal sense
of ‫ נהג‬Piel here and translate according to the more common meaning ‘lead,
drive’, either with Yahweh as subj. (LXX) or with the Egyptians as subj. and
their chariots as the obj. (Tgg, Sy): Vulg ferebantur as passive is imprecise
in this regard. ‫( בכבדת‬for which many SP mss have the spelling ‫בכבודות‬: cf.
the forms in MT at Judg. 18.21 and Ps. 45.14) is variously rendered ‘with
difficulty’ (TgJ), ‘with force, might’ (LXX, TgO, Sy: as a means to counter the
‘difficulty’?), ‘into the deep’ (Vulg) and ‘dragged behind them’ (TgN,F). The
two final renderings appear to be guesses from the context: the latter (based
on TgO?) leads into a long expansion (briefly alluded to in MRI [Lauterbach,
p. 241]) which ends with the chariots being thrown into the sea as in 15.4-5,
on which Vulg probably also draws (cf. in profundum again in 15.5). TgJ has
a much shorter addition ‫והוון מהלכין ושׁריין מן בתריהון‬, which AramB 2, p. 202,
renders ‘and they were gradually leaving them behind’.
278 EXODUS 1–18
‫( ויאמר‬14.25) So also 4QExg: SP and the Vss all have pl. forms to match
the natural understanding of ‫ מצרים‬as a pl. subject. TgJ,N,F add ‘to one another’
and TgN also prefixes ‘When the Egyptians saw this praiseworthy act’.
‫( אנוסה‬14.25) Again the Vss (but not SP) have pl. forms. TgF prefixed
‘But let us go back and’.
‫( ישׂראל‬14.25) TgJ,N,F prefix ‘the people of the children of’; Sy simply
‘those of the house of’.
‫( כי יהוה נלחם להם‬14.25) SP ‫הנלחם‬, so giving the sense found also in the
Tgg (with ‘the power of’ prefixed to the divine name in TgO and ‘the Memra
of’ in TgJ,F(N)): ‘It is Yahweh who fights/fought for them…’ This could be the
original reading, with the initial he lost in MT by haplography (so also Propp,
p. 470). But LXX, Vulg and Sy show no sign of the variant.
‫( במצרים‬14.25) Vulg contra nos assumes that ‫ מצרים‬is the people here,
as do LXX and TgJ, but the other Tgg and Sy render ‘in [or against] Egypt’:
TgN,F clearly intend ‘in’ as they add ‘while they were living (with us)’ and
in different ways TgN and TgF(P) then go on to speak of a new divine act (cf.
MRI [Lauterbach, p. 242]). This way of understanding ‫ במצרים‬avoids the
oddity of the Egyptians referring to themselves by name (which will also
be responsible for Vulg’s nos), but in the wider context of the chapter (esp.
v. 14) ‫ (ה)נלחם‬must surely refer to the event at the sea and not to the earlier
plagues. After the verse both 4QpalExm and 4QExg, like MT and SP, had an
interval.
‫( יהוה‬14.26) TgNmg,F prefix ‘the Memra of’.
‫( נטה‬14.26) Sy reverts to ʾrkn here after its variation in v. 21.
‫( וישׁבו‬14.26) Von Gall printed this as the SP reading, but the vast
majority of mss read the plene form ‫וישׁובו‬. Vulg recognises the final sense of
the waw with ut. LXX rendered with ἀποκαταστήτω (τὸ ὕδωρ) and used
the same verb in v. 27, but Aq found this insufficiently exact and replaced it
with ἐπιστραφήτω (from a verb that is much more common for ‫ שׁוב‬in LXX
itself): cf. Vulg. LXX then added και ἐπικαλυψάτω, perhaps following its
Vorlage, to match the outcome in v. 28, a change more characteristic of SP.
Since LXX did not render the ‫ על‬that precedes ‫ מצרים‬in MT the addition could
be regarded as a fuller representation of its meaning (cf. Propp, p. 470).
‫( על־מצרים‬14.26) Vulg ad Aegyptios is curious, especially as Vulg uses
super twice elsewhere in the verse to render ‫על‬. But perhaps style prevailed
over precision here, as it often does in Vulg.
‫( על־רכבו ועל־פרשׁיו‬14.26) LXX follows Greek idiom in having no equiva-
lent for the suffixes or for the second ‫על‬: the O-text adds ἐπί for the latter.
Vulg is content with one super at the beginning and one eorum at the end.
Third person pl. suffixes are also, understandably, used in Tgg and Sy.
‫( ויט‬14.27) While the other Vss maintain their regular equivalents for ‫נטה‬,
Sy switches again (as in v. 21) to agree with TgO’s ‫וארים‬. On Vulg’s subordina-
tion of this clause see the note on v. 21.
‫( את־ידו‬14.27) As before LXX idiomatically omits ‘his’ (cf. Vulg manum),
and the Three and the O-text add αὐτοῦ.
14.1-31 279
‫( הים‬14.27)2o LXX τὸ ὕδωρ (which it also has for the retrospective
suffix of ‫ )לקראתו‬is a little free but is harmonising with ‫ המים‬in the preceding
instruction to Moses (v. 26). Fb records the correction ἡ θάλασσα without any
attribution, but it is clear from Symm’s version of the next phrase (see below)
that he at least read a fem. noun here and most likely θάλασσα.
‫( לפנות בקר‬14.27) SP reads ‫לפנות הבקר‬, pedantically adding the def. art.
(perhaps from v. 24), but MT has support from 4QExg ‫ [לפנ]ות בקר‬and is
idiomatic. LXX πρὸς ἡμέραν and Vulg primo diluculo render the nuance of
‫ לפנות‬well (see Note yy on the translation), but Tgg and Sy’s ‘at morning time’
(ʿd[w]n) is vague and perhaps reflects unfamiliarity with this idiom.
‫( לאיתנו‬14.27) LXX, Vulg, TgF(VN) (cf. TgNmg) and Sy render ‘to its place’,
which is an apt if minimal equivalent: Symm εἰς τὸ ἀρχαῖον αὐτῆς, ‘to its
ancient, former place’ (cf. Vulg), relates better to other occurrences of ‫איתן‬.59
TgO,J,N,F(P) ‫לתוקפיה‬, ‘to its strength’, is more surprising, but is also supported
in MRI by comparisons with Num. 24.21 and Jer. 5.15. The ἐπὶ στερέωμα
αὐτῆς of Fb might mean ‘to its normal position’ in view of a use of στερεός
in an inscription from Egypt (LSJ, p. 1640).
‫( נסים לקראתו‬14.27) SP reads ‫ נסעים‬for ‫נסים‬, reproducing its reading in
v. 10, and Propp follows it (p. 470). But it is a weak alternative to MT, which
picks up ‫ אנוסה‬in v. 25. The Vss broadly support MT: but LXX ἔφυγον and
TgN ‫ ערקון‬misread the part. as indicating a punctiliar act, and Vulg recasts the
clause by making ‫ לקראתו‬into a verb.
‫( וינער יהוה‬14.27) TgF(P),Nmg prefix ‘the Memra of’ to ‫יהוה‬. For the verb
LXX ἐξετίναξεν and Sy wṭrp give the expected meaning ‘shake off’
(sc. from their chariots) and Aq’s rarer ἀνέβρασεν can probably bear a
similar sense (cf. LSJ, pp. 100, 328), perhaps (as Salvesen, ‘Midrash in
Greek?’, pp. 532-33, suggests) based on an interpretation in MRI (Lauterbach,
p. 246). Vulg involvit, ‘covered, overwhelmed’, and TgO,N(?),F,Nmg ‫שׁניק‬, ‘choked,
drowned’, reflect a search for a sense that would fit the general context better.
TgJ ‫עלים‬, ‘strengthened’, with an explanation that the Egyptians were being
kept alive for future punishment, found its solution in a correlation with ‫נער‬,
‘young man’, which is also found in MRI.
‫( בתוך הים‬14.27) SP reads ‫ תוך‬for ‫בתוך‬, which can presumably be under-
stood as an ‘adverbial accusative of place’ (GK §118g) in the same sense.
LXX μέσον (instead of εἰς μέσον in vv. 22-23) might be a variation to repre-
sent the SP reading, but unlike the apparently unique such use of ‫ תוך‬in Heb.,
μέσον alone is attested elsewhere and may simply reflect MT’s ‫בתוך‬.60 Even
59
ἀρχαῖος is the regular equivalent of ‫ איתן‬in Symm and it corresponds in
meaning and exegetical method to the citation of R. Nathan in MRI (Lauterbach,
p. 245: see Salvesen. ‘Midrash in Greek?’, p. 532).
60
DCH 8, pp. 599-600, does cite Josh. 12.2 and an emendation in Job 22.21 for
‫ תוך‬without a preposition, but neither is a secure case. In Josh. 12.2 the grammar is
obscure and ‫ תוך הנחל‬may be an addition (cf. Deut. 3.16), while the emendation in
question is not a likely solution to the problem of Job 22.21b.
280 EXODUS 1–18
without support from LXX SP should be adopted as the original reading: it

is the difficilior lectio and the other witnesses may be conforming to regular
usage.
‫( וישׁבו המים‬14.28) For LXX ἐπαναστραφὲν τὸ ὕδωρ ([aor. pass.] parti-
cipium coniunctum) see Text and Versions on 13.20. TgJ ‫ גללי ימא‬both expands
(cf. ‫ גללוי‬in v. 27) and assimilates to the more specific wording of v. 27. TgNmg
adds ‘these to these’, presumably thinking of the two ‘walls’ of water on either
side.
‫( לכל חיל פרעה‬14.28) LXX and Sy have ‘and’ for ‫ל‬, treating the ‘army’
as a separate entity as they do in v. 17 and as MT might be understood to do
in v. 9. LXX’s δύναμιν for ‫ חיל‬here (instead of στρατιά, a more obviously
military word, as in vv. 4, 9, and 17 but nowhere else) is the most widely used
equivalent.
‫( בהם‬14.28) LXX, Vulg, TgF(V) (cf. TgNmg) and Sy have ‘from them’, but
need not presuppose a different Vorlage.
‫( עד־אחד‬14.28) LXX, Vulg, TgN,F and Sy have ‘not even one’, which is
exactly the sense of ‫עד־אחד‬.
‫( הלכו‬14.29) The sense is pluperfect (cf. Note bbb on the translation).
LXX’s failure to make this explicit is not surprising, but Vulg might have
been expected to do so in view of its pluperfects in vv. 24 and 28 and often
elsewhere in Exodus.
‫( ביבשׁה בתוך הים‬14.29) Sy inverts the two phrases to agree with v. 22,
while Vulg again combines them, though with sicci before maris in partial
correspondence to the change of order in the Heb. Sy adds ʾyk d, ‘as’, before
bybšʾ and simplifies ‫ בתוך‬to b this time. TgF(P) reads ‫ מיא‬for ‫הים‬.
‫( חמה‬14.29) See the note on v. 22 for the modifications to MT in the Vss
which are repeated here. TgG(J) is extant here and has ‫רמיין‬, ‘high’, like TgF,
adding to the argument that this was the original Pal. Tg reading. Vulg has
quasi pro muro instead of quasi murus – one of several minor variations from
its wording in v. 22, apparently to avoid exact repetition in this verse.
‫( מימינם ומשׂמאלם‬14.29) On the briefer reading of LXX (to which Vulg
corresponds exactly here) and the corrections to it see the note on v. 22 (but
here there is direct evidence only for the addition of the second αὐτῶν in Aq
and Theod). After the verse TgF(P) adds an acrostic poem which is based on
vv. 15-22 and 29 (a tr. is provided in AramB 2, pp. 62-63). Further copies
of this poem appear in two Geniza mss, T (an ‘early’ ms., from the ninth–
eleventh cent.) and X (Klein, Genizah Manuscripts 1, pp. 236-39), and in a
papyrus dated to the fourth or fifth century (ibid., p. xxviii; on this see further
M. Sokoloff and J. Yahalom [eds.], Jewish Palestinian Aramaic Poetry from
Late Antiquity [Jerusalem, 1999 (Heb.)] and B.L. Visotsky in Dozeman et al.
[eds.], The Book of Exodus, pp. 552-54). For other ‘Targumic poems’ from
the Geniza which relate to vv. 29-31, see Klein, ibid., pp. 226-35 and xxviii.
‫( ויושׁע‬14.30) LXX και ἐῤῥύσατο and Vulg liberavitque are rare equiva-
lents to ‫ הושׁיע‬and are probably chosen because of the ‫ מן‬which unusually
14.1-31 281
follows ‫ הושׁיע‬here (cf. LXX at 2.17). A corrector of LXXF provided the more
regular ἔσωσεν. TgO,F and Sy render as usual with prq, but TgJ,N,G reinforced
it with ‫ושׁיזיב‬, which is the regular Targumic equivalent for ‫הציל‬: again the
following ‫ מן‬may have been a factor. The fact that ‫ פרק‬was retained alongside
it may be connected with the frequent addition of )‫ פריק(ין‬to renderings of
‫ יצא‬in the Exodus narrative: here as there ‫ פרק‬could provide the theological
nuance of ‘redemption’.
‫( יהוה‬14.30) TgN prefixes ‘the Memra of’ here.
‫( וירא ישׂראל‬14.30) Vulg disregarded the repetition of ‫ישׂראל‬.
‫( מת‬14.30) The Vss not surprisingly render in the pl. after using pl. nouns
for ‫מצרים‬. MT, with which SP agrees, is defensible (if strange) since ‫ מצרים‬can
be a collective sing. (vv. 10, 25-26) as well as pl. (vv. 17, 23). TgJ has ‫מיתין‬
‫ולא מיתין‬, ‘dying but not dead’ (cf. MRI [Lauterbach, p. 250]), alluding to its
interpretation of ‫ וינער‬in v. 27 and perhaps drawing on the occasional present
sense of the part. ‫ ֵמת‬in BH (for exx. see Ges18, p. 650). Sy kd mytyn may have
the same meaning. TgJ,N,F add ‘thrown up’.
‫( שׂפת‬14.30) LXX and Sy follow the Heb. metaphor, while Tgg and Vulg
use straightforward words for ‘shore’.
‫( וירא‬14.31) TgJ,N,F,G have a pl. verb (as for the same expression in v. 30).
Vulg retains only the ‘and’, attaching what follows as a second object to the
previous occurrence of the verb.
‫( את־היד הגדלה‬14.31) LXX τὴν χεῖρα, Sy ʾydʾ and Vulg manum follow
the Heb. wording exactly: all three words were used metaphorically to mean
‘force, power’. The Tgg bring out this sense in two different ways, either by
prefixing ‘the power of’ (TgO,J) or by rendering ‫ הגדלה‬by ‘mighty’ (TgN,G). TgF
combines the two, as in effect does a freer version of the phrase included in
TgNmg.
‫( אשׁר עשׂה יהוה‬14.31) TgJ (cf. TgNmg) eases the grammar and identifies
the nature of the actions by taking ‫ היד‬literally as ‘the hand’ and rendering
‘with which the Lord performed miracles’. The version in TgNmg attributes
them to ‘the Memra of the Lord’, as does TgG which (like TgF) describes
the action as ‘vengeance’ or ‘punishment’. The other Vss represent the Heb.
straightforwardly.
‫( במצרים‬14.31) Most of the Vss render ‘against the Egyptians’ (Vulg with
typical economy contra eos), but TgJ,N have ‘in Egypt’.
‫( וייראו העם את־יהוה‬14.31) LXX and Vulg have a sing. verb here, with the
sing. subj. adjacent, but are content to follow MT’s pl. in the next clause. Tgg
and Sy found no problem with MT’s pl. here or there, but avoid as elsewhere
(e.g. 1.17, 21) the directness of ‘fearing Yahweh’ and use ‘before’ (‫ )מן קדם‬or
‘from’ (mn) instead. Only Abraham, it seems, could be described as a ‘fearer
of God’ (Gen. 22.12).
‫( ביהוה‬14.31) LXX, as sometimes elsewhere (e.g. 8.25-26), has τῷ θεῷ
for the divine name. Here it may simply be to avoid a third occurrence of
κύριος in one verse. In any case LXX’s solitary variation in such cases hardly
282 EXODUS 1–18
suggests a divergent Vorlage, let alone a superior one. The Tgg (but not Sy)
treat faith as much as fear as something that cannot be directly related to
God: TgO prefixes ‘the Memra of’ (which in TgO probably means no more
than ‘the word of’: cf. Chester, Divine Revelation, pp. 309-10), while the
other Tgg have ‫בשׁ(ו)ם מ(י)מריה די׳‬, where as elsewhere (cf. ‫ )בשׁ[ו]ם‬Memra
will have a somewhat more developed theological connotation (Chester, ibid.,
pp. 310-13).
‫( ובמשׁה‬14.31) The Tgg all insert ‘the prophecy of’, to exclude the idea ‘of
the people displaying faith in a human being’ (AramB 7, p. 40 n. 13).
C h ap t er 1 5 . 1 - 2 1
T wo S ong s C el ebr ati n g th e D e li v e r a nc e

of the I s r ael i tes an d th e D e st r u c t i on
of a n E g y p ti an F or c e
Modern analyses generally see 15.1-18 as a complete unit, but the

Masoretic and Samaritan mss extend it further, in the former case
to the end of v. 19 and in the latter generally to the end of v. 21.1
No intervals of the regular kind occur within the section in either
tradition, but the text is in various ways broken up into short units
by the overall types of layout that are described below. The classic
Masoretic layout, exhibited in both Codex L (B 19a) and BM
Or. 4445, has a very old basis, as divisions which it presupposes
are already found (so far as they survive) in 4QExc and 4QRPc
(365). Neither of these Qumran mss preserves the beginning of the
chapter, but the recognition of a break at that point is indirectly
indicated by the ‘excerpted text’ of 4QExd. In v. 19 the opening
‘For’ links at least this verse to what has preceded, and v. 19 formed
the conclusion to vv. 1-18 in the Greek Odes.2 In most Samaritan
mss the ‘bicolumnar layout’ of vv. 1-18 is continued to the end of
v. 21 (so e.g. in Camb. Add. 713, 714 and 1846). The itinerary-note
in v. 22a indicates a clear change of scene and the beginning of
a new unit, and both 4Q365 and MT have a division here. Their
divisions after v. 19 correspond to the introduction of new characters
(Miriam and the women), even though the subject-matter and (see
the final paragraph of the introduction) the theological emphasis are
essentially the same as in v. 19. Given the fact that v. 19 is itself
closely tied to vv. 1-18 and that the similarity of v. 21b to v. 1b
1
Von Gall records one ms. (B) which has a qiṣṣāh after v. 19.
2
That v. 19 is in prose is the general view of modern scholars, but the classic
layout of the Song is continued into v. 19 in most Masoretic mss and its poetic
character was defended by Ibn Ezra. Nahmanides, however, disagreed. The occur-
rence of short intervals within the verse in 4QExc and 4Q365, as in the preceding
verses, suggests that at Qumran it was regarded as poetry.
284 EXODUS 1–18
can hardly be accidental, the thematic unity of 15.1-21 is the more

evident and it is only the inclusion of two separate (albeit related:
see below) poems, with their different performers, that has made
further sub-division attractive.
The Layout of Exodus 15.1-21 in Jewish and Samaritan Manuscripts

The oldest evidence for the layout of the passage comes from two Qumran
mss, 4QExc and 4QRPc (365). 4QExc survives for parts of vv. 10-21. The
text is not laid out stichometrically, but there are short intervals at five points
in the preserved portions of vv. 14-20: after ‫ וירגזו‬in v. 14, after ‫ יהוה‬in v. 16,
after v. 18, after ‫ הים‬in the middle of v. 19 and after v. 19. Reconstruction of
the lost text strongly suggests that there were other intervals both within and
between the verses as they were subsequently defined (DJD XII, p. 118). But
the sub-division of the text was not as complete as in later mss: there were
no intervals, for example, after v. 12, after v. 15, after ‫ כאבן‬in v. 16, after
v. 16, after ‫ מקדשׁ‬in v. 17 and after v. 17. Verses 20-21 were apparently not
sub-divided at all (one wonders whether this was also true for v. 1 or at least
for v. 1a). 4QRPc preserves parts of vv. 16-20 and has intervals after ‫ נחלתכה‬in
v. 17, after v. 18, in v. 19 after ‫ עליהמה‬and ‫ הים‬and at the end of v. 19 (where
its text is longer than MT and SP). Similar intervals can be reconstructed at
other points in vv. 16-19, but there is no sign of any division within v. 20. In
this ms. there are no places where an interval attested in later sources is known
to have been omitted, but the interval after ‫ עליהמה‬in v. 19 is not found in
Masoretic sources (though it does appear as a mid-line break in the Samaritan
ms. Camb. Add. 1846). So far as it goes, the fragmentary evidence from
Qumran indicates that the process of distinguishing sections within vv. 1-
19 had already begun, with some intervals at least corresponding to later
scribal analysis. The two mss, which are of much the same date, also seem to
represent respectively a more limited and a more far-reaching sub-division of
the text, but the background to this variation can scarcely even be conjectured
from the limited evidence available.3
Well-known Masoretic mss from the tenth cent. A.D. onwards presented
the section 15.1-19 in a special decorative way, which, apart from what is
now recognised as prose in vv. 1a and 19, made use of a complete analysis
of it into short blocks of text consisting of semantically coherent units of
3
Although none of 15.1-21 is extant from 4QpalExm, calculations based on
the surrounding columns have suggested that it ‘could have had the more open
arrangement characteristic of poetry’ (DJD IX, p. 90). For the layout of poetry
in general at Qumran and a comparison with later practice see Tov, Scribal Prac-
tices, pp. 166-76. His classification of 4QRPc as stichometric and 4QExc as not is
curious, but he makes the character of 4QRPc clear on pp. 173 and 175. 4QExd
seems to have been laid out in uninterrupted lines (DJD XII, p. 128), but only a
few words of the passage are extant.
15.1-21 285
between three and five words. Alternating lines of two different patterns
created an artistic arrangement of text and space. One line began at the right
margin with a single word from the end of the previous block of text, had
the next block at the centre and ended at the left margin with a single word
that opened the following block. The remainder of the latter then stood at
the right margin of the next line, with empty space in the centre of the line
and all but the last word of the next block at the left margin. The alternating
pattern repeated itself to the end of v. 19, after which an empty line preceded
vv. 20-21, which were written in the normal way (cf. Tov, Textual Criticism,
pp. 212-13 and pl. 12; Oesch, Petucha, pp. 121-22). This pattern is alluded to
in B.Meg. 16b, which states that ‘All the songs’ were written in this way; in
fact only Judg. 5.2-30 is written exactly like Exod. 15.1-19, as other Talmudic
passages recognise (B.Men. 31b; Soferim 1.11). Manuscripts displaying this
pattern include a tenth- to eleventh-cent. Torah scroll in a private collection,
the Bologna scroll (twelfth–thirteenth cent.), Codex BM Or. 4445 (late ninth–
tenth cent.), Codex Firkovitch II.17 (early tenth cent.), Codex Sassoon 1053
(tenth cent.), the Damascus Codex (late tenth cent.), Codex Leningrad B 19a
(early eleventh cent.) and probably the Aleppo Codex (as reconstructed here
from Maimonides, Mishneh Torah: early tenth cent.).4 Recently it has come
to light in a considerably older scroll fragment which has been dated to the
seventh or eighth cent., the Ashkar-Gilson manuscript, which contains parts
of Exod. 13.19 – 16.1.5 P. Sanders has observed that in addition to the layout
of the Song this older manuscript also anticipates, without the use of artificial
devices, the line-divisions of the preceding prose text which became largely
standard in the major medieval manuscripts with the help, where necessary,
of empty lines and spaces in the text.6 With the support also of other kinds of
evidence, Sanders argues that the scroll from which the London and Ashkar-
Gilson fragments come was used as a model or exemplar for the Torah in the
later vocalised manuscripts.
In Samaritan Pentateuch mss von Gall distinguished two main kinds of
layout of the passage.7 Most of the mss which he used have what he calls
4
On these manuscripts see briefly Tov, Textual Criticism, pp. 46-47; more
fully Penkower, New Evidence, and Sanders, ‘The Ashkar-Gilson Manuscript’,
5-13. Although they are based on the Leningrad Codex, BH3 and BHS do not
follow its layout of the Song: it can, however, be found in the printed edition of
Dotan.
5
See the article of Sanders in the previous note. A further, better preserved,
fragment of what seems to be the same scroll, containing Exod. 9.18–13.2 and
known as ‘the London manuscript’, was published by S.A. Birnbaum, ‘A Sheet of
an Eighth Century Synagogue Scroll’, VT 9 (1959), pp. 122-29. See also Volume
1, pp, 7-8.
6
Art. cit., pp. 13-18, citing the remarks of Maimonides, Mishneh Torah 2.7.10.
7
Hebräische Pentateuch, p. 145 n.
286 EXODUS 1–18
a ‘poetical’ arrangement. His list of these includes Camb. UL Add. 713 and
Oxford, Bodleian Or. 138. Each line of text from v. 1 to v. 21 is divided into
two by a space in the middle, but without attention always being given to the
sense of the passage when the location of the breaks between lines or within
them was being determined. Sometimes a phrase or even a word is split by
the dividing space: the Oxford ms. has more instances of this than the earlier
(late twelfth cent.) Cambridge example.8 Better than either is Camb. UL Add.
1846, which is even older (early twelfth cent.).9 In it most of the verses begin
on a new line. But even here there are some odd divisions. A truly stichometric
arrangement, comparable to that widely used for poetic texts in Masoretic mss
(but not, as we have seen, for Exod. 15), is much rarer and confined according
to von Gall’s survey to two somewhat later mss, Sassoon 30 (c. 1400) and
Gaster 800 (1509/10). In these vv. 1a and 19-21a are written as normal lines
of text, while the remainder of the passage is laid out with two short units of
text in each line, forming two columns with a larger space between them. Von
Gall followed this layout in his edition (as did Sadaqa in part): most of the
units consist of two words, but some contain three or four. Every verse except
v. 18 begins on a new line, and the overall layout is remarkably similar to that
which is favoured today.
A few Samaritan mss do not follow either of these patterns. Among them
are Camb. UL Add. 714 (dated to 1220), which has the Hebrew and its Arabic
translation in wide columns side by side, and Bodleian, Pococke 5 (early
fourteenth cent.), a tiny ms. which has the whole of 15.1-21 on a single page in
a distinctive chequer-board pattern, or more precisely a series of diamonds of
text separated by criss-crossing diagonal lines of space. The blocks of letters
in them usually do not even match word-boundaries. The diamonds seem to
be what A.D. Crown calls ‘lozenges’: he refers to another ms. (Bible et Terre
Sainte, BZ 10: thirteenth cent.) which has this layout for Exodus 35, presum-
ably the list in vv. 11-20, and to a twentith-cent. ms. in Sydney which has it
for Exodus 15.10 The intention is purely decorative and perhaps display: one
wonders whether the pattern may reflect that of some Samaritan synagogue
inscriptions, for example.11
8
For the date of Camb. UL Add. 713 see von Gall, Hebräische Pentateuch,
p. xxxi; Crown, Samaritan Scribes, pp. 169-70.
9
For the date see von Gall, Hebräische Pentateuch, p. lxxxiv; Crown, Samar-
itan Scribes, p. 169.
10
Samaritan Scribes, pp. 58-59.
11
See the catalogue of Samaritan inscriptions by M. Baillet in DBS 11, 860-74
(but there is no reference to such a pattern there). Crown provides some possible
analogies (Samaritan cursive and Arabic mss; synagogue floor decorations) in
the article ‘Art of the Samaritans’, in A Companion to Samaritan Studies, pp. 30,
33. Von Gall’s ϡ (Or.2685 in the British Library) is part of a scroll with its
15.1-21 287
Form and Structure

The section unusually includes two songs, one much longer than
the other, each celebrating Israel’s deliverance from the Egyptians
(vv. 1b-18, 21b) and each with its own short narrative introduction
(vv. 1a, 20-21a). In addition, a brief conclusion (v. 19) has been
added to the first song. The song in vv. 1b-18, sung by Moses and the
Israelites, is the only extended piece of poetry in Exodus (for other
poems incorporated into the Pentateuch see Gen. 49.2-27; Num.
23.7-10, 18-24; 24.3-9, 15-24; Deut. 32.1-43; 33.2-29 and some
shorter passages): it exhibits the standard characteristics of Hebrew
poetry, with balanced lines, examples of parallelism and an absence
of ‘prose particles’ such as the consonantal definite article and the
definite object marker, as well as initial repetition which is found in
some other poems (for details see the Explanatory Notes). A variety
of attempts have been made to sub-divide the poem into stanzas and
strophes (for the terminology and its varied use see Watson, Poetry,
pp. 11-15), generally on the basis of changes in content and genre,
but sometimes according to metrical structure and (presumed)
function. An influential example of the use of the latter criteria was
James Muilenburg’s ‘A Liturgy on the Triumphs of Yahweh’, which
began from what he called the ‘hymnic refrains’ in vv. 6, 11 and
16b and went on to divide up the intervening and surrounding text
into a sequence of sections of equal or approximately equal length
and poetic form: vv. 2-3, 4-5, 7-8, 9-10, 12-14, 15-16a, 17 (vv. 1b
and 18 were treated as a shorter ‘introit’ and ‘coda’ to the poem).12
The approach was taken up and the metrical analysis elaborated by
D.N. Freedman (who was for a time a colleague of Muilenburg) in
two essays on the passage, originally published in 1967 and 1974,
in which he detected a ‘pyramidal structure’ in the poem, with
v. 11 (the longest of the three ‘refrains’) as the apex.13 Much of his
sub-division corresponded to that of Muilenburg, but he ‘improved’
continuation in ms. κ in St Petersburg: it has vv. 1-21 identified as a sub-section

but without any difference in its format from the surrounding narrative. Three
other Samaritan mss are apparently unusual in their presentation of Exod. 15: in
von Gall’s list δ, ι (both also fragments of scrolls) and w.
12
For details of this and other publications specifically on vv. 1-18 see the
special bibliography at the end of this introduction.
13
‘The Song of the Sea’ and ‘Strophe and Meter in Exodus 15’, both reprinted
in Freedman, Pottery, Poetry and Prophecy (Winona Lake, 1980).
288 EXODUS 1–18
the symmetry of the poem by treating vv. 3-5 and 17-18 as its
‘opening’ and ‘ending’ and v. 2 as the original ‘exordium’; vv. 1b
and (controversially) v. 21 were regarded as liturgical additions
from the time when the poem was taken up in worship at Jerusalem
under David or Solomon (Pottery, p. 195).14
The weak point of these analyses has always been their failure
to do justice to the clear change of topic in v. 13. If v. 11 is the
refrain that closes a section, then v. 12 begins a new section. Yet it
is hard to deny that it is related to the subject-matter of vv. 1b-10
(the destruction of the Egyptian army, which must be the antecedent
of the pronoun ‘them’ in v. 12) and not to that of vv. 13-17(18) (the
Israelites’ onward journey and the terror of peoples in Canaan and
Transjordan).15 Muilenburg (pp. 245-46) and Freedman (Pottery,
pp. 190, 209-10) recognise this and seek to overcome the difficulty,
but their arguments are inconclusive and sometimes tend rather
to undermine their case. So while vv. 6 and 11 undoubtedly play
an important role in the rhetoric of the poem and might be called
‘hymnic’ (this is less clearly the case for the subordinate clauses in
v. 16b), this does not mean that they provide the key clues to the
structure of the poem. It is better, in company with those such as
Cross (Canaanite Myth, pp. 126-27), Howell (‘Exodus 15.1b-18’)
and M.S. Smith (Pilgrimage Pattern, pp. 207-12) who give priority
to subject-matter rather than form, to start from a division of the
poem into two parts, vv. 1b-12 and vv. 13-18.16 There are of course
14
Van der Lugt, ‘The Wave-like Motion’, has proposed a division of the poem
into four main sections, with the ‘hymnic’ lines regarded not as refrains but as
openings to a new section. His sub-divisions produce a series of strophes with a
common length but with little relation to the content, especially at the mid-point.
15
The suggestion, occasionally made, that ‘swallowed…up’ relates to the fate
of the Israelite rebels in Num. 16 is totally against the requirements of the context.
16
This was the starting-point in the initial study of Cross and Freedman
(Studies, pp. 50-53; ‘The Song of Miriam’, pp. 241-42). It is surprising to find that
in Zenger’s helpful chart presenting a selection of earlier analyses of the poem
(‘Tradition und Interpretation’, p. 455) only Cross could be cited as supporting this
major division. But Smith was also unable to find any support for it before Cross
(and Childs): Pilgrimage Pattern, p. 207 n. 12 (p. 211 n. 26 should be ignored).
Subsequently it has been taken up by Propp and Schmidt (p. 644, with an emphasis
on cross-links between the two sections). It was also assumed by several scholars
who regarded v. 13 as the beginning of a later addition to the original poem, on
whom see below, p. 293.
15.1-21 289
features that tie these two parts of the poem into a single whole.
The ‘hearing’ (v. 14) which causes peoples to fear is explained by
‘the greatness of your arm’ (v. 16), which can only mean Yahweh’s
defeat of the Egyptians at the sea, as the earlier mentions of his
‘right hand’ (vv. 6, 12) confirm. On the level of assonance the similar
verbal forms ‘you stretched out’ (nāṭîtā, v. 12) and ‘you led’ (nāḥîtā,
v. 13: cf. ‘you brought’, nēhaltā) create a link and a number of lexical
and thematic correspondences have been observed between the two
sections (cf. Smith, Pilgrimage Pattern, pp. 212-14). Both end, as
Smith has noted, with a single four-beat line (ibid., pp. 211-12).
Within each section it is certainly possible to recognise short
sections distinguished by the classic criteria of form criticism,
such as the presence or absence of direct address, the grammatical
subject and different kinds of hymnic language (‘descriptive’ and
‘declarative’ praise) and Howell presents a convincing analysis
along these lines. So Part I comprises v. 1b (poet is subject, in first
person), vv. 2-3 (hymn, with Yahweh as subject in the third person),
vv. 4-5 (narrative), vv. 6-7 (hymn, Yahweh now in second person),
vv. 8-10 (narrative), v. 11 (hymn), v. 12 (narrative). In Part II
v. 13 is narrative with Yahweh as subject (still in the second person)
and Israel as object, vv. 14-16 are narrative with foreign peoples
as the subject or object, with a transition in v. 16b to v. 17, which
is narrative with Yahweh as subject and Israel as the initial object,
and v. 18 is hymnic, with Yahweh as subject in the third person. The
sub-sections are short as well as uneven in length (as Smith, p. 212
n. 31, observes) and continuities between them make it possible
and desirable to join some of them together. Thus in the first two
sub-sections of Part I (vv. 1b-3), which serve to introduce its main
themes, the poet speaks in the first person (vv. 1b, 2) and Yahweh is
mentioned in all three verses in the third person. Verses 6-7 (whose
less specific language, now addressed to Yahweh, should perhaps
be read as narrative rather than as general descriptive praise) arise
out of and respond to the concrete narrative of vv. 4-5 and may be
grouped with them. We are inclined to see v. 8 as a new beginning
(as do both Howell and Propp [pp. 520-21]), not only because of
the renewed focus on more specific details of the episode that is
celebrated but because it probably (like v. 9) backtracks to events
that preceded the destruction of the Egyptian army (see the Explana-
tory Note on v. 8). This narrative section continues to its conclusion
290 EXODUS 1–18
in the summary of v. 12, with the interruption in v. 11 by what is

undoubtedly general descriptive praise responding to what has now
been narrated in fuller detail. So vv. 8-12 comprise a third compo-
nent of Part I.
Part II should probably be seen, apart from v. 18 which is the
hymnic conclusion to both it and the poem as a whole (see more
fully below on the poem’s ancient Near Eastern background and
theological significance), as a single narrative section, framed by vv.
13 and 16b-17 (cf. ‘you led in…brought [them]’ with ‘you brought
them in and planted them’; and ‘the people for whom you inter-
vened’ and ‘the people of whom you had taken possession’). Verses
14-16 present the journey through the wilderness and the settlement
in Canaan as a process to which the neighbouring peoples could
offer no resistance and as such match the detailed narrative sections
of Part I.
Metrically the poem is dominated by a pattern which can be (and
has been) analysed into groupings of cola with either two or four
sound-units each (most often a sound-unit is a single word, but
occasionally two that are closely associated).17 This pattern (which
is also present in three other poems that are likely to be early, the
‘Song of Deborah’ [Judg. 5], David’s Lament [2 Sam. 1.19-27] and
Ps. 29) appears in vv. 1b, 4, 6-12, 13, 15-18 and possibly in vv. 2-3
and 5. Analysis into short cola is preferred by H. Schmidt, Cross and
Freedman18, Muilenburg, Coats, Childs, Smith and B.D. Russell,
whereas longer cola have been identified by e.g. Dillmann, Baentsch
(p. 130), Driver (p. 129), Gunkel, Cassuto (p. 173), Mowinckel,19
Noth, Zenger (apparently) and Propp. One might hope to deter-
mine the correct view by examining verses which can be analysed
according to one alternative but not the other, and by observing
whether parallelism operates between longer or shorter units. The
17
The subject of metre in Hebrew poetry remains highly controversial, with
conflicting views not only about how it is to be described and recognised but about
whether it exists at all: see the cautious review by Watson, Poetry, pp. 87-113,
whose main recommendations have guided what follows.
18
Actually Freedman in his independent work often reckons with 4 + 4 (+ 4)
lines: see Pottery, pp. 207-208, 210, 212-15.
19
So Muilenburg, p. 238: in The Psalms in Israel’s Worship (Oxford, 1967),
2, pp. 164-65, Mowinckel states, rather surprisingly, that this (and not 3 + 3) was
‘the usual metre’ in Hebrew poetry, but he makes no statement about Exod. 15 in
particular.
15.1-21 291
presence of tricola made up of short (‘2’) units would be a decisive

argument in favour of short units. Verse 5 might be read as such,
but the second colon could be seen as a ‘3’, with ‘like a stone’ as a
single unit (so e.g. Muilenburg, p. 241; Freedman, Pottery, p. 203).
Norin divides v. 9 into two tricola (p. 98; cf. Howell, p. 28[?]), but
this is by no means certain, as the verse falls just as easily into three
(bi)cola and this seems to be the more common view. There seems
to be no compelling example of a 2 + 2 + 2 tricolon in the poem.20
On the other side Propp draws attention to vv. 4 and 15b as places
where short cola are problematic (pp. 516, 535). The difficulty
with the overlong first half of v. 4 should probably be resolved by
emendation (see the Explanatory Note), while in v. 15b it is possible
to see a ‘3’ rather than a ‘4’.21 As for parallelism, it is only the
‘synonymous’ variety (in a broad sense) that is likely to be able to
bear the weight of an argument, and here it is striking that, whereas
there is only one case where one ‘2’ parallels another (v. 11b: v. 9
might provide further possibilities, but the senses of the short units
are not really equivalent), there is a whole series of instances where
parallelism occurs between longer units (i.e. 4s: cf. vv. 4, 6, 7, 8, 10,
11a, 13, 15a, 16b, 17aβb). This does seem to provide a good reason
for preferring the analysis into longer cola.
There are a few places where a different metrical pattern appears
or may do so. 3 + 3 metre, which is widespread elsewhere (in
Ugaritic [Watson, pp. 96-97] as well as in Hebrew), appears in
v. 14 and probably v. 16b (on the structural role of these lines see
the Explanatory Note) and possibly in vv. 2-3. In v. 8 the second
and third stichs could be 3 + 3 (so Cross and Freedman) rather than
4 + 4, even though the line begins with a ‘4’. Verse 5 is probably
2 + 3 (see above), with the unusual metre bringing the description of
the Egyptians’ watery fate to a climax (so Howell, p. 20: note also
the chiasmus of verb and noun).22 Finally, v. 15b might be viewed
20
It might be thought that an isolated sequence of four units (as in vv. 12 and
18, and perhaps in v. 3) would have to be divided 2 + 2, but in view of Watson’s
long discussion of monocola with varying lengths and functions (pp. 168-74) this
is not the case.
21
Some describe v. 15b as ‘pure prose’, not poetry at all (Norin, p. 99; Zenger,
p. 464).
22
See further G.B. Gray, The Forms of Hebrew Poetry (London, 1915), pp.
176-82, who described such lines as ‘exceedingly rare’, but occasionally used.
Another possibility is 2 + 4 (ibid., pp. 182-85).
292 EXODUS 1–18
as an isolated ‘3’, in parallel to the preceding ‘4’ stichs. A special

case of the ‘regular rhythm’ of the poem is the so-called staircase
parallelism that has been detected by many in vv. 6, 11 and 16b and
may also appear in v. 13 (see on this feature Watson, pp. 150-56).
But we have seen above that this is not rigidly tied to the functions
commonly attributed to it.
Variations in metre such as those just described are among the
criteria that have been used to identify secondary additions to
the poem (so e.g. Cross and Freedman, Studies, p. 54; Norin, Er
Spaltete das Meer, pp. 98-100; Zenger, ‘Tradition und Interpreta-
tion’, pp. 460-64). While such additions (especially perhaps lines
transferred from other contexts) might betray themselves by their
metrical form, such divergence should not automatically be seen as
a sign of composite authorship. As Watson has emphasised, ‘The
most noticeable aspect of Hebrew metre when described in accen-
tual terms…is that no single poem is consistently written in one
metrical pattern’ (Poetry, p. 98; cf. Noth, p. 98, ET, p. 123). In fact
there has been a striking agreement, even in the modern period,
about at least the substantial unity of the poem, among propo-
nents of both an early and a late (post-exilic) date for it. Often it
is only a single verse or part of a verse that has been thought to
be later or earlier than the rest. A positive case for the unity of the
poem is made by e.g. Baentsch, Muilenburg, Howell and Smith
(Pilgrimage Pattern, pp. 212-14, 219-22) and the literary and
lexical arguments are reinforced by the fact that it is the poem as a
whole which displays a wide range of parallels with the Baal-cycle
of myths from Ugarit, however exactly these are to be explained
(see e.g. Cross, Canaanite Myth, p. 142; Kloos, Yahweh’s Combat,
pp. 150-52, 205-12; and more fully the discussion below). The
identification of later additions was initially due to the desire to find
an older ‘core’ which preceded the poem in its present form. Thus
Wellhausen, who was inclined to view the poem as part of the E
source (Composition, p. 77) but saw what he took to be a reference
to Jerusalem in v. 17 as incompatible with this, suggested that v. 17b
was secondary (cf. Prolegomena, p. 22 n. 1, ET, ibid.). Dillmann,
on the other hand, who had no problem dating the completed poem
to the Judges period, saw vv. 1b-3 as an even older core, perhaps
(as others were to say of the Song of Miriam) contemporary with
the Exodus itself. The close resemblance of v. 1b to v. 21 has often
15.1-21 293
led to it being seen as the original basis of the rest of the poem
(Holzinger, Smend; cf. Carpenter/Harford-Battersby and McNeile,
who assigned the whole of v. 1 to J and dated the rest of the poem
to the post-exilic period). A more substantial addition to an older
core was envisaged by Strack and Driver (vv. 12-17: prior to 1903
[cf. Baentsch, p 128]), G. Adam Smith, The Early Poetry of Israel
in its Physical and Social Origins (London, 1912), p. 51: vv. 13-18;
so also Hyatt, p. 163, with these verses presupposing the erection
of Solomon’s temple) and Dozeman (vv. 13-17: in ‘The Song of
the Sea and Salvation History’, pp. 95-101, alleging evidence of
Deuteronomistic language in an older poem; in his commentary
[p. 334] such suggestions are confined to a footnote). Some poetic
analyses have treated v. 2 (Cross and Freedman, Studies, p. 54) or just
v. 2a (Cross, Canaanite Myth, p. 127 n. 49) or vv. 1b-2 (Freedman,
Pottery, pp. 181-82, 193-95) as outside the original structure of
the poem. Others have gone further: Norin found later additions in
vv. 2, 4-5, 8, 14, 15b and 16b (Er Spaltete das Meer, pp. 96-98: cf.
p. 94 n. 70 for some predecessors), Zenger more boldly saw only
vv. 1b, 5, 6-7, 11-12 as the three original strophes, finding metrical
and historical objections to the rest (‘Tradition und Interpretation’,
pp. 460-64, 472-73). Jeremias by contrast retained all but vv. 2,
4-5, 14 and 15b (Königtum, pp. 98-99) and Spieckermann all but
vv. 14-16 (Heilsgegenwart, pp. 105-107). Most recently Albertz
assigned vv. 1-5 to his ‘Hexateuch redactor’ and the rest to an older
hymn (amplified in vv. 14 and 15b: pp. 231, 235), largely on form-
critical grounds, while A. Klein found the original core of the poem
in vv. 1-3, 6-11a, 13, 17-18 and 21b (‘Hymn and History’), with
specific historical references largely eliminated. In addition to over-
reliance on metrical analyses in some cases, these newer theories
suffer from too great a determination to find (or rather create) a
rigid consistency of form or concept and (like some conclusions
about the date of the poem which will be examined below) an
over-confidence about the direction of dependence between similar
biblical passages. For example, just as good a case can be made for
the dependence of Ps. 118.14 and Isa. 12.2 on v. 2 (cf. B.D. Russell,
pp. 19-20; H.G.M. Williamson, Isaiah 6–12 [ICC; London, 2018],
pp. 727-30) and the influence of vv. 14-16 on Deuteronomy 2 and
Joshua 2 (cf. Russell, pp. 139-41) as vice versa. The sporadic and
by no means universal attempts to disprove the unity of the passage
294 EXODUS 1–18
as a poetic composition can scarcely be said to have produced any

agreed results, as the preceding survey has shown. Our further study
of it can proceed provisionally to treat it as a single unit.
The variations in form within the poem have already been noted
in our earlier discussion of its structure. The opening in the first
person sing. (vv. 1b-2), which is probably to be understood as a
collective representation of the worshipping congregation, is not
sustained in the explicit formulaic language of the poem, but that
is not surprising when the focus quickly turns to the events of an
earlier generation and perhaps the distant past. At first, no doubt
following the format of the (more) ancient hymn (cf. v. 21b) on
which it is based, the poet speaks of Yahweh in the third person
(so still in v. 4, after the emphatic declarations of v. 3). Verse 5, in
which there is no mention of Yahweh, allows a transition. From v. 6
there is a change to direct address to Yahweh, which continues
almost to the end, in v. 17: in v. 18 the congregation revert to the
third person in a bold declaration that probably owes its form as
well as its content to the so-called Enthronement Psalms (cf. Pss.
93.1; 97.1; 99.1). Such variation is not uncommon in Israel’s litur-
gical poetry: see Gunkel, Einleitung, p. 47, who speaks of the turn
to the ‘warmer’ second person form of prayer as expressing more
powerfully the immediacy of the people’s devotion to their God (for
a briefer example of this in an early poem cf. Judg. 5.31). The poem
belongs to the broad category of hymns and, in view of its opening
words, to the sub-category which Gunkel called ‘Hymnen des
Einzelsängers’ (Einleitung, pp. 66-67, cf. 38: the whole section is
relevant [pp. 32-94]). The genre, in its various forms, was primarily
for use in public worship in the temple liturgy, as the reference to
such a place in v. 17 confirms for this example (see the Explanatory
Note there for discussion of which temple or temples are likely to
be meant). The poem is sometimes described as a ‘victory song’,
but despite the focus on victory (by Yahweh!) over the Egyptians
in vv. 1-12 its wider scope, especially in vv. 13-18, makes this
designation inappropriate: it is better suited to the Song of Miriam
(v. 21b), where examples of the genre and its use will be discussed
in the commentary.
It has long been observed that biblical hymns of praise regularly
contain both general statements about the praiseworthiness of the
deity (as in Babylonian and Egyptian hymns) and specific references
15.1-21 295
to the deity’s praiseworthy deeds (Gunkel, Einleitung, pp. 43-47,

51-52, 76-79). These components of the hymn have been well
summed up in the expressions ‘descriptive praise’ (beschreibendes
Lob) and ‘declarative praise’ (berichtendes Lob, which is better
rendered into English as ‘narrative praise’) coined by Claus Wester-
mann (Das Loben Gottes, pp. 25-26: see also Crüsemann, Studien
zur Formgeschichte, esp. Chapter 5 and pp. 304-306). It is only
natural that they should alternate with one another, as they do here,
where it is notable that the ‘narrative praise’ regularly precedes the
‘descriptive praise’: so vv. 2-3 respond to v. 1b, v. 11 responds to
vv. (4-7+)8-10 and v. 18 responds to all the rest of the poem.23
The language of the poem is undoubtedly distinctive. It contains
a number of rare words, including the hapax legomena nāwāh I, in
Hiphil ‘praise’ (v. 2), ʿāram I, in Niphal ‘be gathered together’ (v. 8),
and ṣālal, ‘sink’ (v. 10); it uses the Heb. imperfect as a past narrative
tense alongside the perfect to an extent that may be unparalleled in
any other passage in Biblical Hebrew (the examples in vv. 5, 12 and
14 are beyond question and good reasons can be given for under-
standing those in vv. 6-7 and 15-17 in this way: see the Explanatory
Notes); and there is a concentration of unusual grammatical forms
which again goes well beyond what is found in other poetic texts
(see Notes n, s, t, y, ee, gg, pp, ss, aaa, ccc, hhh on the translation).
Different views have been, and still are, held about the significance
of these and other linguistic features of the poem.
As for vv. 19-21, while v. 19 is in narrative form, it brings no
forward movement in the narrative (in fact it reverts to earlier
events at the end of ch. 14), as its beginning with ‘For’ rather
than ‘And’ (Heb. waw) confirms. Coats calls it a ‘narrative gloss
on the song’ (p. 118): more specifically it is a ‘narrator’s explana-
tion’ of a specific kind that recalls earlier events (cf. 12.39; 13.19).
Verses 20-21a are structured as a succession of clauses which are
introduced and joined together by three waw consecutive verbs
23
The status of v. 6 (and v. 7) has been variously understood. Many, following
Muilenburg and observing its distinctive ‘staircase’ form, view it as descriptive
praise. There is no doubt that, in contrast to v. 4, it is uses general language, but
so does v. 8 (‘The enemy…’) and the imperfects can readily be understood in a
preterite sense (cf. yekaseyumû, ‘covered them’, in v. 5). We therefore take vv. 6-7
as narrative praise using a more general vocabulary.
296 EXODUS 1–18
and further unified by retrospective pronouns (‘after her’ and ‘to

them’). They do advance the narrative and serve primarily as an
introduction to the poem in v. 21b, similar to but more developed
than the introduction to the Song of Moses in v. 1a. The poem itself
conforms precisely to one pattern of the hymn of praise (specifically
‘declarative praise’ in Westermann’s terminology: see further the
Explanatory Note on v. 21). Such hymns were no doubt normally
recited in a temple setting. This is an unusually short example of the
genre, but Psalm 117 is comparable. In its present setting, however,
it functions as a victory song (cf. Weimar-Zenger, Exodus, p. 72).
Such songs were a customary part of victory celebrations after a
military conflict (cf. Gunkel, Einleitung, pp. 311-14). They did not
need to have a specifically religious orientation or be limited to the
battlefield itself, as the example in 1 Sam. 18.6-9 (cf. 21.11) shows.
But such celebrations could naturally begin there and involve the
recognition that ‘the battle is Yahweh’s’ (cf. Ps. 118.15-16, 23-24).
A much fuller example of such a song appears in Judg. 5.1-31a, with
strong elements of a heroic character in it but also a number of refer-
ences to Yahweh, some of which at least may be an original part of
it (for discussion of its analysis and original setting see H.-D. Neef,
Deboraerzählung und Deboralied: Studien zu Jdc 4,1-5,31 [BTS
49; Neukirchen, 2002], pp. 1-114; C.L. Echols, Tell Me, O Muse:
The Song of Deborah in the Light of Heroic Poetry [LHBOTS 487;
London, 2008], pp. 64-92, 169-202). In the Song of Miriam the
role of Yahweh is central and the situation described is of course
exceptional. Metrically it is a single bicolon of 4 + 4, with partial
parallelism between the two cola (see Watson, Poetry, pp. 174-77,
343-44), the second being explicative of part of the first.
Historical Issues
Historical treatments of vv. 1-18 have been based on the language
of the poem, perceptions of its relationship to texts and religious
ideas elsewhere in the Old Testament and more widely in the
ancient Near East, and points of contact with the prose narratives
of the Exodus.
The grammatical features mentioned above (including the use of the imper-
fect) are not simply poetic forms but reflect an older stage of the Hebrew
language than that which is found in most of the Old Testament. So already
in 1880 Dillmann was arguing that they pointed to an early date for the poem
(Exodus und Leviticus, pp. 153-54: Dillmann dated the poem soon after the
15.1-21 297
Israelites’ arrival in Canaan). But, for a variety of reasons, including arguments

that some of the vocabulary of the poem was later, other nineteenth-century
German scholars were concluding that the poem was from a time long after
the Exodus (so first, perhaps, Jülicher, ‘Die Quellen’, pp. 124-26, who attrib�-
uted it to a redactor who knew both J and E). For them the features relied on
by Dillmann were not signs of a truly ancient origin for the poem, but of an
‘archaising’ style used at a late period, and they could point to sporadic occur-
rences of them in other texts which were agreed at the time to be exilic or
later. This has continued to be the most common view in German-speaking
scholarship until the present day. The fullest early examination of the poem’s
language, which continues to be cited, was the article of A. Bender, ‘Das
Lied Exodus 15’, which found extensive evidence for a post-exilic origin
and arrived at a quite precise date c. 450 B.C. on the basis of the poet’s (sup-
posed) knowledge of the Priestly Work and the dependence of Neh. 9.11 on his
account of the Exodus events.
The arguments put forward by Bender are, however, seriously flawed. He
took the view, common at the time but undermined by the work of Gunkel and
Mowinckel, that all psalms were post-exilic, so that there could really be no
doubt that the poem was from that era: ‘Das Lied Ex 15,1b-18 ist nach jeder
Hinsicht ein Psalm. Schon allein dieser Tatsache verweist es in die nachex-
ilische Zeit… und zwar als ganzes (p. 45; cf. p. 1). This dating of the psalms
also affected many of his specific judgements about particular words in the
poem: occurrences of them in psalms pointed automatically to a post-exilic
date. Bender was also obsessed with what he saw as connections between
the poem and messianic hope (understood, it would seem, as eschatology in
general), which again provided him repeatedly with an argument for post-exilic
origin. He wrote, of course, before first Mowinckel and then the discovery of
the Ugaritic texts (on which see below) offered a more convincing alternative
understanding of several passages in the poem. On the specifically linguistic
level Bender drew conclusions from parallels in Aramaic that would no longer
be accepted (already McNeile [p. 89] had doubts). There is also greater aware-
ness today that in the (comparatively) limited surviving literature in biblical
Hebrew the appearance of words that were in use throughout the biblical
period may be patchy in the extant evidence, so that arguments from vocabu-
lary distribution remain inconclusive (cf. Brenner, Song of the Sea, pp. 3-5:
he agrees closely with Bender’s date but relies mainly on a different kind of
argument, on which see below).24 Negatively it is important that the poem
includes no feature that can be ascribed to ‘Late Biblical Hebrew’. Nor have
attempts to find distinctively Priestly or Deuteronomistic expressions been
successful.
24
Nevertheless, arguments for a late origin of the poem on lexicographical
grounds have continued to be mounted: cf. R. Tournay, ‘Recherches sur la
Chronologie’, and ‘Le chant de victoire’; F. Foresti, ‘Composizione e Redazione’.
298 EXODUS 1–18
The argument for an early date based on ancient linguistic forms was
reintroduced by W.F. Albright and his pupils F.M. Cross and D.N. Freedman
in the wake of the publication of the Ugaritic texts from the thirteenth cent.
B.C., which showed conclusively that the alternation of perfect and imperfect
forms (more precisely the suffix-conjugation and the prefix-conjugation) in
past narrative was a standard feature of early North-West Semitic poetry:
see Cross and Freedman, Studies, passim, and their later writings; though
Freedman eventually came to discount the significance of this argument
(Pottery, p. 226). It was certainly criticised by others, on the same grounds
that Dillmann’s similar argument had been earlier, namely that the forms in
question continued to appear in much later writings. In this debate the study
of D.A. Robertson (Linguistic Evidence) has played an important mediating
role. It introduced a more rigorous methodology into the argument, seeking
to identify both the characteristics of standard poetic Hebrew (clearly attested
from the eighth cent. B.C. onwards) and earlier poetic Hebrew (reconstructed
from ancient features in later Hebrew and evidence of Canaanite poetry found
in the el-Amarna glosses and in Ugaritic) before comparing Hebrew poetry
of an unknown or uncertain date with them. Robertson concluded that several
of the ‘ancient’ features that had been appealed to were not of decisive value
for dating because of their occasional appearance in standard poetic Hebrew:
only the ‘verbal patterns’ noted and the longer form of the third person masc.
pl. verbal suffix (‘them’), which entirely replaces the shorter form in Exod.
15.1b-18, were significant in their own right (p. 135). Other features are
only significant when there is a ‘clustering’ of them in a text, which does not
happen in poetry from the eighth cent. or later. The combination of a clustering
of early forms with the occurrence of the corresponding standard forms may
point to a period of transition in the development of the poetic form of the
language (which is to be expected). On this basis the following sequence of
early poetry emerged: (1) Exodus 15 (i.e. vv. 1b-18); (2) Judges 5; (3) (transi-
tion) Deuteronomy 32; 2 Samuel 22 = Psalm 18; Habakkuk 3; Job; (4) Psalm
78. While recognising that turning this sequence into an absolute chronology
was ‘precarious’, Robertson proposed that the United Monarchy was most
likely to be the period of transition, so that Exod. 15.1b-18 (twelfth cent.) and
Judges 5 (late twelfth cent.) would be pre-monarchic and Psalm 78 would
be from the late tenth cent. or early ninth cent. (pp. 154-55). These dates
might well be open to challenge for various reasons and should probably all
be lowered.25 But, as Robertson underlined, the linguistic evidence which he
examined (from syntax and morphology) remains ‘a very strong argument for
dating Ex 15 early’ (p. 155). The preterite use of ‘short prefixed verbal forms’
25
The implication that Job (i.e. the poetic dialogue) is from the tenth cent.
is particularly distant from conclusions reached about its date on other grounds.
Even M.H. Pope (Robertson’s advisor) went back no further than the seventh cent.
15.1-21 299
as a criterion for dating Hebrew poetry has subsequently been re-examined

by Y. Bloch in the light of its continued appearance in later texts such as Isa.
41.1-5 and Psalm 44 (‘The Prefixed Perfective’). While critical of Robertson’s
very early dates for 2 Sam. 22.2-51/Ps. 18.2-51, Exod. 15.1-18 and Deut.
32.1-43, Bloch recognises that all three poems must be ‘more ancient’ than
the later passages examined and can be dated ‘a couple of centuries before
the Babylonian exile – i.e. to the 8th-7th, or perhaps even the 9th centuries
B.C.E.’ (pp. 66-67). Indeed, according to his favoured criterion (the presence
or absence of waw before the verbal form in question) Exod. 15.1-18 should
be the oldest of the three, since it has only one or possibly two cases of a
preceding waw as against ten or eleven without (pp. 54-57). A fuller selec-
tion of texts from the ‘archaic corpus’ (Joosten, Verbal System, pp. 417-19)
and a wider range of verbal forms are studied in the valuable monograph of
T. Notarius, The Verb in Archaic Biblical Poetry. In her analysis of Exod.
15.1-18 (pp. 107-24) Notarius calls in question (perhaps excessively) the
preterite interpretation of the prefix-conjugation forms and notes the greater
prominence of perfect (suffix-conjugation) verbs in the ‘report’ sections. At
the same time, however, she identifies a number of other verbal features of the
poem which make it consistently ‘archaic’ and place it well within the linguis-
tically most primitive group of poems (see pp. 280-97). Since the group also
includes 2 Samuel 22/Psalms 18, which must be from the monarchic period
(v. 50), this placement does not at all exclude a date for Exod. 15.1-18 too
in the period of the monarchy, although it has a closer systemic kinship with
Judges 5 (pp. 308-10).
Similarities between the poem and the Psalms and late prophecy were
observed in early critical scholarship but interpreted in different ways.
Dillmann (p. 154) and Baentsch (p. 129) were confident that the other texts
had borrowed from Exodus 15. Others, such as Bender, Smend (pp. 139,
141-42) and McNeile (p. 89), saw the resemblances as a sign that the poem
belonged to the same post-exilic situation as them. A very different kind of
association of Exodus 15 with psalmody was to be introduced in Mowinckel’s
cult-functional interpretation, which was understandably applied to it as well
as to the Psalter. Here the memory of the Exodus history was seen as being
incorporated into the festival myth, with its original theme of the creation and
renewal of nature being extended to take in the re-experiencing of the ancient
election and deliverance of the people as an assured reality, which had its future
in the ‘eternal kingdom’ of Yahweh (so in The Psalms, pp. 140, 173, 177, 187;
but already in Psalmenstudien II [Kristiania, 1922], pp. 56-58, 214). Even
if it is nowhere explicitly stated (and in ‘Psalm Criticism between 1900 and
1935’, VT 5 [1955], pp. 13-33 [27], Mowinckel pointedly rejected Albright’s
thirteenth-century date for the poem because of the clear reference to Jerusalem
in vv. 17f.), Mowinckel seems to assume that the poem reflects the character
of pre-exilic temple worship in Jerusalem, as opposed to the eschatological
hopes of the post-exilic period (see further below on source criticism). The
300 EXODUS 1–18
shift in interpretation was, like Mowinckel’s ideas in general, largely ignored

in German exegesis of the poem until much later (H. Schmidt, ‘Das Meerlied’,
p. 66, was one exception).26 In America the full force of Mowinckel’s ideas was
not felt at first, because the influence of Albright encouraged a more historical
and linguistic approach to the poem. There is more sympathy with them in
Muilenburg’s essay of 1966 (see above: esp. pp. 236, 248-50; cf. in Britain
Clements, God and Temple, pp. 51-55, to which Muilenburg refers, and his
commentary on Exodus, pp. 90-93).
By this time the importance of parallels to the psalms (and therefore Exod.
15.1b-18) in Ugaritic literature and religion was being increasingly recognised
(for an early assessment of this see Mowinckel, ‘Psalm Criticism’), but before
we discuss this we must consider a further aspect of comparisons between
the poem and particular psalms: the question of interdependence. This was to
be given renewed prominence by M.L. Brenner in The Song of the Sea, with
particular reference to the ‘Asaphite psalms’ and among them especially Pss.
74.12-17; 77.16-21; 78.52-54 (pp. 126-42; 152-60). This group of psalms (to
which Brenner added Pss. 96, 105 and 106 because extracts from them are
included in the song to be sung by ‘Asaph and his brethren’ in 1 Chr. 16.7-37)
is argued to be dependent upon and so later than Deutero-Isaiah because they
(or rather some of them) incorporate the application to the Exodus of mythical
language, which was used for the first time by this prophet (pp. 104-106: see
further pp. 181-86, where evidence from Chronicles [cf. Ezra 3.10] is taken
to show that they were ‘the chief singing class in the post-exilic period’). It
is assumed that all the ‘Psalms of Asaph’ (Pss. 50, 73-83, plus 96, 105 and
106) were composed by these Asaphites in post-exilic times and it is claimed
that Exod. 15.1b-18 is inseparable from them. Some of these psalms are post-
exilic or at least exilic (Pss. 74 and 79 presuppose the fall of Jerusalem), but
others can scarcely be so late. Psalms 77, 80 and 81 are all associated with
the Northern Kingdom and in the latter two cases can scarcely be from after
722 B.C. Psalm 76 is probably pre-exilic and many have taken the same view
of Psalm 78 (see below). It seems clear that this is a collection of psalms that
was assembled over a long period (or at least one that drew on texts written
over a long period), beginning already in pre-exilic times (cf. H.P. Nasuti,
Tradition-History and the Psalms of Asaph [SBLDS 88; Ann Arbor, 1985];
M.D. Goulder’s attempt to date the whole collection in the 720s [The Psalms
of Asaph and the Pentateuch (Sheffield, 1996)] is not credible, but even he
has to account for the present form of it by recognising at least redactional
26
A. Weiser’s commentary on the Psalms seems at various points to imply a
pre-exilic date for the poem: see the 4th ed. (1955), pp. 21, 29, ET, pp. 33, 45. The
1st ed. was published in 1939. In his Einleitung in das Alte Testament (Göttingen,
6th ed., 1966), p. 100, ET, p. 106, Weiser treated the poem as an older component
taken up by J, but no older than the time of David and Solomon: in fact he was
not sure of its exact date and noted that ‘others’ detected Deuteronomic influence
in it.
15.1-21 301
contributions extending into the exilic period). As such, resemblances to it

could not be a decisive argument for dating Exod. 15.1b-18 in the post-exilic
period. The second part of Brenner’s argument, that it is Exodus 15 that is
dependent on the psalms of Asaph and not the other way round, has recently
been addressed by B.D. Russell, in his The Song of the Sea, pp. 113-30; cf.
Norin, pp. 11-14, 119, 128-33. Russell shows that it is much more likely that
the lament in Psalm 74, which uses language about an ancient (cf. qedem, ‘of
old’, in vv. 2 and 12) deliverance, necessarily presupposes ancient traditions
and texts. Psalm 77, another lament, has much in common with Exodus 15 in
vv. 11-16 and is again clearly dependent on older tradition for its argument.
Psalm 78 is a polemical poem which uses the Exodus tradition, including
vocabulary shared with Exodus 15, to justify Yahweh’s favour to Judah
and Jerusalem rather than to the Ephraimites. It is a much longer and more
complex poem and at the same time more specific in its historical references.
Russell argues, following Mark Smith (Pilgrimage Pattern, pp. 225-26), that
in the psalm the somewhat ambiguous language of Exod. 15.17 is applied
first to Shiloh (v. 54; cf. v. 60) and then to Jerusalem (v. 69), reflecting two
different stages in the poem’s use in the Israelite cult. This all seems a much
more likely scenario than that Exodus 15 was pieced together from widely
separated verses in the psalm.27
Russell then seeks to establish the approximate date of the psalms that
he has argued to be dependent upon Exodus 15, in the hope of establishing
a terminus ad quem for the latter. He is unable to reach definite conclusions
for Psalms 74 and 77, but concludes that Psalm 78 is from the late eighth
cent. B.C., in the time of Hezekiah. A pre-exilic setting is favoured by the
eventual optimism of the narrative and the lack of any reference to the fall
of Jerusalem: Zion and the Davidic monarchy are still apparently functioning
as national institutions. The explanation given for the failure of the northern
tribes fits well into the time soon after the Assyrian conquest. Russell also
notes the extensive use of archaic linguistic features alongside standard ones,
which is against a date too late in the monarchy period.28
27
Foresti (‘Composizione’, pp. 59-60) came to the same conclusion from a
comparison between vv. 8, 10, 13 and 16 and related language in the psalm, but
he dated both it and the Song in the sixth century.
28
A similar conclusion about the date of Ps. 78 (or an even earlier date) has
been reached by others (Eissfeldt, Dahood and Day, for example): for a pre-exilic
date see also K. Weingarten, ‘Juda als Sachwalter Israels. Geschichstheologie nach
dem Ende des Nordreiches in Hos 13 und Ps 78’, ZAW 127 (2015), pp. 440-58.
Anja Klein agrees that Ps. 78 has been influenced by Exod. 15.1b-18 but dates
both to the post-exilic period (‘Hymn and History’). In addition to P, she thinks
that the Song of the Sea is later than Ps. 118, Exod. 3 and Ps. 24 (which in itself
need not imply a post-exilic date) and that it never existed apart from its present
context. The relevance of Deut. 2 and Josh. 2 for this date depends on the direc-
tion of influence.
302 EXODUS 1–18
The issue of mythical expressions and their relationship to historical events

lying behind the poem was sharpened by the discovery (first in 1929) and
decipherment (from 1931) of mythological texts from Ugarit, especially a
series (KTU 1.1-6) in which the god Baal is the central character (see Gibson,
CML2, pp. 1-19, 37-81; A.H.W. Curtis, Ugarit [Ras Shamra] [Cambridge,
1985], pp. 28-30, 66-72; also the works cited below and their bibliographies).
One of the first (the first according to Loewenstamm) to observe the parallels
between the poem and the Ugaritic myths was Umberto (M.D.) Cassuto in a
Hebrew essay published in 1943 (‘The Epic Poetry of Israel’: see his Bibli-
cal and Oriental Studies, 2 [Jerusalem, 1975], pp. 69-109 [99-101]; also his
commentary, pp. 177-81): for examples see the Explanatory Notes on vv. 1b,
8, 9, 10, 11, 12, 14-16, 17 and 18, and Note qq on the translation.29 Sometimes
the resemblance is especially close, as with ‘my strength and my protection’
(v. 2: see Note i) and ‘the hill(s) which you own’ (v. 17: see Note iii), and
suggests direct contact. The poem also reflects the key stages of the plot of
the Baal-cycle: Baal’s victory (KTU 1.1-2; cf. vv. 1-12), his temple or palace
(KTU 1.3-4: cf. v. 17) and his kingly rule (KTU 1.2.4.32 (34); 1.6.5.5-6, 6.33-
35: cf. v. 18). This has reasonably been seen as evidence, not only of imitation
but of the unity of the two parts of the poem (cf. Smith, Pilgrimage Pattern,
p. 220, developing an observation of Cross, Canaanite Myth, pp. 141-42). In
numerous details too the poem employs language and motifs that are found in
the Ugaritic myths. Yet the borrowing is selective (other features of the myths,
not found here, are taken up elsewhere in biblical poetry) and a key element
of the mythical plot, the battles between gods, is missing. There is perhaps
a vestige of it in the rhetorical questions in v. 11 that speak of Yahweh’s
incomparability, but Yahweh’s victory is throughout over the human, Egyp-
tian ‘enemy’, and the ‘sea’ (yām) is the main means by which he overcomes
them, not a foe with which he must do battle, as Baal does with the god Sea
(Yammu) in the first episode of the Ugaritic myth (cf. Cross, Canaanite Myth,
pp. 131-32). Maybe it is envisaged that Yahweh has subdued Sea/the sea in
the past, as other biblical poems with a cosmic or cosmological focus would
suggest, but here there is no place for that. The language and pattern of the
myth provide the means to glorify the achievements of Yahweh, the ‘Divine
Warrior’ (v. 3) who leads his people to victory over their human foes, as he
does in other biblical poems where the subject is similar and mythological
themes are also drawn upon (Judg. 5.3-5; Ps. 18.7-15).
This is relevant to an assessment of the debate between those, like Cross
(cf. Canaanite Myth, pp. 143-44) and his pupils, for whom an underlying
historical event is central, and those who see the mythical narrative as having
generated the celebration of an imaginary, pseudo-historical ‘event’ in the
29
For an even earlier study of Ugaritic parallels, especially for Hab. 3, see
Cassuto, ‘Chapter III of Habakkuk and the Ras Shamra Texts’ (1937), in Biblical
and Oriental Studies 2, pp. 3-15.
15.1-21 303
time of Israel’s origins. Wider issues are of course involved in the choice
between these alternatives, but the great difference between the myth and at
least this poem’s presentation of the Israelites’ escape from Egyptian control
make it very difficult to believe that the latter was spun out of nothing more
than the former, as C. Kloos (Yhwh’s Combat with the Sea, pp. 127-214), S.E.
Loewenstamm (Evolution, pp. 256-72, 291-92) and recently M. Leuchter
(‘Eisodus as Exodus’) seem to maintain. It is, of course, possible to recon-
struct an ‘original core’ of the poem much closer to the mythical pattern
by eliminating the references to earthly peoples and armies as secondary
additions (so Norin, Er Spaltete das Meer; Zenger, ‘Tradition und Interpreta-
tion’; Klein, ‘Hymn and History’), but we have already suggested that the
arguments put forward for such a view, on the basis of metrical and other
considerations, are weak.
The relationship of the poem to the prose tradition about what we may
cautiously call ‘the episode at the sea’ has been treated in three different
ways. First, what source or layer of redaction is associated with it? In early
critical scholarship, which sought to relate the whole Pentateuch to its major
sources, the poem was at first assigned to the E source (Knobel [his Rechts-
buch: cf. Num.-Jos., p. 532], Dillmann: Wellhausen also saw the case for this
in Composition, p. 77, but recognised that v. 17 [as a reference to Jerusalem]
required explanation in a different way). But with growing consensus around
a late origin for the poem came the conclusion that it did not belong to any
of the sources (cf. Jülicher, ‘Die Quellen’, above), and such agreements as
existed between them were generally explained by the poet’s dependence
upon the prose narratives (including P in vv. 4, 8 and 9). Even those who
held to an earlier origin for the poem (e.g. Gressmann, Anfänge, p. 58) rarely
attributed it to one of the sources, but Mowinckel went back to the old assign-
ment to E (Psalmenstudien II, p. 191) and Weiser attributed it to J (see above).
Cross maintained that it was preserved in both J and (in what he took to be
its incipit in v. 21b) E, while Propp attributes it to JE (p. 482) and Baden to
J (p. 28). Mark Smith (Pilgrimage Pattern, pp. 214-15) does not rule out the
poem’s presence in an earlier stratum, but finds the clearest link to ‘the priestly
redaction’ in v. 19 (see further below). Among those who see the composi-
tion of the Pentateuch in supplementary terms, Blum acutely observed the
similarity between the introductions in 15.1a and Num. 21.17 and concluded
that here as there it was Kd who incorporated the poem, to create a frame
around the narrative of the journey through the wilderness, which expressed
Israel’s response to Yahweh’s protection and provision of their needs all the
way to Canaan (Studien, pp. 127, 202 n. 437).30 Schmid, by contrast, sees it as
a post-Priestly composition associated with the combination of the patriarchal
30
The similarity between v. 1a and Num. 21.17a was already observed by
Foresti (‘Composizione’, pp. 68-69), who attributed both verses specifically to
DtrN.
304 EXODUS 1–18
and Exodus traditions (like Exod. 1: Erzväter, pp. 238-41; similarly Berner,
pp. 389-400; Albertz, pp. 229-53 [with vv. 1b-5 added to an older hymn by
the ‘Hexateuch redactor’]).
Secondly, there is the question of the history of tradition (which really has a
wider scope): how does the presentation of ‘the episode at the sea’ relate to the
prose versions of it (and others)? Cross and Freedman argued that the narrative
in the poem represented a storm at sea in which Egyptian troops in barges were
drowned and so was very different from and necessarily earlier than the prose
accounts. But this interpretation reads far too much into the poetic description
to be convincing (see the Explanatory Note on v. 8). Others have often claimed
that the description matches the Priestly narrative, but there is no clear parallel
to the two ‘walls’ of water that appear there. The effects of wind, portrayed as
Yahweh’s breath in vv. 8 and 10, rather resemble the non-Priestly account in
14.21 (and for the double action of wind cf. 10.13, 19). There is no real interest
in the Israelites crossing the dry sea-bed, because the poem’s emphasis falls on
the destruction of the Egyptian force, even though it is possible to argue that it
is presupposed in vv. 8-10 (see the Explanatory Notes). This again distances the
poem from the Priestly account (and the redactional addition in v. 19 which is
based on it, as well as a large body of biblical poetry in which this theme became
very central: see Loewenstamm, Evolution, pp. 253-92) and brings it closer to
(what survives of) the non-Priestly account in ch. 14, the Song of Miriam in
15.21 and the brief summary in Josh. 24.6-7. On this basis some have claimed
that this ‘minority tradition’ is the oldest form of the story, with the crossing
motif being imported from the account of the crossing of the Jordan (cf. Noth,
pp. 94-97, ET, pp. 119-22). But it is also possible (see above) that the tradents
in question knew the ‘crossing motif’ but subordinated it to the pattern reflected
in the traditions of ‘Holy War’ (compare the Explanatory Note on 14.11-14).
Larger issues also arise about the history of traditions. The poem embraces
not only an episode from the Exodus tradition (though this is its most promi-
nent theme), but the journey through the wilderness, the settlement in Canaan
(‘conquest’ would be an inappropriate term to use here), the foundation of a
sanctuary (most often identified as the Jerusalem temple) and, briefly in v. 2,
perhaps the patriarchal tradition. Some have also seen the Sinai tradition
alluded to in v. 13 (but see the Explanatory Note). With more justification one
might add the ‘tradition’ of Yahweh’s enthronement/kingship in v. 18. The
scope of this combination of traditions is not as wide as the whole Pentateuch
or the sources/traditions from which it was compiled, but it does extend, even
if only briefly at the end, outside its boundaries. This poses a challenge if
one is seeking to place the poem in a particular context of Israelite literary
activity, for no single literary work covers such a wide range of topics. It
is more probable that the combination was made on the basis of traditions
than literary works and, given the hymnic form of the text, that this took
place in a centre of worship. The cultic centre of which we know most is
the Jerusalem temple and this is where Mowinckel placed the poem’s origin
(cf. The Psalms 1, pp. 125-26, 154-55: already in Psalmenstudien II, p. 58).
15.1-21 305
A possible problem with this is noted by Mark Smith: ‘The Zion tradition,
however, lacks otherwise any theology of the Exodus until Second Isaiah and
later literature’ (Pilgrimage Pattern, p. 222). So either the poem would be
late or it would come from somewhere else, most likely the northern kingdom
where the Exodus tradition was evidently important. Cross thought of Gilgal,
even in pre-monarchic times (Canaanite Myth, pp. 138-43), but this shrine
in the Jordan valley is difficult to reconcile with the ‘mountain’ or ‘hill’ in
v. 17. Another possibility is Shiloh, as Smith himself argues with reference to
Ps. 78.54 and 60, though he also recognises the poem’s later use in Jerusalem
(ibid., pp. 225-26: see also the Explanatory Note on v. 17). One might even
wonder if the poem could have originated at the prime northern sanctuary of
Bethel.31 Yet a Jerusalem provenance, even in the (early?) monarchic period,
should probably not be excluded: Smith’s statement, quoted above, needs
at least some modification to take account of the knowledge of the Exodus
tradition in Jeremiah and Ezekiel and Josiah’s celebration of the Passover
(2 Kgs 23.21-23). Moreover, even though it is the divine election of David
and Yahweh’s dwelling in Jerusalem which are the most obvious themes of
pre-exilic worship in Jerusalem, there are passing references which suggest
that the Exodus story was also known and valued there (Isa. 30.7; Pss. 78 and
114; perhaps Ps. 76.7), as one would expect.
The third kind of relationship of the poem to the prose narrative of Exodus
concerns its role in the present text of the book. It is of course most obviously
an elaboration of what is said in 14.30-31 about the Israelites’ response to
Yahweh’s great act of deliverance from the Egyptians, and it is not difficult
or unnatural to extend this to seeing it as a liturgical conclusion to the whole
narrative beginning in Exodus 1. But, given that the ‘hymnic narrative’ of the
poem continues on beyond the deliverance at the sea, it is also proper to see
it as having a programmatic function in relation to what is still to come. The
ongoing journey through the wilderness, under divine leadership, is clearly
in view in v. 13. But the ‘horizon’ of the poem’s preview has been variously
understood, depending mainly on the identification of the ‘destination’ in
v. 17. Mark Smith has argued, following B. Halpern, that ‘for the priestly
redaction’ the goal of the poem is Mount Sinai (Pilgrimage Pattern, pp.
215-18; cf. Halpern, The Emergence of Israel in Canaan [SBLMS 29; Chico,
1983), pp. 38-39; Russell, Song of the Sea, pp. 45-55, is one of several others
who have adopted this view). This means that the poem can be seen as a micro-
cosm of the whole book of Exodus, and Smith finds some (older) confirmation
of this in the assurance in Exod. 3.12 that after the Exodus Israel will worship
Yahweh ‘on this mountain’, i.e. Horeb/Sinai. It is, however, very difficult to
see v. 17 as a reference to Sinai when it follows the references in vv. 14-16 to
31
Compare Albertz’s suggestion that the Song of Miriam might have been a
Kultruf from the worship there (p. 255). By contrast S.C. Russell, Images of Egypt,
pp. 145-48, makes a strong case for a southern provenance for the Song of the Sea.
306 EXODUS 1–18
the later stages of the wilderness journey.32 The natural interpretation is that it
refers to a sanctuary in the land of Canaan (see the Explanatory Note) and this
fits in well with the passages elsewhere in Exodus (including the key Priestly
passage in Exod. 6.2-8) which see the real destination of Israel’s journey as
the land promised to Abraham, Isaac and Jacob. If so, then the main contribu-
tion of the second half of the poem is to underline that (what we know as) the
Exodus story is, for all its importance, only a stage on the way to Israel’s final
goal and that this final goal has a place of worship at its centre.33
The date of the poem’s origin remains a matter of dispute. An

early date, pre-monarchic rather than Mosaic, has continued to be
supported especially in America (Cross and Freedman, Hyatt [for
vv. 1-12], Norin [for an original core], Kloos, M.S. Smith [tenta-
tively], Propp, B.D. Russell, Leuchter [in oral form]). Archaic
features of the language, the closeness in mythical features and
poetic form to Ugaritic texts, and the historical setting presupposed
are the main arguments for this view. At the other extreme, a wide
range of scholars in Germany and some others have since the 1880s
dated the poem to the post-exilic period (so Carpenter/Harford-
Battersby [except for v. 1b], Holzinger, Smend, McNeile, Fohrer,
Zenger, W.H. Schmidt, Houtman, Brenner, Levin, Van Seters, K.
Schmid, Berner, Klein), sometimes as late as the mid-fifth century.
In support of such a date the use of ‘late’ vocabulary (with reference
to the studies of Jülicher and Bender), dependence upon late writings
(including P) and, at least in earlier scholarship, theological features
have commonly been cited. In between, however, a substantial
number of scholars have supported a date in the monarchy period,
evidently concluding that the arguments for an earlier or later date
are not decisive and that the overall character of the poem fits best in
the time of the monarchy. Some have connected it specifically with
32
For other objections to Smith’s view see my remarks in ‘The Theology of
Exodus’, in E. Ball (ed.), In Search of True Wisdom (FS R.E. Clements; JSOTSup
300; Sheffield, 1999), pp. 137-52 (49-51).
33
So in effect the valuable study of J.W. Watts, Psalm and Story, pp. 41-62,
who considers separately first ‘the narrative role’ of the poem in its present context
and then the purpose of its inclusion (as he believes, at a late stage in the composi-
tion of the Pentateuch: see pp. 55-62). A better case than Smith’s, though still not
a decisive one, could be made for the view that v. 13 refers to Sinai as an interim
goal of the Exodus. This might also have been the view of ‘the priestly redactor’,
for whom the revelation at Sinai occupied such a vast extent of text in Exodus,
Leviticus and Numbers, even if it was not what the original poet intended.
15.1-21 307
the time of Josiah, because they see his policy of cultic centralisation
presupposed in v. 17 and the combination of Exodus and Jerusalem
traditions as fitting in with the Deuteronomic ethos of his reforms
(Baentsch, Beer, Jeremias, Blum[?], Albertz [for vv. 6-18]). Others
have preferred a date earlier in the monarchy period (Gressmann
[Anfänge, p. 58], Mowinckel, Muilenburg, Childs, Loewenstamm,
Spieckermann [except for vv. 14-16], Blenkinsopp[?], Baden, S.C.
Russell), pointing to the absence of clear Deuteronomistic features
and, in some cases, to the poem’s inclusion in one of the older narra-
tive sources. A small but distinguished group of scholars (Eissfeldt,
Gunkel, Coats, Gertz, Dozeman) understandably refrain from
giving any clear indication of when the poem was composed.34
Only a brief discussion of the arguments used can be given here.
It is, first, impossible to prove that the poem was or was not part of
a specific ‘old source’ or layer of composition. Its continuation in
v. 19 (to be discussed below) certainly makes a connection with the
Priestly narrative in 14.22-23, 28-29, but it is not certain that it was
added at the same time as vv. 1-18, or even that it was intended to
refer to it. The introduction in v. 1a does make a connection between
14.31 (cf. ‘At that time’) and the poem and so it is the most secure
clue to the latter’s incorporation into the narrative. Blum has shown
(see above) that a very similar introduction to a poem appears in
Num. 21.17 in a composite itinerary that has some Deuteronomistic
features (see my discussion in ‘The Wilderness Itineraries and the
Composition of the Pentateuch’, pp. 10-11), but they may have
been added to material with an older origin, including vv. 16-19.
So we must remain uncertain whether Exod. 15.1a too is from the
old non-Priestly narrative or from a Deuteronomistic redactional
layer: there is, at any rate, nothing to associate it with P. All that
can be deduced from the form of the introduction is that the poem
probably existed prior to its incorporation into the narrative, as
34
Noth, p. 98, ET, p. 123, says that it ‘is a relatively late piece; we cannot give
a more accurate indication of the time at which it was composed’; he writes that
v. 8 ‘is reminiscent of P’s description of the miracle at the sea’ (p. 124), which
may or may not be meant to imply that it is later than P. Schmidt’s recent treatment
of the problem in his commentary (pp. 640-42) begins with the usual arguments
for an exilic or post-exilic date, but then notes a number of factors (including the
poem’s Nachwirkung) which favour a date which is ‘not…too late’
308 EXODUS 1–18
in Num. 21.17-18 and Josh. 10.12-13: this is also the most likely
explanation for its inclusion of ‘anachronistic’ details in vv. 13-17.
The poem itself, not surprisingly, shares some material and even
vocabulary with the narrative in ch. 14, but not enough to prove
definite dependence on it or any of its sources. Such links as exist
are with the non-P sources rather than with P (see above). There is
of course a very close similarity to the Song of Miriam (v. 21), but
its antiquity is so generally recognised that dependence on it would
not rule out any of the dates that have been entertained. Depend-
ence on other parts of the Hebrew Bible has often been asserted
(especially by those advocating a later date), but remains unproven.
For example, the fear of the nations in vv. 14-16 is compared to
Deut. 2.25 and Josh. 2.9, 11 and said to be dependent on one or both
of these passages, but the relationship may as readily be the other
way around and the verses in the poem may be inspired by a passage
in the Ugaritic Baal myths (see the Explanatory Note), to which the
poem appears indebted in other ways. Again, as discussed above,
Brenner’s linking of the origin of the poem to the Asaphite psalms
overlooks. the likelihood that Psalms 74, 77 and 78 depend upon it
and the fact that the collection is by no means entirely a post-exilic
creation. In fact, the dependence of Psalm 78 on the poem and its
probable pre-exilic origin can serve as an argument that the poem
itself is older than this. The focus on a single shrine in v. 17, most
likely the Jerusalem temple, has been thought to depend upon the
centralisation of worship required in Deuteronomy and enforced by
Josiah, but there is nothing in this verse which could not have been
part of the Jerusalem temple ideology in an earlier period (just as
similar prominence was given to other temples in the ancient Near
East in their liturgical texts without any exclusive claims being
made for the worship there).
As for the presence of late theological ideas, such as eschatology
and messianism, in the poem, the case has been much weakened by
the widespread recognition of a preterite use of the imperfect in vv.
15-17 (though much German scholarship has been unaccountably
reluctant to embrace this) and the acceptance that Israelite worship
made bold claims about Yahweh’s sovereignty and blessing even
in the present. Even the declaration of Yahweh’s enduring sover-
eignty in v. 18 no longer needs to be seen as a product of post-exilic
prophecy when similar beliefs were being maintained about the
Babylonian and Ugaritic gods centuries earlier.
15.1-21 309
The language of the poem has been appealed to, in different

ways, by proponents of both an early and a late date for it. There
is no doubt that it displays a concentration of archaic grammatical
forms (see above), but it has been claimed that this is insignificant
because the same forms also occur in what are agreed to be much
later texts, where they represent ‘archaising’ rather than signs of
a genuinely ancient text, so that in Exod. 15.1b-18 they may also
be explained in this way. This, however, overlooks, as Robertson
pointed out, that in the later texts such features are only sporadic,
whereas in the Song of the Sea they recur frequently and the
standard form of the third person pl. verbal suffix does not occur at
all. This is a strong argument that the Song does come from an early
stage in the history of Biblical Hebrew. Unfortunately Robertson
perpetuated the view of Cross and Freedman that the ‘early stage’
was limited to the pre-monarchic period, on the doubtful basis that
the tenth century was likely to have been the transitional period
in which standard forms gradually became established. It is safer
to acknowledge that on linguistic grounds the poem could have
been composed in the early monarchy period (but before the eighth
century, when prophetic poetry shows that the standard forms were
well established).
The proponents of a late date for the poem, on the other hand,
have given great weight to the occurrence of ‘late’ vocabulary in
the poem. Much of the evidence collected by Bender is discredited
by the fact that he regarded all psalms as post-exilic, so that occur-
rences of words in the psalms were automatically dubbed ‘late’.
Today it is recognised that dating psalms is often uncertain and that
a good many of them are most likely pre-exilic in origin. In general,
arguments based on vocabulary are seen to be precarious because of
the limited quantity of ancient Hebrew text that has survived. The
study of ‘Late Biblical Hebrew’ by scholars such as Avi Hurvitz has,
in the opinion of most, made it possible to identify some psalms as
post-exilic on linguistic grounds, but this does not apply to the Song
of the Sea. More up-to-date studies of the poem by R. Tournay and
F. Foresti have continued to affirm its lateness, but their basis is
weak.35
35
See above, n. 24. In his more recent article Tournay abandons the Josianic
dating for a post-exilic one because of ‘late’ vocabulary, dependence on P in
vv. 8-10 and (p. 527) the ‘second Exodus’ motif in vv. 13-17 (!) which has the
310 EXODUS 1–18
Historical arguments have been used in attempts to date the

poem, most commonly based on what is taken to be a reference to
the Jerusalem temple in v. 17(b). There is some uncertainty about
this (see above and the Explanatory Note on v. 17), but it remains
the most probable view and it requires a date for the poem in the
tenth century or later. The same is probably true for the inclusion of
the Philistines among (at the head of!) the fearful enemies in v. 14.
The omission of any mention of the Ammonites as a hostile power
was cited by Cross and Freedman (Studies, pp. 47, 62; cf. ‘The Song
of Miriam’, pp. 239-40; Freedman, Pottery, p. 226) as an argument
for dating the poem in the twelfth or eleventh century, but it is not
a decisive one, as it is an argument from silence. Moreover, as
Loewenstamm pointed out (Evolution, p. 260), the omission corre-
sponds to the absence of any mention of the Ammonites elsewhere
in the old conquest tradition as even a possible hindrance.36 In
the time of Saul and David the Ammonites seem briefly to have
represented a threat to the Israelites (1 Sam. 10.27–11.11; 2 Sam.
10.1–11.1; 12.26-31), but not thereafter: Amos 1.13-15 may be
a reference to these earlier conflicts. It is just as likely that the
omission of Ammon in Exodus 15 is because it was no longer an
enemy to be reckoned with as because it was not yet such.
Arguments for the antiquity of the poem have also been based on
its treatment of Canaanite myth and ‘its extreme use’ of the poetic
pattern of ‘staircase parallelism’, which is also found in Ugaritic
poetry (Cross, Canaanite Myth, pp. 121, 134-44). Cross places the
use of the Baal-myth in Exodus 15 (on which see above) early in
the history of Israel’s poetic traditions, at a time when (under the
‘tribal league’) the religion of Israel had become strongly historical
in its focus (cf. Judg. 5), prior to the resurgence of the old myths of
creation under the monarchy (cf. Pss. 29; 89.6-19; 93) and again in
return from exile in mind. Foresti envisages a date c. 580–560 B.C. But for an
expression of caution about such arguments see E. Blum, ‘The Linguistic Dating
of Biblical Texts – An Approach with Methodological Limitations’, in J.C.
Gertz et al. (eds.), The Formation of the Pentateuch (FAT 111; Tübingen, 2016),
pp. 303-25 (303-14).
36
In Num. 21.24, Deut 2.37; 3.16 and Josh. 3.10 the Ammonites’ territory is
regarded as ‘off limits’, outside the area which Israelite tribes claimed. Josh. 13.25
appears to be a late attempt to expand the territory claimed for Gad at the expense
of the Ammonites (V. Fritz, Das Buch Josua [HAT; Tübingen, 1994], p. 145; cf.
perhaps M. Avi-Yonah, The Holy Land [Grand Rapids, 1977], pp. 26-27, 40-41).
15.1-21 311
the exile and afterwards. According to him the poem is also earlier
than the Gilgal cult’s association of the Exodus with the crossing
of the Jordan, which is reflected in Psalm 114 and Habakkuk 3, as
well as in Joshua 3–5 and Ps. 78.13. There is a certain plausibility
in outline about this succession of initiatives, but a problem arises
when it is used to establish a precise date for particular compo-
sitions. Since a similar historical focus to Exodus 15 is found in
Psalm 78, which is undoubtedly later than the time of David, and
Psalm 80, which is probably from the divided monarchy, one cannot
rule out a date for the composition of Exodus 15 in the monarchy
period, at a time when others were composing psalms that kept
much closer to the nature-myth language of Canaanite poetry,
including Psalm 77, which combined it with the Exodus story. As
Cross points out (Canaanite Myth, p. 131), the latter poem (‘from
the early monarchy’) also uses ‘staircase parallelism’ (v. 17; again
in a weakened form in v. 20), which shows (as other examples do)
that this criterion too cannot be pressed in support of a very early
date for Exodus 15.
To sum up, the overall character of the poem, including its
climax in vv. 17-18, is most readily compatible with a date in the
monarchy period. There is no need to come down very late in that
period, because what have been taken to be signs of Deuteronomic
influence are perfectly explicable in other terms. A date between the
tenth and the eighth centuries is supported by archaic features in its
language, which became less frequent later, and by the dependence
of Psalm 78 upon it.
As for vv. 19-21, critical scholarship has almost universally
agreed that v. 19, which represents narrative elements from ch.
14 with very little change (see the Explanatory Note), is from a
post-Priestly redactor, according to many (but not all: cf. Blum) the
redactor who inserted the Song of the Sea before it.37 The majority
of scholars have also agreed that vv. 20-21 are part of the earlier
narrative material.
37
A non-Priestly origin is only conceivable if the corresponding verses in
ch. 14 are, against all the indications, attributed to an older narrative source, as
they were by Wellhausen (E/JE: p. 77) and are by Dozeman (non-P: pp. 320, 342),
or if (equally improbably) v. 19 is supposed to have stood in the song-book from
which vv. 1-18 were taken (Dillmann, p. 160).
312 EXODUS 1–18
The early consensus was that they belonged to E, because of Miriam’s

prophetic status and her appearance in other E contexts (especially Num. 12):
so Knobel, Carpenter/Harford-Battersby, Holzinger, Baentsch, Gressmann,
McNeile and more recently Cross (Canaanite Myth, p. 123; cf. Studies,
pp. 45-46, and ‘The Song of Miriam’, pp. 237-38), Propp (p. 482) and Baden
(J, E and the Redaction, p. 182 n. 215). But Smend objected that the descrip-
tion of Miriam as Aaron’s sister could not stand in a source which regarded
Aaron as Moses’ brother (cf. 4.14, which he [unlike most others] attributed
to E) and so, since neither Miriam nor Aaron appeared in J, he assigned these
verses to the additional, older J1 source (Erzählung, pp. 139, 143-44). In this,
as usual, he was followed by Eissfeldt (pp. 268*, 271*), Beer (p. 79) and later
Fohrer (pp. 111-12). Rudolph naturally did not accept an Elohistic origin for
the verses: as elsewhere he deleted the decisive evidence (‘the prophetess’)
and attributed what was left to J (Elohist, pp. 31-32). This then became,
almost by default, the standard view, even among scholars who (unlike
Rudolph) believed that there was an E source, like Noth, Hyatt and Weimar/
Zenger. The last-named made a detailed case for the origin of both the Song
of Miriam and the introductory narrative in a pre-J account composed by
opposition groups in the time of David, which was later used by J (Exodus,
pp. 71-87), but they like others in this period found it necessary to regard the
epithets of Miriam in v. 20 as later additions.
As uncertainty about source-analysis increased, some scholars hesitated
to decide between J and E (Childs, pp. 246-48; Schmidt, Exodus, Sinai und
Mose, pp. 64-65), while others followed Freedman’s lead in regarding the
whole of vv. 1-21 as a literary unit, perhaps reflecting a liturgical setting in
which the Song of Miriam was the (antiphonal) conclusion to the Song of
Moses (cf. Freedman, Pottery, pp. 79, 81 [as part of the original ‘twelfth-
century’ poem], 181-82, 194-95; similarly [without the early date] Houtman,
pp. 240-41, citing Mowinckel, Psalmenstudien II, pp. 111-12, in support).
Alternatively, J.G. Janzen (whose contorted exegesis is followed by Russell,
The Song of the Sea, pp. 17-18, 24-25, and, more surprisingly, by Albertz,
pp. 235-36) read vv. 20-21 as indicating (by way of the analepsis in v. 19)
what preceded the singing of the Song of the Sea by the Israelites (see the
Explanatory Note on v. 19). None of these writers was interested in the prehis-
tory of the present text, except for Albertz (on whom see below).38
At the present time opinion is divided between those who see most of
vv. 20-21 as relatively early in origin and those who assign them to a late
stage of the composition of the Pentateuch. In addition to Propp and Baden,
who have revived the old attribution to E, the former group includes Blum
and Blenkinsopp, who include these verses in their older, Deuteronomistic,
38
Russell recognises such concerns (pp. 132-33), but apparently in a way that
allows Priestly texts to have a pre-exilic origin (p. 197 n. 29).
15.1-21 313
compositions (Studien, p. 106; Pentateuch, p. 158), and others who more

explicitly locate them within a (late) pre-exilic Exodus story (Levin, Jahwist,
pp. 342-43; Kratz, Komposition, p. 331 and passim; Berner, Exoduserzählung,
p. 390 [citing Num. 20.1 in support]; Klein, ‘Hymn and History’). By contrast
H. Spieckermann argued that, while the Song of Miriam itself is much older
(though probably no older than the monarchy period), its redactional frame
contains in vv. 19(!) and 20 elements that are post-exilic in origin (Heilsge-
genwart, pp. 99-102) and Albertz specifically attributes vv. 19-21 to his first
Priestly Bearbeitung (pp. 253-55), though again with the proviso that the Song
itself could come from the monarchy period.39 Dozeman takes a similar view
of vv. 20-21 and also argues that the focus in the Song on the event at the sea
as the climax of the Exodus fits the Priestly conception of salvation history
better than it does the non-Priestly view (pp. 342-44). This is not far from
Brenner’s argument that everything in vv. 19-21 (like vv. 1-18) was composed
by post-exilic Levites (see esp. The Song of the Sea, pp. 42-53, 80-84).
The variety of opinions about the sources and composition of

15.19-21, both now and in earlier times, turns out to have rather
less to do with the great upheavals in Pentateuchal scholarship since
the 1970s than with some curious deviations from a natural under-
standing of the passage, which have had more influence than they
deserve. To begin with v. 19, the introductory ‘For’ (Heb. kî) should
leave no doubt that its primary connection is with the preceding
Song of the Sea, not with the Song of Miriam and its introduction.
There should be no doubt (pace Dozeman) that the verses in ch. 14
which it recapitulates are from the Priestly source, so that it will
have been added by a redactor who knew that account. Verse 19
also seems to presuppose some features of the Song of the Sea (see
the Explanatory Note), so that it cannot have been added prior to the
latter’s incorporation into the narrative. It is unlikely to have been
included by the author or redactor who first introduced the Song of
the Sea here, because v. 1a, which must be from that writer, contains
no distinctively Priestly language. It is best described as an isolated
gloss which served to remind hearers or readers of the events which
(supposedly) gave rise to the Song of the Sea, after their thoughts
had been carried forward to later events by vv. 13-18.
39
In an earlier work Albertz had attributed vv. 20-21 to what he now calls the
Exoduskomposition from the exilic period (Religionsgeschichte, pp. 71-72, ET,
pp. 42-43).
314 EXODUS 1–18
The arguments for a very late date for vv. 20-21 which have
recently been put forward by a minority of scholars are not very
weighty. The description of Miriam as a prophetess and the sister
of Aaron by no means require such a late origin. The former fits
well into the picture of pre-classical prophecy that may be derived
from the Old Testament historical books and the latter is actually
more likely to reflect an early stage in the development of the figure
of Aaron than the later period when he was seen as Moses’ brother
too (see the Explanatory Note on v. 20). Spieckermann cited the
‘anachronistic’ scenario of the women ‘coming out’ as if to greet
victorious warriors, but the parallels to this are all in pre-exilic
literature. Only Brenner has claimed that the Song of Miriam
itself is post-exilic, but the vocabulary items to which he draws
attention are entirely inconclusive (on ‘horse and its driver [sic]’
see Note g on the translation). It is in fact hard to see why a late
redactor would have been concerned to ‘save’ the Song of Miriam
for posterity when it was almost entirely present already in the
Song of the Sea. It is much more likely, as most scholars think, that
it owes its separate preservation to its inclusion in an old Exodus
narrative, like other snippets of poetry that appear elsewhere in
such older texts (cf. 17.16).
But which ‘old Exodus narrative’? For those recent scholars who
believe that there was only one such narrative (or at least that only
one has been preserved), this is not really an issue. But the present
commentary has found good reason to think, as earlier commenta-
tors generally did, that the pre-exilic development of the tradition
was more complex and is still accessible, in part, to us. That the
scholarly discussion about vv. 20-21 is also somewhat complex
is, as we suggested earlier, more due to some false turns than to
actual problems in the evidence. The early consensus was that these
verses came from E, and those who initially departed from this did
so for idiosyncratic reasons. Smend’s objection that to call Miriam
the sister of Aaron was inconsistent with 4.14 only had force
because his peculiar exegesis attributed that verse to E: most of his
contemporaries (and successors) regarded it, as we do, as a part of
a secondary addition to the text. Rudolph of course did not believe
in E and was always looking for reasons to discredit such an attri-
bution. Here, as sometimes elsewhere, he had recourse to arbitrary
procedures, deleting ‘the prophetess’ on the grounds that Miriam
does nothing prophetic in Exod. 15.20-21 and undercutting the
15.1-21 315
appeal to Numbers 12 by citing Gressmann’s view that prophetic

Sagen were ‘secondary’ (Elohist, p. 72). Of course Gressmann
meant ‘secondary in the process of oral tradition’, not in terms of
literary sources, and Rudolph says nothing of the fact that Gress-
mann had been perfectly happy to attribute Exod. 15.20-21 and
most of Numbers 12 to E (Anfänge, pp. 55, 98-99).
Like Rudolph, Noth attributed the verse to J, but it is clear both
from the brief discussion in ÜGP (p. 32 n. 107) and the fuller treat-
ment in his commentary (pp. 97-98, ET, pp. 122-23) that he was
far from confident about this and was inclined to regard the verses
as a secondary addition to J. It is very strange that in the commen-
tary he writes: ‘Because it [i.e. the section] is in all probability of
relatively great age it is most often assigned to the source J, but
there is no conclusive evidence for this’ (p. 96, ET, p. 121).40 He
says nothing about the widespread attribution of the verses to E,
for which strong if not necessarily ‘conclusive’ arguments had been
put forward. He perhaps had Rudolph and Beer (who simply repro-
duced Smend’s view) in mind. But Noth did not agree with either
of their overall views about the sources of the Pentateuch: only a
few pages before in the commentary (pp. 82-84, ET, pp. 105-106)
he had declared his acceptance of the older view that elements of
E’s account appeared in ch. 13 and 14. It is surprising that he did
not at least consider the possibility that 15.20-21 came from E too.
It was probably largely due to Noth’s presentation of the issue that
most subsequent commentators were either unsure what to do with
the passage (Childs, Schmidt; more recently Gertz) or attributed it
to J (Hyatt, Burns [Has the Lord, pp. 15-16: ‘it is often tentatively
attributed to the J source’]). Weimar and Zenger’s reconstruction
of an early pre-J narrative which included vv. 20-21 seems to have
been transmuted in the Einleitung by Zenger and others into an old
Exodus story originating in the northern kingdom (2nd ed. [1996],
p. 119). This is not the E source (cf. pp. 111-12 for the rejection
of that view), but neither is it J. In fact it only extends to ch. 14
and it is not explained there how the poems and their frameworks
in ch. 15 found their way into the seventh-century ‘Jerusalemer
40
My italics. Noth does make the important observation that the idea of
Yahweh ‘throwing’ the Egyptians into the sea is shared by the Song and the J
narrative (14.27: cf. p. 122). But different verbs are used and Noth evidently
supposed that the Song was widely known in early Israel.
316 EXODUS 1–18
Geschichtswerk’, which the authors of this volume had come to

regard as the earliest comprehensive presentation of the traditions
about Israel’s origins.
If, as we have argued, there is evidence earlier in Exodus (includ-
ing ch. 14) of a two-fold non-Priestly literary tradition, it continues
to make sense to ask to which strand or stage of that tradition Exod.
15.20-21 is likely to belong. In addition to the older arguments for
an attribution to E, two further considerations lend some support to
that view. First, these verses must have been attached to the end of
the story of the ‘sea event’ in ch. 14. In our analysis (see the intro-
duction to 14.1-31) vv. 30 and 31b were given to J and v. 31a to
E. The response of 15.21b fits more naturally after the recognition
that the victory is due to Yahweh alone in 14.31a than it does after
the inclusion of ‘Moses his servant’ in 14.31b. The prominence of
Miriam (and to a lesser extent Aaron) in 15.20 is also perhaps easier
after 14.31a. Secondly, the reference to Miriam as Aaron’s sister
would benefit from, even if it does not require, a reference to Aaron
earlier in the narrative to which it belongs. In our analysis of the text
so far this is not the case for J, but we have attributed 4.27-28 tenta-
tively to E. Aaron also appears in 4.14-16 and other verses which
belong with it, but we have followed the common view that these
passages reflect the work of a redactor, most likely RJE. We see no
reason to regard 15.20-21 as redactional in origin.41
Three questions remain to be dealt with about the Song of Miriam
itself:
(i) What is the relation between it and the very similar introduc-
tion to the Song of the Sea in v. 1b? Several different explanations
have been given. (a) As the text stands, it looks like a response to
the Song of the Sea, a short refrain sung by Miriam and the women
to conclude it. The interruption by v. 19 is something of a prob-
lem for this view, since if the two songs are connected one would
expect the explanation to come after the second, not the first, but as
we have seen v. 19 is probably a very late addition to the text. (b) A
variant of this view is to see v. 21b as the title or incipit of the Song
of the Sea (Cross and Freedman) or more probably, since there is
no evidence for the use of incipits in ancient Israel, as an alterna-
tive introduction to it (Houtman). The implication of this would be
41
Equally we see no need to regard either of the epithets of Miriam in these
verses as later additions (see above).
15.1-21 317
that Miriam and the women sang the whole of the Song of the Sea
after Moses and ‘the Israelites’ (presumably the men) had done
so. There is no positive reason to think that this is the case, as the
Song of Miriam is complete in itself and embodies the essential
components of hymnic composition. (c) A third view, which seeks
a historical relationship between the two songs, sees the Song of
Miriam as the basis for the Song of the Sea. This relies to some
extent on the assumption that shorter compositions are always
older than longer ones, which may not always be true. It was a
popular and plausible view (and perhaps still is) when the Song of
the Sea was regarded as a very late, post-exilic, composition and
the Song of Miriam was thought to be an early celebration of the
Exodus events, as in Germany throughout the modern period. But
it is also compatible with an earlier, pre-exilic, dating for the Song
of the Sea, provided that the Song of Miriam is seen as even earlier.
(d) Fourthly, in what is in effect a reversal of the first alternative,
it has recently been argued that the Song of Miriam was seen as
the impulse for the Song of the Sea (J.G. Janzen). This is the least
likely of the four explanations, as the exegesis of vv. 19-21 on
which it is based is forced and improbable (see the Explanatory
Note on v. 19).
(ii) Is there any evidence for an alternative wording of the Song
of Miriam? The main textual witnesses do include some variations
from the Masoretic text, as they do for the corresponding section
of v. 1 (see Text and Versions), but one of the Qumran scrolls, 4Q
(Reworked) Pentateuchc (4Q365), seems to have had a more exten-
sive variant text of the Song. Unfortunately, only the beginnings of
seven lines of text are preserved, but they are sufficient to show that
it shared some wording from the Song of the Sea, but also included
other expressions that are not found there (for the details again see
Text and Versions). The style is hymnic and the language and ideas
find parallels in the biblical psalms, documents from Qumran and
other sources from the Second Temple period (see the study of G.J.
Brooke, ‘Power to the Powerless: A Long-Lost Song of Miriam’,
BAR 20/3 [1994], pp. 62-65).
(iii) How old is the Song of Miriam? It is widely, indeed almost
universally, held that the Song of Miriam is very ancient (Brenner
is a rare exception). Its brevity and its focus on a single episode,
without apparently any need to specify who the defeated enemies
were, have been seen as evidence that its origin may be very close
318 EXODUS 1–18
to the events of the Exodus.42 The form of the hymn certainly corre-
sponds, in a very simple form, to the conventions of temple worship,
but there is no need to suppose that these first became established
in the tenth century B.C., especially when the format corresponds
so closely to what any leader of worship might be expected to say.
The fact that the worship is led by a woman, and one who is not said
to have any connection with the central figure of the Exodus story,
Moses, is a further reason to take very seriously the possibility
that the poem comes from a very early period of Israel’s history. It
does not include any of the archaic grammatical forms which have
been used to argue that the Song of the Sea is very old, but the
expressions ‘greatly exalted’ and ‘threw’ are in different ways most
unusual (they also occur of course in the introduction to the Song of
the Sea, but that is essentially the same text as this one) and could
be rare relics of an older stage of the language.
Theology
Both of the poems in this section, for different reasons, have
made an enduring contribution to the Exodus narrative which
they conclude (on aspects of their later interpretation see Text and
Versions, passim, and my ‘Some Christian Uses and Interpretations
of the Song of Moses’). The theological significance of the Song
of the Sea needs to be assessed both for its original cultic setting
(wherever that was) and for its present literary context. As a hymn
for cultic use it both enriched the ‘epic’ tradition of the Exodus
by exalting Yahweh, the conqueror of the Egyptians, as the most
powerful of the gods (v. 11) and affirmed his faithful commitment
to his people (v. 13) and the temple from which he would rule for
ever (vv. 17-18). It was apparently also the inspiration for several
other poems, both psalms and in prophecy. Placed where it is in the
Pentateuch/Hexateuch, it creates a pause in the narrative of Israel’s
journey out of Egypt for its deeper significance and purpose to be
understood (cf. Durham, p. 210). Theological truths implied by the
story are brought to the fore by the use of expressions like ‘strength
42
The view that it originated in the temple cult as a quite general celebration
of the defeat of enemies (Kratz, Komposition, p. 292; Berner, Exoduserzählung,
pp. 391-92) is most improbable, since it refers not to the normal experiences of the
battlefield but to an extraordinary, so far as we know unique, occasion on which
Israel’s enemies were drowned in a ‘sea’.
15.1-21 319
and protection’ (v. 2; cf. vv. 6, 13), ‘holiness’ (vv. 11, 13), ‘loyalty’
(v. 13), ‘(your) people’ (vv. 13, 17), ‘planting’ (v. 17), ‘dwelling’
(v. 17) and ‘reign’ (v. 18). There is no god like Yahweh (v. 11), the
God who intervened on his people’s behalf (v. 13) and leads them
on, as it seems effortlessly, to the place where he will continue to
be in their midst and receive their worship (v. 17).
Verses 19-21 also have important theological and religious
implications. Even the largely repetitious recapitulation of the sea
narrative in v. 19 shows a theological enrichment, compared with
its pattern in the Priestly sections of ch. 14, in the explicit attribu-
tion of the destruction of the Egyptians to Yahweh’s intervention.
This is no doubt derived from the Song of the Sea, to which v. 19 is
primarily related, but it is also the reason for the call to praise in v.
21. The authors have no hesitation in attributing the violent conse-
quences of natural events to Yahweh, since as Creator he holds them
in his control and as the God of Israel he is expected to protect them
against enemy attack. Verses 20-21 round off the story of the Exodus
in the narrow sense with a song of praise, and it is likely (since the
Song of the Sea is probably a later addition) that they represent the
oldest extant attestation of such a response in the narrative tradition
of the Exodus. It is therefore the more striking that the participants
are all women. As has often been observed, this matches the recur-
rent involvement of women in the early stages of the Exodus story
(1.15-21; 2.1-10, 16-22; 4.24-26: cf. also 3.22; 11.2). Even if from
here on the role of women in the narrative is greatly diminished
and, where it appears, is viewed negatively (Num. 12 and 25), it is
an inescapable fact that in the central narrative of Yahweh’s deliv-
erance of his people in the Old Testament women as well as men
played a crucial part.43 This is not likely to have been intended, as
it is often understood now, as a way of improving the situation of
women in Israelite society: more probably it was a means to impress
on the early hearers of the story that it was not the powerful leaders
among Israel’s ancestors who brought about their deliverance but
Yahweh acting in a most unexpected way. The participation of
43
It is notable of course that none of the passages just listed comes from
the Priestly account of the Exodus. As Aaron’s sister, Miriam was inevitably
‘adopted’ into Moses’ family with Aaron when he was made a full brother of
Moses (7.1-2[P]; cf. Num. 26.59), and so Moses’ hitherto anonymous elder sister
obtained a name and the dispute in Num. 12 became a family row.
320 EXODUS 1–18
women is linked to another feature which is rare in the Pentateuch,

though it is attested elsewhere, especially in the psalms: the accom-
paniment of the words of worship by music and dancing (see on
this especially R.J. Burns, Has the Lord, pp. 18-40). As Houtman
notes, such accompaniment has often been looked down upon in the
religious traditions that are founded on the Old Testament (p. 231),
but again its place in the Exodus story gives it a canonical stamp of
approval.
Special Bibliography for Exodus 15.1-18

Bartlett, J.R. Edom and the Edomites (PEFMS 1; Sheffield, 1989).
Bender, A. ‘Das Lied Exodus 15’, ZAW 23 (1903), pp. 1-48.
Berner, C. Die Exoduserzählung. Das literarische Werden einer Ursprungslegende
Israels (FAT 73; Tübingen, 2010), pp. 389-400.
Bloch, Y. ‘The Prefixed Perfective and the Dating of Early Hebrew Poetry – A
Re-evaluation’, VT 59 (2009), pp. 34-70.
Brenner, M.L. The Song of the Sea. Ex 15:1-21 (BZAW 195; Berlin and New
York, 1991).
Coats, G.W. ‘The Song of the Sea’, CBQ 31 (1969), pp. 1-17.
Cross, F.M. Canaanite Myth and Hebrew Epic (Cambridge, MA, 1973), pp. 112-
44.
Cross, F.M., and D.N. Freedman, Studies in Ancient Yahwistic Poetry (SBLDS 21:
Missoula, 1975 [repr. of Ph.D. dissertation, Johns Hopkins, 1950, with addition
of Postscriptum]), pp. 45-65.
———. ‘The Song of Miriam’, JNES 14 (1955), pp. 237-50.
Davies, G.I. ‘Some Christian Uses and Interpretations of the Song of Moses
(Exodus 15:1-18’, in C. Bultmann et al. (eds.), Vergegenwärtigung des Alten
Testaments: Beiträge zur biblischen Hermeneutik (FS R. Smend; Göttingen,
2002), pp. 179-95.
———. ‘Some Points of Interest in Sixteenth-Century Translations of Exodus
15’, in W. Horbury (ed.), Hebrew Study from Ezra to Ben-Yehuda (Edinburgh,
1999), pp. 249-56.
Dozeman, T.B. ‘The Song of the Sea and Salvation History’, in S.L. Cook and S.C.
Winter, On the Way to Nineveh (Atlanta, 1999), pp. 94-113.
Foresti, F. ‘Composizione e Redazione Deuteronomistica in Ex 15, 1-18’,
Lateranum 48 (1982), pp. 41-69.
Freedman, D.N. ‘The Song of the Sea’ in R. Shukraft (ed.), A Feeling of Celebra-
tion: A Tribute to James Muilenburg (San Anselmo, 1967), pp. 1-10; repr. in
Freedman, Pottery, Poetry and Prophecy: Studies in Early Hebrew Poetry
(Winona Lake, 1980), pp. 179-86.
15.1-21 321
———. ‘Strophe and Meter in Exodus 15’, in H.N. Bream et al., A Light unto my
Path (FS J.M. Myers; Philadelphia, 1974), pp. 163-203; repr. in Freedman,
Pottery, Poetry and Prophecy, pp. 187-228.
Goldin, J. The Song at the Sea, being a Commentary on a Commentary in Two
Parts (New Haven and London, 1971).
Gunkel, H. ‘Mosessegen, Moseslied und Meerlied’, RGG2 4, pp. 245-47.
Howell, M. ‘Exodus 15.1b-18. A Poetic Analysis’, ETL 65 (1989), pp. 5-42.
Jeremias, J. Das Königtum Gottes in den Psalmen. Israels Begegnung mit dem
kanaanäischen Mythos in den Jahwe-König-Psalmen (FRLANT 141; Göttingen,
1987), pp. 93-106.
Jülicher, A. ‘Die Quellen von Exodus VII,8-XXIV,11. Ein Beitrag zur Hexateuch
frage’, JPTh 8 (1882), pp. 79-127, 272-315 (esp. 124-26).
Klein, A. Geschichte und Gebet. Die Rezeption der biblischen Geschichte in den
Psalmen des Alten Testaments (FAT 94; Tübingen, 2014).
———. ‘Hymn and History in Ex 15. Observations on the Relationship between
Temple Theology and Exodus Narrative in the Song of the Sea’, ZAW 124
(2012), pp. 516-27.
Kloos, C. Yhwh’s Combat with the Sea (Amsterdam/Leiden, 1986), pp. 127-214.
Labuschagne, C.J. The Incomparability of Yahweh in the Old Testament (PrOS 5;
Leiden, 1966).
Leuchter, M. ‘Eisodus as Exodus: The Song of the Sea (Exod 15) Reconsidered’,
Bib 92 (2011), pp. 321-46.
Loewenstamm, S.E. The Evolution of the Exodus Tradition (Jerusalem, 1992: tr.
from 2nd Heb. ed., 1987), pp. 256-72.
Lugt, P. van der. ‘The Wave-Like Motion of the “Song of the Sea” (Ex 15,1-18)
and the People of Israel as a Worshipping Community’, ZAW 128 (2016),
pp. 49-63.
Mowinckel, S. Psalmenstudien II (Kristiania, 1922).
Muilenburg, J. ‘A Liturgy on the Triumphs of Yahweh’, in Studia Biblica et
Semitica Theodoro Christiano Vriezen…dedicata (Wageningen, 1966), pp.
233-51.
Norin, S.I.L. Er Spaltete Das Meer. Die Auszugsüberlieferung in Psalmen und
Kult des alten Israel (CBOT 9; Lund, 1977), pp. 77-107.
Notarius, T. The Verb in Archaic Biblical Poetry: A Discursive, Typological,
Historical Investigation of the Tense System (SSLL 68; Leiden, 2013).
Robertson, D.A. Linguistic Evidence in Dating Early Hebrew Poetry (SBLDS 3;
Missoula, 1972).
Russell, B.D. The Song of the Sea: The Date of Composition and Influence of
Exodus 15:1-21 (StBL 101; New York, 2007).
Russell, S.C. Images of Egypt in Early Biblical Literature: Cisjordan-Israelite,
Transjordan-Israelite and Judahite Portrayals (BZAW 403; Berlin and New
York, 2009), pp. 127-76.
322 EXODUS 1–18
Schmidt, H. ‘Das Meerlied Ex 15.2-19’, ZAW 49 (1931), pp. 59-66.

Smith, M.S. The Pilgrimage Pattern in Exodus (JSOTSup 239; Sheffield, 1997),
pp. 205-26.
Spieckermann, H. Heilsgegenwart: Eine Theologie der Psalmen (FRLANT 148;
Göttingen, 1989), pp. 96-115.
Tournay, R. ‘Le chant de victoire d’Exode 15’, RB 102 (1995), pp. 522-31.
———. ‘Recherches sur la chronologie des Psaumes’, RB 65 (1958), pp. 321-57
(esp. 335-57).
Watts, J.W. Psalm and Story: Inset Hymns in Hebrew Narrative (JSOTSup 139;
Sheffield, 1992), pp. 41-62.
Zenger, E. ‘Tradition und Interpretation in Exodus XV 1-21’, in Congress Volume:
Vienna, 1980 (VTSup 32; Leiden, 1981), pp. 452-83.
1 [At that timea Moses and the Israelitesb sangc this song about
Yahwehd, and they said as followse:
I will sing of Yahwehd, for he has become greatly exaltedf:
horse and its driverg he threw into the sea.
2 Yah(weh)j is my strengthh and my protectioni, and he has
beenk a deliverancel for me.
This is my God and I will praise himm, my father’s God and
I will exalt himn.
3 Yahweh is a warrioro, Yahweh is his name.
4 Pharaoh’s chariots and his army he threwp into the sea;
and his choicest officersq sankr in the Yam Suf.
5 The deep waters covered thems, they went down into the
depths like a stone.
6 Your right hand, Yahweh, (who are) mighty in strengtht,
your right hand, Yahweh, shatteredu the enemy,
7 and in your great majestyv you cast downw your foesx.
You let go your wrath, it devoured themy like stubblez.
8 And/for by the breath of your nostrilsaa the waters were
gathered togetherbb,
the flowing waters stood firm like a borecc, the deep waters
congealeddd in the midst of the sea.
9 The enemy said, ‘I will pursue, I will overtake,
I will divide the spoil, my appetite shall be sated with themee,
I will draw my swordff, my hand shall dispossess themgg'.
10 You blewhh with your breath, the sea covered them,
they sankii like leadjj into the mighty waters.
11 Who is like youkk among the godsll, Yahweh,
who is like youkk, mighty in holiness/the holy placemm,
fearsome for praiseworthy actsnn, doing wondersoo?
15.1-21 323
12 You stretched out your right hand, the earthqq swallowed

them uppp.
13 In your loyaltyrr you led (in) the people for whomss you
intervenedtt.
With your strength you broughtuu (them) to your holy
dwellingvv.
14 The peoples heard, they quakedww,
anguishxx seized the inhabitants of Philistia.
15 Then the chiefsyy of Edom were dismayed;
the rulerszz of Moab, trembling seized themaaa,
all the inhabitants of Canaan were helpless.
16 There fellbbb upon them dreadccc and terror,
because of the greatnessddd of your arm they were silenteee as
a stonejj,
until your people passed onfff, Yahweh,
until the people of whomss you had taken possessionggg
passed onfff.
17 You brought them (in) and planted themhhh on the hill(s)
which you owniii,
a fixed placejjj for you to dwell you madekkk, Yahweh,
a sanctuarylll, O Lordmmm, your hands establishednnn.
18 Yahweh will reignooo for ever and everppp.]
19 [Forqqq the horses of Pharaoh, with his chariots and his horse-
menrrr, went into the sea and Yahweh brought back the waters of
the seasss upon them. But the Israelites had gonettt on dry ground
through the midst of the sea]. 20 [Then Miriam the prophetess,
the sister of Aaron, took a hand-drumuuu in her hand, and all the
women went outvvv behind her with hand-drums and dancingwww.
21 Miriam sangxxx to themyyy:
Singzzz of Yahweh, for he has become greatly exalted:
horse and its driver he threw into the sea.]

a. Heb. ‫ אז‬is given two temporal meanings in BDB, p. 23, ‘at that time’
(e.g. Gen. 4.26; Exod. 4.26) and ‘thereupon’ (i.e. ‘next’): the distinction is not
always easy to make as both would fit the context, and some have doubted
whether ‫ אז‬ever meant ‘next’ when followed, as here, by an imperfect (I.
Rabinovitz, ‘ʾāz Followed by Imperfect Verb-Form in Preterite Contexts:
A Redactional Device in Biblical Hebrew’, VT 34 [1984], pp. 53-62; IBHS
§31.6.3; Gibson, Syntax, §62 Rem.1 [allowing ‘next’ with the perfect]; DCH
1, p. 167). On the use of the imperfect here see Note c. The most recent
324 EXODUS 1–18
discussion is by Joosten, Verbal System, pp. 108-13: he usefully observes the

way that ‫ אז‬with the imperfect can introduce a new section of a narrative, as
in Deut. 4.41, 1 Kgs 3.16 and elsewhere. In several such cases the idea of
succession is clear (Josh. 8.30; 1 Kgs 8.1 par. 2 Chr. 5.2; 2 Kgs 15.16 [cf.
v. 14]; Sir. 50.16; 4Q381 f69.6). Elsewhere ‘at that time’, without any idea of
succession, is preferable (1 Kgs 9.11; 11.7; 2 Kgs 8.22 [par. 2 Chr. 21.10]; Ps.
126.2; Job 38.21) and these should probably include cases where the original
occasion for a (well known?) song or poem is indicated (Exod. 15.1; Num.
21.17; Josh. 10.12). Since ‫ אז‬was probably originally a noun meaning ‘time’
this will be the older use, but its weakening to mean ‘then’ in the sense of
‘next’ is readily understandable (as seems to have happened with ‫ (ב)אדין‬in
BAram: cf. Dan. 2.15, 17, 19 etc.).44
b. For the composite subject after a singular verb cf. GK §146f: the
grammatical form is not a reason to omit ‫בני־ישׂראל‬, even though the first
person sing. forms in vv. 1b-2 understandably caused some concern in the
transmission of the text (see Text and Versions).
c. Heb. ‫ישׁיר‬. The use of the imperfect to refer to a single event in the past
after ‫ אז‬is surprising but quite common: there are about 20 examples in the
Hebrew Bible (Joosten, p. 108; cf. the list in DCH 1, p. 167, but Prov. 1.28 and
20.14 are not past). Early attempts to relate this to the iterative or durative uses
of the imperfect (GK §107c) have rightly been abandoned because the actions
concerned are ‘single complete events’ (Joosten, p. 109). It is attractive to see
the phenomenon as due to the preterite use of prefix-conjugations in early
North-West Semitic, which has left traces in Biblical Hebrew, most notably in
the ‘waw-consecutive’ construction (cf. JM §113h-i; 117). The only problem,
as Joosten has pointed out (pp. 110-11), is that almost all the forms which
follow ‫ אז‬are ‘long imperfects’ as here (the only exception is 1 Kgs 8.1 par. 2
Chr. 5.2), whereas it is only the short form of the prefix-conjugation which is
generally thought to have borne a preterite sense. However, it appears that the
long form did so too in Ugaritic (cf. E. Greenstein, ‘Forms and Function of the
Finite Verb in Ugaritic Narrative Verse’, in S. Fassberg and A. Hurvitz [eds.],
Biblical Hebrew in its North-West Semitic Setting [Jerusalem and Winona
Lake, 2006], pp. 75-102; Joosten, Verbal System, p. 110), and it could be the
fact that ‫( אז‬unlike waw) remained a separate word which led somehow to its
use with the longer form in most cases.
d. Heb. ‫ליהוה‬. The ‫ ל‬has generally been rendered ‘to’, as e.g. in Pss. 95.1
and 96.1. But since the Song begins with references to Yahweh in the third
person (as is characteristic of hymns generally: Gunkel, Einleitung, p. 47),
‘about’ (cf. 14.3; and Isa. 1.5aβ, which is similar to here) would seem more
appropriate: on this rendering see P.A.H. de Boer, ‘Cantate Domino: An
44
Rabinowitz’s attempt to eliminate the idea of chronological sequence has
had more influence than it deserves and is combined with an antiquated under-
standing of the imperfect; cf. the gentler critique of Joosten, p. 110.
15.1-21 325
Erroneous Dative’, OTS 21 (1981), pp. 55-67 (58-59); so already Freedman,

Pottery, Poetry and Prophecy, p. 199 (1974). Houtman’s ‘in honour of’
(p. 277) is similar.
e. Heb. ‫ויאמרו לאמר‬. See Note d on the translation of 7.6-13: the sequence
of two forms of ‫ אמר‬is not, however, limited to Priestly texts (cf. 5.10; 32.12)
and the inf. const. occurs immediately after a finite form only here and in
Num. 20.3 in the Pentateuch (elsewhere 2 Sam. 5.1; 20.18; Jer. 29.24; Ezek.
33.10, 24; Zech. 2.4).
f. Heb. ‫גאה גאה‬. The verb is rare, being used (outside the identical expres-
sion in v. 21) only in Ezek. 47.5; Job 8.11; 10.16 in BH and in Sir. 10.9: some
find the Hithpalel by emendation in Isa. 9.8. Only in Exod. 15 is it used of
God; elsewhere it refers to the ‘rising’ of waters (Ezek. 47.5), the growth of
a plant (Job 8.11) and to human pride (Job 10.16 [?]; Sir. 10.9). The meaning
with reference to Yahweh is indicated by the cognate nouns which are used
(as well as in other ways) of his greatness or majesty (‫ גאוה‬in Deut. 33.26; Ps.
68.35; ‫ גאון‬below in v. 7 and e.g. in Isa. 2.10, 19, 21; ‫ גאות‬in Ps. 93.1; Isa. 12.5;
26.10): see further TWAT 1, 878-84 = TDOT 2, pp. 344-50. On the variant
readings of SP here see Text and Versions.45
g. Heb. ‫ורכבו‬, lit. ‘and its rider’, at first glance suggests a reference to
cavalry, which would have implications for the dating of the poem (and
equally for the Song of Miriam in v. 21, which uses the same expression),
since cavalry was only gradually introduced into Near Eastern armies (in
Assyria) from the ninth century onwards and the Egyptians did not have
cavalry until the Hellenistic period (so Mowinckel, ‘Drive and/or Ride’,
p. 280; BRL, p. 254; ABD 4, pp. 826-31). It is clear that ‫ ַר ָכּב‬could mean a
chariot-driver (1 Kgs 22.34 par.) as well as a horse-rider (2 Kgs 9.17), just as
the verb ‫ רכב‬could be used for a charioteer ‘driving’ a team of horses (2 Kgs
9.16; cf. v. 20) and its participle could refer to chariot-drivers (‫)מרכבה ורכביה‬
as well as cavalrymen (‫ )סוסים ורכביהם‬in Hag. 2.22 (cf. Jer. 51.21). But the
last phrase at least suggests that following ‫ סוס‬the expression ‫ רכבו‬here would
mean a cavalryman, since such a suffixed form would be equivalent to )‫רכב(י‬
‫( סוס‬so e.g. Weimar and Zenger, Exodus, pp. 76-78; Brenner, Song of the
Sea, pp. 82-84: note the vigorous response of Propp, p. 510). The question
is whether it (and the fuller expression with ‫ )סוס‬could also refer to chariot-
drivers, perhaps especially at a time and place when chariots and not cavalry
were used. An argument in favour of this is that the more detailed account of
the Egyptian débacle later in the poem does refer explicitly to chariots (‫מרכבת‬:
v. 4), but nowhere specifically to cavalry. Elsewhere the part. ‫ ר ֵֹכב‬is indeed
sometimes used, like the verb, of riding on the back of an animal (Gen. 49.17;
Judg. 5.10; cf. Lev. 15.9), but there are also some cases where ‫ רכב‬followed
45
The dagesh in ‫ ָּג ָאה‬is surprising, like several other cases in this poem (see
GK §20e, 21d). No real explanation has been found for the divergences from the
normal pattern.
326 EXODUS 1–18
by ‫ סוס‬can with some certainty be taken to mean riding in a chariot pulled by

horses (2 Kgs 18.23 [in view of v. 24]; Jer. 6.23 = 50.42; 17.25 = 22.4; Hos.
14.4; Hab. 3.8), as well as others where this is likely (Ezek. 23.6, 12, 23) and
many more where it is possible. In addition, since the verb ‫ רכב‬could (as noted
above) mean ‘drive a chariot’, the participle could in itself mean ‘chariot-
driver’ and allow the translation here ‘(the) horse and its chariot-driver’.46
Given that only on a very late dating of the poem would Near Eastern military
history make a reference to cavalry plausible, the linguistic and contextual
arguments (which in the broader sense include the focus on Egyptian chariots
in both major components of Exod. 14) justify understanding the phrase in
this way as it stands. Some prefer to emend to ‫ ֶר ֶכב‬or ֹ‫ר ְכבּו‬,ִ ‘(its) chariot’ (see
Text and Versions), but this is not necessary or justified.47
h. Heb. ‫עזי‬. The doubled zayin points to a derivation from ‫עזז‬, ‘strong’. For
the alternative vocalisations of the suffixed forms see BL §71i and Mandel�-
kern, p. 839. Since the widespread acceptance that ‫ זמרת‬means ‘protection’
here (see the next note), the suggestion that ‫ עז‬here and in some other passages
means ‘refuge’ (from the root ‫ עוז‬or a by-form ‫ עזז‬II: cf. T.H. Gaster, ‘Exodus
xv.2: ‫’עזי וזמרת יהּ‬, ExpT 49 [1937–38], p. 189; G.R. Driver, ‘Notes on the
Psalms. I.1-72’, JTS 43 [1942], pp. 149-60 [158]; V. Hamp, ‘Ps 8,2b.3’, BZ
16 [1972], pp. 115-20 [117-19]) has become popular (cf. HAL, p. 762; DCH 6,
p. 325; NEB, REB: Ges18, pp. 941-42 is more doubtful). LXX βοηθός might
lend some support to this proposal, but ‘help’ is not quite the same as ‘refuge’,
probably no nearer than it is to ‘strength’. The association with ‫ זמרת‬does
not require that ‫ עז‬mean ‘refuge’ here, as associated words can have similar
meanings rather than identical ones. There is no real reason for envisaging
a homonym here or in the other passages suggested. The suggestion that a
word related to Ar. ǵāzī, ‘warrior’, is present here and in the parallel passages
(Isa. 12.1; Ps. 118.14), presumably with the repointing ‫( ָעזִ י‬Ben Yehuda: cf.
D. Winton Thomas, ‘A Note on Exodus xv.2’, ExpT 48 [1936–37], p. 478),
is scarcely necessary either, though it fits the context (cf. v. 3) and may find
further support in Ug. ǵz (KTU 1.16.6.43: cf. DULAT, p. 328).
46
A similar idiom seems to have existed in Egyptian, to judge from passages
in the accounts of Ramesses II’s victory at Kadesh (B88, R18 [Kitchen, RITA, II
§3, pp. 17, 20]; cf. Kloos, Yahweh’s Combat, p. 128).
47
A choice between a reference to horse-riding or chariot-driving elsewhere is
often difficult (cf. HAL, p. 1149; Ges18, p. 1242; DCH 7, pp. 486-87). Mowinckel
argued one-sidedly for chariot-driving in almost all occurrences of ‫‘( רכב‬Drive
and/or Ride’, pp. 278-99), W.B. Barrick equally one-sidedly for the preserva-
tion almost everywhere of at least an underlying sense ‘mount’ (‘The Meaning
and Usage of RKB in Biblical Hebrew’, JBL 101 [1982], pp. 481-503; cf. TWAT
7, 508-15 = TDOT 13, pp. 485-91), which HAL and Ges18 agree is of only
minor significance in BH. The best discussion is in THAT 2, 777-81 = TLOT 3,
pp. 1237-39.
15.1-21 327
i. Heb. (‫וזמרת )יה‬. Three peculiarities require explanation (as in the identical
phrases in Isa. 12.2; Ps. 118.14). (a) The preservation of the old feminine
ending ‫ת‬- in the abs. form has a few parallels, mainly in poetic texts and
in proper names (GK §80f-g: to the evidence for its being the ancient form
everywhere in §80m may be added the decisive case of Ugaritic). ‫ פרת‬in Gen.
49.22 may be an(other) early example, but the feature also occurs in Jer.
48.36 and Ezek. 28.13. (b) The absence of a pronominal suffix (contrast ‫)עזי‬
has been variously explained. Rashi, Rashbam and Ibn Ezra took ‫ זמרת‬as in
the construct state, but this ignores the Masoretic vocalisation and does not
produce a plausible sense for the line. In modern times it has been attributed
to the closely following ‫( יה‬GK §80g) or to haplography (Cross and Freedman,
Studies, p. 55: cf. the readings of SP and Vulg, also 1QIsa at 12.2). It might
‘preserve early orthography’ (Cross and Freedman, ibid.: cf. GK, ibid.) in
which final vowels were not represented (as in Ugaritic), or it might be a case
of a ‘double-duty suffix’ (cf. Freedman, Pottery, p. 200), where only one of
two associated nouns has the appropriate suffix (M. Dahood, Psalms III [AB
17A; Garden City, 1970], p. 158 [on Ps. 118.14: cf. the lists of other examples
on pp. 429-31]). (c) The long-established rendering ‘(my) song’ (see Text and
Versions: so still Houtman, p. 279), although backed up by occurrences of
‫ זמרה‬elsewhere (Isa. 51.3; Pss. 81.3; 98.5: cf. Amos 5.23) and similar state-
ments about Yahweh as (the object of) Israel’s ‘praise’ (‫תהלה‬: Deut. 10.21; Jer.
17.14), was found inferior by some commentators to the sense suggested by
LXX σκεπαστής, and they emended MT as a result (Graetz ‫ ;ועזרתי‬Gunkel
‫ ;וסתרי‬Beer ‫)וסתרתי‬. Two early indications that MT might itself bear the sense
‘my protection/protector’ were Ben-Yehuda’s definition of it as ‘a mighty man
and strong, conquering and subduing his enemies’, with ref. to Ar. ḏamara
(cf. D. Winton Thomas, ‘A Note on Exodus xv.2’, p. 478) and M. Noth’s
suggestion that some Israelite proper names including the root ‫ זמר‬might best
be explained in this way (Personennamen, p. 176; on the general approach
cf. Barr, Comparative Philology, pp. 181-84). Evidence from comparative
philology for such a meaning has grown with new discoveries: in addition to
its first attestation in Epigraphic South Arabian (ḏmr: already cited in BDB,
p. 75), it has been found in Amorite (zmr: H.B. Huffmon, Amorite Personal
Names in the Mari Texts: A Structural and Lexical Study [Baltimore, 1965],
pp. 187-88) and Ugaritic (ḏmr: DULAT, p. 287), and more tentatively in North
Arabian, Punic, Palmyrene and Eblaite (cf. Ges18, p. 305; DULAT, ibid.;
DNWSI, pp. 254, 601). Of particular significance is the Ugaritic phrase ʿzk
ḏmrk introducing a series of divine epithets in a hymn to Baal (as rpi mlk ʿlm:
KTU 1.108.24 [cf. 21-22]): cf. Dahood, Psalms III, p. 158. This meaning is
especially appropriate in the Exodus context and should probably be accepted
(cf. T.H. Gaster, ‘Notes on “The Song of the Sea” [Exodus xv.]’, ExpT 48
[1936–37], p. 45; ‘Exodus xv.2: ‫’עזי וזמרת יהּ‬, p. 189 [citing also Job 35.10];
NEB, REB; Ges18, p. 305: for further discussion see S.B. Parker, ‘Exodus xv
2 again’, VT 21 [1971], pp. 373-79). The alternative suggestion ‘my strength’
328 EXODUS 1–18
(HAL, p. 263: cf. NRSV) is less well supported from the cognate languages
and seeks an unnecessarily close synonymy with ‫עזי‬. Since ‫ זמר‬I, ‘sing’,
is derived from Proto-Semitic ZMR and ‫ זמר‬II, ‘cut’, perhaps from Proto-
Semitic ZBR (cf. Ug., Ar., Eth.: Ges18, p. 304), a ‫ זמר‬III, ‘protect’, in Heb.,
from Proto-Semitic ḎMR, adds a homonym which would only have become
a potential problem for understanding after consonantal changes in early
Hebrew (such cases are said by Barr to ‘form peculiarly certain examples of
homonymy in comparison with other types’ [Comparative Philology, p. 128]).
j. Heb. ‫יה‬. This shortened form of the divine name is mostly found (over
40x) in the Psalms, especially in the cry ‫ הללו יה‬which often opens and/or
closes psalms which are likely to be post-exilic (e.g. Ps. 150.1, 6). But it also
occurs in Psalms which are probably older (68.19; 89.9; 118.5, 17-19), as well
as in two places where the text may be corrupt (Exod. 17.16 [see the notes
there]; Isa. 38.11). Many theophoric personal names, especially but not only
in later biblical and epigraphic sources, end with this form of the divine name
(cf. J.D. Fowler, Theophoric Personal Names in Ancient Hebrew [JSOTSup
49; Sheffield, 1988], pp. 34-35, 371, 380).
k. Heb. ‫ויהי‬. Cross and Freedman (Studies, p. 55) proposed to take the waw
as the final letter of the previous word, giving ‫ יהו יהי‬as an early spelling of
‫( יהוה יִ ְהיֶ ה‬which would coincide with LXX’s omission of a conjunction at this
point): they took ‫ יהי‬as a preterite, like other imperfects/jussives in the poem
(see Note s on the translation).
l. Heb. ‫לישׁועה‬. On the meaning of ‫ ישׁועה‬see Note gg on the translation of
14.1-31. Here again the threat posed by the Egyptians is prominent in the
context and the sense ‘deliverance’ (or ‘deliverer’) may be preferred.
m. Heb. ‫ואנוהו‬. A verb ‫ נוה‬occurs in BH elsewhere only in Hab. 2.5, where
the meaning would be ‘dwell, abide’ (BDB, p. 627) or ‘succeed’ (Ges18,
p. 791: cf. Ar.) if the text is correct. Neither of these meanings fits the use of
the Hiphil here. Cross and Freedman render ‘admire’ on the basis of Albright’s
speculation about the semantic development of a root attested in Ar. (Studies,
p. 56; still in Canaanite Myth, p. 127), but the sense is any case weak. The
Vss, apart from TgO (on which see Text and Versions), uniformly render
‘praise’ as the context suggests: this was related to MH ‫( נוי‬a later form of
‫)נאה‬, hence ‘beautify, adorn’, already in MRI (Lauterbach 2, p. 25: cf. BDB,
p. 627); cf. Sir. 13.3. Ar. nawwaha (HAL, p. 641, proposing revocalisation as a
Piel here; Ges18, p. 791), from a verb meaning ‘be high’, provides a different
semantic basis for ‘praise’. Freedman, Pottery, Poetry and Prophecy, p. 200,
takes the initial waws in this line as emphatic (cf. p. 205 on v. 7): for this
possibility see Note aa below.
n. Heb. ‫וארממנהו‬. The retention of he after energic nun is rare and mainly
restricted to poetry (GK §58i, k), but it is not limited to early texts.
o. Heb. ‫אישׁ מלחמה‬, lit. ‘a man of war’. The variants in the textual tradition
(see Text and Versions) are due to unease over the apparent description of
Yahweh as a human being (contrast Num. 23.19; Hos. 11.9), and so confirm its
15.1-21 329
originality rather than calling it in question. For the use of ‫ אישׁ‬to form descrip-
tive phrases cf. BDB, pp. 35-36; for this phrase used of human warriors see
e.g. Josh. 17.1; 1 Sam. 16.18.
p. Heb. ‫ירה‬. The verb is mainly used of shooting arrows, but the sense
‘throw’ is attested in Josh. 18.6 and 4Q169 fr.3-4, 4.2 (of casting lots) and
(in Hiphil) in Job 30.19 (of a human being cast into mire [‫)]ח ֶֹמר‬: cf. also the
cognates cited in the lexx. This verb is now generally distinguished from the
homonym used in the Hiphil for ‘teach’ (whence ‫)תורה‬: cf. HAL, pp. 416-17;
Ges18, pp. 494-95; DCH 4, pp. 290-91. Three different rare words are there-
fore used here of the ‘ejection’ of the Egyptian charioteers into the sea (cf.
‫ וינער‬in 14.27; ‫ רמה‬in 15.1, 21): the common ‫ השׁליך‬appears only in Neh. 9.11.
q. Heb. ‫שׁלשׁיו‬. On the meaning of ‫ שׁלישׁ‬see Note t on the translation of
14.1-31.
r. Heb. ‫טבעו‬. In BH the closest parallels relate to sinking in ‘mire’ rather
than water, but the latter association is attested in Akk., later Aram. and Eth.
Propp (p. 517) suggests it is an archaism here.
s. Heb. ‫יכסימו‬. The sense must be preterite, as often with prefixed forms
in this poem (see above, Note c): a perfect follows in the next clause. Such
alternation is common in Hebrew and Ugaritic poetry (see Sivan, Grammar,
p. 107). The form is doubly unusual (‘doubly archaic’ according to Cross and
Freedman, Studies, p. 58), with the third radical yodh retained (cf. GK §75dd)
and the suffix uniquely written ‫מוּ‬-, (which BL §21j deems the more ancient
form of ֹ‫מו‬-, for which cf. v. 7 and Note y below).48
t. Heb. ‫נאדרי בכח‬. The verb is rare and the Niphal (only here and v. 11)
presumably serves as the passive of the Hiphil (Isa. 42.21): cf. GK §51f. The
so-called hireq compaginis occurs regularly in the construct and suffixed
forms of certain nouns (‫אב‬, ‫אח‬, ‫)חם‬, occasionally in construct forms of other
nouns, adjectives and participles and, in late poetic texts, with participles not
in the construct state. It seems likely that that its use was extended from the
first category to the second and then to the third (GK §90k-m; JM §93l-o).
Here the participle is in the construct state before a prepositional phrase
(GK §130a). Despite some similarity to the old genitive case-ending, this is
unlikely to be the source of the phenomenon, since its function is completely
different.49 ‫ נאדרי‬is often taken with ‫( ימינך‬cf. Text and Versions), but its masc.
48
One might say triply unusual, in view of the defectively written pl. ending
(cf. Propp, p. 517), a relic of the ancient spelling. But this feature remained in
widespread use (BL §48i).
49
Possible analogies have been found in some Babylonian forms of the
construct state (cf. GAG §64e, h: so G.R. Driver, ‘The Origin of “ḥireq compag�-
inis” in Hebrew’, JTS 26 [1925], pp. 76-77) – but in Heb. the vowel cannot be
described as a ‘helping vowel’ – and in the endings of what seem to be infinitives
absolute in Amarna Canaanite (W.L. Moran, ‘The Hebrew Language in its North-
West Semitic Background’, in G.E. Wright [ed.], The Bible and the Ancient Near
330 EXODUS 1–18
form is against this and it is best regarded as an attribute of ‫יהוה‬, as in v. 11

(cf. Vulg and possibly TgO and Sy here). As for the meaning, traditionally
‘glorious’ has been preferred (cf. Text and Versions), but the more recent
lexica etc. have given a greater priority to the element of ‘power’ in the
word-group, which fits both biblical usage (esp. the adj. ‫אדיר‬: and cf. SamTg
ḥsynʾ here) and cognates elsewhere (esp. Phoen. [DNWSI, pp. 17-19], but also
Ug. [DULAT, pp. 20-21]). The association with ‫ כח‬supports this. ‘Mighty’ or
‘majestic’ is therefore preferable here and in v. 11. Cross and Freedman argued
that ‘the archaic sense…“the One-to-be-feared” ’ is most appropriate (Studies,
p. 59), citing Akkadian adāru as well as Ug. words, but the sense ‘fear’ seems
to be confined to Akkadian (for which see AHw, pp. 11-12) and in v. 11 that
attribute is expressed separately by ‫נורא תהלת‬.
u. Heb. ‫תרעץ‬. In BH ‫ רעץ‬only occurs elsewhere in Judg. 10.8, together
with (perhaps glossed by) the more common ‫רצץ‬. Both verbs (as well as ‫רעע‬
II, which is commonly regarded as an Aram. loanword: BDB, p. 949) go back
to Proto-Semitic RḌḌ (cf. Ar. raḍḍa). The spelling ‫ רעץ‬may be the result of
dissimilation (cf. Moscati, pp. 58-59). The sense is again preterite (cf. Notes
c and s) and the form most likely third person fem. sing. (although it could be
second person m. sing., with ‫ ימינך‬understood instrumentally).
v. Heb. ‫וברב־גאנך‬. Lit. ‘with the greatness of your majesty’ (cf. GK §128r).
‫גאון‬, like the verb ‫( גאה‬v. 1: see Note f), is used of ‘rising high’ in both positive
and negative (as well as neutral) senses: for the positive sense see e.g. Isa.
2.10; 4.2.
w. Heb. ‫תהרס‬. Again the sense is preterite, as it is with the other verbs in
this verse. The verb ‫ הרס‬is normally used of the demolition of buildings etc.
(hence some variations in the Vss), but in Isa. 22.19 it also refers to human
‘downfall’.
x. Heb. ‫קמיך‬. Cross and Freedman (Studies, p. 59) rightly observed the
use of this nominalised part. in other poems which are likely to be early
(Deut. 33.11; Ps. 18.40, 49 par.): it also occurs once in Ug. (in parallel to ʾib,
‘enemy’: cf. v. 6) of Baal’s foes (KTU 1.10.2.25). But it occurs in later texts
too (Ps. 74.23; Lam. 3.62). On the use of the suffix cf. GK §116g-i.
y. Heb. ‫יאכלמו‬. ֹ‫מו‬- is the more widespread form of the third person pl. suffix
which preserves the original vocalic ending (on the unique ‫מוּ‬- in v. 5 see Note
s) and it occurs eight times with verbs in this poem (simple ‫ם‬- never appears
in it); in texts from the monarchy period or earlier cf. Deut. 33.29; Pss. 2.3,
4, 5; 21.10, 11, 13; 45.17; 80.6. The final vowel could have been lost in the
transmission of other early poems. The longer form certainly continued to
East [FS W.F. Albright; London, 1961], pp. 54-72 [60], comparing also ‫ אסרי‬in
Gen. 49.11). The feature is confirmed by Rainey, Canaanite in the Amarna Rablets
2, pp. 382-88. But this is no reason to repoint the form here as an inf. abs. (‫נֶ ְאדּ ִֹרי‬:
presumably equivalent to a finite verb), since the participial form makes good
sense.
15.1-21 331
appear in later texts, especially with nouns, but there is no doubt that it is an
archaic feature of the language (cf. GK §58g, 91l).
z. Heb. ‫כקשׁ‬, with the def. art. as often in comparisons (GK §126n-p; cf.
Bekins, ‘Non-Prototypical Uses’).
aa. Heb. ‫וברוח־אפיך‬. For the use of the dual to mean ‘nostrils’ (rather than,
as elsewhere, ‘face’) cf. Gen. 2.7; 7.22 etc. For reasons discussed in the
Explanatory Note on v. 8 the initial waw is probably best understood not to
indicate succession but as either causal (JM §170c: e.g. Ps. 7.10) or emphatic
(JM §177n: e.g. Ps. 89.38b). See also IBHS §39.2.3-4 on ‘disjunctive’ and
‘epexegetical’ waw, and Note m above on Freedman’s interpretation of v. 2.
bb. Heb. ‫נערמו‬. The verb occurs only here in BH, but the related noun
‫ ערמה‬is quite common, at least in the later books: in 8.10 ‫ ח ֶֹמר‬is used, cf. ‫חמור‬
in Judg. 15.16 (see also below on ‫)נד‬. There are cognates in Aram., Ar. and
probably Ug., though in the latter case the words in question are rare and their
meaning is debated (DULAT, p. 326). But a verb ǵrm occurs in KTU 1.82.5,
apparently of the sun-god ‘bringing together’ her rays on Mot and a noun ǵrm
is used in KTU 1.3.2.11 of grasshoppers (swarming?) and in KTU 1.16.6.44
of a (unit of?) fighting men (cf. par. ǵzm). The plausibly common element in
these passages is scarcely ‘a heap’, but ‘gathering together’, and it is easy
to see how such a sense might have come to have the specialised meaning
‘heap’ in later BH, Aram. and Ar. (in Syriac there may be some relics of the
more general sense: cf. Payne Smith, pp. 428-29): whether this narrowing
had become fixed in pre-exilic Heb. or not is immaterial, as the broader sense
would fit the present context just as well.50
cc. Heb. ‫כמו־נד‬. The MT accents imply that ‫ נד‬is not in the construct state
and most of the Vss take the same view (only TgN clearly links it to the
next word). In Ps. 78.13 ‫ נד‬is used again with reference to the Exodus and
in Josh. 3.13, 16 of the similar phenomenon at the crossing of the Jordan.
In Ps. 33.7 it appears in a comparison to the ‘gathering’ (Heb. ‫ )כנס‬of the
waters at the creation, but LXX and Sy read it as ‫נאד‬, ‘skin-bottle’, which is
a possible meaning there. In Isa. 17.11 it occurs with ‫קציר‬, but the text there
has long been doubted (cf. RV ‘shall flee away’, Vulg ablata est and BHS).
The meaning ‘heap’ can be supported from Ar. naddun (BDB, p. 622), but it
is striking that all the textually secure uses in BH (where there are no related
words) refer to water. Possibly the fact that the Exodus references involve a
comparison, while the Jordan ones do not, is significant: maybe ‫ נד‬was used
50
The Ar. cognates have ayn rather than ǵayin, as Ug. ǵayin seems occasion-
ally to correspond to Ar. ayn: see the general conclusions of J.A. Emerton, ‘Some
Notes on the Ugaritic Counterpart of the Arabic ghayin’, in G.E. Kadish and
G.E. Freeman (eds.), Studies in Philology in Honour of Ronald James Williams
(Toronto, 1982), pp. 31-50 (47-48), although he thought that ‘The precise meaning
of ǵrmn [the form in KTU 1.3.2.11, the only one of the passages which he
discussed] is still uncertain’ (p. 47).
332 EXODUS 1–18
originally of a river phenomenon, a ‘bore’, to which the movement of the

waters in the Exodus story could be compared (this might also be true in
Ps. 33.7). HAL (p. 634; cf. Ges18, p. 783) proposed ‘dam’, citing cognates
in Ammonite and Akkadian, but neither provides any basis for this meaning:
none of the meanings for nīdu given in AHw, p. 786, is relevant, and the
reading ‫ נד צדק‬originally proposed for line 4 of the Amman Citadel inscrip-
tion has been generally abandoned (cf. W.E. Aufrecht, A Corpus of Ammonite
Inscriptions [ANETS 4; Lewiston, 1989], pp. 154, 161; W.H. Shea, ‘The
Architectural Layout of the Amman Citadel Inscription Temple’, PEQ 123
[1991], pp. 62-66 [63]). W.F. Albright gave a similar interpretation to ‫ נד‬here,
‘dykes (of irrigators)’ (Yahweh, pp. 40-41; cf. ‘From the Patriarchs to Moses.
II. Moses Out of Egypt’, BA 36 [1973], pp. 48-76 [61-62]) on the basis of the
Ar. cognate, but his interpretation of ‫נזלים‬, which is crucial to his suggestion,
has no parallel in its other occurrences. The rendering ‘wall’ in some of the
Vss (see Text and Versions) simply uses the prose account in ch. 14 to clarify
a difficult word: Cross and Freedman’s adoption of it (Studies, p. 51: so also
L.L. Grabbe, ‘Comparative Philology and Exodus 15,8: Did the Egyptians
Die in a Storm?’, SJOT 7 [1993], pp. 263-69 [266], as a serious possibility) is
inexplicable – in Canaanite Myth, p. 128, Cross took ‘hill’ to be the meaning
on the basis of Ar., but it is scarcely more satisfactory.
dd. Heb. ‫קפאו‬. The few other occurrences of the verb in ancient Heb. all
mean ‘thicken’ or ‘solidify’ (Zeph. 1.12; Job 10.10 [of cheese coagulating];
Sir. 43.20 [of ice forming]; cf. ‫[ קפאון‬prob. reading] in Zech. 14.6) and Aram.
cognates can bear a similar sense. Cross and Freedman proposed first ‘churn’
and then ‘foam’ on the basis of MH and later Aram. (cf. Canaanite Myth,
p. 128 with n. 59), but there is no sign of these meanings elsewhere in BH and
the idea that an ‘action’ word is required here (Studies, p. 60) is in fact against
the context of the verse: it is only in v. 10 that the waters begin to move again
and cover the pursuing Egyptians.
ee. Heb. ‫תמלאמו‬. On the form of the suffix (as also with ‫ תורישׁמו‬later in
the verse) see Note y above. Cross and Freedman (Studies, p. 60), following
Albright, saw here an example of ‘enclitic mem’, but this is unnecessary (see
J.A. Emerton, ‘Are there Examples of Enclitic mem in the Hebrew Bible?’,
in M.V. Fox et al. [eds.], Texts, Temples and Traditions (Winona Lake, 1996),
pp. 321-38 [326]).
ff. Heb. ‫אריק חרבי‬. This idiom for drawing a sword occurs otherwise only
in Lev. 26.33; Ezek. 5.2, 12; 12.14; 28.7; 30.11 (with ‫ חנית‬also in Ps. 35.3):
at Hab. 1.17 1QpHab 6.8 reads ‫ )יריק( חרבו‬for MT ‫חרמו‬, ‘his net’, and this
may well be the original reading. Elsewhere the verb ‫ שׁלף‬Hiph. is used (e.g.
1 Sam. 17.51; and cf. the idiom ‫ אישׁ שׁלף חרב‬in Judg. 8.10 etc.). The basis for
the idiom with ‫ ריק‬Hiph. is not certain (cf. HAL, p. 1145): possibly the sword
is thought of as being ‘emptied out’ from its sheath or (perhaps more likely) it
is the sheath which is ‘made empty’ by its removal. It is striking that the other
15.1-21 333
occurrences of the idiom all (except for Ps. 35.3, which has been assigned
various dates) come from the sixth (or late seventh) century.51 This must be a
problem for those who would give Exod. 15.1-18 a very early date.
gg. Heb. ‫תורישׁמו‬. BDB, p. 440, and Ges18, pp. 501-502, give the meaning
of ‫ ירשׁ‬Hiph. here as ‘bring to ruin, destroy, exterminate’ (cf. Houtman,
p. 285). This is strongly supported by the Vss (see Text and Versions: only
LXX rendered otherwise) and some passages in the Heb. of Ben Sira come
close to it (20.22; 32.23). But the latter may rather mean ‘make poor’ (i.e.
from √ ‫ רישׁ‬or a byform of it) and the only other instance proposed for
‘destroy’ is Num. 14.12, where ‘disinherit’ is a plausible alternative. Here, in a
context which refers to ‫שׁלל‬, the common sense ‘dispossess’ is surely adequate
(so HAL, p. 421; Propp, pp. 525-26; Dozeman, pp. 319, 324).
hh. Heb. ‫נשׁפת‬. The spelling with ‫ פ‬occurs elsewhere only in Isa. 40.24: a
form ‫( נשׁב‬which SP has here) is found in Gen. 15.11; Isa. 40.7; Ps. 147.18;
Sir. 43.20; 4Q185.1.1.10. Both are attested in MH and Aram. as well. ‫נשׁם‬
and ‫ נשׁמה‬are almost certainly related. In cognate languages nsb occurs in Ar.
and našāpu in Akk. (AHw, p. 758). For such interchange between labials in
different languages cf. Moscati, Introduction, pp. 25-26; Ges18, p. 118: within
Heb. ‫( בנבשׁכם‬for ‫ )נפשׁ‬in AHI 2.24.18 shows the same variation.
ii. Heb. ‫צללו‬. ‫ צלל‬II occurs only here in BH, but the sense ‘sink’ (cf.
Vss) is found in MH and a root ‫( צול‬whence ‫צולה‬, ‫ )מצולה‬may be a by-form
of it (so HAL, p. 962; Ges18, p. 1119). Akk. ṣalālu, ‘lie down, sleep’, ESA
ḍll, become ill’, and Ar. ḍalla, ‘perish’, are probably related (cf. Cross and
Freedman, Studies, p. 61; G.R. Driver, ‘Hebrew Homonyms’, in Hebräische
Wortforschung [FS W. Baumgartner; VTSup 16; Leiden, 1967], pp. 50-64
[62]).
jj. Heb. ‫כעפרת‬. For the def. art. in a comparison and with a well known
substance cf. Note z above.
kk. Heb. ‫כמכה‬. The final ‫ ה‬is by no means ancient orthography, or even
standard BH. There are some late pre-exilic examples of it in inscriptions with
second person m. sing. perf. verbs (Renz and Röllig, Handbuch, II/2, p. 45),
but none that is certain with the suffix ‫ך‬-. For BH see Bergsträsser, 1, §7e; on
the much wider use of ‫ה‬- at Qumran see Qimron, pp. 43, 58.
ll. Heb. ‫באלם‬. The sing. ‫ אל‬refers to God or a god (the latter several times in
Deutero-Isaiah [e.g. 43.10], where it is also used for an image identified with
a god [44.10, 15, 17], and so the pl. (which is rarer) is naturally understood to
mean ‘gods’ (Pss. 29.1; 89.7 (both with ‫ ;)בני‬Dan. 11.36; perhaps Job 41.17).
The older view that it was from the root ‫ אלה‬is now doubted (cf. TWAT 1,
51
BDB, p. 938 (citing Dillmann), compared the Akk. idiom kakkē tabāku (cf.
now AHw, pp. 1295-96), but it apparently means to ‘throw’ rather than to ‘draw’
a weapon and bears no relation to the Heb. expression here.
334 EXODUS 1–18
261-62 = TDOT 1, p. 244; THAT 1, 142 = TLOT 1, p. 107). Another pl. form,
used of human leaders (cf. v. 15), is separate and probably a metaphorical
use of ‫‘ = איל‬ram’ (cf. HAL, p. 39). At Qumran ‫ אל‬was also used to refer to
angels (cf. DCH 1, p. 253; TWQ 1, pp. 179, 182), a later use which probably
accounts for some early interpretations of the biblical instances of the pl. (see
Text and Versions).
mm. Heb. ‫בקדשׁ‬. ‫ ק ֶֹדשׁ‬is sometimes used alone in a concrete sense to mean
‘holy place, sanctuary’, i.e. as an equivalent to ‫מקדשׁ‬: the instances are concen-
trated in Ezekiel, P, Chr. and other later texts. This could theoretically be the
sense here (cf. ‫ב‬, which might have a literally local meaning), but the abstract
‘holiness’ is more likely in a context that at this stage is concerned with divine
action and qualities. On the exact sense implied (which might almost be
‘God-ness’, ‘divinity’) see the Explanatory Note.
nn. Heb. ‫נורא תהלת‬. ‫תהלת‬, lit. ‘praises’, can mean ‘praiseworthy acts’ (BDB,
p. 240, s.v. 4), so it can give the reason why Yahweh is ‫( נורא‬cf. GK §116h-l for
a part. in the constr. st. being defined more closely in this way).
oo. Heb. ‫פלא‬. As the preceding verb suggests, the abstract noun is probably
used for concrete ‘wonderful acts’ here.
pp. Heb. ‫תבלעמו‬. On the preterite sense and the form of the suffix see above
Notes s and y. Here a preterite imperfect follows a perfect (the reverse of v. 5).
qq. Heb. ‫ארץ‬. The mention of the ‘earth’ is at first surprising and caused
difficulty to the Vss (see Text and Versions). The only other places where
‫ ארץ‬is the subject of ‫ בלע‬are Num. 16.32, 34 and other references to the same
episode, but there it is clear that an opening up of the (dry) ground is meant.
While here ‫ ארץ‬might refer to the mud at the bottom of the sea, in a poetic
passage it is more likely equivalent to ‘death’ or the grave, as more clearly
in the expressions ‫ ארץ תחתית‬in Ezek. 31.14, 16, 18 and ‫ ארץ תחתיות‬in Ezek.
26.20; 32.18, 24, where words like ‫מות‬, ‫קבר‬, ‫ בור‬and ‫ שׁאול‬are present close by,
and ‫ ארץ‬in Jon. 2.7; Ps. 22.30. This is not quite to endorse the view, (over-)
popularised by M. Dahood (but first proposed by Gunkel in 1895 according
to Childs, p. 243), that ‫ ארץ‬alone could actually be synonymous with ‫שׁאול‬
(which is used with ‫ בלע‬in Prov. 1.12) and mean ‘the underworld’ (see his
Proverbs and North-West Semitic Philology [Rome, 1963], p. 52; N.J. Tromp,
Primitive Conceptions of Death and the Nether World in the Old Testament
[Rome, 1969], pp. 23-46 [25-26]; so here already Cross and Freedman, ‘The
Song of Miriam’, p. 247 n. 39 [cf. Canaanite Myth, p. 129]; TWAT 1, 430-31
= TDOT 1, pp. 399-400, and Propp, pp. 530-31, are sympathetic to this view),
but it is close to it. In Ug. ʾrṣ certainly can mean ‘the underworld’ (cf. DULAT,
pp. 106-108) and passages like this one at least show the influence of such a
usage.
rr. Heb. ‫בחסדך‬. The more recent study of the meaning of ‫ חסד‬has shown
that it means neither the fulfilment of the requirements of a covenant (except
in Deuteronomy and books dependent upon it) nor any act of (exceptional)
‘kindness’ (see the overview in G.I. Davies, Hosea [NCB; London, 1992],
15.1-21 335
pp. 94-97; more fully K.D. Sakenfeld, The Meaning of Hesed in the Hebrew
Bible: A New Inquiry [HSM 17; Missoula, 1978], and Faithfulness in Action:
Loyalty in Biblical Perspective [OBT; Philadelphia, 1985]; THAT, 1, 600-21
= TLOT 2, pp. 449-64; TWAT 3, 48-71 = TDOT 5, pp. 44-64). On the human
level, as examples in Genesis show (19.19; 20.13; 21.23; 24.49; 40.14; 47.29),
it refers to supportive action ‘that could be expected of an individual within
a definite social context, whether this be the family or hospitality towards
a guest or even a context created by the generous initiative of one party
towards the other’ (Davies, p. 95). In its religious use, which alone is present
in Exodus (cf. 20.6; 34.6-7), Yahweh’s ‫‘ חסד‬does not of itself presuppose
the explicit formulation of the covenant (berît) concept…it does, however,
assume the special relationship between Yahweh and Israel (or the individual
worshipper), which eventually came to be defined as one of election or
covenant, but which may have received its original formulation simply in the
designation of Israel as Yahweh’s people’ (ibid., pp. 95-96): cf. Exod. 3.7, 10;
5.1. etc. A precise English equivalent does not exist, but ‘loyalty’ expresses
the relational context best of the suggested renderings (Luther ‘Barmherzig-
keit’, AV, RV ‘mercy’ [from LXX, Vulg: see Text and Versions]; RSV, NRSV,
ESV ‘steadfast love’; NJPS ‘love’; JB ‘grace’; NEB, REB ‘constant love’; EÜ
‘Güte’; NIV ‘unfailing love’).
ss. Heb. ‫זוּ‬, as again in v. 16. Although in origin a demonstrative pronoun
(like the rarer ֹ‫)זו‬, ‫ זוּ‬is more commonly found as a poetic form of the relative
particle (see also Isa. 42.24; 43.21; Pss. 9.16; 10.1; 17.2; 31.5; 32.8; 68.29;
132.12 [ֹ‫ ;]זו‬142.4; 143.8; ‫ הז‬is only occasionally so used: Isa. 25.9; Pss. 78.54;
104.8, 26; Prov. 23.22; Job 15.17; 19.19 [probably not in Exod. 13.8: see Note
l on the translation of 13.1-16]). The recognition of this sense here was long
forgotten and seems only to have been recovered in the sixteenth century (see
Text and Versions).
tt. Heb. ‫גאלת‬. Where it is associated with a verb like ‫( הציל‬as in 6.6), ‫גאל‬
may be given the sense ‘set free’ (cf. Note l on the translation of 6.1-9), but it
has a more general meaning, seen in family law, in which it refers to various
kinds of intervention by a kinsman on behalf of a person who has fallen on
hard times (see the Explanatory Note on 6.6-8), and this probably informs its
theological use here and elsewhere.
uu. Heb. ‫נהלת‬. The verb occurs ten times in BH, all but one (Gen. 33.14) in
the Piel; in most cases the meaning is ‘lead’ or (2 Chr. 28.15) ‘bring’, which the
parallel with ‫( נחה‬again in Ps. 31.4) and the following ‫ אל‬would suggest here.
BDB, pp. 624-25, gave the basic meaning as ‘lead, guide to a watering-place
or station and cause to rest there’ on the basis of Ar. (cf. ESA nhl, ‘watering-
place’, in Ges18, p. 788), which makes possible an easier accommodation of
some rarer senses (Gen. 47.17; 2 Chr. 32.22) as well as an explanation for the
possibly related hapax ‫ נהלל‬in Isa. 7.19. But although some BH occurrences
are connected with water (Isa. 49.10; Ps. 23.2: not so clearly Isa. 40.11), this
need not be because ‫ נהל‬had such a specialised meaning (similarly Propp,
336 EXODUS 1–18
p. 532) and other lexica deal with the range of meaning in different ways.
For more detailed study of the root see R.D. Wilson, ‘Hebrew Lexicography
and Assyriology’, Presbyterian Review 6 (1885), pp. 319-28 (319-21: two
roots: cf. DCH 5, pp. 630-31); P. Haupt, ‘The Hebrew Stem nahal, to Rest’,
AJSL 22 (1905), pp. 195-206 (followed by TWAT 5, 279-80 = TDOT 9,
pp. 260-61, and Houtman, p. 287).
vv. Heb. ‫נוה־קדשׁך‬. For the nomen rectum as equivalent to an adjective cf.
GK §128p. BH ‫נָ וֶ ה‬, like the fem. form ‫נָ וָ ה‬, is sometimes used for animal pasture
(2 Sam. 7.8 par.; Isa. 65.10; cf. Jer. 33.12), but more often for places or areas
of human habitation, with some instances where a people is compared to a
flock of sheep perhaps forming the bridge between these senses (Jer. 23.3;
49.20 par.; 50.19; Ezek. 34.14). Akk. nawûm, ‘pasture land, steppe’, with its
derivatives (AHw, pp. 729-30, 771), strengthens the argument that the general
use of human habitations was a secondary development from this: see further
A. Malamat, Mari and the Early Israelite Experience (Oxford, 1989), pp. 43-
47. Like other words for human dwelling-places (cf. the notes on v. 17), ‫נוה‬
could also be used for sacred sites where God was believed to be present in
a special sense, even in a city such as Jerusalem (2 Sam. 15.25; Isa. 33.20;
perhaps also Jer. 31.23 [cf. the parallelism]).52 MRI (Lauterbach 2, pp. 26,
70) cited Ps. 79.7 as a further case, but ‘Jacob’ is the natural antecedent for
‘his’ there. On the debate over whether Jerusalem, Sinai or somewhere else is
meant here see the Explanatory Note.
ww. Heb. ‫ירגזון‬. After the perfect ‫ שׁמעו‬a preterite use of the imperfect is
required, as several times already in this poem. ‫ רגז‬is often used of the effects
of an earthquake (e.g. Amos 8.8), and then metaphorically of people ‘quaking’
in fear (cf. Ps. 99.1). The asyndeton of MT is a frequent feature of the poem
(cf. vv. 5, 7, 8, 9, 10, 12).
xx. Heb. ‫חיל‬. The noun is most often used of the pains of childbirth (cf.
LXX, Vulg), but significantly as a picture of the terror caused by a powerful
enemy, so the use here of mental anguish is not surprising: the more so as the
verb ‫ חול‬is frequently so used (BDB, p. 297).
yy. Heb. ‫אלופי‬. This term is used for Edomite leaders or clans (see below)
in Gen. 36.15-18, 29-30, 40-43 (par. 1 Chr. 1.51-54) and in connection with
Judah in Zech. 9.7; 12.5-6. The Vss uniformly support the rendering ‘chiefs,
leaders’ and this has been the view followed in lexica and EVV until recently
(cf. also Bartlett, Edom, p. 90). NJPS and NRSV, however, have ‘clans’, an
interpretation previously suggested for the verse in Zechariah with emenda-
tion of the text to ‫א ֶלף‬/‫י‬
ֶ ‫א ְל ֵפ‬,ַ which apparently bears this meaning in a few
52
But to say ‘It is a specific designation of a tent-shrine’ (Cross, Canaanite
Myth, p. 125; cf. Studies, p. 52, ‘encampment’) claims too much. A further occur-
rence of ‫ נוה‬referring to the divine dwelling in Jerusalem was once found in an
inscription from Kh. Beit Lei (cf. AHI 15.007), but the reading ‫ נקה‬is now gener-
ally preferred (cf. Renz and Röllig, Handbuch 1, p. 248).
15.1-21 337
places (cf. Mic. 5.1). Cognates with the meaning ‘chief, commander’ seem to
exist in Ug. ʾulp (KTU 1.66.11, 33; perhaps also present in PNs, cf. DULAT,
p. 63)53 and Off. Aram. (Persepolis ritual texts [ed. Bowman] 52.3, 118.3: cf.
DNWSI, p. 65), and an instance from Gen.R. is cited by Jastrow (p. 68). In
BH ‫ אלוף‬can also mean ‘tame, docile’ (Jer. 11.19) or ‘friend, associate’ (Jer.
13.21; Mic. 7.5 etc.) and even ‘husband’ (Jer. 3.4; Prov. 2.17), and the sense
‘companions’ might conceivably be the basis for its application to the ruling
elite of Edom. But the other ‘Edomite’ contexts are rather ambiguous and the
sense ‘clan’ is possible there. A choice between ‘chiefs’ and ‘clans’ is difficult,
but the agreement of the Vss and at least some support from related languages
probably tips the balance in favour of ‘chiefs’.54 It would also make a closer
parallel to ‫אילי מואב‬, though that it is not a decisive consideration.
zz. Heb. ‫אילי‬. The general sense ‘ruler, commander’ is clear from the other
occurrences (2 Kgs 24.15Q [the K is ‫אולי‬, which would be a hapax with the
same sense]; Ezek. 17.13; 31.11; 32.21).55 There are two possible etymolo-
gies: the Vss give strong support to a connection with ‫אל‬,ֵ ‘strength’ (see Text
and Versions), while the modern lexica see it as a metaphorical development
from ‫‘ = ַאיִ ל‬ram’ (‘as leader of the flock’: BDB, p. 18): cf. the use of ‫עתוד‬,
‘he-goat’ in Isa. 14.9 and Ezek. 34.17. Cross and Freedman (Studies, p. 62)
drew attention to similar phrases in KTU 1.15.4.6-8: see now the fuller survey
of P.D. Miller, ‘Animal Names as Designations in Ugaritic and Hebrew’, UF
2 (1970), pp. 177-86 (with discussion of Heb. [and possibly Phoen.] ‫ איל‬on
pp. 181-82).
aaa. Heb. ‫יאחזמו‬. For the preterite sense (cf. ‫ אחז‬in v. 14) and the suffix ‫מו‬-
see above Notes s and y.
bbb. Heb. ‫תפל‬. Up to this point the Vss were content to render the imperfect
verbs in the poetic narrative as preterites, no doubt encouraged by the alter-
nation with perfect tenses in vv. 4-5, 8-10, 12-15. But from here to the end
of the poem, where perfect forms occur only in the latter parts of vv. 16-17,
which are or may be subordinate clauses, the imperfects are consistently
treated as future or jussive in sense. This is strange, because the third person
pl. pronouns in v. 16 must refer to the peoples who have been the subject of
verbs translated in the past tense in vv. 14-15, so that v. 16 hardly looks like
a place where the poem would change its horizon from the past to the future.
English and Latin versions of the later sixteenth century solved this problem
53
The older view that ʾulp in KTU 1.40, 1.84 and 1.154 meant ‘prince’ has
been superseded (cf. DULAT, p. 2) in favour of its analysis into three components
ʾu, l and p, meaning ‘and/or like’.
54
Since ‫ אלוף‬in this sense is nearly always used of Edomites (only otherwise in
Zech. 9-14), it may well have been an originally Edomite word (J. van der Ploeg,
‘Les chefs du people d’Israël et leurs titres’, RB 57 [1950], pp. 40-61 [51]).
55
Further occurrences have been conjectured at Jer. 25.34; Ezek. 30.13; Job
41.17 (HAL, pp. 38-39; Ges18, p. 46), but MT is preferable in each case.
338 EXODUS 1–18
by placing the change from past to future in v. 14 (or even, in the case of the
Geneva Bible, in v. 13), which also had the advantage of fitting the narrative
setting, by the Red Sea, where ‘Moses and the Israelites’ (v. 1) were supposed
to have uttered this song. Theological considerations may also have played
a part (see my essay ‘Some Points of Interest’, pp. 253-55). But in a poem
where the preterite use of the imperfect is so widespread, it is surely preferable
to make connected sense of it by extending this understanding of the verbs
as far as v. 17 (for the implications of this for the date of the poem see the
introduction to this section).
ccc. Heb. ‫אימתה‬. Such forms (mainly confined to poetry) in which a
‘paragogic he’ is added without change to the meaning (with fem. nouns to
the older form ending in ‫ת‬-) are apparently due to the weakening of an ending
which originally expressed direction towards a place (e.g. ‫)ארצה‬.56 With the
discovery of Ugaritic the older view (GK §90c-i; still in BL §65w) that this
was a remnant of the ancient accusative case-ending -a had to be given up,
because forms like ʾarṣh showed that the he was originally consonantal and
not just a vowel-marker (UT §8.56; Sivan, pp. 178-79). Possible survivals of
the accusative ending and examples of the paragogic he, including its original
directional use, are therefore carefully distinguished in JM §93 (for the closest
parallels to ‫ אימתה‬here, such as Pss. 3.3; 44.27; 120.1, see para. j) and in IBHS
§10.5. Although the morpheme is ancient, examples of it are found in poetry
of all periods.
ddd. Heb. ‫ ִבּגְ דֹל‬, i.e. the constr. st. of the adj. ‫( גָּ דוֹל‬the pointing is confirmed
by plene writings in 4QExc and SP: see Text and Versions) is used in place
of the regular noun ‫( גּ ֶֹדל‬a word popular in Deuteronomy and Ezekiel): cf. the
very similar use of ּ‫( גּבֹ ַה‬in place of ּ‫ )גּ ַֹבה‬in 1 Sam. 16.7 and GK §132c, where
some related examples are given. Possibly both here and in 1 Sam. 16.7 (cf.
Driver ad loc.) the forms could be read as infinitives construct of the related
verbs: ּ‫ ִכּגְ בֹ ַה‬in Ps. 103.11 shows that o-forms of the inf. cons. can occur with
verbs which have an imperfect with -a- as the second stem-vowel.
eee. Heb. ‫ידמו‬. For the (‘Aramaic’) form, based on a sing. ‫יִ דֹּם‬, cf. GK
§67g. Cross and Freedman (Studies, p. 63) read the Niph. ‫יִ ַדּמּוּ‬, ‘they are
struck dumb’ (Canaanite Myth, p. 130: ‘they were…’). It is not clear what
they mean by ‘The niphal is the preferred form’, when the Qal is over three
times as common, and the change is quite unnecessary. In any case the Niph.
forms may well belong to a different root (cf. Ges18, pp. 254-55, and more
fully G.R. Driver, ‘A Confused Hebrew Root [‫דום‬, ‫דמה‬, ‫’]דמם‬, in M. Haran and
B.Z. Luria [eds.], Sepher N.H. Tur-Sinai [PSSSI 8; Jerusalem, 1960], pp. 1*-
11*). BDB, p. 199, gives ‘be struck dumb, astounded’ for ‫ דמם‬here and in Isa.
23.2, but there is no need for this departure from the regular meanings ‘be
silent, still’ and the sense proposed (which is more characteristic of the Eth.
56
Propp (p. 536) thinks ‘a double feminine suffix’ is involved; but such forms
occur with masc. nouns too (see examples in GK §90f).
15.1-21 339
cognate) scarcely fits the comparison with a stone. The latter makes ‘were
silent’ preferable to ‘were still’.57 On the past tense translation adopted here
see Note bbb above.
fff. Heb. ‫עד־יעבר‬. ‫ עד‬may mean either ‘until’ or ‘while’ (BDB, pp. 724-25).
In either case the perf. usually follows when past time is referred to, but
there are other examples of the imperfect: Josh. 10.13 (again in a purportedly
ancient poem); Ps. 73.17; 2 Chr. 29.34. Here and in Josh. 10.13 it may be used
in a preterite sense (like other imperfects in the present context), in the other
cases it has a modal sense. ‫ יעבר‬is translated ‘passed over’ (sc. the Jordan) by
Cross and Freedman (Studies, p. 53), but this does not really fit the context.
ggg. Heb. ‫קנית‬. Among the possible meanings of ‫קנה‬, ‘acquire, buy, create’
(BDB, pp. 888-89), ‘acquire’ is the most widespread and the most likely here:
‘which you acquired’ makes a good (expanded) parallel to ‘your (people)’ in
the previous clause. A case could be made for ‘buy’, which is a specialisation
of ‘acquire’, on the basis that ‫גאל‬, which is used in the equivalent clause in
v. 13, involves the payment of a price when it is used of a legal transaction;
but it is not certain (or even likely) that it is so used there. Cross and Freedman
(Studies, p. 64) favoured ‘created’ on the basis of Gen. 4.1; 14.19; 22; Deut.
32.6 (the last of Israel as here) and the use of its cognate in Ugaritic (on which
see now DULAT, p. 706). Some of these occurrences really mean ‘procreate’,
but AHI 4.201.3 provides an epigraphic instance of the sense ‘create’. Never-
theless the latter is most secure in cosmological contexts and the idea of
Yahweh’s taking possession of his people is firmly grounded in early Israelite
tradition: the creation language (with different verbs) is more characteristic of
Deutero-Isaiah (e.g. Isa. 43.1, 15).
hhh. Heb. ‫תבאמו ותטעמו‬. These verbs too are probably preterite imperfects
(cf. Note bbb above), as the use of the perfect later in the verse would most
naturally suggest. The pointing of ‫ ותטעמו‬with simple waw may, however,
have been designed to encourage a future or jussive reading of them, as in the
Vss (cf. Text and Versions and JM §177l; for a selection of the many cases in
poetry where the meaning must be future see B. Johnson, Hebräisches Perfekt
und Imperfekt mit vorangehendem we [CBOT 13; Lund, 1979], p. 57, irrespec-
tive of the flaws in Johnson’s interpretation; also Joosten, Verbal System, pp.
425-29, with the same proviso). Propp thinks the context here rules out a
preterite sense (p. 541). For the unusual form of the suffix see Note y above.
iii. Heb. ‫בהר נחלתך‬. ‫ הר‬is used both of an individual peak and of ‘hill
country’ more generally, the latter apparently with particular frequency of the
hills of Ephraim (Josh. 17.15 etc.). The question of the place that is meant here
(for which see the Explanatory Note) thus has to include consideration of the
possibility that a region is in view. ‫ נחלה‬is usually translated ‘inheritance’, as
57
Driver (p. 5*) assigned this occurrence to ‫ דמם‬II = (in Niph.) ‘be struck
down, motionless’. For him (pp. 9*-11*) it was only ‫ דמה‬II that means ‘be (made)
silent’.
340 EXODUS 1–18
in the Vss, but it has this specific sense in only about a quarter of its occur-
rences (c. 45 acc. to BDB, p. 635, out of a total of 222 in BH). Elsewhere
‘possession’ in a wider sense is involved (cf. THAT 2, 55-59 = TLOT 2, pp.
731-34; TWAT 5, 342-60 = TDOT 9, pp. 319-35 [which makes ‘inheritance’
the primary meaning]). In theological usage it most often refers to Israel as
Yahweh’s own people (e.g. Deut. 4.20; 1 Sam. 10.1; Ps. 78.71), but it also
designates their land as Yahweh’s own land (mainly in Jeremiah, e.g. 2.7;
but also Ps. 68.10). Its use of the Jerusalem temple in Ps. 79.1 (and perhaps
in Jer. 12.7: cf. the par. ‫ )ביתי‬can be traced back to the description of Mount
Zaphon as the ǵr nḥlt of Baal (KTU 1.3.3.30, 4.20), along with other features
of the Jerusalem cult tradition which have a Canaanite origin (cf. Clifford,
Cosmic Mountain, pp. 66-79, 131-60 [though he overlooks Ps. 79.1 itself];
W.H. Schmidt, Alttestamentliche Glaube in seiner Geschichte, pp. 215-28,
ET, pp. 207-20; THAT 2, 58 = TLOT 2, p. 733; TWAT 5, 358-59 = TDOT 9,
p. 333). Here too the reference must be to a mountain (or range of hills) which
‘belonged’ in a special sense to Yahweh.
jjj. Heb. ‫מכון‬. Although often, as here, translated ‘place’, ‫ מכון‬is not a
simple synonym of ‫מקום‬, but derives its meaning from the sense of the verb
‫כון‬, ‘to be firm, fixed, established’, which occurs in the parallel line here.
It can mean ‘foundation’ (Ps. 104.5; probably also, metaphorically, in Pss.
89.15; 97.2), but otherwise it is used in relation to God’s dwelling-place,
either in heaven (1 Kgs 8.39, 43, 49 par.; Isa. 18.4; Ps. 33.14) or on earth
(1 Kgs 8.13 par. [||‫ ;]בית‬Isa. 4.5; Dan. 8.11; Ezra 2.68). It presumably conveys
either the ‘sureness’ or solidity of the divine dwelling-place (as in Isa. 2.2;
Mic. 4.1) or the idea that its location has been fixed by divine decree (cf.
Ps. 93.5). It is probably no coincidence that the verb can be used of the perma-
nence of a royal throne, esp. in the Niphal (2 Sam. 7.16 par.; 1 Kgs 2.45;
Pss. 89.38; 93.2 [God’s throne]) but also in the Hiphil (Ps. 103.19 [God’s
throne]) and the Polel (2 Sam. 7.13 par.; Ps. 9.8 [God’s throne]), but that
does not imply that ‫ מכון‬means ‘throne’, as Propp seems to affirm (p. 542):
rather perhaps that ‫ לשׁבתך‬carries this connotation, as it does elsewhere (BDB,
p. 442). ‘Fixed place’ allows for both the possible nuances noted above.
kkk. Heb. ‫פעלת‬. Cross and Freedman observed (Studies, p. 65) that
‘This verb is quite common in early Yahwistic poetry, though rare in later
materials’, citing Num. 23.23; Deut. 32.4, 27; 33.11; Pss. 68.29; 77.13; Hab.
3.2. As a generalisation this is seriously misleading, for the verb occurs at
least as often in later parts of Isaiah (26.12; 41.4; 43.13; 44.12 [2x], 15) and
twelve times in Job. Among the purportedly early instances many would date
Deuteronomy 32 and Psalm 77 to the late pre-exilic or exilic period, and in
Hab. 3.2 it is the noun ‫ פּ ַֹעל‬and not the verb which appears. The 57 occur-
rences of the verb in BH seem to be spread over most periods of the history
of the language. What is worth observing is that the verb occurs in both Ug.
(mainly in the form bʿl, and with a derivative nomen agentis: cf. DULAT,
pp. 205-206) and Phoenician (and Punic), including early inscriptions, where
15.1-21 341
it is the regular word for ‘make, do’ (DNWSI, pp. 924-27). Although this is
certainly not proof of an early origin for the poem or its Canaanite character,
it could reinforce arguments of other kinds for a proximity to Canaanite
culture.
lll. For the dagesh with qoph see GK §20h.
mmm. Heb. ‫אדני‬. As a divine name or title ‫ אדני‬occurs several times
in Genesis 18–20 (but 18.3; 19.2, 18 should probably be excluded, as the
speakers at these points are not aware of the divine identity of their visitors)
and elsewhere in the Pentateuch in Exod. 4.10, 13; 5.22; 34.9; and Num.
14.17 (BDB, p. 11, lists occurrences elsewhere in BH and in longer phrases:
for its use in pre-exilic worship cf. Pss. 2.7; 68.12 etc.; 110.5). On the form
and meaning of the expression see Note b on the translation of 4.10-17. Here,
as in some other passages, there is doubt over whether it or ‫ יהוה‬is the original
reading (see Text and Versions). Either the original poet or a later scribe may
have used it here for variation from the successive occurrences of ‫ יהוה‬in
vv. 16-18.
nnn. Heb. ‫כוננו‬. On the meaning and use of the verb ‫ כון‬in Heb. see above
in Note jjj: it is the causative Polel stem that is used here. In Phoenician ‫ כון‬is
the regular verb for ‘to be’, but it is never used causatively.
ooo. Heb. ‫יהוה ימלך‬. In view of the following phrase ‫ ימלך‬must be taken
as future, as in Ps. 146.10. Elsewhere statements about Yahweh’s kingship
generally employ the perfect tense (Pss. 47.9; 93.1; 96.10; 97.1; 99.1; Isa.
52.7) or the noun ‫( מלך‬Pss. 10.16; 24.7-10; 29.10; 47.3; 95.3; 98.6; Isa. 6.5;
Zech. 14.9, 16-17). In view of expressions like those in Pss. 10.16; 29.10; 93.5
the use of the future tense is not inconsistent with such formulae, but it does
lay particular emphasis on the future hope that such confidence inspires. The
similar expressions used of Baal in KTU 1.2.4.10 and 1.108.1, 19-20, 21, 22
indicate that the language of this verse had a long history: cf. also DNWSI,
p. 639. The emphatic word-order here, with the divine subject preceding the
verb, corresponds to the classic formula in Pss. 93.1; 96.10; 97.1; 99.1 rather
than the (later?) P-S order in Isa. 52.7; Pss. 47.9; 146.10.
ppp. Heb. ‫לעולם ועד‬. For the formula cf. Mic. 4.5; Pss. 9.6; 45.18; 119.44;
145.1-2, 21; Dan. 12.3; without ‫ ל‬Pss. 10.16; 21.5; 45.7; 48.15; 52.10; 104.5;
inverted Pss. 111.8; 148.6. The noun ‫( ַעד‬in pause ‫עד‬:ֶ cf. GK §29r, JM §32c,
both without examples outside this phrase or any specific explanation [cf. BL
§ 69z]; Bergsträsser §29e also includes cases where seghol is preferred to
another vowel in pause, which may be significant), which also occurs alone
in the sense ‘perpetuity’(not ‘eternity’), is either derived from a root ‫( עדה‬Ar.
ʿadā), ‘advance’ (cf. Prov. 25.20; Job 28.8), or since J. Barth, Etymologische
Studien zum semitischen insbesondere zum hebräischen Lexicon (Leipzig,
1893), pp. 64-65, perhaps more plausibly related to Ar. ǵad, ‘tomorrow, later
future’ (cf. HAL, p. 742; Ges18, p. 921). Among the eighteen occurrences of
such language in Exodus (many of which are in Priestly legislation) the most
similar are 3.15 (the divine name) and 17.16 (state of war with Amalek).
342 EXODUS 1–18
qqq. Heb. ‫כי‬. The sense ‘when’ has also been proposed (Rashi: cf. NIV,
REB, NRSV), with the main clause beginning at ‫וישׁב‬: the verse would then
be a resumptive repetition of 14.22-29 after the ‘interruption’ by the Song of
Moses, setting the scene again for Miriam’s song in vv. 20-21 (cf. Childs, pp.
242, 248; also Watts, Psalm and Story, p. 44: the verse ‘re-establishes the
temporal…and physical…setting’). One might compare 6.28-30 where there
is a similar repetition after a genealogy has interrupted the Priestly narrative
and 12.51 after the addition of vv. 43-49. But in prose when ‫ כי‬means ‘when’
in the past, it is usually preceded by ‫( ויהי‬Gen. 6.1 etc.). On other medieval
Jewish interpretations see Propp, p. 546.
rrr. Heb. ‫ברכבו ובפרשׁיו‬. On the meanings of ‫ פרשׁ‬see Note x on the transla-
tion of 14.1-31. ‫ ב‬can mean ‘with’, of accompaniment (BDB, p. 89): whether
it is used here ‘of what one takes or brings with one’ (ibid.), with the oddly
chosen subject ‘Pharaoh’s horses’, is less certain.
sss. Heb. ‫את־מי הים‬. This combination of ‫ מים‬and ‫( ים‬and the short form of
the constr. st. of ‫ )מים‬is not found elsewhere in Exodus (with ‫ ים־סוף‬in Deut.
11.4; Josh. 2.10; in different contexts ‫ מי הים‬occurs in Amos 5.8; 9.6; Ps.
33.7), but it is a natural development from the association of the two words
in 14.26-28.
ttt. For the translation of these words see Note bbb on the translation of
14.1-31.
uuu. Heb. ‫את־התף‬. The def. art. is idiomatic (cf. Note o on the translation of
2.11-22) and carries no special significance. ‘Hand-drum’ is Russell’s transla-
tion (Song, pp. 12, 27; cf. ABD 4, p. 936, ‘frame-drum’; C.L. Meyers, ‘Of
Drums and Damsels’, pp. 18, 21 [= Feminist Companion (1994), pp. 213-14,
220-21]; Schmidt, p. 636); this is preferable to the usual ‘tambourine’ or
‘timbrel’, since the iconographical evidence (see the Explanatory Note) shows
no signs of jingling metal discs. It is also now clear from Old Aramaic that the
related verb meant ‘strike’ (DNWSI, p. 1226).
vvv. Heb. ‫ותצאן‬. For the defective spelling cf. 1.18-19 and GK §47l.
www. Heb. ‫בתפים ובמחלת‬. For the combination cf. Judg. 11.34; 1 Sam. 18.6;
Jer. 31.4; Pss. 149.3; 150.4 (the last three with the masc. form ‫)מחול‬. On ‫בתפים‬
see Notes rrr and uuu above; the ‫ ב‬with ‫ מחלת‬denotes rather an accompanying
action, which is close to BDB’s ‘concomitant (or surrounding) conditions’
(p. 89, III.1.c).
xxx. Heb. ‫ותען‬. This is one of the clearer cases for the meaning ‘sing’
(‫ ענה‬IV, for which there are cognates in Ar. [ǵanay, 2nd stem], Sy and, less
certainly, Ug. [cf. DULAT, p. 173; Emerton, ‘Some Notes’ (above n. 7 on
Note bb), pp. 42-43]), in view of the association with ‫( שׁיר‬similarly Num.
21.17; 1 Sam. 18.7 [21.12; 29.5]; Ps. 147.7 [par. ‫ ;)]זמר‬likewise probably Ps.
119.172 and Ezra 3.11 in contexts of praise. Whether the same is true of the
utterances in battle in Exod. 32.18 (first two exx.) and Jer. 51.14 is less clear
in view of the uses for a harvest ‘shout’ (‫ )הידד‬in Jer. 25.30 and perhaps in
Isa. 27.2 (Piel), Hos. 2.17 and the third occurrence in Exod. 32.18 (Piel) and
15.1-21 343
for an animal’s cry in Isa. 13.22 (with which cf. DAPT 1.10).58 More specific
renderings such as ‘led the singing’ (see Text and Versions, Luther, Tyndale,
EÜ; Schmidt, p. 636: cf. JB) or ‘sang…this refrain’ (NEB, REB) probably
read too much into the verb: neither fits the other occurrences. A derivation
of the form from ‫ ענה‬I = ‘answer’ (cf. Ug. ʿny: AV, RV) is unlikely and rarely
advocated, but see Houtman 1, p. 388, and Watts, Psalm and Story, p. 43; also
Text and Versions on TgO,N for this word.
yyy. Heb. ‫להם‬. For the masc. suffix referring to a fem. antecedent cf. 1.21;
2.17 and GK §135o. It is scarcely possible that the suffix refers back to the
Israelites in v. 19 (or v. 1), as Houtman (p. 295), Janzen and Russell seem to
suppose (see the Explanatory Note, where other difficulties in their interpreta-
tion of the section are discussed).
zzz. Heb. ‫שׁירו‬. Again the masc. form is preferred (see the previous note).
On the remainder of v. 21b see Notes d, f and g on the translation of 15.1-21.
Explanatory Notes
1a. The introduction to the poem is separated from 14.31 by
a division in MT and SP. No manuscript with continuous text at
this point survives from Qumran.59 The connection to 14.31, while
explicable in terms of the poem giving expression to the people’s
restored faith in Yahweh, with Moses once again taking the lead
(cf. 14.13-14) is not entirely smooth, as the people are now referred
to as ‘the Israelites’ (Heb. benê yiśrāʾēl, last in 14.29) rather than
‘Israel’ or ‘the people’ as in 14.30-31. We translate Heb. ʾāz ‘At that
time’ as in other places where it leads into the citation of a song or
poem (Num. 21.17; Josh. 10.12), rather than ‘then’ (e.g. NRSV),
which might imply a direct consequence of what has preceded (see
further Note a on the translation). In any case the connection is not
made in the most simple way by ‘and’ (Heb. waw).60 The word for
‘song’ (Heb. šîrāh rather than šîr) is one which seems to be reserved
58
ʿnyh in DAPT 1.13, ‘oracle-priestess’, is cited as a possible cognate by
Ges18, p. 990, but it is probably related to the homonym ‘answer’ (so J. Hoftijzer
and G. van der Kooij, Aramaic Texts from Deir ʿAlla [DMOA 21; Leiden, 1976],
p. 212, with ref. to Akk. āpil[t]u, ‘answerer’).
59
In 4QExd, where 15.1 follows 13.16 in what seems to be a collection of
extracts from Exodus, there appears, not surprisingly, to have been an open section
between the verses.
60
Blum keenly observes that Num. 21.17 almost exactly matches the wording
here and thus concludes the wilderness narrative in the same way as it began (in
Kd: Studien, p. 127).
344 EXODUS 1–18
for compositions which were attached to narratives (Propp, p. 508)

and perhaps implies that they would already have been well known
as independent texts.
1b. The opening words of the poem correspond closely to the
‘Song of Miriam’ in v. 21b, the only difference being between the
initial ‘I will sing’ here and the plural imperative ‘Sing’ there. On
the significance of this repetition see the introduction to the present
section. In their present position these words summarise the theme
of vv. 1-12, as v. 13 probably does for vv. 13-17. The metre in
this verse (based on the number of separable Hebrew words) is 4
+ 4, with each ‘4’ easily subdivided into two units of ‘2’.61 These
patterns, although not as widespread as 3 + 3, are relatively frequent
in Biblical Hebrew poetry, and predominate in the poems in this
chapter: W.G.E. Watson associates this with the need for a ‘victory
song’ to be ‘quick and lively’ (though as a whole 15.1-18 is better
described as a hymn: see the introduction to this section).62 Paral-
lelism, in the sense of correspondence (which need not be exact) in
form and/or meaning, is partially present here: the second half of
the line ‘seconds’ (to use J.L. Kugel’s term) the expression ‘he has
become greatly exalted’ by spelling out the action through which
this dominance was achieved.63
In its present context it is natural to understand the subject
of the first-person verb, like the first-person pronouns in v. 2,
as a reference to Moses. No doubt this is why the Versions here
render ‘Let us sing’ to conform more closely to the inclusion of
the people in the introduction to the poem (likewise the Pal. Tgg
also in v. 2a). Similar expressions occur elsewhere at or near the
beginnings of other poems where an individual singer has already
been mentioned (Deut. 32.1-2; 1 Sam. 2.1: cf. Judg. 5.3). But as
a stylistic feature it is widespread at the beginning of individual
thanksgiving psalms (e.g. Pss. 9.2-3; 30.2), of the so-called hymns
of the sole singer (e.g. Pss. 89.2; 108.2; 145.2) and elsewhere
61
The short Heb. word for ‘for’ (kî) is taken together with the following word.
62
Poetry, p. 111. See his whole chapter on metre (pp. 87-113), which makes it
clear that the subject remains one of much debate and uncertainty. The comments
made here are deliberately cautious and limited to observations which do not
depend upon a particular theory.
63
On the idea of ‘seconding’ see Kugel, The Idea of Biblical Poetry: Paral-
lelism and its History (New Haven and London, 1981), pp. 51-58. In using the
expression here I do not subscribe to all the uses that Kugel makes of it.
15.1-21 345
(Isa. 5.1; 12.1; 63.7).64 Gunkel already observed that such intro-
ductions were common in Babylonian hymns and in the Homeric
Hymns, and Ugaritic examples can now be cited from the hymns
to ‘the gracious gods’ and to Nikkal and the Kotharat.65
At least at its beginning the poem speaks of Yahweh in the third
person, like many biblical hymns, and so the translation ‘I will sing
of, or about, Yahweh’ (not ‘to Yahweh’) is preferable (cf. Note d
on the translation). Since Yahweh’s triumph is said to be over a
chariot-force rather than cavalry in both the poem (v. 4) and the
prose account(s) (14.6-7, 9, 17-18, 23, 25-26, 28; cf. 15.19), we
render ‘horse and its driver’ (rather than ‘…and its rider’): this also
corresponds to the military use of horses in Egypt and the Near East
generally in the earlier biblical period (see further Note g on the
translation).
2. The verse, which has close parallels in Isa. 12.2b and Ps.
118.14, can be analysed as two 2 + 2 lines if the short Hebrew words
for ‘Yah’, ‘for me’ and ‘This (is)’ are not counted separately and
the phrase ‘my father’s God’ is taken as a single unit. Alternatively
the lines may be in 3 + 3 metre, like some lines later in the poem.66
Both lines exhibit strong parallelism and even, unusually, internal
rhyme. On the translation ‘protection’ rather than ‘song’ see Note
i. The series of attributes of Yahweh in v. 2a are closely related in
meaning (the abstract ‘deliverance’ is equivalent to ‘deliverer’) and
all have the same temporal reference, which ‘has been’ identifies as
the specific event that is celebrated in the poem. The continued use
of first-person pronouns may suggest that Moses’ own vindication
against the criticism of the people is especially in view. Certainly
the expressions ‘my God’ and more especially ‘my father’s God’
are ones that are used with specific reference to Moses elsewhere
in Exodus (3.6; 4.10, 13; 18.2). The latter expression is particularly
characteristic of the stories of the patriarchs in Genesis (26.3 etc.)
and points to a society in which the family rather than the people
64
Cf. Gunkel, Einleitung, pp. 38, 267; F. Crüsemann, Studien zur Form
geschichte, pp. 210-306 (Chapters 4 and 5). The use of such expressions as an
introductory formula is to be distinguished from their appearance later in a psalm
as a concluding formula (Ps. 104.33) or in the ‘vow of praise’ in a psalm of lament
(e.g. Ps. 7.17).
65
Gunkel, Einleitung, p. 38; KTU 1.23-24 (cf. CML2, pp. 123-29).
66
So Cross and Freedman, Studies, p. 54.
346 EXODUS 1–18
as a whole is the horizon for religious life and practice (see the
Explanatory Note on 3.6). Thus this verse, like 3.13-15, represents
the identification of Yahweh, the God of the Exodus and now the
God of the people as a whole, with the god(s) of the individual
families.67 An alternative to the view suggested above would take
the ‘I’ here as collective, referring to the people as a whole. One
way or the other, the placing of these apparently personal words
of faith in the context of a national hymn of praise need not be
regarded as intrusive.68
3. The line can be analysed as 2 + 2 if the two Heb. words (ʾîš
milḥāmāh, lit. ‘man of war’) that lie behind ‘a warrior’ are treated as
a single unit. Both the repetition of ‘Yahweh’ and the word ‘name’
itself give special prominence to the identity of Israel’s God, as reg-
ularly happens at the beginning of a hymn (Gunkel, Einleitung, pp.
40-41; for the use of a nominal clause to express Yahweh’s attrib-
utes, as here, see pp. 48-49). The bold description of Yahweh as ‘a
man of war’ fits the context perfectly, but was too bold for most
of the ancient textual witnesses, which softened it (see Text and
Versions). Similar phrases are rare (cf. Ps. 24.8; Isa. 42.13), but
the idea expressed, that Yahweh ‘fights’ on his people’s behalf, is
widespread in Israel’s early traditions (cf. 14.14, 25; Num. 10.35-
36; 23.21-22; 24.8; Deut. 33.26-29; Judg. 5), in the narratives of
Joshua, Judges and Samuel, and in Deuteronomy (e.g. 7.17-24):
see further von Rad, Heilige Krieg; Smend, Jahwekrieg. Such ideas
were by no means limited to ancient Israel: the gods of other peo-
ples were believed to fight on their behalf, as well as against other
gods in mythical conflicts (B. Albrektson, History and the Gods
[CBOT 1; Gleerup, 1967]; Miller, Divine Warrior; KTU 1.119 =
COS 1, pp. 283-85).
The formula ‘Yahweh is his name’ recurs three times elsewhere
(Jer. 33.2; Amos 5.8; 9.6), but similar formulae incorporating
the expression ‘(the God) of hosts/armies’ occur in Amos 4.13;
67
‘My God’ is naturally a more widely attested expression: on it see
O. Eissfeldt, ‘ “Mein Gott” im Alten Testament’, ZAW 61 (1945–48), pp. 3-16,
and H. Vorländer, Mein Gott: Die Vorstellungen vom persönlichen Gott im Alten
Orient und im Alten Testament (AOAT 23; Neukirchen, 1975).
68
Contra, e.g., Cross and Freedman, Studies, p. 54; cf. the response of Muilen-
burg in ‘A Liturgy’, p. 239. In Canaanite Myth, p. 127 n. 49, Cross regarded only
v. 2a as an addition; in Pottery, pp. 179-227, Freedman had given up the idea of
secondary material altogether in favour of a liturgical function for the verse.
15.1-21 347
5.27, in (Deutero-)Isaiah (4x) and in Jeremiah (8x, all probably

in secondary contexts): in Hos. 12.5 the formula is ‘Yahweh the
God of hosts, Yahweh is his memorial (name: Heb. zēker)’. The
whole series of passages was examined by J.L. Crenshaw in his
study of the ‘doxologies’ of Amos (Hymnic Affirmation of Divine
Justice [SBLDS 24; Missoula, MT, 1975], pp. 75-114; cf. id.,
‘YHWH ṣebaʾôt šemô: A Form-Critical Analysis’, ZAW 81 [1969],
pp. 156-75). The occurrences in Deutero-Isaiah and Jeremiah, as
Crenshaw showed, appear in contexts with several common themes
and can be traced to a literary and perhaps cultic development of the
exilic period, and Crenshaw found enough similarity in the Amos
doxologies to justify a similar dating for them. Crenshaw actually
says that, although its literary attestations were (mostly at least) late,
‘The phrase “Yahweh of Hosts is his name” is an ancient hymnic
liturgical formula from the time of the ark at Shiloh’ (p. 111),
and the association of the similar formula in Hos. 12.5 with the
Bethel cult may give some support to this conclusion.69 Its use of
zēker in place of šēm (‘name’) links it to Exod. 3.15 (see further
below). It is also possible that the Amos doxologies are pre-exilic
in origin, whether or not they were present in the original collection
of the prophet’s sayings. Crüsemann argued that the ‘hymnic parti-
ciples’ which are their main structural feature represented a distinct
form of early Israelite hymn, most likely modelled on foreign
prototypes, since their content tends not to be distinctively Israelite
(Studien, pp. 83-154). The formula ‘Yahweh is his name’ was a
way of ‘reclaiming’ the theology of creation for Israelite worship,
already in pre-exilic times (ibid., p. 153). H.W. Wolff associated
the inclusion of the doxologies with a ‘Bethel-redaction’ of the
book of Amos in the time of Josiah and suggested that they origi-
nated as ‘syncretistic hymnic material from the Bethel sanctuary of
the seventh century’ (Joel und Amos [BKAT; Neukirchen, 1969],
pp. 135-36, 254-56 [136]). They appear to be much closer to real
liturgical texts than the passages in Isaiah and Jeremiah where the
formula occurs and are therefore likely to be older, perhaps much
older, than them.70 As a compositional element in Exodus 15 the
formula is probably not ‘reclaiming’ a foreign concept but serves,
69
Crenshaw himself, however, retracted it in ZAW 81 (1969), p. 168.
70
The words ‘in Yah, (it is) his name’ (byh šemô) in Ps. 68.5, which follow
a participial expression, may be a corruption of a further instance of the formula
348 EXODUS 1–18
perhaps with reference to v. 2b, as a way to give added prominence

to the God who is worshipped in this poem (compare the further
occurrences of the divine name alone at key points within it [vv. 6,
11] and at the end [vv. 16-17]).71
4. The verse introduces a note of explicit specificity into the poem,
regarding both the enemy (‘Pharaoh’) and the location of his defeat
(‘the Yam Suf’). Both these expressions and some others (‘chariots’,
‘army’, ‘the sea’, ‘choicest’ and ‘officers’) recur in the prose narra-
tive of chs. 13–14 (cf. 13.18; 14.4, 5, 7, 21, 25), all but one of them
(‘his army’, Heb. ḥêlô) in the verses attributed to the non-Priestly
account.72 One might deduce from this that the prose has influenced
the poem or the poem the prose, but all the expressions concerned
are common, so that a direct relationship between the passages need
not be presumed. The metre, like v. 1b, is 4 + 4 with each ‘4’ easily
subdivided into 2 + 2, if ‘Pharaoh’s chariots’ and ‘in the Yam Suf’
are each counted as a single unit. This is perhaps a little difficult in
the former case: in consequence it has been suggested that either the
text combined two alternative formulations at the beginning of the
line (‘Pharaoh’s chariots’ and ‘Pharaoh and his army’) or ‘and his
army’ is an addition from the prose account (or from Ps. 136.15).73
Only here and in 13.18 and 15.22 in the surrounding prose narra-
tive is the location of the ‘sea event’ identified as the Yam Suf in
Exodus (in 10.19 and 23.31 Yam Suf is mentioned without any
connection to this episode). In 13.18 and 15.22 journeys to and
from the Yam Suf are being described and the sequence of the prose
narrative certainly implies that the ‘sea event’ took place there. But
15.4 is the only verse in Exodus that explicitly states this, so it is
appropriate that our most detailed consideration of Yam Suf as the
location of the ‘sea event’ should appear here.
(cf. BHS, with LXX and Sy) or at least be related to it. Ps. 68 certainly contains
much old material, even if its present form is late (cf. Kraus, Psalmen, pp. 631-32).
71
Outside the realm of cultic and theological usage the same type of formula
(proper name + ‘(is/was) his name’ often serves to identify a person who is being
introduced into a narrative (see TWAT 8, 129-31 = TDOT 15, pp. 134-37, where
other uses are also noted).
72
The word for ‘chariots’ (markebôt) is the one used in 14.25, not the word
rekeb, which recurs several times in both accounts in ch. 14 (e.g. vv. 6-7 [non-P];
vv. 17-18 [P]).
73
So respectively Cross and Freedman, Studies, p. 56 (citing W.F. Albright);
and (e.g.) Gertz, Tradition, p. 192 n. 14.
15.1-21 349
At the end of the Excursus that follows the Explanatory Note on

14.1-4 we stated that the most likely location for the place-names in
14.2 (P) was near el-Qantara, near the northern end of Lake Ballah,
but that the non-Priestly portion of the prose narrative, including
the mention of the Yam Suf in 13.18, more probably envisaged the
latter as being to the east of Wadi Tumilat, that is at either Lake
Timsah or (one of) the Bitter Lakes. The main reason for this was
the mention of Sukkoth in the itinerary-notes in 12.37 and 13.20,
since it appears to correspond to Egyptian Ṯkw in the eastern part of
Wadi Tumilat. What needs closer examination now is whether the
term Yam Suf itself and its likely Egyptian equivalent fit best with
such a location or with a different one.74
If the evidence of the Old Testament is taken by itself, it is hard
to escape the conclusion that Yam Suf stands for the Gulf of Aqaba
(which is too far away to be part of the Exodus narrative) and the
Gulf of Suez (which the contexts in 10.19 and 13.18 would seem
to support).75 The latter alternative could be more easily combined
with the mention of Succoth in Exod. 13.20 if, at the time when
the non-Priestly narrative was written, the Gulf of Suez extended
further northwards, to the region of the Bitter Lakes, than it does
now.76 But since the work of H. Brugsch, L’Exode et les monuments
égyptiens (Leipzig, 1875) it has become increasingly common to
base conclusions about the location of Yam Suf on an equation
of ‘Suf’ with Egyptian ṯwf(y), ‘papyrus’, a fresh-water plant that
points to an inland stretch of water being meant. For a long time this
equation was recognised to be generic and was cited in support of a
74
This way of formulating the issue presumes that Yam Suf is everywhere
a real toponym and not, for example here in Exod. 15.4, a term derived from
myth, as N.H. Snaith suggested (‘‫ים סוף‬: the Sea of Reeds: the Red Sea’, VT 15
[1965], pp. 395-98): for others who have supported such a view see H. Lamberty-
Zielinski, Das “Schilfmeer” (BBB 78: Frankfurt, 1993), pp. 17-19.
75
At one time Noth took the view that the Gulf of Aqaba was meant here (as
did some others: see my Way of the Wilderness, p. 71 with n. 45), but he later
recognised that this was improbable because of the distance from Egypt (p. 85,
ET, p. 108). The idea has occasionally been revived (e.g. C.J. Humphreys, The
Miracles of Exodus [London, 2003], pp. 172-260; M.D. Oblath, The Exodus
Itinerary Sites [StBL 55; New York, 2004], pp. 98-106), but without good reason.
76
See more fully my Way of the Wilderness, pp. 70-74. The ‘northerly exten-
sion’ hypothesis has recently attracted renewed support: see Hoffmeier, Israel in
Egypt, pp. 207-10; Van Seters, ‘Geography’, pp. 272-73, who is right to point out
(n. 50) that the idea goes back to E. Naville.
350 EXODUS 1–18
number of different locations for Yam Suf, although A.E. Gardiner

and H. Cazelles deduced that pa-ṯwf, ‘the papyrus region’, was a
specific area in the eastern Nile Delta (Gardiner, Ancient Egyptian
Onomastica [Oxford, 1947] 2, pp. 201-202; Cazelles, Autour de
l’Exode, pp. 208-209), implying that Yam Suf was either the eastern
part of what is now Lake Menzaleh (Gardiner) or ‘a kind of bay
of the Mediterranean to the east of Tanis’ (Cazelles, with refer-
ence to Isa. 11.14-15: he does not mean Lake Sirbonis): there is not
much difference between the two.77 This latter approach was taken
up again by M. Bietak in 1975 in the context of his excavations
at Tell el-Dabʿa (Tell el-Dabʿa II. Der Fundort im Rahmen einer
archäologisch-geographischen Untersuchung über das ägyptische
Ostdelta [Vienna, 1975], pp. 135-37 [p. 137: ‘mit grosser Vorsicht’],
217-21 [with map]; see also the map in his Avaris: The Capital
of the Hyksos [London, 1996], fig. 1; and recently ‘Der Aufenthalt
“Israels” in Ägypten und der Zeitpunkt der “Landnahme” aus
heutiger archäologischer Sicht’, ÄuL 10 [2000], pp. 179-86, esp.
186). Bietak concluded that the Shi-Hor and Yam Suf (twf) must be,
respectively, the salt lakes north and south of the Qantara isthmus,
the latter being the former lake(s) of El-Ballaḥ. This view has been
adopted by P. Weimar, H. Donner (with qualifications) and, in a fresh
examination of the issues (2005), J.K. Hoffmeier.78 Hoffmeier now
acknowledges that the Egyptian name for the lakes in the southern
part of the isthmus of Suez was not pa-twf(y) but km wr. While
some of the Egyptian references to pa-twf(y) are indeed imprecise
geographically (a point underlined by J.R. Huddleston in ABD 5,
pp. 633-42), its occurrence in town-lists (Onomastica) indicates a
specific location and apparently one close to Tjaru/Sile. Hoffmeier
also points out (pp. 88-89) that the modern name Tell Abu Sefeh
(at the north end of Lake Ballaḥ) could preserve the consonants of
ancient pa-twf(y). If this interpretation of Yam Suf on the basis of
77
Cf. Albright’s view that Yam Suf was near el-Qantara: ‘Exploring in
Sinai with the University of California African Expedition’, BASOR 109 (1948),
pp. 5-20 (15-16). Van Seters (art. cit., p. 273) takes Isa. 11.15 to refer to ‘a narrow
gulf of the Red Sea’.
78
P. Weimar, Die Meerwundererzählung (ÄUAT 9; Wiesbaden, 1985), pp.
258-68; H. Donner, Geschichte des Volkes Israel und seiner Nachbarn in Grund
zügen 1 (Göttingen, 2nd ed., 1995), pp. 109-110; Hoffmeier, Sinai, pp. 75-89. In
Hoffmeier’s earlier studies he had favoured Lake Timsah.
15.1-21 351
the Egyptological evidence is followed, the Song of Moses placed

the sea-crossing at Lake Ballaḥ. Bietak and his followers hold that
it was somewhere near its northern extremity, but this depends on
their evaluation of the toponyms in Exod. 14.2, 9 and that, while
plausible, is not the only possibility (see the Explanatory Note on
14.1-4). Lake Ballaḥ extended some 25 km from north to south and
its southern tip was only about 10 km from the eastern end of Wadi
Tumilat, where Succoth (13.20) seems to have been located. If it
was in the southern part of Lake Ballaḥ that the ‘Yam Suf’ tradi-
tion located the sea-crossing, this would not be difficult to reconcile
with a ‘southern’ location for the toponyms in Exod. 14.2, 9 and the
association of Yam Suf with Succoth in Exod. 13.18-20. However,
13.18, 13.20 and 14.2, 9 probably belong to three different compo-
nents of the Exodus narrative and it remains possible that they
reflect two or even three different views about the location of
the event (see the introductions to 13.17-22 and 14.1-31). In that
case our argument above might only show how their combination
in the final text of Exodus could have seemed intelligible to the
redactor(s) involved. A residual problem for the equation of Yam
Suf with Lake Ballaḥ is that it is then not easy to see the rationale
for its well attested use to refer to the northern gulfs of the Red
Sea: the ‘northern extension’ hypothesis can scarcely be stretched
to include a lake so far to the north.
5. The two parts of the verse exhibit parallelism of sense, without
verbatim correspondence, and are closely connected to v. 4 by the
pronouns ‘them’ and ‘they’. Metrically the second half is longer
than the first (2 + 4 or 2 + 3, depending on whether ‘like’ [kemô]
is treated as a separate unit [in v. 8b it is probably not]): Cross and
Freedman (Studies, p. 58) thought that a word had been lost after
‘covered them’. Another possibility is that ‘like a stone’, which
is not essential, was secondarily added from v. 16 (though the
shorter form of ‘like’ is used there and the sense is different). But
no change is needed: longer cola sometimes appear (cf. v. 15b).
‘The deep waters’ (Heb. tehōmōt) are here those of the sea, in the
plural (like ‘the depths’) to indicate their expanse and power. There
need be no connection with the subterranean waters (Gen. 49.25)
or the primeval waters (Gen. 1.2), for which tehôm is also used.
The description of the Egyptians’ fate in the poem, as in the prose
narrative, is comparable to ancient portrayals of other similar catas-
trophes (for which see the Explanatory Note on 14.26-29).
352 EXODUS 1–18
6. A new, hymnic section begins here, introduced (as in v. 11)

by verbal repetition: ‘Your right hand, Yahweh… Your right hand,
Yahweh…’ Such repetition is more frequent in early Hebrew poetry
and has been described as ‘staircase parallelism’: it is a feature
which this poem shares with Ugaritic poetry (cf. Watson, Poetry,
pp. 150-56). The metre is clearly 4 + 4, with each half further
broken down into 2 + 2. From here to v. 17 the address is directly
to Yahweh, whereas hitherto he has been referred to as ‘he’. His
‘right hand’ (cf. v. 12) is no doubt mentioned because it holds a
weapon (compare the famous Baal stela from Ugarit in ANEP, pl.
490), which fits well with the description of Yahweh as a warrior in
v. 3 (Propp, p. 519, citing Lohfink). For ‘the enemy’ compare v. 9:
similar expressions are a commonplace of battle records, including
those of ancient Egypt (e.g. ANET, pp. 235-37, 256).
7. The first half of the verse may be the conclusion of v. 6.79 The
initial ‘and’ suggests a connection (though Cross and Freedman
regarded it as secondary because it is uncharacteristic of early
poetry),80 ‘you cast down your foes’ is a close parallel to ‘(it)
shattered the enemy’ in v. 6, and metrically v. 7a could be a third
‘4’ (2 + 2) to go with the two lines of v. 6: then v. 7b and v. 8a (to
‘gathered together’), which are also linked by ‘and’, would make
another 4 + 4 line, with the remainder of v. 8 being 3 + 3, most
probably. The problem is that this analysis produces an odd line
in vv. 7b-8a, because it changes topic in mid-line, where in fact a
recapitulation of earlier events seems to begin (see below on v. 8).
So the Masoretic verse-division is best retained, with v. 7 as another
4 + 4 line.
‘Your majesty’ (Heb. geʾônkā) picks up ‘has become greatly
exalted’ (gāʾōh gāʾāh) in v. 1 (see Note f on the translation). Such
language was frequently used of Egyptian kings on their campaigns
(cf. ANET, pp. 234-41, 245-48, 253-58). It is also found in Assyrian
records, with reference both to the king himself and to the national
god Ashur (e.g. ANET, pp. 286-88, 292, 294) and other gods (ANET,
pp. 300-301): on the expression melammu see A.L. Oppenheim,
79
See C. Cohen, ‘Studies in Early Israelite Poetry I: An Unrecognized Case
of Three-Line Staircase Parallelism in the Song of the Sea’, JANES 7 (1975),
pp. 13-17; cf. Propp (p. 505), who includes the whole of v. 7 in the first of his
three stanzas.
80
Studies, pp. 59-60.
15.1-21 353
‘The Golden Garments of the Gods’, JNES 8 (1949), pp. 172-93; E.

Cassin, La splendeur divine (Paris, 1968). Here there is no human
victor and Yahweh’s majesty alone is celebrated.81
In the context of the poem the prefixed verbal forms here and
in v. 6 are best understood as preterites (see Notes u and w on the
translation), but the fact that such forms can also indicate repeated
or future action would have made the generalisation of the state-
ments a possibility (cf. Text and Versions).
The imagery of the final stich must compare Yahweh’s ‘wrath’
(ḥrwn) to fire, as its etymology (cf. TWAT 3, 182-83 = TDOT 5,
p. 171-72) and associations elsewhere (Num. 11.1; Nah. 1.6; Zeph.
3.8; Lam. 4.11) would suggest. ‘Devour’ (ʾkl) is literally ‘eat’, but
is often used of consuming by fire (e.g. Lev. 6.3 [EVV. 6.10]; 10.2;
Num. 16.35), and ‘stubble’ is common in imagery associated with
fire (Isa. 5.24; 33.11; 47.14; Joel 2.5; Obad. 18; Nah. 1.10; Mal.
3.19). Here no actual fire is implied – as is clear elsewhere in the
poem Yahweh’s weapon is water – and the expression is simply a
vivid way of describing total destruction.
8. The three parts of this verse are all parallel, repeating the same
sense in different ways. In each part the subject is a word for the ‘sea’:
‘waters’, ‘flowing waters’, ‘deep waters’. The first part is certainly a
‘4’, possibly 2 + 2, and the second and third could be 4 + 4, but also
3 + 3, if ‘like’ and ‘in the midst’ are not counted as separate compo-
nents (as MT’s maqqephs imply). Cross and Freedman (Studies,
p. 51) favoured the latter alternative. The ‘and’ with which the
verse begins (though Cross and Freedman again thought it was not
original, as in v. 7) is important for its interpretation, for it seems
to associate what it describes with the destruction of the Egyptians
(but see below). ‘The breath of [Yahweh’s] nostrils’ is an exception-
ally striking anthropomorphic expression for the divine intervention
and all the main textual witnesses except MT and the Peshiṭta
avoid it by paraphrasing (see Text and Versions). But it must, like
the slightly less bold expression in v. 10, be intended to identify
Yahweh very closely with the effects of a strong wind on the sea
(more so than in 14.21). A very similar feature can be seen in Psalm
29, where ‘the voice of Yahweh’ is identified with thunder: an exact
correspondence to this appears in the Ugaritic myth of ‘The Palace
81
Even in the royal psalms such expressions are rare, but Ps. 21.6 provides
one example.
354 EXODUS 1–18
of Baal’ (KTU 1.4.7.27-32; CML2, p. 65). In ‘Baal and Mot’ (KTU

1.5.5.7; CML2, p. 72) ‘your [sc. Baal’s] breath/wind’ (rḥk) occurs
in a list of meteorological terms and forms a close parallel to the
present passage. The effect of the wind is to pile up the waters and
seemingly to solidify them (see Note dd on the translation for the
sense ‘congealed’: ‘churn’ or ‘foam’, which Cross and Freedman
preferred, is not the meaning in Biblical Hebrew). This has been
interpreted as a reference to a storm at sea in which the Egyptian
forces, using ships to cross the water and (presumably) cut off the
escaping Israelites, were drowned (so Cross and Freedman, Studies,
p. 47; ‘The Song of Miriam’, pp. 238-39; Cross, Canaanite Myth,
pp. 131-32). But there is no support for this view in the language of
the poem, which probably refers to something much more like the
prose account in Exodus 14.82
In fact v. 8 may not refer to the destruction of the Egyptian forces
at all. It is certain that in v. 9 the poet, having spoken of the climax
of the episode in the drowning of the Egyptians and Yahweh’s
involvement in it (vv. 4-7), steps back to an earlier stage in the
sequence of events when the Egyptians resolved to pursue the Isra-
elites and rob them of their possessions (cf. Propp, pp. 520-21).
This is then naturally followed in v. 10 (and v. 12) by a brief reca-
pitulation of the disaster itself. It is possible that v. 8 also forms
part of the prelude to the catastrophe. Indeed, since it says nothing
about the Egyptians and would otherwise place the cause of their
downfall in vv. 4-7 after it rather than before it, it is more likely
that this is the case.83 The only problem is the ‘and’ at the begin-
ning of the verse, which at first sight implies a close connection
with, and even a continuation of, v. 7. It might be a secondary addi-
tion, as Cross and Freedman thought, but the only ancient evidence
for its omission is the Peshiṭta. It may be better to understand it
as an instance of one of the rarer uses of the Hebrew conjunction
to express causality (‘For’) or emphasis (‘Indeed’), both of which
82
See the full refutation by L.E. Grabbe, ‘Comparative Philology and Exodus
15,8: Did the Egyptians Die in a Storm?’, SJOT 7 (1993), pp. 263-69, and already
in Kloos, Yhwh’s Combat, pp. 136-37.
83
Propp, p. 521, supports this view (which was also advocated by Smend,
Erzählung, p. 143, and others before and after him: so apparently Beer, p. 81),
but explicitly on the basis of the prose narrative. Cf. Kloos, Yhwh’s Combat,
pp. 136-38.
15.1-21 355
are suited to the resumption of an earlier part of a narrative (see

more fully Note aa on the translation). A similar transition (this
time without an ‘and’) can probably be seen in vv. 13-14.84
Even if the poem does presuppose the same general sequence of
events as the prose in ch. 14, the two accounts of the story are not
identical. There is nothing here about two walls of water, one on
each side of the escaping Israelites, as in 14.22, 29 (from P), only
something like a single ‘bore’ (or ‘heap’ or ‘wall’). This is actually
much closer than the familiar account to the phenomenon of ‘wind
setdown’ which has sometimes been suggested as the scientific
explanation for what happened at the Re(e)d Sea.85
84
For such ‘delay’ in recounting earlier parts of a narrative see J. Licht, Story-
telling in the Bible (Jerusalem, 1978), pp. 109-13. Earlier M. Weiss, ‘Weiteres über
die Bauformen’, had called it ‘flashback’, as in 2 Kgs 9.14-15.
85
Wind set down involves the effect of a powerful offshore wind blowing for
a period of hours, which causes the water to recede from the normal shore-line
to an unusual degree. A striking example occurring early in 1882 was reported by
a British surveyor working near the Suez Canal, who compared it to the Exodus
account (A.B. Tulloch, ‘Passage of the Red Sea by the Israelites’, JTVI 28 [1896],
pp. 267-80). More recently the possible relevance of the phenomenon to locations
in the Gulf of Suez and the Gulf of Aqaba has been explored respectively by
D. Nof and N. Paldor (‘Are there Oceanographic Explanations for the Israelites’
Crossing of the Red Sea?’, Bulletin of the American Meteorological Society 73
[1992], pp. 305-14; ‘Statistics of Wind over the Red Sea with Application to
the Exodus Question’, Journal of Applied Meteorology 33 [1994], pp. 1017-25)
and C.J. Humphreys (Miracles of Exodus, pp. 248-57), but these are among the
least satisfactory possibilities from a geographical point of view. A much more
plausible scenario, because unlike the others it involves a wind from the east (cf.
Exod. 14.21) blowing in very much the area which is currently most favoured for
the location of the biblical toponyms (see above on 14.1-4), is presupposed in the
computer simulation of C. Drews and W. Han (‘Dynamics of Wind Setdown at
Suez and the Eastern Nile Delta’, Public Library of Science One 5[8]; see also
Drews, ‘Could Wind Have Parted the Red Sea?’, Weatherwise [Jan.–Feb. 2011],
pp. 30-35). Alternative natural causes such as a tsunami (e.g. B.J. Sivertsen,
The Parting of the Sea: How Volcanoes, Earthquakes and Plagues Shaped the
Story of the Exodus [Princeton, 2009]; for evaluation see M. Harris, ‘The Thera
Theories: Science and the Modern Reception History of the Exodus’, in T.E. Levy
et al. [eds.], Israel’s Exodus in Transdisciplinary Perspective: Text, Archaeology,
Culture and Geoscience [Cham, 2015], pp. 91-99) or storm surge fit the portrayals
in the biblical texts less well. The same applies to the various attempts to connect
them with phenomena in and around Lake Sirbonis (see the Explanatory Note on
14.26-29). It is of course not to be expected that a cultic hymn and a traditional
(and composite) narrative would provide data sufficiently precise to determine the
356 EXODUS 1–18
9. Here there is no doubt that the verse is to be analysed as 4 +

4 + 4, with each ‘4’ comprising two ‘2’s: in the first stich the speech-
introduction is separated from the two asyndetic verbal forms
which begin the speech, while in the other two initial verbal forms
open each hemistich, with a first-person form being followed by
a third-person form both times. In these stichs the accompanying
noun denotes first the object and then the subject (with a first-person
suffix which preserves the focus on the speaker): the parallelism
between them is therefore very close. The ‘enemy’ who speaks (or
‘thinks’, as Houtman suggests [pp. 283-84], for Heb. ʾāmar can
have this sense even without the addition of ‘in his heart’ [cf. 2.14])
could be Pharaoh (see Text and Versions), but ‘them’ and ‘they’
in v. 10 probably require an antecedent and this favours a collec-
tive reference to the Egyptian force as a whole, as for the same
word in v. 6 (compare the pl. form in v. 7). It is, significantly, unlike
the Canaanite myths about Baal, not the sea, as Albertz observes
(p. 250: cf. Childs, p. 251). The repeated ‘them’ in this verse is
the first, tangential, reference to the Israelites in the poem; until
v. 13 all the emphasis falls on the destruction of the Egyptian army.
The citation of their confident boast can be compared to Pss. 35.21;
83.5 (cf. Childs, ibid.). Its content expresses a typically military
viewpoint, as in Gen. 49.27 and Prov. 1.12-13, where it is mixed
with the imagery of a predatory animal (Propp, pp. 524-25): perhaps
this is also implied in ‘my appetite shall be sated’. Houtman (p. 244)
cites this verse as evidence that an Israelite sea-crossing is assumed
by the poet, but on its own it need not imply this. Following v.
8, however, and if the latter is understood (see the note above) as
describing a wind-driven withdrawal of the sea prior to the Egyptian
pursuit in v. 9, the latter would certainly point to an Egyptian pursuit
across the dried-up sea-bed and, most obviously, to a previous flight
of the Israelites in the same direction.
10. The human enemy changes from being the subject to the
object (‘them’) of the leading verb, and from proud boast to a
humiliating disappearance from the scene, as the divine subject who
dominates the poem reasserts his power. The metre is again 4 + 4,
with sub-division of each stich into 2 + 2 that has a clausal basis
scientific character of what (if anything) they portray, and it is understandable that
many historians regard such speculation as of little value (e.g. de Vaux, Histoire,
pp. 358-64 [359], ET 1, pp. 381-88 [382]; Hoffmeier, Israel in Sinai, pp. 108-109).
15.1-21 357
in the first but not the second stich, the latter being an expanded
parallel to the second hemistich of the first.86 The Egyptians’ catas-
trophe is, like the piling up of the waters in v. 8, attributed to the
divine breath/wind. Here their descent to the depths is intensified,
compared with v. 5, by the comparison to the heavier material of
lead. The ‘mighty waters’ (Heb. mayim ʾaddîrîm: for the sense
‘mighty’ see Note t) recall the mythological language of Ps. 93.4,
where ʾaddîr describes ‘the waves of the sea’ (or, if a popular
emendation is adopted [cf. BHS], Yahweh’s superiority to them),
and the more frequent mayim rabbîm, ‘many/mighty waters’ (ibid.;
cf. Pss. 18.17; 29.3 etc.).87 But even if this lies in the background, it
is not the central focus here: the (physical) sea is Yahweh’s instru-
ment of judgement, not his enemy. At most one might envisage that
Yahweh’s use of the sea presupposes an earlier, primeval, defeat of
the god Sea as implied in Psalm 93.
11. The metre of this reflective interlude (cf. v. 3), which does not
advance the description of Yahweh’s specific act of deliverance, is
4 + 4 + 4, with clear subdivisions into 2 + 2 in the second and third
stichs: the first by contrast forms a single unit, though the repetition
of the first two Heb. words, ‘Who is like you…?’, encourages the
idea that here too they could be viewed as a sub-unit. The repeti-
tion is similar to vv. 6 and 16, but unlike them does not constitute
an example of ‘staircase parallelism’ in the strict sense, because the
first stich is complete in itself. The poet apparently felt free to vary
his use of this traditional stylistic feature. The incomparability of a
god was a regular theme of the repertoire of praise in the Old Testa-
ment and more widely in the ancient Near East: already Gunkel
listed parallels in 1 Sam. 2.2; Jer. 10.6-7; Pss. 18.32; 35.10; 71.19;
77.14; 86.8; 89.7, 9; 113.6; 135.5 as well as in Babylonian litera-
ture (Einleitung, pp. 72-73). Some Egyptian examples are added in
the much fuller study of C.J. Labuschagne, The Incomparability of
Yahweh in the OT (Pretoria Oriental Series 5; Leiden, 1966). The
rhetorical question implies a negative: there is no other god like
Yahweh. In this verse, as in some of the other biblical examples
(1 Sam. 2.2; Pss. 77.14; 86.8; 89.7, 9; 113.6; 135.5) and also in
86
There is no need to think of ‘a number of variant narratives’ being ‘combined
together’ here or in v. 5 (Noth, p. 99, ET, p. 124).
87
On which see recently A.R. Gray, Psalm 18 in Words and Pictures (BibInt
127; Leiden, 2014), pp. 106-108.
358 EXODUS 1–18
Babylonia and Egypt, such a statement could be made where the

existence of other gods was still recognised; elsewhere, presumably
in later texts (so especially in Deutero-Isaiah), it became an expres-
sion of monotheistic belief.88
The focus is first on Yahweh’s power: ‘mighty’ is Heb. neʾdār
as in v. 6 (and ʾaddîrîm in v. 10 is related to it). The exact sense
of ‘in holiness’ (Heb. baqqōdeš) is not certain. The combination
with ‘mighty’ makes a reference to divine holiness in the sense
usually understood unlikely, although the word is used in that way
in v. 13 (here tr. ‘holy’: see Note vv). Some other possibilities are
noted by Propp (pp. 527-28) and Labuschagne (pp. 97-98), the
latter including an unexpected association with victory in war (as
in Ps. 77.14). ‘Holiness’ is certainly sometimes used concretely to
mean a ‘holy place’ or ‘shrine’ (see Note mm on the translation),
and there are Ugaritic parallels for this (KTU 1.3.3.30 and par.: for
several more see DULAT, p. 697). But this is hardly appropriate
here, where Yahweh’s power is not shown in a shrine or sanctuary.
The Septuagint renders ‘among holy ones’, either interpreting the
sing. qōdeš so or on the basis of a Heb. text that read the pl. of the
related adjective qedôšîm (as MT has in Ps. 89.6, 8). With ‘holy
ones’ understood as divine or semi-divine beings this reading would
make a good parallel to ‘gods’ earlier in the verse and it has been
adopted by some scholars (see Text and Versions). But it is too
easy a way out of the difficulty. Since, as just noted, the adjective
‘holy’ was sometimes used as the equivalent to ‘god, divine being’
(compare also 1 Sam. 2.2; Zech. 14.5; Job 5.1; 15.15 and the phrase
‘the Holy One of Israel’ in Isa. 1.4 etc.), the corresponding abstract
noun could perhaps have had the sense ‘divinity, divine power’,
which would fit the context well here (so in effect BDB, p. 871,
which finds the same sense in Pss. 68.18; 77.14 and in the expres-
sion ‘holy arm’ in Isa. 52.10; Ps. 98.1).89
88
As Blum observes (Studien, p. 159), the pre-eminence of Yahweh appears
again in direct association with the Exodus in Jethro’s words in 18.10-12: see
further the notes there.
89
Alternatively one might, with Russell, Song, pp. 16, 155 n. 42, suppose that
qōdeš is itself collective in meaning here, as has been suggested for some other
occurrences (Deut. 33.2; Ps. 77.14).
15.1-21 359
The list of Yahweh’s attributes which make him incomparable

continues with his ‘fearsomeness’, which is due to his ‘praise-
worthy acts’ (see Note nn on the translation). Such fearsomeness
is a theme of other hymnic passages (Pss. 89.8; 96.4). Here it is
especially other nations who fear Yahweh, as vv. 14-16 will illus-
trate, but 14.31 has related this as the reaction of his own people
too. The final descriptive phrase, ‘doing wonders’, also evokes a
human response to Yahweh’s acts of power: the identical phrase
occurs in Ps. 77.15 (cf. v. 12; 78.12; 86.10), again with reference
to the deliverance at the sea. Here at least it is Yahweh’s actions in
history, and specifically the Exodus story, which make him unique
among divine beings (see further Labuschagne, ibid., pp. 91-108).
12. Taken alone this verse would comprise a single ‘4’, divided
by its two verbal clauses into 2 + 2. It is therefore tempting to see it
as either the conclusion of the larger unit in v. 11 (cf. BHS), restating
the event which exemplifies the incomparability of Yahweh, or the
opening of v. 13 (cf. Freedman), with which it shares verbal forms
with a past reference. Yet in the former case the shift from question
to narrative is abrupt, while in the latter the join would cross the
boundary between the two parts of the poem’s narrative, which treat
respectively the event at the sea (vv. 4-12) and the onward journey
to the mountain of Yahweh (vv. 13-17). In fact isolated ‘4’ lines (or
2 + 2) do occur elsewhere in the poem in vv. 3 and 18, each of which
concludes a distinct component of it, and v. 12 plausibly serves a
similar function.90
In the prose narrative it is Moses who stretches out his hand or
his staff to bring about divine intervention (cf. 14.16, 21, 26-27), but
in the non-Priestly story of the plagues there are traces of Yahweh
doing the same (3.20; 6.1; 7.17; 9.15) and his ‘outstretched arm’
became a formulaic motif of Deuteronomic and Priestly language
about the Exodus (cf. 6.6 and the Note there). It is a familiar expres-
sion for divine judgement (Isa. 5.25; 23.11), which may be the
sense here, but in Isa. 49.22 Yahweh raises his hand as a signal and
this would also fit the context: the earth is commanded to take the
drowned Egyptians to their final destiny.
90
The comparison between vv. 12 and 18 in this regard was made by Smith,
Pilgrimage Pattern, pp. 211-12.
360 EXODUS 1–18
The expression ‘the earth swallowed them up’ is at first sight

anomalous in a context where death comes by drowning. The lan-
guage is paralleled only in the story of Dathan and Abiram (Num.
16.25-34), to which Houtman sees a reference here (p. 286).91 But
this is impossible, as that episode concerned Israelites, not Egyp-
tians. Some commentators have thought that the sea-bed specifically
is meant, but most likely a more general reference to the earth as
‘the realm of the dead’ is intended (Noth, p. 99, ET, p. 125), as in
the passages cited in Note qq on the translation. This is not quite to
adopt the view that Heb. ʾereṣ could mean the underworld (Sheol),
as its Ugaritic cognate sometimes did (see the Note). But it is close
to it, and the expression ‘swallowed’ is certainly reminiscent of pas-
sages that describe Sheol as devouring men (Isa. 5.14; Prov. 1.12;
Job 18.13), just as Mot the god of Death does in Ugaritic mythology
(KTU 1.5.1.4-8; 1.6.2.17-23 = CML2, pp. 68, 76-77).
13. Metrically this verse is a straightforward 4 + 4 (since both
‘for whom’ [Heb. zû] and ‘to’ [Heb. ʾel] undoubtedly belong with
the adjacent words) but, while further subdivision of each ‘4’ into
2 + 2 (cf. BHS) is possible, it is not supported by the clause struc-
ture. The two halves of the verse display an overlapping parallelism
(ab: ac). It begins a new section of the poem (vv. 13-17[18]), in
which the topic is no longer Yahweh’s intervention at the sea and
the fate of the Egyptians (though these are probably alluded to in
vv. 14 and 16), but the Israelites themselves (‘the people’ here and in
v. 16) and their onward journey to the mountain of God. The scope
of this verse and its relationship to vv. 14-17 are much debated and
depend on whether Yahweh’s ‘holy dwelling’ is the same as the
dwelling-place and ‘sanctuary’ in v. 17. If it is, then v. 13 speaks
of the whole journey from Egypt to Canaan as a summary of what
is to come; if it is not, then v. 13 refers only to a first stage on the
journey. We will defer discussion of this issue until the end of the
treatment of this verse. It is in any case full of theological terminol-
ogy, which is concerned no longer with Yahweh’s power and control
of the elements of nature but with aspects of his relationship to his
people. Two different verbs are used to describe his ‘leading’ of
them on their journey: the first (Heb. nāḥāh) has already appeared
in 13.17, 21 (see Notes a and o on the translation of 13.17-22),
91
Berner on this basis sees the verse as a secondary and late addition to the
poem (pp. 396-97).
15.1-21 361
while the second (Heb. nāhal Piel: see Note uu on the translation
of this passage) has often been thought to have the specific sense of
‘give rest’ or ‘bring to a watering-place’, like its Arabic cognates.
This is probably to read too much into the meaning of the verb (so
also Propp, p. 532). But both verbs have a strong association with
shepherds leading their flocks (cf. Isa. 40.11; 49.10; Pss. 23.2-3;
77.21; 78.52-53, 72), caring for them and providing their needs. As
attributes of Yahweh in this connection the poet mentions first not
his ‘strength’ (for which cf. v. 2 and, with a different Heb. word,
v. 6) but his ‘loyalty’ (Heb. ḥesed: see Note rr on the translation)
or ‘steadfast love’, as later in 20.6 and 34.6-7. This was a precious
quality in human communities and summed up Yahweh’s ongo-
ing care for his people. The expression ‘for whom you intervened’
(Heb. gāʾal, ‘often translated ‘redeem’: see Note tt on the transla-
tion) comes from the realm of family law (as best illustrated in Ruth
3 and 4), where it denotes the intervention of a ‘kinsman’ to help
someone in the family who has fallen on hard times, usually by the
(re-)purchase of property or a debt-slave. In the reference to the
Exodus deliverance here (and in 6.6) no purchase is involved, and
the verb presupposes, rather than expressing, the creation of a rela-
tionship between Yahweh and Israel, just like Yahweh’s reference to
Israel as ‘my people’ already in 3.7, 10 before the Exodus has taken
place (see also the Note below on ‘had taken possession’ in v. 16).92
The final phrase of the verse, ‘to your holy dwelling’, is the most
controversial (see Note vv on the translation). ‘Dwelling’ (Heb.
newēh, from nāweh) probably originally had the sense ‘pasture land’,
like its Akkadian cognate, and this meaning is occasionally found in
BH (e.g. 2 Sam. 7.8). But it generally refers to a dwelling-place of
human beings or, by a metaphorical transfer that is found with other
words such as ‘house’, a place where a god has his earthly ‘home’
(2 Sam. 15.25; Isa. 33.20). According to Cross ‘It is a specific desig-
nation of a tent-shrine’ (Canaanite Myth, p. 125), but he gives no
evidence for this claim (one might think of Isa. 33.20 where ‘tent’
occurs alongside nāweh, but both are used metaphorically for the
92
Schmidt (Exodus, Sinai und Mose, p. 66 n. 94) and others have seen the use
of the expression here as a sign of dependence of the poem on P (6.6). But the
theological use of gāʾal is now well attested in pre-exilic inscriptions, esp. in the
PN gʾlyhw, Gealiah, as well as in some probably older biblical passages, as Albertz
(p. 251) recognises.
362 EXODUS 1–18
‘city’ of Jerusalem) and it contradicts the widespread use of nāweh

for town dwellings. The location of this ‘holy dwelling’ of Yahweh
has been variously identified (see the reviews in Houtman, p. 288,
and Propp, pp. 532-33, 568-69). Among the suggestions that have
been made are Sinai/Horeb (Cassuto, Hyatt), Kadesh-Barnea,
Mount Casios (Zaphon) by Lake Sirbonis (Norin, Er Spaltete das
Meer, p. 90), Shittim (Cross, Canaanite Myth, p. 141), the whole
land of Canaan (Noth, p. 99, ET, p. 125; Albertz, p. 251), Shiloh
(Smith, Pilgrimage Pattern, pp. 223-26, cf. p. 220, acknowledging
the prior contribution of J. Goldin, The Song at the Sea, pp. 51-55)
and Jerusalem (an ancient view [see Text and Versions]: recently
Childs, p. 252; Houtman, p. 288; and Blenkinsopp, Pentateuch,
pp. 159-60). Shittim can scarcely be meant, as it is not known to
have been a Yahwistic holy place (Num. 25.1-5 suggests something
very different),93 and while nāweh might indeed refer to the whole
land of Canaan (cf. Jer. 10.25; Ps. 79.7), its description as a ‘holy
dwelling’ is hard to parallel. The strongest contenders are Sinai,
Shiloh and Jerusalem. Against an identification with Sinai it might
be said that (unless v. 17 refers to it too, which we shall argue below
to be unlikely) this would be an uncharacteristically slight allusion
to an episode of such importance (though its content would match
3.12 in the prose narrative, and indeed ch. 18). More telling, perhaps,
is an argument from the structure of the poem: just as vv. 1-12 begin
in v. 1b (or 1b-3) with a summary of what is to come, so v. 13 should
summarise vv. 13-17(18) and therefore refer to the final destination
of the journey in Canaan. A choice between Shiloh and Jerusalem
as the specific location will depend on how v. 17 is understood and
may be deferred until after the consideration of that verse.
14-16. These verses portray, with a rich vocabulary of terror, the
reactions of neighbouring ‘peoples’ (the plural form of the same
Heb. word ʿām as is used for Israel in vv. 13 and 17) to Yahweh’s
powerful onslaught on the Egyptians (cf. ‘the greatness of your arm’
at the end of v. 16).94 The metre is mixed, with lines (and even one
93
It also depends on Cross’s unlikely view that Gilgal is referred to in v. 17
(see the Note there).
94
The theme is taken up in Deut. 2.25 and Josh. 2.9, 14. There is a close
similarity to the reaction of Baal’s (apparently human) foes to his ‘voice’ in KTU
1.4.7.29-41, which may have been the model for this part of the poem (cf. Cassuto,
pp. 179, 181, who reconstructed an Israelite myth [‘epic poem’] that might have
15.1-21 363
half-line) of three stresses alternating with the dominant 4 + 4 (+ 4)

rhythm of the poem. Verse 14 is probably 3 + 3, with ‘inhabitants
of’ going with ‘Philistia’, while 15 is 4 + 4 followed by a half-line
which on the same principle is best analysed as a ‘3’, though it might
be an undivided ‘4’ (see Propp, p. 535, on the metrical issues raised
by this half-line). Verse 16 begins with a 4 + 4, with weak inner
clause divisions in each half, followed by another repetitive line
exhibiting ‘staircase parallelism’ (though in a weaker sense than v.
6). The latter could be a 4 + 4 (with ‘of whom’ taken with the next
or preceding word, like ‘for whom’ in v. 13: in both cases the Heb.
is the short particle zû), but this depends on taking ‘until’ (Heb.
ʿad) as a separate unit, against the maqqephs of MT, and it may
be another 3 + 3. If it is, vv. 14-16 would have a chiastic metrical
pattern, perhaps an appropriate way of marking them off from the
surrounding poetic narrative (see below) in which Yahweh (in the
second person) is the subject.
The translation of the verbs in the main clauses of these verses (as
also of v. 17) is a problem of their interpretation which also affects
wider issues such as the date and original setting of the poem (see
on a particular stage in the history of interpretation my essay ‘Some
Points of Interest’, pp. 252-55, and on the linguistic background to the
problem Note bbb on the translation). The main verbs in this section
alternate between forms which normally have a past meaning (the
‘perfect’ or suffix-conjugation) and those which most often have a
future meaning (the ‘imperfect’ or prefix-conjugation): in vv. 16-17
the main verbs are all of the latter form. But the situation is compli-
cated by the fact that the ‘perfect’ can also be used of the future (the
‘prophetic perfect’), while the imperfect can have a preterite sense
(as was known in antiquity and understood by Hebrew scholars in
the late Middle Ages: the comparative evidence for this has been
significantly increased by the discovery of the Ugaritic texts). The
ancient translators of these verses rendered vv. 14-15 in the past,
no doubt because perfects predominate in them (and occur in both
halves of the preceding v. 13), but then translated v. 16 as a prayer
or wish (another possible use of the imperfect, or more precisely
served as an intermediate stage in the transmission of the motif). Further ancient

Near Eastern parallels are cited by N. Waldman, ‘A Comparative Note on Exodus
15: 14-16’, JQR 66 (1976), pp. 189-92: cf. Russell, The Song of the Sea, pp. 78-79.
364 EXODUS 1–18
the jussive). Early versions of the sixteenth century followed this

(rather inconsistent) pattern, but from the Geneva Bible (1560)
onwards a future translation was adopted for all three verses (as
can be seen in the AV). Apart from linguistic considerations this
may have seemed more appropriate to the place of the poem in the
biblical narrative. In the Revised Version (1885) the future interpre-
tation was abandoned and replaced by a mixture of past tenses (for
the perfect forms) and presents (for the imperfects): this odd combi-
nation probably reflects the influence of S.R. Driver, whose study
Hebrew Tenses had appeared in 1875 and who was a member of the
translation panel, and it remained dominant in the new versions of
the 1950s and 1960s (RSV, NJPS, JB). But S.R. Driver’s account
of the Hebrew tenses was already being superseded by that of his
son G.R. Driver, who could make use of the comparative data from
Akkadian and Ugaritic, and this provided the basis for a much
more widespread recognition of the preterite use of the imperfect
(see his Problems of the Hebrew Verbal System, pp. 132-44).95 In
the NEB (for which the younger Driver was the Chairman of the
Old Testament translation panel) past tenses were used throughout
vv. 14-16 and the German Einheits-Übersetzung of 1980, the NRSV
and the REB followed suit. Only the two conservative evangelical
translations of recent years have demurred: the NIV goes right back
to the future tenses of the Geneva Bible, while the ESV reverts to
the ‘mixed bag’ of the RV and the mid twentieth-century versions.
Modern study of the Hebrew verbal system continues to recog-
nise a variety of uses for both the perfect and the imperfect ‘tenses’,
so that in theory any of the translation patterns proposed for Exod.
15.14-16 could be correct. But the rediscovery of the preterite use
of the imperfect has shifted the balance of the argument consid-
erably: no longer need there be any hesitation, from a linguistic
point of view at least, about translating imperfect forms in this
way if that is what the context implies – there may still be (and
arguably is in the recent evangelical translations) a hesitation based
on a reluctance to contemplate a rendering in the past tense (even
more for v. 17, on which see below) which threatens the early date
95
As Driver readily acknowledged, he was only building on earlier studies by
other scholars, especially H. Bauer and G. Bergsträsser (cf. pp. v-vi, 9-20).
15.1-21 365
or the Mosaic authorship of (this part of) the poem.96 The context
in the poem sets the time reference firmly in the past, both by the
predominance of the perfect tense in v. 14 and by its exclusive use
in v. 13, which is either a summary of vv. 14-17 or (less likely) an
action which precedes it. Although v. 16 uses only the imperfect, it
cannot be separated in its time-reference from vv. 14-15, because it
is linked to them by the pronouns ‘them’ and ‘they’, as well as by
the continuing theme of fear at the mighty power of Yahweh.
The verses are therefore most likely to describe reactions to
Yahweh’s action which are anterior to the poet’s situation (or ‘implied
situation’, the position in which he places himself to compose the
poem). The scope of those reactions is first general (v. 14a), then
specific to particular peoples, although the reactions are the same in
each case (vv. 14b-15), and then general again (v. 16). The Philis-
tines are mentioned first (v. 14b), perhaps because their territory
lay closest to Egypt. But the mention of them raises a historical
problem, since their arrival in southern Canaan occurred after the
most likely time for the events, whatever they were, that lay behind
the Exodus story (see the Note on 13.17-18). Other biblical passages
place the Philistines in Canaan before the arrival (or emergence) of
the Israelites (Gen. 26.1-33 [cf. 20.1-18]; Josh. 13.1-3; Judg. 3.1-4),
but traditions about conflict between them and the Israelites begin
to appear only in the period of the Judges and the early monarchy
(Judg. 3.31; 10.6-7, 11; 13-16; 1 Sam. 4ff.). By contrast Moab,
Edom and Canaan all occur in Egyptian texts of the thirteenth cent.
B.C. (for references see e.g. POTT, pp. 32, 231). It is notable that no
kings of Moab or Edom are mentioned here, in contrast to other texts
(Gen. 36; Num. 20.14; 21.26; 22.4, 10; 23.7) but in agreement with
Deut. 2.1-19. The words ‘chiefs’ and ‘rulers’ in v. 15 (see Notes yy
and zz on the translation) point to a polity of tribal or local leaders
which probably preceded the establishment of a unified monarchy
in Edom and Moab (cf. Bartlett, Edom, pp. 81-82 [which needs to
be supplemented in the light of the discoveries at Kh. en-Naḥas,
on which see T.E. Levy et al., ‘Lowland Edom and the High and
Low Chronologies: Edomite State Formation, the Bible and Recent
96
It is noteworthy, to say no more, that in vv. 1-12, where there is the same
mix of perfect and imperfect forms as in vv. 14-16, the NIV and the ESV have no
such hesitation in translating imperfects by past tense English verbs.
366 EXODUS 1–18
Archaeological Research in Southern Jordan’, in T.E. Levy and T.

Higham (eds.), The Bible and Radiocarbon Dating: Archaeology,
Text and Science (London, 2005), pp. 129-63]; J.M. Miller, ‘Moab’,
ABD 4, pp. 882-83 [889-90]; O. LaBianca and R.W. Younker,
‘The Kingdoms of Ammon, Moab and Edom: The Archaeology of
Society in Late Bronze/Iron Age Transjordan (ca. 1400–500 BCE)’,
in T.E. Levy [ed.], The Archaeology of Society in the Holy Land
[New York, 1995], pp. 399-415). In the case of Philistia and Canaan
it is ‘the inhabitants’ who are afraid. The switch between leaders
and led is probably just for variation. Cross and Freedman (‘The
Song of Miriam’, pp. 242, 249 n. 49; cf. Canaanite Myth, p. 130)
followed Albright in rendering Heb. yōšebê by ‘the enthroned’ the
second time (in Pottery, pp. 142-43, Freedman extended this to the
first occurrence) to improve the parallelism, but this is unnecessary
and not supported by Amos 1.5, 8.
The sense and reference of ‘pass on’ (Heb. yaʿabōr) in v. 16 have
been variously understood (see Propp’s review, pp. 537-38). The
Heb. verb ʿābar can mean ‘to cross’, ‘to pass by’, ‘to go through’ or
simply ‘to pass on’ (cf. BDB, p. 716). Propp takes its primary refer-
ence to be to the crossing of the sea by the Israelites (p. 537), but
this scarcely fits the reference to the fear of the nations of Palestine,
since that only began after the crossing of the sea and so the latter
cannot be introduced by ‘until’ as if it was still to happen (the verbal
parallel in Josh. 4.23, cited by Propp, is in a different context). Cross
thinks that the crossing of the Jordan is meant (Canaanite Myth,
p. 141; cf. Studies, p. 53), but again it is difficult to see why this
should limit the nations’ fear (especially the Canaanites!). The verb
is used several times elsewhere of the Israelites ‘passing through’
(or desiring to do so) the territory of other peoples on their journey
(e.g. Num. 20.17-21; Deut. 2.4, 8) and this provides the most
natural explanation. What is most striking then is that no account
is taken (even in Canaan) of the battles that the books of Numbers
and Joshua say that the Israelites had to fight along their way: their
journey is remembered as the unimpeded progress of Yahweh’s
people to the ‘sanctuary’ (v. 17) that was their destination.97
17. The metre of this verse reverts to the dominant 4 + 4 (+ 4) of
the poem. Three parallel expressions, one in each of the three stichs,
97
For ‘had taken possession of’ as the sense of Heb. qānāh see Note ggg on
the translation.
15.1-21 367
designate the destination of the journey as a place that belongs to

Yahweh and has been ‘made’ and ‘established’ by him. It is possible
to take ‘a fixed place’ and ‘a sanctuary’ as in apposition to ‘the hill(s)
which you own’ and presume the ellipse of a relative particle (as is
common in Hebrew poetry) before the verbs that follow: ‘a fixed
place for you to dwell, which you made; a sanctuary, O Lord, which
your hands established’ (cf. NRSV). The sense is little different,
but the latter rendering makes the identity of reference of the three
expressions more explicit.
To ‘plant’ a people (the pronoun ‘them’ clearly refers back to
Yahweh’s people Israel in v. 16) is a metaphor for giving them a
settled place to live (so with explicit comparison to a vine in Jer.
2.21; Ps. 80.9, 15-16; elaborated in the ‘Song of the Vineyard’
in Isa. 5.1-7). This is preceded by Yahweh’s ‘bringing them’ (or
‘bringing them in’, as Heb. bôʾ often implies) to their future home.
This place is first designated as the ‘hill’ (or ‘hills’: see Note iii
on the translation) which Yahweh owns. ‘Hills’ might well seem
the more appropriate translation, as the highlands of central Canaan
were the heartland and original limit of the territory in which the
Israelites were settled (so Noth, p. 100, ET, pp. 125-26; and for this
phrase Albertz, p. 252). But the expression ‘which you own’ points
in a different direction. It is literally ‘(the hill[s]) of your posses-
sion’ (Heb. naḥalātkā; not ‘of your inheritance’: see Note iii again),
and the parallelism in Ps. 79.1 shows this to mean the Jerusalem
temple or rather the hill on which it stood. This hill was sometimes
known as ‘the hill/mountain of Yahweh’ (Isa. 2.3; 30.29; Mic. 4.2;
Zech. 8.3; Ps. 24.3; perhaps Gen. 22.14: in Num. 10.33 it is the title
for Mount Sinai), more often (21x) as his ‘holy hill’ (the reference
to Jerusalem is especially clear in Isa. 66.20; Joel 2.1; 4.17; Ps. 2.6;
Dan. 9.16). The same terminology is used in the Ugaritic texts for
Mount Zaphon, the sacred mountain on which Baal had his palace
(ǵr nḥlt: see the references in Note iii on the translation).
The phrase ‘a fixed place for you to dwell’ also has strongly
temple-oriented language. ‘Fixed place’ (Heb. mākôn), when it
does not refer literally or figuratively to a ‘foundation’ (Pss. 89.15;
97.2; 104.5), has a specific reference to Yahweh’s dwelling-place,
either in heaven or on earth (see Note jjj for occurrences), while ‘to
dwell’, or possibly ‘to sit, be enthroned’ (Heb. yāšab), is a widely
used expression for the divine presence in the temple (e.g. 1 Kgs
8.13; cf. Ps. 132.7, 13). Likewise a ‘sanctuary’ (Heb. miqdāš),
368 EXODUS 1–18
except when the word is used metaphorically (Ezek. 11.16: perhaps

also Isa. 8.14, but the text may be corrupt), is always a building
dedicated to sacred use.98
All this emphatically suggests that at (or near) its climax the
poem envisages a particular place of worship as the destination
of the journey of Yahweh’s people through the wilderness, rather
than the hill-country of Canaan as a whole. Much of the language
used has a specific connection with the Jerusalem temple (though
not the word ‘sanctuary’) and it has commonly been seen to derive
from the cultic traditions of that temple (cf. Clements, God and
Temple, pp. 52-55; J. Jeremias, ‘Lade und Zion: Zur Entstehung
der Ziontradition’, in H.W. Wolff, Probleme biblischer Theologie
[FS G. von Rad; Munich, 1971], pp. 183-98 [196-97]; Blenkin-
sopp, Pentateuch, pp. 158-60; Houtman, p. 292; Albertz, p. 252,
for the second and third clauses of the verse). Some scholars have
maintained that other holy places could be meant (see the review in
Propp, pp. 564-68). The verbs at the beginning of v. 17 must refer
to the Israelite settlement in Canaan (especially after the ‘passing
by’ of other peoples in v. 16), which excludes a holy place in the
desert, including Mount Sinai: the reference to a (built) ‘sanctuary’
also conflicts with what we know of the worship there (unless the
tent or tabernacle were meant, but this seems unlikely). F.M. Cross
argued that the sanctuary was that at Gilgal, by the Jordan, which
was certainly an important place of worship from early times (e.g.
Josh. 4.5; 1 Sam. 11.13, 15; Amos 4.4; 5.5) and would have been
an appropriate place to celebrate the settlement in Canaan and the
history that led up to it (Canaanite Myth, pp. 138-44, followed
e.g. by Miller, Divine Warrior, pp. 116-17). But, despite Cross’s
attempt to evade it by pleading that it ‘could apply to any Yahwistic
sanctuary’ symbolically (pp. 142-43), the fact that Gilgal was low
in the Jordan valley and in no sense on a ‘hill’ must rule it out.
M.S. Smith has made the more plausible suggestion that the use of
similar terminology to refer to Shiloh, which was certainly in the
highlands and was itself situated on a low hill, in Psalm 78 means
that it could be referred to here (Pilgrimage Pattern, pp. 225-26,
developing ideas from J. Goldin, The Song at the Sea, pp. 51-55).99
98
‘Lord’ here is not the divine name but Heb. ʾadōnāy: see Note mmm on the
translation.
99
According to Brenner, Song of the Sea, p. 16, Ewald already suggested this.
15.1-21 369
Shiloh was an important sanctuary in pre-monarchic times, at least

for the tribes of central Palestine, according to 1 Samuel 1–4 and
it was where the ancient sacred ark symbolising the presence of
Yahweh was kept for a time (1 Sam. 3.3; 4.3-9, 17-18, 21-22). It
could be argued that much of the Jerusalem cultic tradition came
from Shiloh with the ark, and Ps. 78.60 certainly speaks of Shiloh
as a divine dwelling-place (cf. Jer. 7.11). It is less certain that v. 54,
which shares theological ideas with Exod. 15.17, is meant to refer
to Shiloh: the psalm is clearly from Jerusalem (cf. vv. 67-72) and
it may well have Jerusalem already in view in v. 54 (so e.g. Kraus,
Psalmen, p. 710). It is in fact to Jerusalem that v. 69 applies the
extraordinary idea (see further below) that Yahweh built his own
temple. The preservation of the song is easier to understand if it
came from Jerusalem and there is no need to envisage any earlier
Heimat for it. The early view in post-biblical tradition (see Text and
Versions) that the Jerusalem temple was meant is probably correct.
This brings us to the question of the tenses of the verbs in this
verse. The issues are essentially similar to those already discussed
in relation to vv. 14-16. But it is notable that there has been a much
greater reluctance to adopt a past tense translation for the main
verbs here, which are again in the imperfect tense and therefore
in themselves capable of bearing a future meaning. The ancient
Versions all take up this possibility and in modern times even trans-
lations which avoid a future in vv. 14-16 have one here (RV, RSV,
JB, NJPS, REB, ESV). This is no doubt due to the present placing
of the poem, where to use the past tense (as NEB, the EÜ and NRSV
do) creates a sharp dissonance. Even Cross and Freedman used the
future in 1955 (‘The Song of Miriam’, p. 242), although Studies
(p. 53) has the past and so does Cross in 1973 (Canaanite Myth,
p. 131), and Childs from his canonical perspective not surpris-
ingly did the same (p. 241), although he was content to follow the
preterite interpretation in vv. 15-16 (and earlier).
But the dissonance created (or rather preserved) is not a strong
argument against a past tense rendering if it is likely on other
grounds. The poem is dissonant in other ways from the prose narra-
tive that precedes it, such as its concentration on the fate of the
Egyptians in vv. 1-12 and the inclusion of a continuation of the
story in vv. 13ff., and so represents a somewhat different perspec-
tive on the ‘sea-event’ anyway, which the compiler saw no need to
modify. Three reasons internal to the poem, at least, favour a past
370 EXODUS 1–18
tense translation of the verbs which open v. 17: (i) this is what a
careful reading of vv. 14-16 leads the reader to expect, and there
is nothing at the beginning of v. 17 to suggest a sudden change to
the future; (ii) the perfect tense verbs in the second half of v. 17 are
most naturally rendered in the past and this suggests, even if it does
not prove, that the actions which provide their context are prior to
them and therefore also in the past; (iii) if v. 13 is an outline of what
follows, which seems most likely, the fact that it too has perfect
tense verbs implies that v. 17, which corresponds to its second half,
should also be in the past.
A final comment is needed on the unusual attribution of the
building of the ‘sanctuary’ to Yahweh himself. As has already been
noted, this is not entirely without parallel in the Old Testament, as
Ps. 78.69 says the same and even elaborates it with a comparison
to the creation of heaven (‘the heights’: cf. Ps. 148.1) and earth.
Generally, both in the OT and elsewhere in the ancient Near East,
the role of human builders of temples for the gods is freely recog-
nised (no doubt because most of the texts derive from circles
close to those [royal] builders!). But sometimes in mythical texts
a leading god has a temple built for him by other gods, as when at
the end of Enuma Elish the Anunna-gods build Esagila for Marduk
and then shrines for themselves (VI.45-68: COS 1, p. 401) and
when at Ugarit Baal has his temple built for him by Kothar-and-
Khasis (KTU 1.4.5-6 = CML2, pp. 61-63): Propp, p. 544; Hurowitz,
I Have Built You an Exalted House: Temple Building in the Bible
in the Light of Mesopotamian and Northwest Semitic Writings
(JSOTSup 115; Sheffield, 1992), pp. 332-34, gives a fuller list.
Baal could even claim that he had built the temple himself (KTU
1.4.6.35-38 = 8.32-37), presumably in the same sense that kings in
their inscriptions say that they have built structures which were in
fact constructed by their subjects. This, or something like it, will
be what is meant in these two laudatory biblical passages: Yahweh
himself is given the credit for what has actually been done by his
human subjects, perhaps because he was seen as the instigator of the
project and its guide and enabler (cf. Ps. 127.1). Goldin (The Song at
the Sea, pp. 47-55) saw what is said here as anti-Solomonic polemic
from descendants of the Shilonite priesthood, but the fact that it
could be taken up in a strongly pro-Judaean and pro-monarchist
psalm like Psalm 78 shows that it was not understood in that way:
15.1-21 371
as Houtman aptly comments, ‘in poetry YHWH is often subject,

where prose has human protagonists’ (p. 292).
18. The concluding verse of the poem is a single four-beat line,
like verse 12 at the end of its first main section. The imperfect verb
here is shown to have a future sense by the adverbial phrase that
follows it. Declarations of Yahweh’s kingly rule occur more often
at the beginning of a hymn and in the (timeless) perfect tense (Pss.
93.1; 97.1; 99.1): here it is placed where it is to show the connec-
tion with the foundation of his temple, a connection which is closely
paralleled in Babylonian and Ugaritic mythology (see further the
introduction to this section). The parallel in tense and word-order
to bʿlm yml[k], ‘Baal shall be king’ (the mem is ‘enclitic’), in
KTU 1.2.4.32 at the end of the Ugaritic myth of Baal’s battle with
Yam is particularly striking. H. Spieckermann deduces from this
that Yahweh’s victory celebrated here is also over a ‘mytisch [sic]
gefährliche Feindesmacht’ (Heilsgegenwart, p. 110), but this is to
underestimate the contrast between this poem, where the enemy is
portrayed in firmly earthly terms, and the Ugaritic myth. It is in the
portrayal of Yahweh and his ‘weapons’ that the correspondences to
mythological language occur (vv. 6, 8, 10, 12).
19. After the long poem in vv. 1-18, this verse in prose (it is
not part of the poem) recapitulates the climax of the narrative in
ch. 14, in close dependence on the wording of the Priestly sections
in vv. 22-23, 26, 28-29. There are small variations from that
wording (in addition to much that is omitted in this summary)
which are probably due to the impact of the Song of Moses on
the writer: specific mention is made (at the very beginning!) of
Pharaoh’s horse(s) (v. 1: cf. v. 21),100 and Yahweh is no longer only
indirectly involved (through his instruction to Moses in 14.26) in
the ‘return’ of the waters, but actively brings it about (cf. 15.10, 12).
The effect of the verse is to recall the events which called forth the
hymns of praise in vv. 1-18 and 21, after the closing verses of the
Song of Moses have ‘digressed’ on to the foundation of a temple
in Canaan. In the first place (cf. ‘For’) it does this for vv. 1-18,
but the redactor who inserted v. 19 was also able thereby to relate
100
In 14.9 and 23 the word ‘horse(s)’ (Heb. sûs) is probably an addition to the
basic Priestly text, again to harmonise it with the poetic texts: see the Explanatory
Notes there.
372 EXODUS 1–18
Miriam’s intervention explicitly to the earlier narrative (Schmidt,

p. 655). Similar Wiederaufnahmen appear after secondary inser-
tions in 6.28-30 and 12.51. Whether the verse is meant to explain
the singers’ reason for singing the Song or the validity of what it
says is not entirely clear, but it makes little difference.
Excursus
J.G. Janzen has argued that the syntax of vv. 19-21 is intended to make
Miriam’s song prior in (narrative) time to the Song of Moses, so that it is
the impulse that gave rise to it, despite its position in the present text of
Exodus (pp. 108-11; more fully in ‘Song of Moses, Song of Miriam: Who is
Seconding Whom?’, CBQ 54 [1992], pp. 211-20). He accepts the translation
‘For’ (not ‘When’: cf. Note qqq on the translation) for Heb. kî at the begin-
ning of the verse, so that it introduces an explanation for the Song of Moses.
But he observes that the Hebrew allows the explanation to continue into
vv. 20-21 (there is a waw at the beginning of v. 20, often translated ‘Then’,
but it is really ‘And’). He then gives two reasons why this way of reading the
verses should be preferred to the usual one: (i) the Hebrew for ‘to them’ in v. 21
is the masculine form (lākem, not lāken) and so should refer not to the women
but to ‘(Moses and) the Israelites’ in v. 19 (cf. v. 1); (ii) the recapitulation in
v. 19 ‘places the actions and words of 15:20-21 back behind 15:1-18, back to
the point reached in 14:29’ (p. 109). An obvious objection to this interpretation,
which Janzen anticipates, is that a much more straightforward and effective
way to make Miriam’s song precede that of Moses (and the Israelites) would
have been to place vv. 20-21 immediately after 14.29, where there would be
no need for any recapitulation. Janzen’s reply is that the narrator’s aim was
to accentuate the involvement of women at the beginning and end of the
Exodus story (cf. 1.15-21; 2.1-10; 4.24-26) in an example of the ‘envelope’
or ‘chiastic’ pattern (the former expression is more appropriate here) which
is a common trope in biblical literature (p. 110: cf. Exod. 14.13-14). His
argument has been accepted by some (Russell, Song of the Sea, pp. 32-39 [esp.
36-39: but he prefers to view v. 19 as a temporal clause]; Albertz, pp. 235-
36, 253-54 [in a modified way, which takes account of the process of literary
composition]), but rejected as ‘unnatural’ by others (Propp, p. 547; cf. Berner,
Exoduserzählung, p. 391 n. 174; more tentatively, Dozeman, p. 342). There
is, it must be acknowledged, a certain logic in having the command ‘Sing’
(v. 21) precede the declaration ‘I will sing’ (v. 1) in the two closely parallel
verses of poetry. But Janzen’s argument from the masculine form of ‘them’
is weak, since this was often used to refer to women (see Note yyy on the
translation) and here, just a few words before in v. 20, ‘the women’ is the
obvious antecedent for it. The words of v. 19, while certainly recalling the
earlier events, can scarcely be said on their own to indicate that of the two
15.1-21 373
responses Miriam’s was the prior one. In fact the long recapitulation in
v. 19 makes it unlikely that what follows in v. 20 is still viewed as part of the
explanation of the Song of Moses. It was not without good reason that both
the MT and 4Q365 placed a clear break between vv. 19 and 20 (cf. LXX: and
see the introduction to this section). Finally, Janzen’s response to the objection
that there was a much more obvious way to achieve the sequence he advocates
overlooks the fact that the ‘envelope pattern’ is just as forceful if vv. 19-21 are
understood in the usual way.
20. Miriam is named here for the first time, so it is fitting that
epithets (both unique) follow which identify what and who she was.
Female prophets or prophetesses appear only rarely in the Bible
(only Judg. 4.4 [Deborah]; 2 Kgs 22.14 par. [Huldah]; Neh. 6.14
[Noadiah]: in Isa. 8.3 Heb. nebîʾāh may mean ‘wife of a prophet’,
just as malkāh means ‘wife of a king, queen’). They were more
common in Mesopotamia (muḫḫūtu[m]/maḫḫūtu[m]; āpiltu[m];
raggim/ntu: texts in M. Nissinen, Prophets and Prophecy in the
Ancient Near East [SBL Writings from the Ancient World 12;
Leiden, 2003]; for discussion see now J. Stökl, ‘Female Prophets
in the Ancient Near East’, in J. Day [ed.], Prophets and Prophecy
[LHBOTS 531; London, 2010], pp. 47-61; id., Prophecy in the
Ancient Near East: A Philological and Sociological Comparison
[CHANE 56; Leiden, 2012], pp. 67-69, 121-27, 186-92, 216-17;
id. and C.L. Carvalho [eds.], Prophets Male and Female: Gender
and Prophecy in the Hebrew Bible, the Eastern Mediterranean and
the Ancient Near East [AIL 15; Atlanta, 2013]). When the evidence
from ancient Israel is so slight, conclusions about their activities and
their place in society can only be tentative, but the case of Deborah
(whether wholly historical or not) illustrates the varied functions
that a prophetess might have had. When Miriam appears again (with
Aaron) in Numbers 12 it is implied that she (like him) was an inter-
mediary for divine oracles, but here (like Deborah in Judg. 5) she
is a singer of a patriotic song and apparently a leading figure in the
community (cf. Mic. 6.4). Her description as ‘the sister of Aaron’,
and not also of Moses, suggests a stage of tradition prior to the late
Priestly genealogy in Num. 26.59, one in which Aaron had not yet
been made into the ancestor of all legitimate priests. This could be
associated with the portrayal of Aaron in Exodus 32 as a rival to
Moses and the fabricator of the ‘golden calf’, as well as the role
he plays in ch. 4 as Moses’ ‘fellow Levite’ (lit. ‘brother Levite’).
374 EXODUS 1–18
Apparently in the older, pre-Priestly tradition, Moses, Aaron and

Miriam were all leaders in early Israel, sometimes united with each
other, sometimes at loggerheads. For attempts to reconstruct the
history and nature of this tradition see Noth, ÜGP, pp. 195-200;
H. Seebass, Mose und Aaron, Sinai und Gottesberg (Bonn, 1962);
Norin, Er Spaltete das Meer, pp. 189, 192, 206; Valentin, Aaron,
esp. pp. 375-84, 412-18; and the introduction below to 17.8-13.
Miriam herself is listed as the sister of Moses and Aaron in
Num. 26.59 (a late Priestly passage) and 1 Chr. 5.29 (the Miriam in
4.17 is a Judahite, and there is no reason to regard her as the same
person). Elsewhere she appears only in Numbers 12, Deut. 24.9
(a reference to the same episode), Mic. 6.4 (linked to Moses and
Aaron) and (most interestingly) in a notice of her death and burial
in Num. 20.1b, which is generally attributed to the older source-
material.101 No agreed explanation for the name ‘Miriam’ has been
found: the closest Heb. roots produce uncomplimentary attributes.
W. von Soden tentatively proposed the meaning ‘gift, present (of a
god)’ on the basis of Akk. râmu III/riāmu and its derivatives (AHw,
p. 952), which might also explain the origin of the Heb. sacrificial
term ‫‘( תרומה‬Mirjām-Maria “(Gottes-)Geschenk” ’, UF 2 [1970],
pp. 269-72): the root may be Amorite in origin. The suggestion that
the name has an Egyptian origin (perhaps mry/t-ymn, ‘Beloved of
Amun’: cf. M. Görg, ‘Mirjam – ein weiterer Versuch’, BZ NF 23
[1979], pp. 287-89) is also worth considering, given the appearance
of other Egyptian names in the tribe of Levi. Propp notes Ugaritic
mrym as a noun for ‘height’, so perhaps ‘eminence’ (p. 546: for
other suggestions see HAL, p. 601; Ges18, p. 741).102
‘Hand-drum’ (or ‘frame-drum’) is the appropriate translation for
Heb. tōp (see Note uuu on the translation: so LXX, Vulg, Luther
[Pauke], EÜ). EVV since Tyndale have employed ‘timbrel’ or its
101
E.g. Noth, Numeri, p. 128, ET, p. 145. In ÜGP, p. 200, Noth had opposed
this view and attributed Num. 20.1b to P.
102
For a careful and thorough study of all the biblical traditions about Miriam
see R.J. Burns, Has the Lord Indeed Spoken only through Moses? (SBLDS 84;
Atlanta, 1987). Understandably Miriam has become the focus of much feminist
scholarship: see e.g. P. Trible, ‘Bringing Miriam out of the Shadows’, BR 5/1
(1989), pp. 170-90; the two volumes of A Feminist Companion to Exodus to
Deuteronomy (ed. A. Brenner; Sheffield, 1994 [pp. 166-254] and 2000 [pp. 104-
73]); and U. Rapp, Mirjam: eine feministisch-rhetorische Lektüre der Mirjamtexte
in der hebräischen Bibel (BZAW 317; Berlin, 2002).
15.1-21 375
more familiar synonym ‘tambourine’, but its jingling discs do not

appear in the iconographical evidence from Old Testament times,
including the ‘Canaanite band’ on a pottery stand of the tenth cent.
B.C. from Ashdod (illustrated in T. Dothan, The Philistines and
their Material Culture [New Haven, London and Jerusalem, 1982],
pp. 249-50; other illustrations from Egypt, Assyria and Canaan
in Keel, Die Welt, pp. 314-18, ET, pp. 337-40 [figs. 450-54, with
commentary]).103 The hand-drum consisted of a shallow disc of
clay or wood with skin stretched over it, which was beaten with
one hand while the other held the rim. Most of the evidence is of
women players, as here (cf. Judg. 11.34; 1 Sam. 18.6-7; Ps. 68.25),
celebrating a victory.104 But hand-drums were also used in regular
worship (1 Sam. 10.5 [male prophets]; 2 Sam. 6.5; Pss. 81.3; 149.3;
150.4), often with dancing, and in general merrymaking (Isa.
5.12; 24.8; 30.32; Jer. 31.4; Job 21.12). The involvement of ‘all
the women’ in the celebration here corresponds to their apparently
regular role in welcoming victorious warriors after a battle, as in
Judg. 11.34 (the focus on one woman is dictated by the plot there)
and 1 Sam. 18.6-7; Ps. 68.24-26 points to a cultic occasion in which
the praise is given to God, which is closer to what we have here (see
the next note and Burns, Has the Lord Indeed Spoken, pp. 11-40)
21. The translation ‘Miriam sang’ is virtually certain and the only
real alternative, ‘Miriam answered’, is rarely advocated today (see
Note xxx on the translation and Text and Versions). Two different
Hebrew roots ʿNH are involved: originally their first consonants
were distinct. Meanings like ‘led in song’ (LXX, Vulg) and ‘sang
(as) a refrain’ (NEB, REB: cf. Mowinckel, Psalmenstudien II, p. 111
n. 1) or ‘sang antiphonally’ (McNeile, p. 93) have been read into the
sentence, but there is no linguistic justification for such senses in
Hebrew: the second and third suggestions mistakenly combine the
meanings of the two different roots. At best support for them has to
103
Baentsch, unusually for a German, comments ‘unter dem Klingen der mit
Schellen besetzten Handpauken’ (p. 137: my italics), indicating something like a
tambourine – he cites E.C.A. Riehm, Handwörterbuch des biblischen Altertums2
(Bielefeld, 1893–94), p. 1053, which says ‘oft mit dünnen runden Metallscheiben
behängt’. Cf. Beer, p. 84.
104
For a full discussion of women’s musical performance in the Bible and
neighbouring cultures, which centres on the use of the hand-drum, see Meyers,
‘Of Drums and Damsels’.
376 EXODUS 1–18
be found in the wider context (and the wording of Miriam’s song as

a command, a call to worship, does make her into a cantor in effect),
but the simple translation is to be preferred, as it is what fits the other
occurrences of this root. ‘To them’ is sometimes taken to mean ‘the
Israelites’ in v. 19, but the closest referent is ‘the women’ in v. 20
and they must be intended as the addressees (see the Excursus on
v. 19 and on the linguistic issues Notes yyy and zzz on the transla-
tion).
Apart from the first Hebrew word (‘Sing’, a plural imperative,
instead of ‘I will sing’), Miriam’s song is identical to the beginning
of the Song of Moses in v. 1b (see the Notes there on the words
that are repeated). The opening imperative makes it correspond
to the most common type of hymn of praise, in which a choir or
congregation in the temple were called upon by a cantor to join in
the words of the psalm (see Gunkel, Einleitung, pp. 33-38; Crüse-
mann, Studien, Chapter 1, with Exod. 15.21b as the oldest surviving
example). Here ‘the women’ occupy that role. On the many expla-
nations that have been offered for the virtual repetition of v. 1b here
see the introduction to this section.
Text and Versions

‫( אז‬15.1) LXX τότε, Sy hydyn and probably TgG(J)(, N?) ‫ בכד(י)ן‬support ‘at
that time’ as the meaning, while TgO,J,G(FF),F ‫( בכ(י)ן‬to which TgJ,G(FF),F(P) prefix
)‫ה(א‬, ‘behold’) suggest the weaker ‘then, next’. Vulg tunc could be taken in
either way.
‫( ישׁיר‬15.1) SP writes defectively, ‫( ישׁר‬cf. GSH §82aβ), Sy (except 5b1)
šbḥw assimilates to the compound subj. and ‫ ויאמרו‬later in the verse. Tgg also
use ‫שׁבח‬, more often ‘praise’, but with an obj. presumably ‘sing in praise’.
LXX and Vulg follow MT.
‫( השׁירה‬15.1) TgO and Sy simply use a cognate noun of ‫שׁבח‬, meaning ‘a
song of praise’ (cf. its use in the various titles prefixed to the section in mss of
Sy), but the other Tgg combine this with )‫שׁירת(א‬, with the latter preceding in
TgN,G but more awkwardly following in TgJ,F.
‫( ליהוה‬15.1) LXX has τῷ θεῷ for the first occurrence but τῷ κυρίῳ for the
second (cf. the alternation – for variety? – in 14.31). The Tgg as usual have
‫ קדם‬for ‫ ;ל‬TgG(J) writes ‫ אדני‬for the divine name the first time.
‫( לאמר‬15.1) Vulg and Sy have no equivalent, presumably regarding it as
otiose. Aq renders pedantically with λέγειν, as do Syh (i.e. the O-text) and
LXXA.
15.1-21 377
‫( אשׁירה‬15.1) SP ‫( אשׁירו‬the plene reading is better attested than von Gall’s

‫ )אשׁרו‬is anomalous and must (unless [J. Joosten, pers. comm.] it is meant
to be a Hiphil imper.) be a mechanical hybrid due to the reading ‫ שׁירו‬in the
otherwise identical line in v. 21 (SamTg has the pl. imper. ‫ שׁבחו‬here too). The
Vss all have first person pl. forms to agree with the compound subj. in the
(secondary) introduction to the poem (the Jewish Tgg again render with ‫שׁבח‬
[as does Sy] and add ‘let us give thanks’ either before [TgJ,N,G,F] or after [TgO]
it). MT’s reading fits the first person sing. forms that follow in v. 2 and must
be more original.
‫( כי‬15.1) Sy has no equivalent (also in v. 21) and apparently takes the
next word as a title of Yahweh (‘the glorious One’): possibly, like TgJ, its
following d is a substitute for ‫ כי‬with only a slight difference in sense.
‫( גאה גאה‬15.1) MT’s text (including the vocalisation) is clearly reflected
in LXX and Vulg (with adverbs unusually corresponding to the inf. abs., as
in Gen. 32.13; Exod. 8.24; Num. 22.17: cf. Lemmelijn, Plague, p. 144), but
SP reads ‫( גוי גאה‬also in v. 21), ‘an arrogant nation’ (cf. SamTgJ: other mss
read it as a verb), i.e. Egypt, which it must take as the primary object of ‫רמה‬.
The understanding of the second word as the adj. ‫ נֵּ ֶאה‬probably lies behind
the expansions of the phrase in the Jewish Tgg too, which all include a refer-
ence to human arrogance, but their interpretations also assume that the first
word is a verbal form with Yahweh as subject in the sense ‘was high (over),
overcame’. This approach is recorded (at length) in MRI (Lauterbach 2,
pp. 13-19), but before it are other interpretations which are based more closely
on MT (pp. 8-13).105 Sy (as noted above) takes the first ‫ גאה‬as a title of Yahweh
(cf. TgJ,G(J)) and reads the second as a verbal form like MT, but linked with ‘the
horses’, supplying ʿl before the latter. MT (with early support from LXX and
[in part] MRI) should be preferred to the other readings attested or implied,
which can be attributed to puzzlement over the repeated consonants and/or
interpretive ingenuity.
‫( סוס ורכבו‬15.1) The words are preserved in 4QExd but only the first
two letters and the final waw are certain. Tgg (except TgO) and Sy render
both words in the pl. but may simply be unpacking the collective meaning of
MT. In general the Vss support the pointing of ‫ רכבו‬as a participle: the sense
‘(their) chariot(s)’, implying ‫ר ֶכב‬,ֶ is found only in a few witnesses (LXXFb
ἅρμα; TgF(P) ‫ ;ורתיכיהון‬the Sam. oral tradition rikbu [J. Joosten, pers. comm.:
105
The Targumic expansions incorporate up to three elements: (1) two-part
formulae acclaiming Yahweh’s general superiority to arrogant humans; (2) a
general statement about the punishment (‫ )פרע‬of the arrogant; (3) application of
this principle to Pharaoh or the Egyptians. TgO omits (2) and has a distinctive
version of (3); TgN omits 1); TgJ and the other Pal Tgg texts have all three, but
with variations among themselves.
378 EXODUS 1–18
cf. GSH §55bα-β]; perhaps TgF(V) ‫)וריכביהון‬. The emendation of ‫ ורכבו‬to ‫וָ ֶר ֶכב‬
was proposed by P. Haupt, ‘Moses’ Song of Triumph’, AJSL 20 (1903–1904),
pp. 149-72 (153, 158-59) and adopted by Gressmann (in fact ֹ‫וְ ִר ְכבּו‬: Anfänge,
p. 55), Mowinckel (‘Drive and/or Ride’, pp. 278, 284) and Cross and
Freedman, Studies, pp. 50, 52. The argument for it is historical rather than
philological and it is probably unnecessary: see Note g on the translation. The
lack of an equivalent to the suffix in LXX* and Vulg is scarcely evidence for
a different Vorlage, especially when all the Heb. witnesses (including 4QExd)
have the suffix.
‫( רמה‬15.1) Tgg, other than TgO, add either ‫( וטמע‬TgJ,F,G(J),Nmg) or ‫וטבע‬
(TgN,G(FF)) to match v. 4: ‫ טמע‬is presumably the original Aram. supplement and
‫ טבע‬an assimilation to the Heb. form.
‫( בים‬15.1) TgJ,N,G(FF),F give the specific location ‘Yam Suf’ from v. 4. TgG(J)
‫ בגו ימא‬may have been based on 14.27 or 29.
‫( עזי‬15.2) The expected sense ‘my strength’ is given in TgO and Vulg;
TgJ,N,F(V) have ‘our strength’ in accordance with their renderings of ‫ אשׁירה‬in
v. 1, while Sy tqypʾ, ‘strong, mighty’, ignores the suffix and generalises the
statement (cf. TgF(P)). LXX βοηθός does likewise with a slight shift of mean-
ing, for which there are several parallels in the Psalms (28.8; 59.18; 81.2).
‫( וזמרת יה‬15.2) The absence of a suffix on ‫ זמרת‬in MT is surprising after
‫( עזי‬LXX [except the O-text] and Sy also show no knowledge of one, but after
their treatment of ‫ עזי‬this may not be significant). Most SP mss read ‫( וזמרתי‬and
omit ‫)יה‬, including the Rylands ms. used by Crown and Camb. 1846, but a
few (inc. those used by Tal and Sadaqa and Crown’s other three) have ‫וזמרתיה‬,
some with one or more letters detached at the end (see von Gall’s apparatus).
This reading is likely to be due to contamination from MT. TgO and Vulg
have ‘my praise’ (cf. LXX at Isa. 12.2; Ps. 118.14), the other Tgg ‘our praise’.
MT could have lost a final yodh by haplography: for other possible explana-
tions of its reading see Note i on the translation. Sy’s reading mšbḥ, ‘(to be)
praised’, is a free variation of the sense given by Tgg and Vulg, but LXX has
σκεπαστής, ‘protector’ (cf. σκεπάζω in the papyri: LSJ, p. 1606), which also
occurs at Deut. 32.38 (for ‫ )סתרה‬and Ps. 71.6 (for ‫גוזי‬: read ‫)?עוזי‬. At one time
this rendering was seen as evidence for a reading ‫ סתרתי‬here (so still BHS), but
it now seems likely to preserve an alternative (and probably original) meaning
for )‫וזמרת(י‬: see Note i on the translation. LXX also had no equivalent to ‫;יה‬
the O-text adds κύριος. Tgg expand the text, TgO with ‫דחילא‬, ‘the fearsome
One’ (no doubt based on ‫ נורא‬in v. 11, where ‫ תהלת‬follows), which TgJ,N,F(V)
extend by ‘in all ages’; TgJ,N,F with ‫רוב‬, ‘magnitude of’ before ‘our praise’;
and TgF(P) with ‘he is the Master of all worlds’, a liturgical formula (Jastrow,
p. 1440). Sy adds mryʾ, its usual equivalent to the divine name, to yh, perhaps
from Isa. 12.2.
‫( ויהי לי‬15.2) LXX ἐγένετό μοι ignores the waw, making ‫ עזי וזמרת‬the
subject or (Wevers, Notes, p. 228) complement of the verb. All the Tgg prefix
15.1-21 379
‘he said by his Memra’: the reason is not clear, unless it is to recall that God
promised that he would deliver Israel (through Moses presumably, in 14.13-
14; or perhaps implied in 14.4, 17-18). TgO,J and Vulg render ‫ לי‬by ‘to me’
(like LXX), but the Pal Tgg and Sy continue their first person pl. forms.
‫( לישׁועה‬15.2) LXX εἰς σωτηρίαν and Vulg in salute reflect the ‘action
noun’ of the Heb. idiom; Tgg and Sy use ‘agent nouns’ to fit the personal
subject and the intended sense.
‫( זה אלי‬15.2) LXX, TgO and Vulg render MT (with which SP agrees)
straightforwardly, as does Sy, which now follows MT’s first person sing.
forms, whereas the other Tgg once more have first person pl. Most of the latter
also contain, in various forms, a Tosefta whose nucleus (based on Ps. 8.2b)
appears in MRI (Lauterbach 2, pp. 11-12): see the detailed study of M.L Klein,
‘The Targumic Tosefta to Exodus 15:2’, JJS 26 (1975), pp. 61-67 (= Michael
Klein on the Targums [2011], pp. 133-40), where a fuller version from an
Ashkenazi maḥzor (MS Parma 2887: fourteenth cent.) is presented. In TgN,F(P)
most of the Tosefta is missing, but the superfluous words ‘the Israelites said’
must derive from an exemplar (like TgF(V)) which included a form of it: TgNmg
restores it in that form.
‫( ואנוהו‬15.2) So also SP: almost all the Vss have a verb meaning ‘praise’,
which the context requires (see also Note m on the translation), but TgO ‫ואבני‬
‫ ליה מקדשׁ‬found here a reference to the future temple, probably on the basis
of the use of the noun ‫ נוה‬in v. 13 and other refs. in the Note there: this inter-
pretation appears, among others, in MRI (Lauterbach 2, pp. 25-26). Sy does
not represent the initial waw here or with the following verb (Propp, p. 472).
‫( אבי‬15.2) Tgg (including TgO ‫ )אבהתי‬make explicit a reference to the
‘fathers’ of the people as a whole (cf. MRI [Lauterbach 2, p. 29]); Sy, as in
most of the second half of the verse, follows MT precisely.
‫( וארממנהו‬15.2) TgO (but not the other Tgg here) renders ‫ואפלח קדמוהי‬, with
the common ‘distancing’ of God from human words and perhaps an allusion
to temple worship.
‫( יהוה‬15.3) TgJ prefixes ‘The Israelites said’ (cf. TgN in v. 2). Two ‘litur-
gical’ (Klein) Pal. Tg mss (TgG(U) and TgF(P)) introduce their rendering of the
verse with the midrash which TgJ and the other Pal. Tgg have in its original
place at 14.13 (see Text and Versions there).
‫( אישׁ מלחמה‬15.3) SP and the Vss (except LXXFb and TgN: see below)
found various ways of reproducing the sense of MT (or at least providing
an appropriate substitute) without giving the impression that Yahweh was
a human being (‫)אישׁ‬: for the concern see MRI (Lauterbach 2, p. 34). In
most cases ‫ גבור‬or its Aram. equivalent is used in place of ‫אישׁ‬, probably as
an adj. = ‘mighty’, and ‫ מלחמה‬is paraphrased in various ways. Thus SP has
‫ גבור מלחמה‬and Sy has gnbrʾ wqrbtnʾ, ‘mighty and warlike’, while TgJ and
most of the Pal Tgg begin with ‘mighty, who fights our/your wars’ which is
then elaborated in various ways. TgN ‫ גוברא עבד קרביא‬is closer to MT and
380 EXODUS 1–18
perhaps the result of assimilation to it. TgO ‫ מרי נצחן קרביא‬and Vulg quasi
vir pugnator (cf. MT at Isa. 42.13: also MRI [Lauterbach 2, p. 30-31, where
‫ גבור‬is also used]) exhibit different ways of achieving the same end. LXX
συντρίβων πολέμους, ‘breaking [ending?] wars’, as also in Isa. 42.13,
makes a larger change to the meaning, but one that is based on other biblical
passages where συντρίβω is used (Hos. 2.20; Ps. 76.7: cf. 46.10): cf. also
Fritsch, p. 9. The mg of LXXF preserves the correction ἀνὴρ πολέμου, which
may derive from Aq.
‫( יהוה שׁמו‬15.3) Vulg renders the divine name here with omnipotens,
which is usually its equivalent for ‫( שׁדי‬cf. 6.3: in 34.23 it renders ‫)האדון‬:
Jerome seems to have regarded all these expressions as synonymous, presum-
ably as being implied by Dominus, his usual rendering for ‫יהוה‬. TgJ and the
Pal. Tgg have ‘as his name is, so is his power (‫)גבורתא‬, may his name be
praised for ever and ever’, which the two liturgical versions (TgG(U) and TgF(P))
elaborate further.
‫( פרעה‬15.4) TgF(P) adds the epithet ‘wicked’ (cf. TgN in v. 9, TgJ in vv. 1,
9 and 21, TgF in v. 9 and TgG(U) in v. 9).
‫( וחילו‬15.4) Although von Gall gives the same reading for SP, many mss
(including those used by Tal, Sadaqa and, in three cases, Crown) read the pl.
form ‫וחיליו‬, and this is supported by TgJ and the Pal. Tgg (see also Text and
Versions on 14.9). The pl. became frequent in exilic and later BH and in post-
biblical Heb. and Aram., but older texts use only the sing. and that is likely to
be the original reading here (cf. LXX, TgO, Sy, Vulg).
‫( ירה‬15.4) SP reads ‫ירא‬, due to its common confusion of the gutturals (cf.
GSH §12a). In TgN,F(V) ‫ ירה‬is taken as a reference to shooting arrows (of fire
in TgN), but the additional obj. required shows that this was not the original
meaning.
‫( ומבחר שׁלשׁיו‬15.4) LXX (which like Vulg has no equivalent to the waw:
Cross and Freedman [Studies, p. 58] saw it as a secondary addition) here
equates the ‫( שׁלשׁים‬τριστάτας as in 14.7: see Text and Versions there for this
and what follows) with the ‘rider(s)’ of v. 1 by prefixing ἀναβάτας. Only
Vulg (principes, instead of duces in 14.7) gives the sense ‘commanders’: in
Tgg and Sy ‘warriors’ appears throughout here, but with the specification
‘young’ added in TgJ,N,F,G (due to a double rendering of ‫ מבחר‬as in the mg of
LXXF?). TgN specifies that the warriors were Pharaoh’s.
‫( טבעו‬15.4) MT’s pl. is also attested by SP, LXXB, TgO, Vulg and two of
the oldest mss of Sy (5b1 and 7a1). But LXX* κατεπόντισεν (cf. most mss
of Sy) points to a sing. active form (with Yahweh as subj.: so more ‘anthropo-
morphic’ than MT, as also in v. 5 [Fritsch, p. 62]), which Cross and Freedman
thought could go back to an early spelling in which final waw was not written
(Studies, p. 58). But the sing. may be due to the influence of v. 4a and also
v. 1: in TgJ and most of the Pal. Tgg (not TgF(V),Nmg) ‫ רמא‬is actually imported
from v. 1 alongside a sing. ‫ וטמע‬or ‫וטבע‬.
‫( תהמת‬15.5) Because LXX saw Yahweh as the subj. of ‫( יכסימו‬see the next
note) it rendered ‫ תהמת‬by an instrumental dative πόντῳ. Aq, Symm and Theod
15.1-21 381
ἄβυσσοι conformed to the straightforward understanding of the Heb., which

all the other Vss exhibit. LXXO (cf. Syh) seems to have read πόντος, perhaps
an earlier correction.
‫( יכסימו‬15.5) So also some mss of SP, but most read ‫יכסמו‬, avoiding the
unusual retention of the third root-letter (on which see Note s on the trans-
lation). LXX ἐκάλυψεν, which keeps Yahweh as the subj. (as in v. 4a and
in LXX also v. 4b), may have been based on the same reading (this could
explain why LXX is more ‘anthropomorphic’ here, as Fritsch, p. 62, noted):
but it must be secondary. LXX’s aorist, like all the other Vss, recognises the
preterite sense of the Heb. (cf. vv. 6, 7, 12, 14, 15). A Genizah ms. (cf. BHS)
reads ‫תכסיומו‬, with an additional mater lectionis making the pl. form apparent
and a fem. prefix to match the more common gender of ‫( תהום‬but it can be
masc.: cf. BDB, p. 1062). The two ‘liturgical’ Pal. Tgg mss (TgG(U) and TgF(P))
add ‘the waters of the sea’ (cf. v. 19), as a gloss on ‫ תהמת‬presumably.
‫( ירדו‬15.5) TgJ,N add ‘and sank’ (cf. v. 10) and TgG(U),F(P) have this instead.
Sy’s addition of wtbʿw draws instead on v. 4 to enrich the description.
‫( במצולת‬15.5) TgJ adds ‫דימא‬, another of its (here hardly necessary) explan-
atory additions. The addition of ‫ דימא רבא‬in TgG(U),F(P) is more interesting, as
it appears to be based on an explanation of the verse in MRI (Lauterbach 2,
p. 38), which sees the ‘deep waters’ as resulting from the ‘Great Sea’ (i.e.
the Mediterranean) breaking through into the Yam Suf. This phrase already
appeared in TgF(P) in 14.23: see Text and Versions there for its other occur-
rences in connection with the Egyptians’ fate. Propp (p. 473) suggests that
LXX’s sing. may imply a sing. Heb. here, but βυθός is mostly used in the sing.
and so cannot be pressed in this way.
‫( כמו־אבן‬15.5) The sing. is understood collectively, ‘stones’, in TgJ,F(V) and
Sy (note the seyame). TgJ,N,F(V) add a verb ‘were made like’ or ‘made them-
selves like’, possibly an allusion to an interpretation in MRI (Lauterbach 2,
p. 39): ‫על שׁהקשׁו את לבם כמו אבן‬.
‫( נאדרי‬15.6) Many SP mss (inc. all of Crown’s and Camb. 1846) read
‫ נדארי‬by metathesis, which is found elsewhere involving aleph (cf. 2.24; 6.5
and GSH §20a) and probably involves no lexical difference (cf. SamTg).
LXX, TgJ and the Pal. Tgg take it with ‫( ימינך‬likewise Rashi), despite the
gender difference, but Vulg magnifice and possibly TgO and Sy construe it
as a vocative with ‫יהוה‬. ‘Glorious’ is the sense generally given, but TgJ,Nmg,F(V)
have ‘praiseworthy’.
‫( בכח‬15.6) TgG(G) gives two renderings of the verse, one close to TgO and
the other expanded with a list of Israel’s deliverances similar to what MRI
(Lauterbach 2, pp. 45-46) has on v. 7. This unusual ms., which continues to
v. 16, is characterised throughout by a ‘base text’ close to TgO and expansions
which have few if any parallels in other Tg witnesses for this passage: see
further the Introduction to the Commentary 2 (ii) c.
‫( ימינך‬15.6)2o Here, for ‘poetic variation’ (Wevers, Notes, p. 229), LXX
adds χείρ to its previous rendering ἡ δεξιά σου: Syh, following Aq, Symm,
Theod, omits it.
382 EXODUS 1–18
‫( תרעץ‬15.6) Most of the Vss render in the past tense, but TgJ ‫ תכרית‬is
either future (cf. MRI [Lauterbach 2, p. 42]) or iterative (AramB). TgF(V),Nmg
neatly combine the two interpretations: ‫דרצצת ומרצצת‬.
‫( אויב‬15.6) TgO and Vulg follow MT’s sing., LXX ἐχθρούς takes it
collectively. TgN,G(U),F(P) add ‘and the adversary’, a word which some Tgg
have in the next verse (cf. Sy ‘your adversaries’ here), and TgG(G) ‘the son
of iniquity’; TgN names him as ‘Pharaoh’ (cf. v. 4), TgG(U),F(P) with the same
identification in view prefix ‘the army of’. TgJ and TgF(V),Nmg generalise here
too but in different ways.
‫( גאונך‬15.7) The variant ‫ גאוניך‬in many SP mss (inc. Crown’s) is phonetic
(GSH §55bγ) and does not affect the meaning. TgO,G(G) and Sy render ‘your
might’ from the context.
‫( תהרס‬15.7) TgO,G(G),G(U),F(P) render ‘you broke, shattered’ (cf. AramB 7,
p. 42 n. 13 and ‫ תרעץ‬in v. 6) and LXX συνέτριψας takes the same view (as it
does in Ps. 58.7), apparently also finding a metaphorical use of ‫ הרס‬too bold.
A similar difficulty may lie behind the treatment of its object here in some Tgg
(see the next note). TgJ,N relate the verbs in this verse to the future (cf. MRI
[Lauterbach 2, pp. 47-48]).
‫( קמיך‬15.7) TgJ,G(U),F(P) prefix ‘the walls of’, presumably supposing that
‫ הרס‬implies the removal of such structures. This supplement also found its
way, less obviously, into TgF(V),Nmg in v. 6. Only Sy lsnʾyk among the Vss
rendered the suffix precisely: LXX ignored it (Greek usage perhaps being
reinforced by exegetical considerations here: Fritsch, p. 11, compares 23.27
and Deut. 32.10 for similar safeguarding of God’s majesty), the Tgg have ‘the
adversaries of your people’, in line with the explanation justified at length
in MRI (Lauterbach 2, pp. 42-47), and Vulg’s original adversarios meos (of
Moses?) is probably an adaptation of the same approach. TgG(G) has a longer
addition here.
‫( חרנך‬15.7) TgJ,N prefix ‘the might of’ and add ‘against them’; TgG(U),F(P)
read ‘the angel of (your) wrath’, distancing the intervention from God himself
and recalling the ‘destroying angel’ found in 12.13, 23 by TgJ (cf. TgNmg).
TgG(G), like TgO, avoids these elaborations but adds ‘to weaken their armies’.
‫( יאכלמו‬15.7) LXX, TgG(U),F(P) and Sy avoid the asyndeton by adding ‘and’
(cf. Vulg quae).
‫( כקשׁ‬15.7) The Tgg clarify the comparison by inserting references to
‘fire’ (so more briefly Sy ʾyk dlḥbtʾ, ‘as [for] stubble’).
‫( וברוח אפיך‬15.8) Most of the Tgg paraphrase with references to divine
speech (‘of your mouth’ in TgO, ‘from before you’ in the remainder) to avoid
the strong anthropomorphism of MT, which they evidently presuppose (as
does Sy’s literal rendering). LXX τοῦ θυμοῦ σου and Vulg furoris tui (cf.
TgG(G)) suggest a Vorlage ‫ אפך‬rather than ‫אפיך‬, and this is exactly what most
SP mss have (inc. Tal, Sadaqa, 3 of Crown’s mss and Camb. 1846): it, rather
than von Gall’s ‫( אפיך‬for which cf. GSH §55bγ), is likely to be the oldest SP
reading (cf. SamTg). Though early (cf. LXX), it is the easier reading, as it too
15.1-21 383
avoids the physical anthropomorphism (cf. Deut. 33.10 and Fritsch, p. 14),
and MT must be original. TgN,G(U),F(P) add ‘O Lord’ to leave no doubt about who
is addressed (cf. MT in v. 6).
‫( נערמו‬15.8) The sense of this rare verb is represented by the Three, TgJ,
the Pal. Tgg and Sy as ‘were heaped up’. Vulg congregatae sunt is less specific
(it is used again for ‫ )קפאו‬but fits the context. LXX διέστη, ‘were separated’,
is (like its treatment of ‫ )נד‬based on 14.21-22, while TgO ‫ חכימו‬fancifully
related the form to ‫ערום‬, ‘crafty’, a view also attested in MRI (Lauterbach 2,
pp. 50-51, where ‘he made them like heaps’ is also found).
‫( נצבו‬15.8) TgJ,G(W),F(P),Nmg (TgG(U) is no longer extant) add ‫להון‬, ‘for them’.
This could refer to the enemies of v. 7 (as does TgJ’s ‫ עליהון‬later in the verse)
and accords with the interpretations given in MRI (Lauterbach 2, pp. 51-52:
see the next note). LXX ἐπάγη, ‘solidified’ (which it also has, more precisely,
for ‫)קפאו‬, is from a word that is used only here to represent the common verb
‫( נצב‬which caused the other Vss no difficulty): presumably the translator was
looking for a word to fit his rendering of ‫ נד‬by τεῖχος. TgG(G) has ‘separated
this way and that’.
‫( כמו־נד‬15.8) TgO (on which TgG(G) seems to depend here) has ‘like a wall
(‫’)שׁור‬, which has no etymological basis. The other Tgg related ‫ נד‬to ‫נֹאד‬, ‘skin,
bottle’, adding ‘tied up’ (‫ )שׂרירין‬to clarify the image (cf. Sy). This interpreta-
tion was known to MRI but there it is applied to the drowning Egyptians, not
the waters, and the ‫ נזלים‬are taken to be ‘sweet waters’ for the Israelites. Vulg
adopted neither of these devices and gave no equivalent at all here, as unda
fluens is presumably its rendering of ‫נזלים‬.
‫( נזלים‬15.8) TgN has ‫דמיא נזליא‬, ‘(like skins) of flowing waters’, whereas
other Vss correctly take ‫ נזלים‬as the subj. of ‫נצבו‬.
‫( בלב־ים‬15.8) For ‫ ים‬TgJ and the Pal. Tgg (except TgG(G)) all have
‫ דימא רבא‬here (see the note on ‫ במצולת‬in v. 5). For ‫ בלב‬they present what
look like two textually related readings: TgN,F(P) read ‫בפלגות‬, ‘in the midst of’,
which is a close rendering of MT, but TgJ,G(W) have ‫פילגוס‬, preceded by ‫ בגו‬or
‫ב‬, which introduces a Greek loanword for ‘the high sea’ that occurs in the Tgg
elsewhere only in Ps. 46.3 (also followed by ‫)דימא רבא‬. (See Jastrow, p. 1163,
and CAL for further references.)
‫( אמר‬15.9) TgO,J,G(G) (but not the main Pal. Tgg texts) prefix ‫ד‬, which is
probably to be translated ‘because’ (not ‘so that’ as in AramB 7): the clue to
what is intended is in MRI (Lauterbach 2, p. 54), where this verse is related to
the beginning of the episode (i.e. 14.5-8), not (as is in fact more likely to be
meant) the Egyptians’ entry into the sea.
‫( אויב‬15.9) TgJ,N,G(W),F identify the enemy as ‘wicked Pharaoh…the adver-
sary’ (Sy uses the final word as its equivalent to ‫ אויב‬as in v. 6): cf. MRI
(Lauterbach 2, p. 55).
‫( ארדף‬15.9) Each of the verbs and short phrases in the remainder of the
verse is amplified in TgJ and the main Pal. Tgg witnesses. Here TgJ,N,G(W),F add
‘after the people of the Israelites’.
384 EXODUS 1–18
‫( אשׂיג‬15.9) LXX καταλήμψομαι (cf. LSJ, p. 897), TgO and Sy wʾdrk

simply give the sense of MT; Vulg comprehendam, ‘I will seize’, is probably
based on a misunderstanding (via the OL, which had the same rendering)
of the meaning of LXX (in 14.9 it has a better, if not ideal equivalent for
‫ נשׂג‬Hiph.). The other Tgg (except TgF(P), ‘confound’) also render the verb
accurately but amplify it either with ‫( בתקוף ידי‬TgG(G)) or with a much longer
addition, in various forms, which draws on 14.9 and anticipates the capture of
the ‘spoil’ which is presupposed in the next phrase of MT (TgJ,N,G(W),F).
‫( אחלק שׁלל‬15.9) LXX, Vulg, TgO and Sy render MT straightforwardly,
but other Tgg have explanatory additions: TgJ,N,G(W),F specify the recipients of
the spoil as Pharaoh’s army, while TgG(G), again differently, identifies the spoil
with what the Israelites had taken from the Egyptians (3.22; 11.2; 12.36).
‫( תמלאמו נפשׁי‬15.9) LXX ἐμπλήσω [the O-text adds αὐτῶν] ψυχήν
μου assimilates the expression to the first person sing. verbs in the context
(possibly on the basis of a Vorlage ]‫)אמלא[מו‬, a secondary variation. Sy renders
the verb with tblʿ, ‘will devour’. SP, TgO and Vulg agree with MT, which is
also the basis for the elaborations in the other Tgg, whether the prefixing of
‘when’ or ‘until’ in TgJ,N,G(W),F (TgJ also amplifies ‘them’ to ‘the blood of their
slain’) or the addition of ‘when I surround them with my armies and troops’
in TgG(G), which relates this element of MT to what follows rather than what
precedes, as is also presupposed in one comment in MRI (Lauterbach 2, p. 57).
‫( אריק חרבי‬15.9) Most of this phrase survives in 4QExc; the omission
of an equivalent to ‫ חרבי‬in TgF(V) is no doubt accidental. LXX ἀνελῶ (τῇ
μαχαίρᾳ μου) shows that the translator did not recognise the Heb. idiom
(as again in Lev. 26.33 and 3x in Ezek.): he replaced it with a plausible
sense (‘destroy, kill’) that the other Vss found in the next verb. They had no
difficulty with the idiom here, but TgJ,N,G(W),F prefixed ‘after this’ to underline
the sequence of events as indicated in their renderings of the previous phrase.
TgF(P) added ‘from its sheath’ for additional clarity.
‫( תורישׁמו‬15.9) LXX κυριεύσει (with αὐτῶν added in the O-text), ‘shall
dominate’, is a rare (cf. Jer. 30.3) but defensible equivalent for ‫ירשׁ‬. Or
perhaps, as Propp suggests (p. 473), LXX derived the form from ‫רשׁי‬, ‘have
power’. The other Vss all give the sense ‘destroy, kill’, which is less appro-
priate. In TgJ,G(W),F,Nmg the first sing. verb (with ‫ ב‬prefixed to the next word) will
again be due to assimilation to the previous verb(s): the other Tgg follow MT.
‫( ידי‬15.9) TgJ,N,G(W),F specify Pharaoh’s ‘right’ hand, perhaps to match
his boast more closely to the role of Yahweh’s right hand in vv. 6-12, on
the ‘measure for measure’ principle spelt out for other expressions in MRI
(Lauterbach 2, pp. 57-58).
‫( נשׁפת ברוחך‬15.10) SP spells the verb ‫נשׁבת‬, so ensuring consistency with
the only other Pentateuchal occurrence of the verb in Gen. 15.11: it is also
the more common spelling in MT (see Note hh on the translation). As the
difficilior lectio MT is to be preferred here. The strong anthropomorphism,
which must be original, was removed in various ways in the Vss, the boldest
15.1-21 385
variation being TgO,G(G) ‫אמרת במימרך‬, to which TgG(G) added a reference to the
death of magicians and sorcerers. LXX ἀπέστειλας τὸ πνεῦμά σου (cf.
Fritsch, p. 15) also adopted a quite different verb (which was ‘corrected’ by
the Three). The Tgg and Sy used causative forms of ‫נשׁף‬/‫ נשׁב‬as a less drastic
way of overcoming the problem, some (like LXX) ignoring the preposition ‫ב‬
and making ‫ רוחך‬the object. Vulg (cf. TgNmg) also evaded it and made ‫ רוחך‬the
subject. TgJ and the Pal. Tgg, as elsewhere, make use of the formula ‘(from)
before you’ to distance the ‫ רוח‬from God himself. TgJ,N,F(P) added ‘O Lord’ as
in v. 8 to identify the ‘you’.
‫( כסמו‬15.10) TgG(W), which is an idiosyncratic member of the main Pal.
Tg group in this verse, reads ‘you covered’.
‫( ים‬15.10) TgN,F(P) prefix ‘the waters of’, anticipating ‫ מים‬later in the verse
(and cf. v. 19). TgJ has the puzzling ‫ גירין‬instead: CAL derive it from a by-form
of ‫גילא‬, ‘clay’, with reference to the wet clay at the bottom of the sea. TgG(G)
simply added ‘and their mouths with which they uttered blasphemies were
shut up’.
‫( צללו‬15.10) TgJ,N prefix ‘they went down and’ on the basis of v. 5.
‫( מים אדירים‬15.10) The adj. is variously rendered ‘violent, mighty’ (LXX,
TgO,N,G(G), Vulg, Sy) or ‘glorious’ (TgJ,G(W),F(P)).
‫( מי כמכה‬15.11) SP and 4QExc (which preserves only the first occurrence)
read the older spelling ‫כמוך‬, without the final vowel letter: cf. Note kk on the
translation and GK §9d. TgO completely rewrites this phrase in both its occur-
rences, turning the questions into statements, not about incomparability but
about existence: ‘There is none except you…There is no (other) god’.106
‫( באלם‬15.11) In place of this TgO has ‘You are (the only) God’; Sy simply
omits the troublesome suggestion that other gods exist. LXX ἐν θεοῖς takes
the Heb. at face value, and the plene reading ‫ באלים‬of SP and 4QExc confirms
the MT vocalisation as a pl. noun. MRI, however, knew the consonantal
text of MT (which has the appearance of an archaic spelling) and, in one
of its explanations (Lauterbach 2, p. 60) derived it from the MH verb ‫א ַלם‬,ָ
‘be strong’, hence ‘among the strong’: cf. here LXXFb (Aq?) (ἐν) ἰσχυροίς,
Symm ἐν δυναστείαις, Vulg in fortibus. TgJ and the Pal. Tgg, on the other
hand, agree with MT, only adding ‫מרומא‬, ‘on high’ (perhaps from Ps. 89.7),
which left no doubt that heavenly beings were intended, but perhaps angels,
as such expressions were often understood in Judaism (cf. P.S. Alexander,
‘The Targumim and the Early Exegesis of “Sons of God” ’, JJS 23 [1972],
pp. 60-71): certainly this is how another interpretation of ‫ באלם‬in MRI under-
stood it (Lauterbach 2, p. 61: ‫)באלו שׁמשׁמשׁין לפניך במרום‬. TgG(G) begins like TgJ
and the Pal. Tgg but then adds words of praise and answers the question with
a strong negative statement, similar though not identical to that found in TgO:
‘there is none like you’.
106
For others who interpreted the verse in this way see Salvesen, Symmachus,
p. 94.
386 EXODUS 1–18
‫( נאדר בקדשׁ‬15.11) SP reads ‫ נאדרי‬as in v. 6, no doubt a secondary assimi-

lation. 4QExc does not survive for this word. Most of the Vss render ‫ נאדר‬as
they do in v. 6, but TgJ has ‘glorious’ here (the other two witnesses for ‘praise-
worthy’ there offer nothing) and TgG(G), which has ‫ אדיר‬like TgO, again adds
an extra question, ‘Who is like you, revealing mysteries before the sun?’, and
answers it like the previous ones. Most of the Vss render ‫ קדשׁ‬as an abstract
(this may even be the case with Symm [ἐν] ἁγιασμῷ: so Salvesen, p. 94), but
LXX ἐν ἁγίοις personalises it, ‘among holy ones’ (cf. Ps. 89.8), most likely
understood as angels, and Cross and Freedman (Studies, p. 61), among others
(cf. BHS), read ‫ קדשׁ(י)ם‬on the basis of this. It makes a good parallel to ‫באלם‬,
but perhaps too good: the less obvious sing. of MT, SP and the other Vss (and
apparently 4QExc) should be preferred.
‫( נורא תהלת‬15.11) Tgg follow MT (with which SP agrees), and Vulg
and Sy ‘fearsome and praiseworthy’ give a recognisable unpacking of its
meaning. LXX θαυμαστὸς ἐν δόξαις takes liberties with both Heb. words:
θαυμαστός anticipates ‫ פלא‬in the next phrase; and δόξα, which only here
stands for ‫תהלה‬, suggests objective ‘glory’ rather than ‘praise’.
‫( עשׂה פלא‬15.11) SP ‫ פלאה‬uses a unique fem. form (for such variation in
gender cf. GSH §145). Most of the Vss (not TgG(G), which keeps the sing.)
rightly take ‫ פלא‬to mean ‘wonderful deeds’ in the pl., with some amplification
in TgJ,N,G(W),F(P).
‫( נטית ימינך‬15.12) TgJ,N,G(FF,W),F prefix a midrash to this verse (with some
variations of wording) to explain why it was the earth and not the sea which
‘swallowed up’ the Egyptians: there was an argument between sea and earth
which was resolved when God swore an oath (‘stretched out his hand’: cf.
the idiom in 6.8 etc.) that no guilt would be attached to the earth for this on
the day of judgement. The same explanation appears in MRI (Lauterbach 2,
pp. 67-68) and elsewhere (for refs. see AramB 2, p. 204 n. 23); but not in TgO
or TgG(G) (on which see below). 4QExc is damaged but seems to have read
]‫( [הרי]מו[ת‬DJD XII, p. 118), which corresponds to the rendering of TgO,G(G)
and Sy (for earlier instances of such a rendering of ‫ נטה‬see Text and Versions
on 14.16, 21, 26). The other Vss agree with MT (and SP) here, except that
Vulg has manum instead of the expected dextram and TgG(G) adds some
general statements about the activity of God’s hand for good and evil before
repeating its rendering of MT.
‫( תבלעמו ארץ‬15.12) The preterite sense of the verb is recognised in all
the Vss: TgJ,N,G(FF,W),F prefix ‘opened its mouth’ (from Num. 16.32) to it. SP
reads ‫הארץ‬: such variations from MT in the use of the article are frequent (cf.
GSH §166), but this is the only place in the poem where SP adds an article.
4QExc uses a final mem in its writing of ‫תבלעמו‬, as it does in the similar form
in v. 15: since it omits the archaic waw altogether twice in v. 17, this may be
a deliberate signal that the shorter form was to be preferred. SP retains the
long form with verbs, but avoids it with nouns in Deut. 32 (cf. GSH §55bγ).
15.1-21 387
‫( נחית‬15.13) Most of the Vss give a straightforward rendering of MT

(only Vulg’s dux fuisti[…populo] occasions some surprise, but cf. 13.21;
Deut. 32.12; Pss. 31.4; 48.15; 78.72); TgO,G(G) ‫( דברהי‬cf. TgF(P)), however, must
surely mean either ‘he led (it, sc. the people)’ or ‘lead (it)’ (imper.).107 R.P.
Gordon (personal comm. 19 February 2014) suggests to me that it is an imper.,
expressing a prayer for the Jews of the Targumists’ own times (cf. Tg to Zech.
2.16: Gordon, Studies in the Targum to the Twelve Prophets, from Nahum to
Malachi [VTSup 51; Leiden, 1994], pp. 130-31).
‫( בחסדך‬15.13) LXX τῇ δικαιοσύνῃ σου employs a surprising equivalent
for ‫ חסד‬which (apart from Prov. 20.28; Isa. 63.7: cf. 57.1) only appears in the
Pentateuch (5x Genesis; 2x Exodus [cf. 34.7]). For its 13 other occurrences
in the Pentateuch (from Gen. 24.12 on) LXX has the expected ἔλεος, which
Aq introduced here. The implications of the wider use of δικαιοσύνη (and
δίκαιος) by the LXX translators deserve further study (cf. D. Hill, Greek
Words and Hebrew Meanings: Studies in the Semantics of Soteriological
Terms [SNTSMS 5; Cambridge, 1967], pp. 104-106). Here the other Vss
render in their normal way: Vulg by misericordia (cf. ἔλεος) and Tgg and Sy
by ‘kindness, good deeds’.
‫( עם זו גאלת‬15.13) SP replaces the rare ‫ זו‬with ‫זה‬, as it does again in v. 16,
clearly preferring the more familiar form. 4QExc does not preserve the word.
Most of the Vss adopt the interpretation ‘this’, which is also suggested by the
maqqeph in MT, and presumably supply the rel. pron. (and in some cases the
possessive ‘your’) from the context: Vulg (quae alone) looks like an excep-
tion, but in view of its iste in v. 16 it may simply offer a freer version of the
analysis in the other Vss. The correct understanding of ‫ זו‬here (see Note ss)
seems only to have become established in the course of the sixteenth century
(Luther, Tremellius; then the AV: see my ‘Some Points of Interest’, pp. 251-
52). TgG(G) continues the sentence with ‘from darkness, to sustain them with
quail and manna and to shade them from the heat of your sun’.
‫( נהלת‬15.13) For SP von Gall gives the MT reading, but nearly all SP mss
(inc. Tal, all of Crown’s and Camb. 1846; also SamTg) read ‫נחלת‬, presumably
Piel in the sense ‘you caused to inherit/possess’ (cf. TgJ ‫)ואחסינת‬. The reading
probably originated in the frequent interchange of the gutturals in SP mss (cf.
GSH §12h), but owed its popularity to its straightforward and welcome sense.
The following ‫אל‬, however, only really fits with the MT reading, which must
be original. The Vss seem to have had serious difficulty with the sense (and
not only here): only Sy gives a precise equivalent. TgO has the imper. again;
TgN has a future form; LXX renders with παρεκάλεσας, ‘you invited’, which
Symm ‘corrected’ to διεβάστασας, ‘you carried through’ (cf. Vulg); and
107
Both Grossfeld (AramB 7, p. 42) and Klein (1, p. 250; Fragment-Targums,
2, p. 47) translate ‘you led’, without comment.
388 EXODUS 1–18
TgG(G),F(P) have the imper. of ‫סובר‬, ‘carry’. The use of verbs for ‘carry’ may be
guided by the expression ‫ אשׁר נשׂאך יהוה‬in Deut. 1.31; the other variations are
probably due to puzzlement at the use of the perfect of a stage in the journey
which still lay in the future at the point when the song was supposed to have
been sung (cf. AramB 2, p. 204 n. 24).
‫( אל־נוה קדשׁך‬15.13) The required sense ‘dwelling’ for ‫ נוה‬is conveyed
in various ways in the Vss, with the elaboration ‘the house of (your holy)
Shekinah’ in TgJ,N,G(W),F(P); TgJ also prefixes ‘the mountain of your sanctuary’,
to make the reference to the (Jerusalem) temple even clearer (cf. MRI [Lauter-
bach 2, p. 70]; this may well be correct: see the Explanatory Note).
‫( שׁמעו‬15.14) Vulg adtenderunt, ‘gave heed’ (as apparently only elsewhere
in Jer. 29.8), gives ‫ שׁמע‬a sense which it can certainly have, but its reason for
preferring it here is elusive.
‫( עמים‬15.14) LXX ἔθνη reflects its tendency to avoid the more frequent
equivalent λαός when ‫ עם‬refers to a foreign people (so especially clearly in
19.5): 23.11 is a rare case where ἔθνος is used for ‫ עם‬referring to Israel. A
marginal note in one ms. (Fb) records the variant λαοί, which might be from
Aq.
‫( ירגזון‬15.14) SP reads ‫וירגזו‬, with waw added to make the connection and
the paragogic nun omitted (cf. GSH §189bβ and 63b respectively: in both
cases the variation from MT can go either way), and 4QExc has the same
reading. LXX, Vulg, TgO,G(G,W),F(P) and Sy also have ‘and’, but the doubly
unusual reading of MT (cf. TgJ,N) is most likely original. LXX and Vulg give
‫ רגז‬the sense ‘be angry’ (which it can have, esp. in Aram.: cf. ‫ כעס‬here in MRI
[Lauterbach 2, p. 71]), but are probably only following the interpretation they
gave it (more plausibly) on its previous occurrence in Gen. 45.24: here it does
not fit the context as well as ‘trembled, were fearful’, which Tgg and Sy have
(Schleusner’s view, reported by Wevers, Notes, pp. 232-33, that ὀργίζομαι
could mean ‘be frightened’ is most unlikely). Aq ἐκλονήθησαν, ‘rushed
wildly’, and Symm ἐταράχθησαν, ‘were confused’, showed their dissatisfac-
tion with LXX here.108 After this word TgG(G) has a plus which expands the
description of the peoples’ distress and attributes it (in a literal rendering of
Isa. 59.17bα) to God’s intervention.
‫( חיל‬15.14) LXX ὠδῖνες and Vulg take ‫ חיל‬in its more frequent sense
of ‘(birth-)pains’, but Tgg and Sy give the contextually more probable sense
‘trembling, fear’ (see Note xx on the translation).
‫( ישׁבי פלשׁת‬15.14) LXX, Vulg and TgJ use the gentilic ‘Philistines’,
despite its oddness after ‘inhabitants of’, apparently influenced woodenly by
the fact that ‫ פלשׁת‬often refers (or could refer) to the people. The addition of
108
TgJ is the only version to render the verb as a future (cf. its handling of
the verbs in v. 7), but its reading may be corrupt, as it uses the perfect for ‫יאחזמו‬
in v. 15.
15.1-21 389
‘the land of’ in TgJ,N,G(W),F(P) will have a similar origin. For ‫ ישׁבי‬TgJ has ‫כל עמודי‬
‫דיירי‬, for which AramB 2, p. 204, gives ‘All the pillars of the inhabitants of’:
TgJ has a similar equivalent for ‫ כל ישׁבי כנען‬in v. 15 (from which the ‫ כל‬here no
doubt derives), and cf. Gen. 46.28 (with n. 28 on it in AramB 1A) and 49.19.
While, however, ‫ עמודי‬might be an expression for ‘leaders’ here, one wonders
whether it is a (consistent) miswriting of ‫( עמורי‬cf. Sy’s ʿmwrʾ/ʿmwryh for
‫ ישׁבי‬in v. 15), which is a common word in Syriac for ‘inhabitant’. The verb ‫עמר‬
occurs in TgProv 2.21; 25.24; 30.28 according to some editions (cf. Jastrow,
p. 1090; CAL). Jastrow also recognises the possibility that ‫ עמד‬may have had
the sense ‘stay, dwell’ in Aram. (p. 1086; pres. as a Hebraism). Either way
‫ עמודי‬will represent part of a double rendering of ‫ ישׁבי‬here and in v. 15: ‫ דיירי‬is
the equivalent given in TgN,G(W),F(P).
‫( אז‬15.15) In general the Vss render ‫ אז‬in the same ways as in v. 1. Here
TgN certainly has ‫בכד(י)ן‬, ‘at that time’, as does TgG(W). TgG(G) joins the wit-
nesses to ‫בכ(י)ן‬, agreeing with TgO as it does throughout its rendering of MT
in this verse.
‫( נבהלו‬15.15) LXX ἔσπευσαν, ‘hastened’, follows a sense which ‫ בהל‬can
have in LBH and TgAram., but the context requires its other meaning, ‘were
agitated, disturbed’, which Vulg conturbati sunt (cf. ἐθορυβήθησαν in an
unnamed Gk. version) and probably the Tgg’s use of the Aram. cognate attest.
Sy’s dḥlw, ‘feared’, is guided too much by the context. Walters (pp. 144-48)
argues that ‘were frightened’ was the meaning intended by LXX.
‫( אדום‬15.15) TgJ,N,G(W),F(P) have the gentilic ‘of the Edomites’.109
‫( אילי מואב‬15.15) Here the same Tgg have the gentilic, as does LXX.
LXX and Sy prefix ‘and’, no doubt as a secondary addition. For ‫ אילי‬LXX
has ἄρχοντες, as it does again in Ezek. 31.11. Sy gbrʾ, probably in the sense
‘strong or mighty man’ (Payne Smith, p. 59), would belong with Tgg ‫תקיפי‬
and Vulg robustos here (cf. LXX at 2 Kgs 24.15 [in a καιγε section] and the
Three at Ezek. 31.11; Vulg at Ezek. 17.13; 31.11; 32.21; Tg at Ezek. 31.11;
32.21; Sy at Ezek. 31.11), which evidently related ‫ איל‬to ‫ ֵאל‬in the sense of
‘strength’ (BDB, p. 43, sect. 7; cf. the use of ἰσχυρός for ‫ ֵאל‬as a divine title
in some books of LXX and more widely in the Three). There seems to be no
trace in the Vss of the modern view that ‫ איל‬in the sense of ‘leader, ruler’ is
a metaphorical development from ‫‘ = איל‬ram’ (on which see Note zz on the
translation).
109
In 4QExc there may be an empty space (at the end of a line) after this
word: there are ‘dark spots’ in it, but DJD XII, p. 119, is doubtful if they are the
remains of writing which would form an addition to MT. TgG(G) has another of its
expansions at this point, attributing the disturbance to ‘the report of the dwellers of
Zoan’, who are less likely to be the Israelites (Klein 1, p. 250 n. 8) than Egyptian
informants.
390 EXODUS 1–18
‫( יאחזמו‬15.15) 4QExc again uses a final form of mem: see Text and
Versions on ‫ תבלעמו‬in v. 12.
‫( רעד‬15.15) TgG(G) adds ‘and they drank the cup of wrath and wormwood’
(cf. Jer. 25.15).
‫( נמגו‬15.15) The (metaphorical) sense ‘melted’ is reproduced in LXX and,
with the addition of ‘their hearts (within them)’, in TgJ,N,G(W); TgO,G(G) and Sy
‘were frightened’ (cf. CAL on ‫ )תבר‬will be an explanation of this (TgF(P) has
elements of both these renderings). Vulg obriguerunt, ‘became stiff, hard’, is
more puzzling, but since Vulg elsewhere renders ‫ מוג‬by prostrari, ‘be thrown
to the ground’ (e.g. Josh. 2.24, with timore), the sense intended may be ‘numb,
motionless’.
‫( כל ישׁבי כנען‬15.15) On TgJ see above on ‫ ישׁבי פלשׁת‬in v. 14. TgN,G(W),F(P)
again have ‘the land of (Canaan)’ here.
‫( תפל‬15.16) All the Vss render as a future/jussive rather than a preterite:
TgJ,G(W),F have a second person sing. Aphel form, making Yahweh the subject
(this reading also appears in some TgO sources, and consequently in TgG(G)).
At first sight MRI seems to continue its past tense interpretation, since it corre-
lates both this and the next verb with events in the Exodus-conquest narrative
(cf. the tr. in Lauterbach 2, pp. 74-75). But it does refer to prayers here and
it may cite the events because they were seen as the fulfilment of the text as
a prayer.
‫( אימתה‬15.16) SP has ‫ אימה‬and this standardisation of the unusual form
in MT (cf. Note ccc on the translation) was made at an early date, as it also
appears in 4QExc. Aq oddly gave it the sense ‘amazement’, probably by using
κατάπληξις (cf. 23.27). Some Tgg elaborate, either by reading ‘fear of death’
(TgJ,Nmg,F(V): cf. Ps. 55.5) or by adding ‘of you’, i.e. God (TgN,G(W)).
‫( ופחד‬15.16) TgN,G(W) again add ‘of you’, while TgG(G) has a distinctive
addition, ‘upon every people and tongue’ (for the expression cf. Isa. 66.18
and Tg there). ‫ תברא‬in TgNmg,F(V) is probably just an alternative, if more intense,
word for ‘fear’ (see the note on ‫ נמגו‬in v. 15). TgG(W) adds ‘O Lord’, as other
Tgg do in vv. 8 and 10.
‫( בגדל‬15.16) SP mss all (inc. Tal, Sadaqa, Crown and Camb. 1846) have
the plene form ‫בגדול‬, as does 4QExc, confirming MT’s vocalisation (on which
see Note ddd on the translation). Von Gall’s bracketing of the waw, implying
that it should be deleted (cf. p. lxix), exemplifies his tendency to prefer MT
readings even against massive SP evidence and should be ignored. The use
of nouns in the Vss to render this word is not evidence for a reading ‫ ְבּג ֶֹדל‬: as
modern EVV. show, a noun may just as well stand for MT.
‫( זרועך‬15.16) TgO,G(G),F(P) avoid the anthropomorphism by substituting
‫ ;בת(ו)קפך‬the other Tgg bring out the same meaning by additions.
‫( ידמו כאבן‬15.16) Again the Vss give the verb a future/jussive sense (cf.
above on ‫)תפל‬. LXX ἀπολιθωθήτωσαν, ‘let them become like stone’, appar-
ently derived ‫ ידמו‬from ‫דמה‬, ‘be like’, and Sy nṭbʿwn will be modelled (quite
inappropriately) on the stone-imagery in v. 5, where Sy uses the same verb.
15.1-21 391
The Tgg (inc. SamTg) recognised the verb as ‫ דמם‬in the sense ‘be silent’, with
‘be still, stupefied’ in addition in TgN and instead in TgF(V): Aq and Theod
taceant (for σιγάτωσαν?) and Symm immobiles fiant (cf. Vulg: for ἀκίνητοι
γενηθήτωσαν?) correct LXX accordingly. TgJ,N,G(W),F(V) and Sy hve the pl.
‘like stones’ (cf. v. 5 and the note).110
‫( יהוה‬15.16) Tgg (except TgF(V), which omits the whole of v. 16bα) add
‘(the streams of) the Arnon’ to specify which river was ‘crossed’. TgG(G) also
adds ‫בניסין‬, ‘with miracles’.
‫( עם‬15.16) LXX and TgJ,F(P) repeat the possessive suffix from the previous
clause, unnecessarily because of the words that follow.
‫( זו‬15.16) Here Vulg joins the other Vss in having a double rendering iste
quem (see the note on ‫ זו‬in v. 13). SP again reads ‫זה‬, displacing the less familiar
form. 4QExc has ֹ‫[ז]ו‬, but little survives.
‫( קנית‬15.16) LXX, Vulg and TgJ (cf. MRI [Lauterbach 2, pp. 75-76])
have the expected sense ‘you acquired, possessed’, but all the other Tgg
and Sy have ‘you redeemed’, which is clearly drawn from v. 13. AramB 7,
p. 43 n. 23, attributes this to influence from the Talmudic interpretation of the
passage as attesting the double redemption and entry into the land after the
Exodus and the Babylonian exile (B.Sanh. 98b). The fact that ‘redemption’
in the strict sense is a kind of ‘purchase’, which ‫ קנה‬can mean, may also have
played a part. Afterwards TgO,N,G(W) add ‘(the ford of) the Jordan’ and TgJ,F(P)
‘the ford of the Jabbok’: TgF(V), which has a conflated rendering of the end
of the verse ending ‘which you acquired for your name’, mentions both. The
Jordan, which is also mentioned in MRI (Lauterbach 2, pp. 75), fits the Exodus
story well, but the Jabbok in central Transjordan suits a journey from the north
into (future) Israelite territory better (cf. Deut. 3.16), like Jacob’s in Genesis
32. TgG(G) adds ‘with great (or many) marvels’ after ‘redeemed’ and nothing
more survives: but, given its basic similarity to TgO, it most likely ended with
‘the Jordan’.
‫( תבאמו ותטעמו‬15.17) 4QExc has the regular form of the suffix with both
verbs; compare its use of a final mem where it has the longer form in vv. 12
and 15. LXX and Vulg idiomatically represent the suffix with only one of the
verbs, and for the first LXX as often uses the Greek participle (cf. Aejmaleus,
On the Track, pp. 7-16; Lemmelijn, pp. 146-48). As in v. 16 all the Vss
render these verbs in the future tense. ‫ ותטעמו‬is rendered straightforwardly by
LXX, Vulg, TgJ,F(V) and Sy. TgO ‘you will cause them to dwell’ unpacks the
metaphor, while TgN,G(W),F(P) elaborate it with ‘you will give them an inherit-
ance’ on the basis of the following phrase.
‫( בהר נחלתך‬15.17) TgJ ‫ בית מוקדשׁך‬interprets the ‘inheritance’ to mean
the temple here (cf. MRI [Lauterbach 2, pp. 77-78]). TgN,F(V) add ‫ בית‬with the
same intention. TgG(W) omits the phrase altogether. Neither SP nor SamTg
110
Most SP mss read ‫( ידמאו‬or ‫)ידאמו‬, which is a merely orthographic variation
(cf. Num. 21.18 and GSH §16a).
392 EXODUS 1–18
contains any sectarian addition here, but M.Marqa 2.10 clearly brings out the
connection with Gerizim/Bethel in Samaritan understanding of the verse (on
Gerizim see further Crown, Companion, pp. 99-103, and the addition in SP
after Exod. 20.17).
‫( מכון לשׁבתך‬15.17) These words can be treated as in apposition to the
previous phrase, with a rel. pron. supplied before ‫( פעלת‬so LXX, Vulg,
TgG(W),F(P)), or they can be seen as the obj. of ‫( פעלת‬so TgO,J,N,F(P), Sy), making
a close parallel with ‫ מקדשׁ‬in the second half of the verse. The ambiguity of
the verb ‫ כון‬and a desire to play down the reality of a divine dwelling on earth
led to considerable variations in the translation of this phrase. LXX ἕτοιμον
related ‫( מכון‬viewed as a part.?) to the sense ‘prepare’ for ‫כון‬, which also
influenced the renderings of TgN,F(P) and one of the two which appear side by
side in TgJ. Alongside it TgJ has ‘a place which corresponds to (pass. part.
Pael of ‫ ’) ָכּוַ ן‬God’s heavenly dwelling (‘the throne of your glory’), a view of
the temple deduced in MRI (Lauterbach 2, p. 78) from Ps. 11.4 and 1 Kgs
8.13. Other Vss have no such qualms and render ‫ מכון‬as ‘a (strong: cf. ‫= כון‬
“establish” – TgO [cf. Vulg]) place’ (TgG(W),F(V),Nmg, Sy) and expand ‫ לשׁבתך‬to
‘a house for your dwelling’/(holy) Shekinah’ (TgO,N,Nmg,G(W),F and the other
version in TgJ) or represent it without any modification (LXX, Vulg, Sy). Aq
(and possibly Symm), with ἕδρασμα εἰς καθέδραν σου, seems more intent
on etymological correctness than theological niceties (cf. his use of ἑδράζω
for ‫ כוננו‬later in the verse).
‫( פעלת‬15.17) MT’s reading is supported by SP (in 4QExc and 4Q365 there
are lacunae) and the renderings of LXX, Vulg and Sy (cf. MRI [Lauterbach 2,
pp. 78-79]), but the Tgg use ‫ תקן‬Aph. (TgO,J,N,F(P),G(W)) or ‫( זמן‬TgF(V/N),Nmg), both
words that correspond to the meanings of ‫כון‬. Since they are so used by some
Tgg elsewhere in the verse, it is probably not necessary to assume a different
Vorlage here: the Tgg avoid the idea that God ‘made’ the temple and substitute
a more general word attributing it to his oversight.
‫( מקדשׁ‬15.17) The MT accents and the Vss indicate a break in sense
before this word, but the SP mss put the break after it, presumably construing
it as the obj. of ‫ פעלת‬and also implicitly of ‫כוננו‬. SamTg does not help. Most
Sy mss add the suffix ‘your’ to match those earlier in the verse, but 5b1 and
the other witnesses agree with MT’s reading, which must be original. TgJ
prefixes ‫בית‬.
‫( אדני‬15.17) SP, 4QExc and 4Qflor(174) have ‫ יהוה‬again and according to
BHS so do many Masoretic mss and at least one from the Cairo Genizah. The
Vss are no help in determining their Vorlage. Cross and Freedman read )‫יהו(ה‬
here, and the Qumran attestation of this reading gives it strong support. But it
is easy to see how an original ‫ אדני‬might have been changed to ‫ יהוה‬to match
its uniform use elsewhere in the poem (vv. 1[, 2], 3, 6, 11, 16, 18, as well as
the preceding phrase in v. 17), so certainty is impossible.
‫( כוננו‬15.17) LXX, TgO and Vulg render according to the standard
meanings of ‫ כון‬on which their treatments of ‫ מכון‬were based (see above:
similarly Aq); also Sy, although its tqnyhy bʾydyk turns the words into a prayer
15.1-21 393
(reading MT as ‫כונן‬, sing. imperative, plus ‫ו‬-, object suffix). The other Tgg use
here forms of ‫שׁכלל‬, which can mean ‘establish’ as well as ‘finish’.
‫( ידיך‬15.17) At first sight it seems that there is evidence for a sing. form
which might be intended to resist the bold anthropomorphism of MT. But SP’s
‫ ידך‬probably represents the pl. (cf. GSH §55bγ), and in 4QExc the kaph of
‫ ידך‬is said by DJD XII, p. 119, to have been overwritten on ‘two descending
strokes, perhaps yod-waw’, so that the scribe may have intended the yodh
to be retained (see Pl. XIX). In TgO,N,F ‫ ידך‬is of course a possible way of
writing the dual noun with the suffix (Stevenson, §13), as their verbal forms
confirm. TgJ,N,F.G(W) actually emphasise the dual sense by adding ‫תרתן‬, ‘two’,
no doubt reflecting the view that it showed how precious the temple was to
God (MRI [Lauterbach 2, p. 79]). The normal dual ending is found in 4Qflor
and reflected in LXX, Vulg, Sy and TgJ,G(W).
‫( ימלך יהוה‬15.18) The imperfect verb is rendered as a future in Aq, Symm,
Vulg and Sy (cf. MRI [Lauterbach 2, p. 80]), but by a present participle in
LXX (rendered qui regnas in some OL witnesses) and Theod and by nominal
periphrases in the Tgg, presumably to avoid any suggestion that God is not yet
the King (cf. Salvesen, Symmachus, p. 96). Other LXX translators had no such
qualms about using the future of βασιλεύω of God (Ezek. 20.33; Mic. 4.7; Ps.
145.10), even where the Heb. had the perfect (Isa. 24.23; 52.7) or the noun
‫( ֶמ ֶל‬Ps. 10.16); but the Exodus translator has already shown his preference
for the timeless present and participle in 3.14. Here as there the Tgg take the
opportunity to give a full doctrinal exposition of the text. The transliteration
of the divine name as ΠΙΠΙ in Aq, Symm and Theod is recorded here (cf. the
lists of other attestations in Hatch and Redpath, p. 1135): this, or rather the
original form in the palaeo-Heb. script, was apparently the general practice
in Jewish Greek biblical mss (cf. Jellicoe, Septuagint, pp. 131, 271-72: to the
evidence mentioned 8HevXIIgr should now be added). The Syhex presentation
of Symm has dylk after it, which Salvesen (ibid., noting possible parallels in
Isa. 51.22; 52.7) finds ‘unclear’ and Wevers, (Notes, p. 235 n. 25) views as
an error. Perhaps it is an indication that Symm at least expected ΠΙΠΙ to be
read as (ὁ) κύριος.
‫( לעלם ועד‬15.18) SP reads ‫( עולם ועד‬or ‫עולם ועוד‬: see below), an alterna-
tive expression that appears in the Pss (see Note ppp on the translation). It is
an old reading, attested in all the three Qumran witnesses which survive at
this point (4QExc, 4Q365 and 4Qflor) and it could be the Vorlage for LXX
(see below). TgO, Sy and Vulg probably presuppose the ‫ ל‬like MT, but the
paraphrases of the other Tgg (see below) are harder to penetrate. Since in
general ‫ לעלם‬is such a common expression, it is more likely to have arisen as
a secondary reading and ‫ עולם ועד‬is probably original here. LXX τὸν αἰῶνα
καὶ ἐπ’αἰῶνα καὶ ἔτι is a unique expression, requiring explanation in several
respects. τὸν αἰῶνα is less likely the obj. of βασιλεύων (Salvesen, ibid.) than
an accusative of duration ‘for ever’ (NETS), as it is several times elsewhere
in LXX (e.g. Ezek. 43.7, 9). It need not be a gloss on the unique ἐπ’αἰῶνα, as
such expanded renderings of ‫ (ל)עלם ועד‬are frequent in the Psalms (e.g. 9.6;
394 EXODUS 1–18
10.16), presumably for emphasis. ἔτι, like ἐπέκεινα in Mic. 4.5 and ultra in
Vulg here, will be due to the misreading of ‫ וָ ֶעד‬as ‫וָ עֹד‬. TgJ and the Pal. Tgg see
in the double expression here a reference to Yahweh’s kingship both in this
world and in the world to come (and in TgN before creation as well: cf. 3.14):
possibly TgO’s ‫ ועלמי עלמיא‬for ‫ ועד‬has the world to come in view too. The SP
variant ‫ ועוד‬noted above occurs in a significant number of early mss (9 out of
the 12 consulted for Schorch’s edition: cf. also von Gall’s apparatus) and is
presupposed in mss of SamTg and SamArab, which confirm that it was under-
stood in the sense ‘and beyond’. Probably the Samaritan tradition had both
this reading and ‫ ועד‬from its beginning and ‫ ועוד‬shows the impact of a two-age
eschatology at this point there too, as also in a related variation in Gen. 8.22
(S. Schorch, ‘In aeternum et ultra. Die Vorstellung eines Zeitenendes nach
Gen 8,22 und Ex 15,18’, in J. Kotjatko-Reeb et al. [eds.], Nichts Neues unter
der Sonne? Zeitvorstellungen im Alten Testament [FS E.-J. Waschke; BZAW
450; Berlin and New York, 2014], pp. 371-82). In addition TgJ and the Pal.
Tgg make this verse a separate utterance of the Israelites and TgJ (to which
TgF(V) is very close) and TgN,G(FF) recapitulate the events at the sea which gave
rise to it. TgF(P) has the ‘Four Nights’ midrash here instead of at 12.42 (see
Text and Versions there).
‫( כי‬15.19) All the Vss understood the sense to be causal rather than
temporal, but a minority reading in LXX is ὅτε, which Wevers describes as
‘much simpler’ than the unclear connection of ὅτι (Notes, p. 235).
‫( סוס‬15.19) The evidently collective sense was handled in different
ways. LXX’s ἵππος could be intended as the fem. form = ‘cavalry’, as the
def. art. makes clearer in 14.9, 23 (cf. v. 7). Vulg equus (the reading eques is
secondary) departs from its equitatus in 14.9, 23 (and OL here) and renders
the Heb. in its normal sense (as in 15.1, 21), no doubt intending a collective
sense for the sing., which Tgg and Sy bring out more clearly by using pl.
forms here.
‫( ברכבו‬15.19) Here all the Vss use the pl. Sy and TgG(W) ignore the ‫ ב‬and
LXX the suffix: an equivalent to the latter is added in the Three and the O-text.
‫( ובפרשׁיו‬15.19) Not only Sy and TgG(W) but LXX, Vulg and TgJ,F pass
over the ‫ ב‬here, and LXX again does not represent the suffix: the Three and
the O-text add αὐτοῦ. As for the previous word, there is no need to envisage
a Vorlage different from MT and SP. LXX ἀναβάταις continues to use the
equivalent which it introduced in 14.23.
‫( בים‬15.19) Sy bgw ymʾ assimilates to ‫ בתוך הים‬at the end of the verse.
‫( וישׁב‬15.19) LXX ἐπήγαγεν (cf. OL adduxit) is, as Wevers observes
(Notes, pp. 235-36), a free and rare match for ‫ שׁוב‬Hiphil: ἐπάγω is more often
used for ‫ בוא‬Hiphil. The other Vss (inc. Vulg reduxit) follow the Heb. attested
by MT and SP, which retains the verb used in the Qal in 14.26-28 and is no
doubt original.
‫( יהוה עלהם‬15.19) Most Sy mss, including 7a1, invert the word-order,
probably for stylistic reasons: 5b1 and 8b1 follow MT.
15.1-21 395
‫( את־מי הים‬15.19) SP and 4QExc agree with MT, but 4Q365 has replaced
‫ מי‬with ‫מימי‬, which is the form of the constr. st. elsewhere in Exod.
‫( ביבשׁה בתוך הים‬15.19) 4Q365 continued with the words which follow
this phrase at the end of 14.29 to complete the recall of the earlier narrative.
TgJ has a different addition: ‘and there sweet springs came up, and fruit trees
and green plants and choice fruit at the bottom of the sea’, partly anticipated
in MRI (Lauterbach 2, pp. 51-52, on 15.8). For the fruit AramB 2, p. 205 n. 38
compares Exod.R. 21.10.111
‫( ותקח‬15.20) Vulg sumpsit ergo makes the kind of causal connection
between vv. 19 and 20 which some have found to be implied by ‫ כי‬in the
former verse. But for Vulg enim (‫ )כי‬there clearly made a connection with the
preceding poem.
‫( את־התף‬15.20) The Vss all use words for ‫( תף‬also later in the verse)
which mean ‘drum’, not ‘tambourine’, according to the most recent authori-
ties: cf. Note uuu on the translation. TgN has the pl. here, probably a mistake
caused by the pl. form later in the verse: the other Pal.Tgg have the sing.
‫( ותצאן‬15.20) The (secondary) plene form appears in SP, 4QExc (probably)
and 4Q365 ([‫ו]תֹצינה‬: for the spelling cf. GK §74k and Qimron, pp. 22-23
(100.61).
‫( בתפים ובמחלת‬15.20) SP and 4QExc agree with MT (4Q365 has only the
first ‫ ב‬before a lacuna) and LXX and Vulg render as expected, perhaps also
TgO (but see below) and TgF(P),G(W). TgN,F(V) understandably paraphrase with
‘they were dancing with drums’. But Sy wbrbyʿʾ (pl.) took ‫ מחלה‬to mean
another musical instrument (‘a square or oblong tabor hung from the neck’
[Payne Smith, p. 526]), and a similar understanding of it is presupposed in
MRI (Lauterbach 2, pp. 82-83). This may also be intended by TgO’s ‫בחנגין‬:
although CAL gives only the meanings ‘(circle) dance, fair’, both Levy (1,
p. 265) and Jastrow (p. 458) cite occurrences of this rare root which seem to
involve musical instruments. Certainly this is how TgJ understood it in 32.19,
where it adds ‘which were in the hands of the wicked’, and its expansion of
the paraphrase in TgN,F(V) here by the addition of ‘and with ‫ חנגייא‬they were
‫( ’מחנגין‬AramB ‘and playing the hingas’) points the same way.
‫( ותען להם‬15.21) The meaning ‘sing’ is clearly represented in TgJ,F(P), Vulg
and Sy. In Sy the Aphel gives the special sense ‘lead in singing’, which is also
indicated in Vulg and in LXX ἐξῆρχεν (cf. MRI [Lauterbach 2, pp. 7-8, 83]).
TgF(P), Sy and Vulg make explicit that ‫ להם‬refers to the women in the previous
verse (cf. Note yyy on the translation), and LXX αὐτῶν probably does too.
TgO ‫( ומעניא‬cf. TgNmg) and TgN ‫ וענת‬use forms of Aram. ‫עני‬, ‘answer, respond’
(cf. the renderings in AramB), which Jastrow (p. 1093) and CAL agree can
mean ‘sing a response, refrain’ in the Pael. Whether such a special usage is
involved here or not, the use of ‫ עני‬suggests that ‫ להם‬is being taken to refer
111
Two Sy mss (7h13, 8b1) have ‘through the sea as on dry land’, harmonising
with the Peshiṭta’s tr. of 14.29.
396 EXODUS 1–18
not to the women but to those who sing vv. 1b-18 (TgO,N both use the masc.
form ‫ )להון‬and that Miriam alone sings (?) a response to them. Aq and Symm
κατέλεγεν (followed in Aq at least by αὐταῖς) is a word which stands for ‫ענה‬
I in its special legal sense in LXX at Deut. 19.16: it can mean ‘recount, repeat,
recite’ (but not ‘answer’) and it is also used by Aq and/or Symm for ‫ ענה‬IV
(= ‘sing’) in Num. 21.17, 1 Sam. 21.12 and Ps. 147.7, as well as in Jer. 25.30;
51.14 where ‫ ענה‬IV may have a related but different meaning (see Note xxx
on the translation). These revisers of LXX were evidently seeking a consistent
equivalent that fitted the passages where ‘answer’ was not appropriate, but
they seem to have either not known or ignored the possibility that ‘sing’ was
meant. In all three approaches to the sense here there was a strong preference
for presenting the action as continuous: only TgJ,N have the perfect tense that
most closely corresponds to MT.
‫( מרים‬15.21) LXX added λέγουσα, providing an introduction to the
direct speech (followed by Vulg dicens after OL) as it does occasionally
elsewhere (e.g. 3.12) in imitation of Heb. ‫( לאמר‬see THGE, p. 255).
‫( שׁירו‬15.21) Here too, as with ‫ אשׁירה‬in v. 1 (see the note), the Vss gener-
ally substitute a first pl. form for the Heb. (in the Tgg, as in v. 1, with two
verbs). MT finds support in SP, 4QExc, TgO (except for its extra verb), Sy (cf.
Brockelmann, §173) and the mg of LXXF (ᾄσατε, perhaps from Aq).
‫( כי גאה גאה סוס ורכבו רמה בים‬15.21) LXX, Vulg, TgO and Sy reproduce
their renderings of the same words in v. 1 and SP has the same variant ‫גוי‬
for the first ‫ גאה‬as there: on the details see the notes on v. 1. The other Tgg
are broadly but not exactly similar to what they have there. TgF(P) is the only
surviving text to have all three of the expansionary elements (see the footnote
on ‫ גאה גאה‬in Text and Versions on v. 1); TgJ,N omit the general statement
about punishment here, but they have (different) introductory formulae of
praise (cf. Sy, and TgJ in v. 1) which TgF(P) lacks. Minor variations between
these Tgg reflect a complex pattern of exegetical development which cannot
be explored here.
A different and perhaps older expansion of the Song of Miriam appears,
sadly in a fragmentary form, in 4Q365 fr. 6a ii + 6c, 1-7 (cf. DJD XIII, pp.
269-71). Only the right-hand portion of these lines at the top of a column
is preserved. The text of 15.22-26 lower down on the fr. suggests a ‘letter
space average’ of 63 per line (DJD XIII, p. 270), so a substantial amount of
text has been lost between the sections that survive. Probably not much is
missing from the beginning of the expansion, as the end of v. 20 is likely to
have been close to the bottom of the previous column (47 lines reconstructed
including it [DJD XIII, pp. 256, 269], which seems an unusually high figure:
see the comparative data in Tov, Scribal Practices, pp. 84-99).112 The state of
112
Cf. A. Feldman, ‘The Song of Miriam (4Q365 6a ii + 6c 1-7 Revisited’,
JBL 132 (2013), pp. 905-11 (907, where his partial reconstruction and translation
of the text can be found).
15.1-21 397
the fragments makes the reconstruction of connected sense-units impossible.

Second sing. masc. forms appear in ll. 1, 3 and 6, suggesting direct address
to God as in vv. 6-17: the second pl. fem. form ‫ רוממנה‬in l. 6 perhaps intro-
duces words which the women of v. 20 are to say in response to Miriam’s
own praise. The language used parallels that of the Song of Moses at several
points (‫ גאות‬in ll. 2, 7; ‫גדול‬
֯ and ‫ ]מושיע =[ ֯מו֯ שיא‬in l. 3, ‫ במים אדירים‬in l. 5, ‫מרומם‬
in l. 6, ‫ ]עו[שה‬in l. 7), so the supplementer may have been seeking to create a
fuller song for Miriam which echoed the canonical poem. The nominal use
of ‫ )שונא =( שונה‬for ‘enemy’ is found in BH (BDB, p. 971), in Ben Sira and at
Qumran (DCH 8, pp. 169-70), and several other words appear in the Qumran
hymns (‫בזה‬, ‫גאות‬, ‫גדול‬, ‫)מים אדירים‬, while ‫ פדות‬is common in the War Scroll.113
This is ‘the largest preserved expansion of 4QRPc’ (DJD XIII, p. 270), but the
preserved text provides no evidence for a direct connection with the expan-
sionary material in the Tgg: the few parallels of language that there are can be
explained in other ways.114 No other ancient text appears to preserve a similar
composition or a reference to one.115
113
For these and other comparisons of the vocabulary with texts from Qumran
see H. Tervanotko, ‘ “The Hope of the Enemy has Perished”: The Figure of Miriam
in the Qumran Library’, in A. Lange et al., From Qumran to Aleppo (FRLANT
230; Göttingen, 2009), pp. 156-75 (166-67), and more fully and with wider refer-
ence to other ancient literature in Denying her Voice: The Figure of Miriam in
Ancient Jewish Literature (JAJSup 23; Göttingen, 2016), esp. pp. 147-61.
114
Tervanotko, art. cit., pp. 170-73, thinks a connection is possible, in the light
of her proposal to read ‫ נש[בח‬at the end of l. 4: this verb is known in Heb. in the
sense ‘praise’, and also occurs in the Targumic expansions of the Song of Miriam.
115
There is a broad similarity of purpose to the poems in LXXDan. 3.26-45,
52-90, as was pointed out to me by Dr N.A. Wormell, who also drew my atten-
tion to Franz Schubert’s cantata The Song of Miriam, a setting of Grillparzer’s
Miriams Siegesgesang. Further comparisons are made by G.J. Brooke, ‘Power to
the Powerless: A Long-Lost Song of Miriam’, BAR 20/3 (1994), pp. 62-65, who
renders l. 6 ‘and he exalted her to the heights’: Judith 16; 1QM 11 and 14; Luke
1.46-55 (the Magnificat); more generally Philo, De Vita Cont. 83-88. See also
especially the works of Tervanotko and Feldman referred to above for details of
other references to Miriam in Second Temple literature (including Qumran) and
some alternative readings of the 4Q365 text.
THE JOU R N E Y TO T H E M O U NTAIN
OF GOD
(15.22–18.27)
C h ap t er 1 5 . 2 2 - 2 7
Sw e e t Wate r an d L aws at M ar ah a nd Elim
This section is clearly marked off from the preceding verses by the
conclusion of the episode at the sea and the introduction of themes
that belong to the wilderness journey. For that reason 15.22 is the
beginning not only of a new pericope but of a fresh major section of
the Exodus narrative, and the break is reflected in both the Jewish
and the Samaritan manuscript traditions (MT; 4Q365; SP).1
This new main section extends to the end of ch. 18 and further
‘wilderness narratives’ appear in Num. 10.11–21.20 (see section 3
[ii] of the Introduction to the Commentary, where their distinctive
characteristics are outlined). Here it is possible only to note some of
the main topics of modern scholarly study of these chapters. At one
time interest in the route of the Israelites’ journey was prevalent,
after the opening up of the Sinai peninsula and neighbouring regions
to modern exploration and mapping (see the classic work of E.H.
Palmer, The Desert of the Exodus, for example). This subsequently
generated a literary and comparative study of the texts describing
the purported route, initiated by Noth’s article ‘Der Wallfahrtsweg
zum Sinai’, PJ 36 (1940), pp. 5-28 (see further the Excursus on
the wilderness itinerary in the introduction to 12.28-42, 50-51).
But well before this studies of the origins of the wilderness tradi-
tion had suggested that the idea of a journey by stages through the
desert was artificial. Instead, it was proposed, the individual stories
reflected a long period which Israel’s ancestors had spent at Kadesh
1
No other Qumran evidence is available at this point.
15.22-27 399
(cf. Deut. 1.46), during which major developments in their religious

life had taken place.2 Noth, who had identified ‘the leading through
the wilderness’ as an old element of the Pentateuchal tradition, was
sceptical about this hypothesis, allowing only that Exod. 17.1-7 had
its roots at Kadesh, and it no longer enjoys much support.3 More
attention is given to the recurring motif of ‘murmuring’ or, as Coats
prefers, ‘rebellion’ by the people against Moses, of which examples
have already appeared earlier in the narrative (5.20–6.1; 14.11-
14).4 Important as this motif is, it should not be overlooked that it is
interwoven with the more positive theme of Yahweh’s provision for
his people’s needs of sustenance and protection on their journey.5
Finally, on a different front, the past sixty years have also seen fresh
advances in the study of the geography and archaeology of the Sinai
Peninsula, which contribute in various ways to the understanding of
the wilderness narratives.6
2
See already Wellhausen, Prolegomena4, pp. 348-49, ET, pp. 342-43, with
further developments of the theory in Meyer, Israeliten, pp. 60-82, and especially
Gressmann, Mose, pp. 386-92, 419-24, 431-48; Anfänge, pp. 77-116. A brief
summary is given by Schmidt, Exodus, Sinai und Mose, pp. 106-109.
3
Noth, pp. 111-12, ET, 140; more fully in ÜGP, pp. 62-63, 180-82. Schmidt,
op. cit., p. 109, is similarly unconvinced: see further the criticisms of H.F. Fuhs,
‘Qades – Materialen zu den Wüstentraditionen Israels’, BN 9 (1979), pp. 54-70,
and L.E. Axelsson, The Lord Rose Up from Seir: Studies in the History and Tradi-
tions of the Negev and Southern Judah (CBOT 25; Stockholm, 1987), pp. 113-18.
Blenkinsopp, on the other hand, gives Kadesh a prominent place in his history of
tradition (pp. 137-38, 162-63, 179 n. 6).
4
Cf. Coats, Rebellion, passim, with further contributions from Childs,
pp. 256-64, Van Seters, Life, pp. 165-75, and Frankel, The Murmuring Stories.
5
On this, and its greater emphasis elsewhere, see C. Barth, ‘Zur Bedeutung
der Wüstentradition’, in Volume du Congrès, Genève 1965 (VTSup 15; Leiden,
1966), pp. 14-23. Another pertinent qualification is that in the non-Priestly narra-
tive the murmuring motif appears only in the sections which we will conclude (on
other grounds) come from the ‘J’ strand: it is not present in 15.25b; 16.4-5, 21,
26-31; 17.8-16, 18.1-27.
6
E.g. B. Rothenberg et al., God’s Wilderness: Discoveries in Sinai (London,
1961); T.L. Thompson, The Settlement of Sinai and the Negev in the Bronze
Age (Beiheft B8 to TAVO; Wiesbaden, 1975); Lipschitz, Sinai; Z. Meshel and
I. Finkelstein (eds.), Sinai in Antiquity: Researches in the History and Archaeology
of the Peninsula (Heb.; Tel Aviv, 1980); Meshel, Sinai: Excavations and Studies
(BARIS 876; Oxford, 2000), esp. ‘An Explanation of the Israelites’ Wanderings in
the Wilderness’, pp. 152-61.
400 EXODUS 1–18
The end of the present section has traditionally been placed

after 15.26 (see Text and Versions there), without any break
between 15.27 and 16.1. This accords with the text’s geographical
connections: 15.27 is a notice of the arrival at Elim which has no
explicit connection with the preceding narrative, whereas Elim is
mentioned again (twice) in 16.1. It was only the introduction of
the Latin chapter-divisions in the medieval period that placed a
division after 15.27, probably because both it and the Marah-story
shared a concern with the provision of water, which is absent from
ch. 16.7
The standard Masoretic and Samaritan textual traditions have
no divisions within vv. 22-26, but the enlarged waw after v. 25 in
4QpalExm indicates a break there, apparently absent from 4Q365,
which might preserve a memory that v. 26 was a later addition
to the story (see below). 4Q365 seems to have had a vacat after
v. 22 (see Text and Versions), i.e. before the Marah-story proper
begins. Verse 22 is a transitional narrative of the beginning of
the wilderness journey, with three verbs of movement and a final
clause indicating a general absence of water. The arrival at Marah,
with its undrinkable water (v. 23), leads to the people’s complaint
(v. 24) and, through the intercession of Moses, a God-given
solution to the problem (v. 25a). The reader expects that the journey
will continue, as it does with the movement to Elim in v. 27, but
first comes a puzzling statement about legislation and a test for
the people ‘there’, i.e. apparently at Marah (v. 25b). The subject
of the two verbs is not explicit but is probably Yahweh (see the
Explanatory Note). The reference to legislation is then picked up
in an explanatory speech by Yahweh with a quite general scope
and with no expressed addressee, though the second sing. verbs
and pronouns presumably refer to the people (like the third person
sing. pronouns in v. 25b). Thus the framework of the section is
provided by a journey narrative (vv. 22-23aα, 27), into which have
been set (a) a short narrative of divine help (vv. 23aβb-25a) and
(b) a brief account of divine instruction (vv. 25b-26). The passage’s
7
The Latin division might be due to a homiletic association of Marah with
Elim through Christian symbolic interpretation in terms of law and gospel (cf.
Jerome, Ep. 69.6; Maximus of Turin, Serm. 67-68: ACCS 3, p. 84). By contrast
Philo, VM 1.181-91, and Josephus, AJ 3.9-32, seem to have associated 15.27 with
ch. 16.
15.22-27 401
diversity of form, as well as other features, has led to a variety of

views among modern scholars about the process of its composi-
tion, especially with regard to (b).
Wellhausen (Composition, pp. 77-79) attributed the whole section to JE,

rejecting the older view (e.g. of Knobel, pp. 157-58; cf. Num.-Josh., p. 532)
that the itinerary-sections in vv. 22-23 and 27 were from P. Dillmann agreed
for vv. 22-23 but not for v. 27 (pp. 161-64). On this issue Wellhausen’s view
was followed by most later commentators until Noth and it continues to be
influential (cf. Childs [p. 266], Levin [pp. 348-51], Van Seters [Life, pp. 155,
175-76], Dozeman [pp. 359-60] and Albertz [pp. 264-65]). Noth (p. 101, ET,
pp. 127-28; ÜGP, p. 18) limited the itinerary-frame to v. 22aα (thus only a
note of departure) and 27 (a note of arrival) and attributed this to P because of
the ‘stereotyped phrases’ which ‘continue to occur…within the framework of
the narratives in P’: he presumably had 16.1, 17.1 (although it is not attached
to a P narrative) and 19.1-2a in mind.8 Some subsequent scholars have agreed
with Noth (Coats, Rebellion, p. 47 [including also v. 22aβ]; Hyatt, pp. 171,
173; Blenkinsopp, Pentateuch, p. 170), and some others have taken the same
view only about v. 27, because of its formal link to 16.1 (Fritz, Israel, p. 7
and n. 4; Graupner, Elohist, p. 90 n. 290: cf. Gressmann, Anfänge, pp. 77, 82).
A more widespread and lasting departure from Wellhausen’s unified
view of the passage concerns vv. 25b and 26. Verse 25b, which Wellhausen
had described as ‘a poetic fragment’, was attributed to E by some for a
time (Carpenter/Harford-Battersby, Holzinger, Gressmann, McNeile), but
Jülicher’s early assignment of it to a Deuteronomistic redactor (‘Die Quellen’,
pp. 275-76) gradually became the dominant view. The main alternative has
been that it is not in fact to be separated from the main Marah narrative (so
Smend, Eissfeldt, Rudolph [in part], Childs, Baden [p. 139]), a view that is of
course easier to maintain for those who regard the J source as a whole as exilic
in origin (Van Seters, Life, pp. 177-78). For v. 26 a Deuteronomistic origin
has been generally accepted, though its final clause (and sometimes more) is
often attributed to the main narrative (Carpenter/Harford-Battersby, Smend,
Gressmann, Eissfeldt, Rudolph, Childs, Baden [ibid.]).9 The discussion of
these verses in older commentaries was complicated by the then popular view
that ‘he tested them’ in v. 25 was a relic of a tradition about Massah, hence
Kadesh (on which see above).
8
Noth appears to forget that he had been content to attribute 12.37 and 13.20
to J.
9
Lohfink (‘Ich bin Jahwe, dein Arzt’, p. 41, ET, p. 63) concludes that the style
of the verse (and v. 25b) has Priestly features too, so that it belongs to the ‘Penta-
teuch redactor’ or later, but this view has found little support.
402 EXODUS 1–18
Apart from some early advocates of E (Knobel [i.e. his Rechtsbuch, cf.
Num.-Josh., p. 532]; Dillmann, Baentsch), most source-critics assigned the
main Marah narrative to J, with more or less conviction (see Baentsch, p. 141,
for the difficulty in resolving the issue on the evidence of vocabulary). Fritz
contents himself with the observation that the passage attaches itself well to
the J strand in chs. 13–14 (p. 8), but when he comes to discuss the geograph-
ical terminology he recognises that ‘Yam Suf’ does not occur there, so that
this location of the crossing of the sea comes ‘after the event’ (nachträglich
[p. 38]; cf. Levin, p. 348), which is surely surprising. Smend (pp. 145-46) and
those who followed him attached the passage to J1/L/N because they thought
that it would duplicate the narrative in 17.1-7 if it occurred in J(2) or E, but this
was never a strong argument and was weakened further when doubts arose
about that passage’s division between two parallel accounts (see the intro-
duction to 17.1-7). Newer kinds of analysis of the Pentateuch have generally
attributed the nucleus of the section to its oldest narrative layer, whatever they
call it, but Lohfink’s study left open the possibility that it was only inserted by
the Pentateuch redactor or a later supplementer (‘Ich bin Jahwe, dein Arzt’,
p. 31, ET, pp. 53-54). Propp, in what is admittedly a very tentative exploration
of the issues (pp. 574-76: see the further ‘Speculation’ in vol. 2, pp. 749-50),
seems to be the only scholar who has been attracted by this possibility.
The account of the ‘sweetening’ of the water of Marah (vv. 23-

25a), with its introduction in v. 22, should probably be regarded
as a literary unity. There are no awkward connections or signs
of parallel narratives in these verses, and arguments based on an
‘itinerary style’ or Priestly characteristics overlook the differ-
ences from real itineraries (and extracts from them) and the regular
features of Priestly narrative. Those who seek to separate off an
‘itinerary introduction’ have not been able to agree where it ends.
Each word of v. 22aα (‘Then Moses…Yam Suf’), which features
in all such proposals, is distinct from Priestly and/or itinerary style:
the verb for ‘departure’ is in the causative form with Moses as the
subject; ‘Israel’ is an expression of the older source(s); and ‘Yam
Suf’ is not used to describe the sea that was crossed in P or in the
itinerary in Numbers 33 (see further below on its relevance for the
attachment of the section to the earlier narrative). There is a general
similarity in vv. 22-23aα to what is found in Num. 33.8, but even
in the expression that is most similar, ‘they went…for three days’,
the Numbers parallel has the additional word derek, ‘a journey of’,
and the name given to the wilderness is ‘Etham’ and not ‘Shur’.
There is insufficient agreement to establish a literary dependence in
15.22-27 403
either direction. Already the end of v. 22 and the beginning of v. 23

are shaped to prepare for the divine intervention that will make the
water of Marah drinkable: there is no interest in an encampment at
Marah as such (though it is not excluded by the word ‘came’). Apart
from its lack of Priestly characteristics, the passage is difficult to
assign to a particular narrative strand. The use of the divine name
Yahweh need not exclude an attribution to E after the revelation
and explanation of the name in 3.13-15 (contrary to Graupner’s
argument in Elohist, pp. 89-90). On the contrary, the use of ‘Yam
Suf’ as the first goal of the departing Israelites in 13.18 might seem
to support such a view (so also Graupner, p. 90). ‘Yam Suf’ also
occurs, in a non-Priestly context, in 10.19, which is generally attrib-
uted to J, but we have argued (see the introduction to 10.1-20) that
the verse belongs to E as part of the main plague-narrative that has
been preserved in Exodus. Nevertheless even this indicator is not
decisive. No mention of the Israelites’ arrival at ‘the sea’ that can be
attributed to J seems to have been preserved – ‘the sea’ first appears
in J (if we are right to assign 14.21aβ to E) in 14.27b – and that is
where a specific name for it would most likely have been given,
perhaps ‘Yam Suf’.10 For the moment we must be content with the
description of the narrative as ‘non-Priestly’, but the following
discussion may provide some indirect evidence to resolve the issue.
At first sight (and this is no doubt intentional) the ‘There’ at
the beginning of v. 25b makes a strong connection with what has
preceded, as if v. 25b (and v. 26) formed the conclusion to the
Marah narrative. But what it says bears no obvious relation to the
main story and its very vagueness, both grammatically and in its
context, only adds to its strangeness (see the Explanatory Note). No
similar conclusion appears at the end of any of the other wilderness
narratives in Exodus and Numbers. Even if it is possible to imagine
a reason why the statement about legislation and testing was placed
here (see below), it is virtually certain that this was not originally
the conclusion of the Marah story. As is suggested in the Explana-
tory Note, it seems most probable that ‘there’ was first intended
to mean the wilderness as a whole, including ‘the mountain of
10
Some other vocabulary items might be cited in favour of a J origin for the
narrative, but in most cases they occur in passages attributed to E as well: see
Baentsch, p. 141.
404 EXODUS 1–18
God’, and that it stood in a different context at the beginning of

an account of the wilderness period of Israel’s early history, to
pick out what the author of that account saw as the most important
themes of the wilderness narrative. Possibly that is its purpose in its
present position, but the natural inference that ‘There’ now means
‘at Marah’ implies rather that the redactor who placed it here had in
mind a specific act of legislation and testing at the beginning of the
wilderness journey, as soon as the Israelites began their ‘march of
freedom’. His intention may have had something in common with
the institution of a permanent ritual remembrance that was intro-
duced in chs. 12–13. There is a clue to where the redactor found this
summary of the wilderness journey in the reference to God ‘testing’
Israel here, as in the ‘test’ of Abraham’s ‘fear of God’ in Genesis 22
(cf. v. 1), which is picked up again in Exod. 20.20 – both passages
being normally attributed to E in source-critical analyses (the ‘fear
of God’ is another well-known characteristic of E’s theology: cf.
14.31a). It is true that Yahweh’s testing of Israel in the wilderness
is also the focus of Deut. 8 (cf. vv. 2 and 16: later 13.4) and that in
v. 26 which follows the language is strongly Deuteronomistic (see
below). But the use of the nouns ‘statute’ and ‘ordinance’ in the
singular here, albeit probably in a collective sense (see Note l on
the translation), is most uncharacteristic of Deuteronomic style and
belongs rather to the sources which the Deuteronomistic Historian
used (Josh. 24.25; 1 Sam. 30.25) and to the non-Priestly narrative in
Genesis (47.26). The idea that anyone, whether God or Moses, gave
the people laws in the wilderness is also quite alien to Deutero-
nomic thinking. Verse 25b is then best regarded, as several early
critics saw, as an extract from (the beginning of) E’s wilderness
narrative and the story to which it was then rather abruptly attached
will presumably be from J, as many have thought (see above). Van
Seters, who takes this view, went on to argue that ‘the motif of
murmuring and intercession in vs. 23-25a’ supports his argument
that J is of exilic origin (Life, pp. 177-78, cf. 167-75). Moses’ role
as an intercessor in Exodus is, however, expressed in quite different
language from that found in Deuteronomistic literature and the
prophetic intercession of Amos (7.2, 5) is a pre-exilic parallel that
is not so easily discounted as Van Seters supposes.
As has just been said (and as is agreed by nearly all scholars), the
literary affinities of most of v. 26 are with the Deuteronomic tradi-
tion (for details see the Explanatory Note). It provides, in classic
15.22-27 405
Deuteronomic terms, a homiletic elaboration of the reference to

legislation in v. 25b by indicating the rewards for obedience to the
law (cf. Deut. 28.1-14). The explicitly positive emphasis of this
verse corresponds closely to the longer exhortation that concludes
the Book of the Covenant (23.20-33), where both the frequent
reinforcement of the opening of the conditional clause (‘If you will
indeed obey…’: Heb. infinitive absolute) and the very rare word
used for diseases here (cf. Note m on the translation) recur (23.22,
25). But it has often been observed that the final clause of v. 26,
with its reference to Yahweh as the ‘healer’ from disease, does not
share the general Deuteronomic character of the language of the
verse. The verb ‘to heal’ (Heb. rāpāʾ) is also used in a metaphorical
sense of making unpleasant water drinkable in a story about Elisha
located at Jericho in 2 Kgs 2.19-22 (cf. vv. 21-22: NRSV ‘[made]
wholesome’), that is in a narrative taken up by the Deuteronomistic
Historian. There is therefore in the use of this verb at least a loose
connection with the Marah story and presumably the redactor either
introduced it here to correlate his parenesis more closely with the
narrative context or found it as the original conclusion of the Marah
story and modified it into a conditional promise to take account
of the reference to legislation which he or an earlier redactor had
included by the addition of v. 25b. It is possible that the other diver-
gences from Deuteronomic style in v. 26 which have been carefully
identified by Lohfink (‘Ich bin Jahwe’, pp. 33-39 = ET pp. 55-61)
are not, as he thinks, due to Priestly influence but to v. 26 being from
an early stage of the Deuteronomic tradition, to which the redactor
who combined J and E (and the author of 23.20-33 if they are not
the same) has sometimes been thought to belong (cf. A. Reichert,
‘Der Jehowist und die sogenannten deuteronomistischen Erweiter-
ungen im Buch Exodus’ [unpublished Ph.D. dissertation, Tübingen,
1972], pp. 89-95, esp. 90-91).
Verse 27 is linked by its reference to Elim to 16.1 (and through
it to 17.1 and 19.2a). It is a separate itinerary-note, which in fact
corresponds only to the second (arrival) element of the standard
form, and it has no connection beyond the references to water with
the Marah-story. Its wording is almost identical to Num. 33.9,
where it also follows Marah in the list of stopping-places (see the
Explanatory Note for minor variations between the two verses). The
widespread attribution of the verse and those associated with it to
J (or the oldest layer of narrative in some recent writing) has no
406 EXODUS 1–18
real justification. A Priestly origin has more in its favour because

of the connection to 16.1 (and 17.1), where some features of
Priestly language appear in the present text, but it will be argued
in the commentary there that these features are secondary additions
deriving from the dominant Priestly version of the manna-story in
ch. 16. Like 12.37 and 13.20 all these verses (and 19.2a) most likely
derive from an ‘itinerary-redaction’ which has a Deuteronomistic
background (see the Excursus on the Wilderness Itinerary in the
introduction to 12.28-42, 50-51 and especially my ‘The Wilderness
Itineraries and the Composition of the Pentateuch’).
With this section of the narrative the landscape changes and a
new group of themes is briefly introduced. The first stage of Israel’s
onward journey quickly brings them face to face with the hardships
of travel through the desert which lies between Egypt and the
promised land of Canaan. The first is the shortage of water, or at
least drinkable water, and it leads the people to complain to their
leader Moses, who had secured their release from Egypt (14.31).
True to his role as Israel’s mediator with Yahweh, he intercedes for
them and through Yahweh’s help the problem is solved (just as a
later prophet will do in 2 Kgs 2.19-22). A new attribute of Yahweh
as the ‘Healer’ of his people is made known (v. 26). This short
‘story of deliverance’ has been elaborated to introduce here, right
at the beginning of the journey, two further ways in which Yahweh
will deal with his people in the future: by providing guidelines for
their behaviour which must be followed if they are to continue to
know his protection, and by testing them to ensure that they recog-
nise his authority over them (cf. 16.4; 20.20). But the journey is not
all hardship, as appears from another addition (v. 27) which inserts
a halt at an oasis before further travel through the desert begins in
16.1.
22 Then Moses made Israel departa from the Yam Suf and they
went outb into the wilderness of Shur. They went through the
wilderness for three days and found no water. 23 Then they
came to Marah, but they could not drink the water from Marahc,
because it was bitter – that is why its namee was calledd Marah.
24 So the people complainedf to Moses, saying,g ‘What can we
drink?’h 25 He cried out to Yahweh and Yahweh showed himi a
piece of woodj. He threw it into the water and the water became
sweet. [There khe made statutes and ordinancesl for them, and
15.22-27 407
there he tested themk]: 26 [he said, If you will indeed obey the
voice of Yahweh your God and do what is right in his sight
and hearken to his commandments and keep all his statutes, I
will not inflict on you all the diseasesm which I inflicted on the
Egyptians, for I am Yahweh your healer.] 27 [Then they came
to Elim, wheren there were twelve springs of water and seventy
palm-treeso, and they camped there by the water.]

a. Heb. ‫ויסע‬, Hiph. of ‫נסע‬, for which see earlier 12.37 and 13.20, in
itinerary-notes of a more regular kind, and the notes there. In the absence of
any further specification the sense ‘depart’ is implied by the following ‫מן‬. The
Hiph. is used with a divine subject in a similar sense in Pss. 78.52; 80.9 (and
in Ps. 78.26 with a wind as the object [cf. Num. 11.31]), but this is the only
case in BH of its use with a human ‘leader’: for other verbs which express this
idea in relation to the Exodus see the Explanatory Note.
b. Heb. ‫ויצאו‬. On the textual variants see Text and Versions. ‫ יצא‬is more
often followed by ‫ מן‬than ‫( אל‬and often so in the Exodus traditions), but occur-
rences with ‫( אל‬or a word with he locale) are by no means infrequent, both of
persons and of places (for the latter cf. 16.3).
c. Heb. ‫ממרה‬. The sense ‘because of bitterness’ is theoretically possible (cf.
Prov. 14.10) but unlikely, as the reason is already expressed in a different way
in the following words. Although ‘the water of’ is more commonly expressed
by the constr. state of ‫מים‬, ‫ מן‬is occasionally used (like Eng. ‘from’) even
where there is no preceding verb that requires it (7.18, 21; 2 Sam. 23.15 par.
1 Chr. 11.17).
d. Heb. ‫קרא‬, an indefinite use of the third person m.s., which is equivalent
to a passive as in Gen. 11.9 etc. (cf. LXX ἐπωνομάσθη) here. GK §144d-e
and JM §155d-e follow the medieval view that in such cases a cognate parti�-
ciple (which appears e.g. in Deut. 17.6) is understood (cf. Driver, Samuel2,
p. 132, on 1 Sam. 16.4).
e. Heb. ‫שׁמה‬. The omission of the object marker ‫ את‬here is especially
common in this formula (cf. 2.10; 17.7, 15), but by no means universal (e.g.
2.22).
f. Heb. ‫וילנו‬. For the plural verb with a collective sing. subject see GK
§145b. The verb ‫( לון‬which is to be distinguished from ‫לין‬, ‘lodge, pass the
night’) occurs only in the Hexateuch (Exod. 15.24; 16.2, 7-8; 17.3; Num. 14.2,
27, 29, 36; 16.11; 17.6, 20; Josh. 9.18) and, probably, in Ps. 59.6. In Exodus
and Numbers it appears chiefly in Priestly passages (only Exod. 15.24; 17.3
are normally regarded as non-Priestly): cf. the related noun ‫( תלנה‬Exod. 16.7,
8 [2x], 9, 12; Num. 14.27; 17.20, 25). Both Niphal (as here) and Hiphil forms
are attested in MT, sometimes with variations within a passage and between
408 EXODUS 1–18
Kethibh and Qere, but without any apparent difference in meaning.11 Such
synonymity between the Niphal and the Hiphil is unusual (more often [e.g.
‫ ]סתר‬the Niphal represents the passive or reflexive of the Hiphil): the Niphal
is found with other verbs representing sound (e.g. ‫אנח‬, ‫)אנק‬, and possibly
the Hiphil was meant to introduce a note of wilfulness (cf. its use with ‫זעק‬
and ‫)רוע‬. The common translation ‘murmur’ (Tyndale, AV; and see Text and
Versions) has been criticised by G.W. Coats, Rebellion, pp. 21-28, on the
ground that the frequent connection with ‫ על‬implies stronger opposition or
‘rebellion’ (cf. TWAT 4, 528-30 = TDOT 7, p. 510-12; contra Childs, pp. 266,
268), but there are other words for this. The milder ‘complain’ would fit the
contexts just as well (so REB, NRSV). If the form ‫ וילינו‬in Ps. 59.16 is derived
from this verb (with some versional support), the meaning there is probably
‘growl’ (so Gunkel, Psalmen, p. 252, noting the comparison to dogs in the
previous verse): a similar sense has been found in the Phoenician Kilamuwa
inscription (KAI 24.10; J.C.L. Gibson, Syrian Semitic Inscriptions, 3, Phoeni-
cian Inscriptions (Oxford, 1982), p. 34: ‘whimper’), but the reading is not
certain and interpretations vary (J. Tropper, ‘ “Sie knurrten wie Hunde”. Psalm
59,16, Kilamuwa:10 und die Semantik der Wurzel lwn’, ZAW 106 [1994],
pp. 87-95; DNWSI, p. 575).
g. Heb. ‫לּאמר‬: see Note b on the translation of 6.10–7.5. Unusually the verse
has no athnach: the tiphchah under this word serves as the principal disjunc-
tive accent here (GK §15f, no. 5; Yeivin, Introduction [Chico CA, 1980],
pp. 177-79): likewise in v. 18.
h. Heb. ‫נשׁתה‬. The imperfect is probably modal (cf. ‫)לא יכלו לשׁתת‬: JM §113l,
Joosten, Verbal System, pp. 273-74.
i. Heb. ‫ויורהו‬. For the sense ‘show’ cf. Gen. 46.28; Ps. 45.5; Prov. 6.13
(‫ ;)באצבעתיו‬and perhaps Job 6.24. Earlier in Exodus (4.12, 15) it had the sense
‘teach, instruct’.
j. Heb. ‫עץ‬. ‘A tree’ is unlikely to be the meaning, as what is ‘shown’ is also
‘thrown’. The sense ‘a piece of wood’ is mainly found in the pl. (e.g. 1 Kgs
17.12) or with reference to a wooden artefact (Deut. 19.5; 2 Sam. 21.19), but
presumably a broken branch of a tree is meant here. BDB’s ‘shrub?’ (p. 781)
is less likely.
k. Heb. ‫שׂם לו…נסהו‬. The antecedents of the subjects of the verbs and the
pronominal suffixes are not as clear as they could be: the last sing. verbal
subject is Moses (‫ )וישׁלך‬and the last pronominal object (‫ )ויורהו‬also referred to
him. The speech introduced by ‫ ויאמר‬in v. 26 refers at first to Yahweh in the
third person, but in the latter part of the verse the speaker is clearly identified
as Yahweh himself (see also the Explanatory Note). While it is just possible
that Moses is the subject of the verbs in v. 25b, the content is much more
11
On the unusual vocalisation of some Hiphil forms in Exodus and Numbers
see Note o on the translation of 16.1-36.
15.22-27 409
appropriate to Yahweh (but see below), who has been mentioned twice earlier
in v. 25. The sing. pronominal suffixes might then refer to Moses, but again
the content and parallels elsewhere (see the Explanatory Note) make this less
likely and the people as a whole are probably meant (cf. ‫ העם‬in v. 24, even
though the verbs of which it is the subject there are plural), as they are by the
second sing. verbal forms and suffixes in v. 26. The lack of clarity may well
be due to the clumsy insertion of material that once stood elsewhere (cf. 18.5
[where ‫ שׁם‬also occurs] and the commentary there; and the introduction to this
section).
l. Heb. ‫חק ומשׁפט‬. ‫ חק‬has occurred in a special sense in 5.14 and in its more
common legal sense in 12.24 (cf. ‫ חקה‬in 12.14, 17, 43; 13.10); for occurrences
later in Exodus see 18.16, 20; 29.28; 30.21 (‫ חקה‬in 27.21; 28.43; 29.9). ‫משׁפט‬
occurs in Exodus as a word for a ‘rule’ also in 21.1, 9, 31; 24.3; 26.30 (the
meaning in 23.6; 28.15, 29, 30 [2x] is different). The exact definition of this
usage is debated: F. Horst and R. Hentschke think that civil laws are meant,
as distinct from the cultic laws denoted by ‫חק‬, while Alt saw the ‫ משׁפטים‬as
those with a casuistic form (‘Die Ursprünge des israelitischen Rechts’ [first
published in 1934], in his Kleine Schriften 1, pp. 278-332 [289, ET, p. 92]),
from which others such as G. Liedke and Hentschke have deduced that ‫חק‬
stands for ‘apodictic legal rules’ (cf. THAT 1, 629-30 = TLOT 2, p. 470; THAT 2,
1005-1007 = TLOT 3, p. 1397). Others again have questioned whether there is
sufficient basis for such distinctions and it is widely agreed that in lists of legal
terms the words become virtually synonymous (cf. TWAT 3, 152-53 = TDOT
5, pp. 142-43; TWAT 5, 103-105 = 9, pp. 94-96). It might seem that in 21.1
‫ משׁפטים‬means civil or casuistic laws in view of the sequence that follows (cf.
also vv. 9, 31), but it may be arbitrary to limit the scope of the title to 21.2–22.26
and not include the apodictic and (in part) cultic laws in 22.17–23.19 as well.
A stronger case can perhaps be made on the basis of other meanings of
‫ משׁפט‬and their association with court procedure, but such connections need
not mean that ‫ משׁפטים‬were not enacted or imposed as much as ‫חקים‬. Indeed
a verse like Exod. 15.25 suggests that they were (cf. Josh. 24.25; 1 Sam.
30.25; Ezra 7.10). So the distinction between the terms may be more between
different aspects of laws than different types of laws. A final issue is whether
the sing. expressions here mean a single ‘statute and rule’ or have a collective
sense, as we have translated them. The similar expressions just mentioned
support both possibilities: in 1 Sam. 30.25 (cf. Gen. 47.26 with ‫ חק‬alone) a
single regulation about the distribution of booty seems to be meant, while in
Josh. 24.25 and Ezra 7.10 a more comprehensive corpus of laws is involved.
m. Heb. ‫כל־המחלה‬. In view of ‫ כל‬a collective sense equivalent to a plural
is presumably intended. ‫ מחלה‬is a rare equivalent to the much more common
‫( חלי‬24x in BH): it occurs elsewhere only in 23.25 (part of the parenesis in
23.20-33 which is similar to 15.25b-26 in other ways) and in 1 Kgs 8.37 and
its par. in Chronicles. It is not used in Deuteronomy (or anywhere else in
Deuteronomistic literature), where the words for sickness are ‫חלי‬, ‫ מכה‬and ‫מדוה‬
410 EXODUS 1–18
(cf. Deut. 7.15; 28.27, 59-61). Some of the other language used here is also
not characteristically Deuteronomic: ‫רפא‬, ‘heal’, with a divine subj. belongs
rather to JE (Gen. 20.17; Num. 12.13: cf. Deut. 32.39) and (metaphorically) to
Hosea (6.1; 7.1; 11.3; 14.5) and ‫ האזין‬only occurs in Deuteronomy twice (1.45,
with God as subj.; 32.1 of the heavens), but widely elsewhere, including two
probably northern psalms (Pss. 77.2; 80.2).
n. Heb. ‫ושׁם‬. ‫ שׁם‬often stands at the beginning of a clause (cf. v. 25 and
BDB, p. 1027) to highlight the connection with what has preceded: with waw
and a noun-clause as here (cf. Num. 13.22; 1 Sam. 1.3; 24.4; 2 Kgs 4.8; Jer.
37.13; Neh. 10.40; 1 Chr. 11.4) it in effect introduces a relative clause (cf.
GK §156b).
o. Heb. ‫תמרים‬. The date-palm is meant: MH ‫ ָתּ ָמר‬and Ar. tamrun are both
used for the fruit, which is not explicitly mentioned in BH; but cf. Song
7.8-9, and it may be the source of the ‘honey’ (Heb. ‫ )דבשׁ‬in 3.8, 17; 13.5 and
elsewhere (ABD 2, p. 807). Jericho, known as ‘the city of palms’, was the
classic palm-oasis in the OT (Deut. 34.3 etc.).
Explanatory Notes
22. The celebrations of their deliverance completed, it is time for
Israel (the name as in 14.30-31a) to move on from the Yam Suph
(on which see the notes on 13.17-18 and 15.4). Unusually for such
notes of movement (but cf. 14.15) Moses’ leadership is emphasised
by the use of the causative form of the verb ‘depart’ (see Note a
on the translation). Elsewhere this idea is expressed by similar
forms of the verbs ‘go out’ (yāṣāʾ) and ‘go up’ (ʿālāh): cf. 3.10-12;
14.11; 17.3; 32.1, 7, 23, 33.1, 12 (in 6.13, 26-27; 16.3 together with
Aaron), and SP and LXX extend it to to the following verb here (see
Text and Versions). The people first enter ‘the Wilderness of Shur’,
which is mentioned only here.12 But ‘Shur’ occurs in other contexts
relating to the northern part of the Sinai peninsula, and in three of
them it is close to or ‘before’ (perhaps ‘east of’) Egypt (Gen. 25.18;
1 Sam. 15.7; 27.8: cf. Gen. 16.7; 20.1). So ‘the Wilderness of Shur’
was most likely a broad term referring to the desert of north-western
Sinai. ‘Shur’ (Heb. šûr) is a poetic word for a ‘wall’ in Gen. 49.22
and 2 Sam. 22.30 = Ps. 18.30 (here apparently of a defensive wall)
and it may have been used of the well defended eastern border of
12
In the full itinerary in Num. 33.1-49 the corresponding region is called ‘the
Wilderness of Etham’ (v. 8), after a place on the edge of it which is named in
vv. 6-7 and also in Exod. 13.20.
15.22-27 411
Egypt (cf. Lipschitz, Sinai [Part I] [Tel Aviv, 1978], pp. 27, 44,
51-52; ABD 5, p. 1230): Egyptian texts refer to this as ‘the Wall of
the Ruler’ (cf. ANET, pp. 19, 446).13 On ‘three days’ as a short but
imprecise period see the Explanatory Note on 3.18.
23. In Hebrew the name Marah sounds like a feminine form of
the word for ‘bitter’ (mārāh), and this is the basis for the connec-
tion made between its ‘bitter’ waters and the name that had been
given to it by travellers. (There is no suggestion here that the name
was due to the specific experience or behaviour of the Israelites,
unlike 17.7 and Num. 11.3 and 34.) On the meanings of this and
other ‘taste’ words in Hebrew and related languages see P.D.
King, Surrounded by Bitterness (Eugene OR, 2012), pp. 326-29:
while some of its applications do relate to substances which are
scientifically ‘bitter’ (Exod. 12.8; Deut. 32.32; Isa. 24.9), Heb. mar
could be used, as here, more generally of anything unpleasant to the
taste or harmful, just as ‘sweet’ need not mean ‘sugary’ but rather
‘pleasant’ (cf. Isa. 5.20; Prov. 27.7). The common inference that the
water of Marah was salty or brackish is therefore fully justified.14
Ora Lipschitz, describing the hydrology of the Sinai peninsula,
notes that high salinity is a general problem in the region south-
east of Suez, affecting such places as Bir Murrah, Uyun Musa,
Ain Sudr, Wadi Gharandel and Wadi Tayibeh (Sinai 1, pp. 10-11).
Earlier travellers to the area observed the same widespread phe-
nomenon and identified particular places with Marah on the basis
of the especially unpleasant water there and their own views about
the location of the sea-crossing: from north to south Bir el-Murr
(Murrah), east of Suez (M. Harel, Masaʿe Sinay [Tel Aviv, 1968],
pp. 46, 232-33), Uyun Musa (Deacon Peter [Egeria?]: cf. Wilkin-
son, Egeria’s Travels, pp. 207-208) and most commonly among
nineteenth-century travellers Ain Hawwara c. 40 mi. SSE of Suez
(Robinson, Biblical Researches 1, pp. 89-90, 96-98; Stanley, Sinai,
pp. 37, 68; Palmer Desert 1, pp. 35, 40, 272-73). Gressmann, by
contrast, but in line with his overall theory, identified Marah with
Ain el-Qudeirat in northern Sinai (Anfänge, p. 78). It would seem
13
šûr is more widely used as a word for ‘wall’ in Aramaic (all periods). For
other suggestions see HAL, p. 1348; also Text and Versions for the Targumic
renderings.
14
In fact, as King points out, a cognate Akkadian word (marratu) was
similarly used of sea-water (ibid., p. 328: cf. AHw, p. 612).
412 EXODUS 1–18
that the placing of this narrative, whatever its ultimate origin, at the
beginning of the wilderness journey at least corresponds to some
knowledge of natural conditions in the west of the Sinai peninsula.
24-25a. The verb ‘complain’ (Heb. lwn: see Note f on the transla-
tion, where its other occurrences are listed) occurs here for the first
time in the wilderness narratives, in which it represents a regular
response of the Israelites to the hardships of their journey (cf. the
classic study of Coats, Rebellion, and the general introduction to
the wilderness tradition above). The motif is especially prominent,
in distinctive forms, in the Priestly narrative (see e.g. the commen-
tary on 16.1-12 and Frankel, The Murmuring Stories), but signs of
Priestly authorship are lacking here and in 17.1-7* and these stories
have generally been attributed to the non-Priestly wilderness narra-
tive (see the introductions to these sections). The complaint here is
addressed simply to Moses and there is no explicit indication at this
stage that such complaints are a criticism of Yahweh’s provision for
the people (contrast 16.7-8; 17.2, 7). Nor is there the nostalgic look
back to the better conditions in Egypt which appears elsewhere (cf.
16.3; 17.3: also 14.11-12), but simply the understandable and even
desperate question, ‘What can we drink?’. The sequel is similarly
straightforward and briefly narrated: Moses cries out to Yahweh for
help, Yahweh directs his attention to the branch of a tree (Heb. ʿēṣ
can mean a piece of wood as well as a whole tree: see Note j on the
translation), he throws it into the water, the water becomes drink-
able and (this is evidently presumed, as it is in 17.6) the people
drink it. In other words, put in its simplest terms, the people face a
crisis, their representative appeals for Yahweh’s help on their behalf
and the crisis is overcome (Childs’s ‘Pattern I’: pp. 258-59). This is
not a narrative of rebellion, like some others set in the wilderness;
nor is it an etiological narrative, for the etiology in v. 24 is only
incidental to the main plot line. It is a short ‘story of deliverance’,
expressing the same assurance of Yahweh’s readiness to help his
people when in need as inspired their simplest psalms of prayer and
thanksgiving.
25b-26. ‘There’, so according to the narrative sequence at the
place where Yahweh had shown his ability and willingness to
help his people in the face of their need for water on their journey
through the desert, ‘he’ made laws for ‘them’ (literally ‘him’ or ‘it’,
but the people have been referred to in the singular as well as the
15.22-27 413
plural in vv. 22-24) and ‘he tested them’; and ‘he’ goes on to say
that if the people (now ‘you’) are obedient they will be spared the
diseases which have afflicted the Egyptians, because ‘I am Yahweh
your healer’. Only in the final words of v. 26 is it made unambigu-
ously clear that the words spoken are Yahweh’s words (though the
two preceding first person verbs leave little doubt that this is the
case). From this it is possible to work backwards and deduce that
the ‘he’ at the beginning of v. 26 and in v. 25b is also Yahweh. But
the reader is at first faced with some uncertainty about the actor
in v. 25b and the speaker in v. 26, and even with some counter-
indications to the ‘obvious’ inference that both are in fact Yahweh.
The last preceding singular subject is not Yahweh but Moses (‘He
threw it…’) and in v. 26 Yahweh is at first referred to in the third
person rather than the first person, as though someone else were
the speaker. Moreover, surprising as it may seem, the first expres-
sion in v. 25b, ‘he made statutes and ordinances for them’, is used
elsewhere with a human subject (Josh. 24.25 [Joshua]; 1 Sam. 30.25
[David]: cf. Gen. 47.26 [Joseph] and Ezra 7.10 [Ezra ‘teaching’]),
and ‘tested’ (Heb. nissāh) is used at least once of what one human
does to another (1 Kgs 10.1 [the queen of Sheba to Solomon]).15 It
is understandable therefore that some commentators have thought
that Moses is the subject both in v. 25b and at the beginning of
v. 26.16 The change to divine first-person speech in v. 26b can in fact
be paralleled in Deuteronomy (7.4; 11.13-15; 17.3; 28.20; 29.45;
the reverse shift in 1.8 [Houtman, p. 312]), where the speaker is
normally Moses.17 However, the association of ‘made (laws)’ and
‘tested’ must tip the balance the other way: they must have the same
subject and it is only Yahweh who can do both. The association is
paralleled in 16.4 (probably non-P); 20.20 as well as in Deut. 8.2;
13.4-5; 33.8-10; Judg. 2.22; 3.4. Yahweh’s ‘testing’ of Abraham in
15
In later biblical Heb. also in Qoh. 2.1; 7.23; Dan. 1.12, 14.
16
So Coats, Rebellion, pp. 49-50; V. Turgman, De l’autorité de Moïse: Ex
15,22-27 (Eilsbrunn, 1987), pp. 15-18, 20; Houtman, pp. 308, 312, 314 (only
v. 26: cf. Nachmanides); Van Seters, Life, pp. 178-79, 181.
17
Albertz (p. 262 and n. 1) observes that a similar inconsistency appears in
Exod. 23.23-33 and 34.11-26, though this involves the occasional change to third-
person reference to Yahweh in what is otherwise a divine speech, i.e. the opposite
of Deuteronomy. It is also sometimes found in speeches attributed to angels: Gen.
16.10-11; 22.12.
414 EXODUS 1–18
Gen. 22.1 (cf. v. 18) is also not greatly different. Israel is sometimes
said to ‘test’ Yahweh (17.2, 7; Num. 14.22; Pss. 78.18, 41, 56; 95.9;
106.14) or warned against it (Deut. 6.16), but that is not the sense
here.
Much remains obscure about what exactly is referred to here.
There is no indication of what the ‘statutes and ordinances’ (on the
meanings of these terms see Note l on the translation) are and specu-
lation about them is fruitless. As for the ‘test’, it might be the lack
of drinkable water, in which case v. 25b provides ‘interpretation and
commentary’ (Houtman, p. 314) on the preceding story. Or (and this
seems more likely in view of what precedes and follows) it might
be the choice between obedience and disobedience to the laws of
Yahweh. The mention of law-giving at this point in the narrative is
itself a surprise, when the tradition lays so much weight on Mount
Sinai/Horeb as the place of law-giving. True, the authors of Deuter-
onomy felt at liberty to relocate the main giving of the law to the
people on the edge of the Promised Land (though with an explana-
tion to ‘bridge the gap’ in 5.28–6.3) and according to Josh. 24.25 a
covenant with ‘statutes and ordinances’ was made at Shechem. But
what was so special about Marah?18 Most likely it was not Marah
or the episode located there as such that was important but the fact
that it marked the beginning of the wilderness journey (or, to put
it differently, the nearest possible occasion to the liberation of the
Israelites from Egypt): here surely some instruction about their new
life of freedom was needed.19
But the change of topic in the middle of v. 25 is, as many commen-
tators have noted, very abrupt and both it and the uncertainties that
have already been discussed could well be due to the connection not
being original, with v. 25b either having been moved from another
point in the narrative or being an extract from a different account
18
Some have found the key in ‘Yahweh showed him’ earlier in the verse (Heb.
hôrāh can mean ‘taught’ [see Note i on the translation] and is related to the word
tôrāh): see Houtman, p. 313, for some examples of this view, which Houtman
rightly rejects as a distortion of the real meaning. Gressmann (Mose, pp. 121 n. 1,
414) saw it, with v. 25b, as a relic of an underlying ‘Massah tradition’ originating
at Kadesh.
19
Levin (p. 351) thinks it was specifically the need to back up the reference
to a law in Exod. 16.4.
15.22-27 415
of the wilderness journey.20 The possibility that ‘There’ originally

referred not to a specific place but to the wilderness as a whole,
hinted at by Propp (p. 579), deserves serious consideration: the
short statements in v. 25b would make an apt prologue (or ‘exposi-
tion’, to borrow Coats’s term) for the wilderness narrative, picking
out what seemed to be its two most memorable themes. They were
certainly taken up in Deuteronomy (esp. 8.2-5, 15-16: cf. 13.4), but
they are also evident elsewhere in the non-Priestly narrative of the
Pentateuch (cf. Gen. 22.1; Exod. 16.4; 18.16, 20; 20.20).21
In v. 26 the language is much more distinctively Deuteronomic
and the introduction of a conditional element into the promise of
Yahweh’s protection corresponds to one of its central emphases.
Even if the latter can also be found later in Exodus itself (19.5-6;
23.22), the wording of this verse is (except for its final words,
interestingly) strongly characteristic of Deuteronomy and Deuter-
onomistic editing elsewhere. This is particularly true of ‘Yahweh
your God’ (Deut. 1.10 etc.; 282 x in Deuteronomy), ‘obey the voice’
(18x Deuteronomy),22 ‘do what is right in his sight’ (5x Deuter-
onomy and often in editorial passages in Kings), ‘commandments’
and ‘statutes’ together (15x Deuteronomy, otherwise mainly in
Kings) and ‘keep’ in the sense of ‘observe’ (with ‘statutes’ in Deut.
4.40; 26.17: in other similar expressions c. 50x in Deuteronomy).
‘The diseases of Egypt’ also appear in Deut. 7.15 and 28.60 (and
only there), even though the Heb. is different (see Note m on the
translation). The verse is clearly an exposition of the implications of
v. 25b and was probably added by an editor close to Deuteronomy.
20
Lohfink makes the important observation that elsewhere in Hebrew prose
narrative an introductory ‘there’ is almost always preceded by ‘and’ (Heb we: cf.
v. 27): the only real exceptions are Gen. 49.31 and Deut. 10.6 (‘Ich bin Jahwe, dein
Arzt’, p. 19 n. 19, ET, p. 41 n. 19).
21
Since Wellhausen (Prolegomena, p. 349, ET, p. 343; cf. Noth, p. 102, ET,
p. 129) ‘tested’ (Heb. nissāh) has been associated with the place-name Massah
(on which see the Note on 17.7) and hence with Kadesh. But this is a false trail
(see e.g. Childs, p. 268, and the general introduction above to the wilderness
narrative).
22
There is a minor difference in the Heb. idiom, since Exod. 15.26 has
the preposition le, ‘to’, before ‘voice’, whereas Deut. consistently has the more
common be.
416 EXODUS 1–18
Its closing words, however, have no such associations. The idea

of Yahweh as ‘healer’ is widespread in the Old Testament but (apart
from the poem in Deut. 32) absent from Deuteronomy (see M.L.
Brown, Israel’s Divine Healer [Carlisle, 1995]; TDOT 13, pp. 593-
602; for a possible Ugaritic parallel see DDD, 558-60; DNWSI,
pp. 742-43). Yahweh heals Abimelech and his family (Gen. 20.17-
18) and Miriam (Num. 12.10-15), and he is Israel’s true healer in
Hosea (6.1; 7.1; 11.3; 14.5) and other prophetic passages, as well
as (without the word being used) in Num. 21.8-9. ‘Healing’ was
also an expression for making foul water safe and drinkable (2 Kgs
2.19-22; Ezek. 47.8-11), as in the story of Marah, and this may
explain the choice of this expression here, although it was clearly
intended in its normal sense.
27. The next stopping-place recorded, with no narrative attached,
is an oasis, whose name Elim (like that of the more frequently
mentioned Elath or Eloth at the head of the Gulf of Aqaba: Deut.
2.8 etc.) probably alludes to the number of palm-trees there.23 These
would have been date-palms, which are the most common type in
the Sinai peninsula and serve the needs of the Bedouin in a variety
of ways (Lipschitz, Sinai, p. 15). Large oases are few and far
between in the peninsula: the oasis in Wadi Feiran was particularly
notable and was the site of a town in antiquity (Rothenberg, God’s
Wilderness, p. 136). Otherwise the oases are smaller and gener-
ally on the coastal plain, east and west, where streams come down
from the mountains. Early Christian travellers located Elim either at
Wadi Gharandel (Arandara: cf. Wilkinson, Egeria’s Travels, p. 208)
c. 50 mi. SSE of Suez (which has remained a popular identification
in modern times (cf. Abel, Géographie 2, p. 210) or at a place called
Rhaethou (Ammonius, Forty Martyrs; Cosmas Indicopleustes,
Christian Topography 5.14: see Davies, Way of the Wilderness,
pp. 38, 43), which was probably at et-Tur in the far south-west of
the peninsula. The latter received renewed support from E. Meyer,
Israeliten, pp. 100-103, citing Agatharchides, On the Erythraean
23
So BDB, pp. 18-19; HAL, p. 40; Ges18, p. 48. The older view that Heb.
ʾayil, ʾēlāh and ʾēlôn meant ‘terebinth’ in BH, as they do in later Heb., has now
been generally abandoned in favour of the less specific ‘mighty tree’, which might
apply to large palm-trees as well as to other species (cf. HAL, pp. 39, 50; Ges18,
pp. 46, 60, 62; ABD 2, p. 808).
15.22-27 417
Sea (ap. Diodorus Siculus 3.42-43; Strabo 16.4.18)24 for the name
Phoenikon in the second cent. B.C., but it is too far south and the
ancient identification will no doubt have derived from the early
monastic community there, which still survived in the 1920s (D.J.
Chitty, The Desert a City [Oxford, 1966], pp. 168-73; cf. F.S.
Bodenheimer and O. Theodor [eds.], Ergebnisse der Sinai-Expedi-
tion 1927 [Leipzig, 1929], pp. 1-3).25
The brief and isolated note of movement to Elim, which is
continued in 16.1, is unlikely to have formed part of a discursive
narrative tradition and must have been based on a geographical
source of some kind (like 12.37, 13.20 and 14.2); cf. Noth, p. 101,
ET, p. 128: ‘surely based on local knowledge’. This could well have
been the complete itinerary in Num. 33.1-49, where v. 9 contains
almost identical wording, so that a literary relationship between the
two texts becomes almost inevitable.26 Modern scholars have tended
to prefer the view that Num. 33.1-49 is a compilation from the main
narrative in Exodus and Numbers and some other source(s) (cf.
Blenkinsopp, Pentateuch, pp. 161-62). But it then becomes neces-
sary to assume the existence of another source for passages like
15.27 and 16.1.
Text and Versions

‫( את־ישׂראל‬15.22) LXX τοὺς υἱοὺς Ἰσραήλ and Sy ldbyt ʾysryl expand
the designation as often elsewhere (cf. 12.21; 14.25).
‫( מים־סוף‬15.22) 4Q365 has just ‫מים‬, which conforms to v. 19 (and Num.
33.8), though oddly without the def. art., as if this were still poetry (cf. ‫ ים‬in
v. 10).
‫( ויצאו‬15.22) SP ‫( ויוציאהו‬von Gall’s ‫צאהו‬- is supported by only four of his
mss) and LXX ἤγαγεν αὐτούς read the verb as a sing. causative like ‫ויסע‬,
24
See further S.M. Burstein, Agatharchides of Cnidus, On the Erythraean Sea
(with ET of the surviving fragments; London, 1989), pp. 147-50.
25
The location of Elim at ‘Surandala’ in the Itinerary of ‘Antoninus’ (para. 41:
cf. Way of the Wilderness, p. 46) may have had Wadi Gharandel in view.
26
The differences are: (i) here there is no mention of the departure from
Marah; (ii) Num. 33.9 has ‘and in Elim’ instead of ‘where’; (iii) Num. 33.9 lacks
‘by the water’ at the end. The Exodus redactor may have tailored his source to
the narrative context by departing from the rigid pattern of the itinerary, as in
12.37.
418 EXODUS 1–18
with ‫ו‬- taken as the sing. obj. suffix. The other Vss assume MT’s reading. Its
transition to a pl. reference to Israel is not a problem (GK §145c) and if SP’s
unambiguous reading were original it is hard to see why it would have been
changed to ‫ויצאו‬. 4Q365 continues with [‫וילכו במדבר ֯ש‬, apparently omitting
‫ ויצאו‬and its whole clause: [‫ ֯ש‬is probably the first letter of ‫( שלשת ימים‬with
a variation in the word-order), not of ‫( שור‬as reconstructed in DJD XIII,
pp. 270-71). The line will then have been noticeably short and may have
ended in a vacat before v. 23.
‫( אל־מדבר־שׁור‬15.22) For ‫ מדבר‬TgF(VN) has ‫אורחא‬, probably an assimilation
to its rendering of ‫ דרך שׁור‬in Gen. 16.7: TgNmg ‫ ארעא‬might be a corruption
of this. LXX, Vulg and Sy transliterate the toponym ‫ ;שׁור‬but the Tgg give,
as elsewhere, an interpretation of it by a contemporary place-name. TgJ,N,F
have ‫חלוצא‬, ‘Halutsa’ or Elusa, the name of an originally Nabataean settle-
ment about 30 mi. S of Gaza which became the chief city of the western part
of Palestina Tertia in the fourth cent. A.D. (cf. Avi-Yonah, The Holy Land,
pp. 121, 123, 125; EAEHL 2, pp. 359-60; Avi-Yonah, Gazetteer of Roman
Palestine [Qedem 5; Jerusalem, 1976], p. 54). The ‫ חגרא‬of TgO is nothing
to do with el-Hejra in Arabia but a name for the line of Roman forts nearby
which was probably given a formal status as the limes Palaestinae by Dio-
cletian in the early fourth cent. (cf. Avi-Yonah, The Holy Land, pp. 119-21,
162-64; Davies, ‘Hagar, el-ḥeǧra and the Location of Mount Sinai’, VT 22
[1972], pp. 152-63 [154-58]). Both these renderings define ‘the wilderness of
the wanderings’ from the point of view of Palestine, unlike ‫ שׁור‬itself (see the
Explanatory Note).27
‫( שׁלשׁת ימים‬15.22) As noted above, these words were probably placed
after ‫ במדבר‬in 4Q365. SP prefixed ‫( דרך‬on its omission by von Gall see
Baillet, ‘Corrections’, p. 30: it is included in the mss used by Sadaqa, Tal
and Crown and in Camb. Add. 1846) and equivalents appear in TgO,N and Sy.
The obvious explanation is assimilation to Num. 33.8, but the longer phrase
is also frequent elsewhere for defining the length of a journey (e.g. Exod.
3.18; 5.3; 8.23).
‫( ולא מצאו מים‬15.22) TgJ prefixes ‫בטלין מן פקודייא‬, a technical expression
for neglecting (the study of) the Law (cf. TgJ on 17.1 etc.; AramB 2, p. 206
n. 41). MRI (Lauterbach 2, pp. 89-90) shows that this expansion was based
on what was taken to be a figurative use of ‘water’ for the Law in Isa. 55.1.
After ‫ מים‬LXX adds ὥστε πιεῖν (cf. ‫ לשׁתת‬in 17.1 and Num. 20.5), perhaps to
distinguish drinkable water from the water of Marah in v. 23 (Wevers, Notes,
p. 237). Syh notes the variation from MT with the obelus.
‫( מרתה‬15.23) Some early translations suggest uncertainty about the
vowel in the first syllable: LXX (εἰς) Μέρραν, Symm (in Syh) mwrt, Sy (l)
27
MRI (Lauterbach 2, p. 87; cf. Jastrow, p. 616) describes it as ‘the wilderness
of ‫’כוב‬, i.e. thorns, a reminder of the difficulties of the journey (so also in other
rabbinic writings cited by Jastrow).
15.22-27 419
mwrt, Theod (in Syh) mwrʾ.28 TgO ‫ ָמ ָרה‬and Vulg marath have the ‘a’ of MT.
Sy, Symm and Vulg (cf. TgN,F(P)) evidently did not recognise that the taw was
the result of morphological variation before the directional ending (GK §90i).
‫( יכלו‬15.23) 4Q365 ‫ י]כולו‬exhibits the full orthography for this form
which is also found in 4Q401 f14 ii 4 and 11Psa at 129.2 (compare the pausal
form in MT at 8.14; 2 Kgs 3.26; Jon. 1.13 [J. Joosten, pers. comm.]): for a
brief comment see Qimron, p. 42 n. 60.
‫( מים‬15.23) Rahlfs followed LXXB and a few other witnesses in omitting
any equivalent, but the other uncials have ὕδωρ and Wevers judges this to be
original (cf. his explanation in Notes, p. 238; for another Propp, p. 574).
‫( קרא‬15.23) LXX and TgF(P) render idiomatically in the passive (cf. Note
d on the translation); the other Vss reproduce MT’s sing. active form (except
for Sy8b1 qrw), either regarding Moses as the subject (so apparently Vulg) or
treating it as indefinite.
‫( שׁמה‬15.23) LXX and Sy have ‘the name of that place’, clarifying with
the help of the language of 17.7.
‫( מרה‬15.23) LXX Πικρία, providing an explanation for those with no
knowledge of the Semitic languages, as it did in 17.7. Vulg evidently did not
find this sufficient and provided an extended paraphrase of v. 23b.
‫( וילנו‬15.24) SP and 4Q365 both read the sing. (4QpalExl is not extant
here); likewise LXX and Vulg, but λαός generally takes a sing. verb in
Exodus (Wevers, Notes, p. 238) and Latin grammar is even stricter over agree-
ment in number, so their evidence is inconclusive. Tgg and most mss of Sy
reproduce MT’s plural, which as the difficilior lectio is to be preferred. Most
of the equivalents used in the Vss mean ‘murmur, grumble, complain’, but TgF
has ‫( ואדיינו‬from ‫דין‬, ‘dispute, quarrel’), which is closer to Heb. ‫ריב‬.
‫( לאמר‬15.24) Sy adds lh as in v. 26.
‫( ויצעק‬15.25) This is also the spelling in SP, but 4Q365 uses the by-form
‫זעק‬, as it did in 14.15 (see Text and Versions there): 4QpalExm does not survive
here. Tgg and Sy render with ‫צלי‬, ‘pray’, as in 14.15. SP, 4Q365, LXX, TgF
and most mss of Sy add ‘Moses’ as the subject.
‫( ויורהו‬15.25) SP reads ‫ויראהו‬, presumably Hiphil: one might conjecture
an older reading ‫ וירהו‬behind both this and MT. Since both ‫ ירה‬and ‫ ראה‬in the
Hiph. can bear the sense ‘show’ (see Note i on the translation), the render-
ings of LXX, Vulg, TgJ,N,F(VN) do not reveal which appeared in their Vorlagen.
Neither 4QpalExm nor 4Q365 survives at this point, but MRI (Lauterbach 2,
p. 92) evidently knew the variant preserved in SP and rejected it in favour of
‫ ויורהו‬taken in the sense ‘taught’, which opened the way to the ‘discovery’ of
an allusion to the Torah in the word ‫( עץ‬see the next note). The use of ‫אלף‬
28
Might the (?) ‘o’-vowel be the result of an interpretation which, on the basis
of ‫ויורהו‬, explained the name as ‘a place of teaching’ (cf. the modern views in
Houtman, p. 313)? Or is it (cf. Propp, p. 576) a reflection of ‫מ ָֹרה‬, ‘bitterness’, in
Gen. 26.35 (cf. LXX later in the verse)?
420 EXODUS 1–18
by TgO,F(P) picks up this interpretation. 4Q365 has a lacuna at this point, with
room for some extra text. TgF,Nmg expand the subject of the verb to ‘the Memra
of Yahweh’.
‫( עץ‬15.25) The Tgg (other than TgO) draw on several strands of early
interpretation of the verse, especially those preserved in MRI (Lauterbach 2,
pp. 92-94: see further the sources referred to in AramB 2, p. 69 n. 22, and
J. Bienaimé, Moïse et le don de l’eau dans la tradition juive ancienne:
Targum et Midrash (Rome, 1984), pp. 11-16, 42-44). TgJ,F(VN),Nmg identify the
tree with a ‘bitter’ creeper; TgJ,Nmg add that Moses (or God?) wrote the divine
name on it (perhaps to equate it with the staff mentioned in 4.20: see Text
and Versions there);29 TgN,F(P) make associations with the Law and the ‘tree of
life’ (cf. Prov. 3.18). Josephus, AJ 3.7 understands a branch of an unspecified
tree to be meant.
‫( שׁם‬15.25) 4Q365 has the longer form ‫ שׁמה‬with the same meaning, as it
does elsewhere: this use of ‫ שׁמה‬without a directional sense, which is already
found in BH (BDB, p. 1027), was widespread at Qumran (cf. DCH 8, p. 421;
more generally Qimron, p. 69). SP agrees with MT in what is certainly the
older reading.
‫( שׂם‬15.25) TgJ,Nmg specify the subject as ‘the Memra of Yahweh’. The verb
itself is rendered as one might expect in most Vss, but ‘showed’ (TgF(VN),Nmg),
‘read/proclaimed’ (TgF(P)) and ‘taught’ (Sy) avoid the idea of enactment.
‫( חק ומשׁפט‬15.25) SP, 4QpalExm, 4Q365 and TgO support MT’s sing.
expressions, but all the other Vss have plurals (on the possibility of a collec-
tive interpretation of the Heb. see Note l on the translation). The equivalents
used for ‫ משׁפט‬here all have a judicial background, but Aram. ‫ דין‬was often
used for ‘law’ and this sense may have been adopted in the use of κρίσεις
and iudicia in LXX and Vulg: it seems not to be paralleled in the secular use
of these words. MRI (Lauterbach 2, p. 94) presented the different views of
R. Joshua and R. Eleazar about which laws were meant (apparently preferring
the former) and TgJ gives a list which draws on both (cf. ‫ סדרי דינין‬in TgN,F). At
16.28 MRI (p. 121) follows R. Joshua’s view that 15.25 refers to the sabbath
law.
‫( ושׁם‬15.25) The early SP ms. Camb. Add. 1846 does not have the waw:
probably an error due to the occurrence of ‫ שׁם‬earlier in the verse, as all other
SP mss (and 4QpalExm) have it.
‫( נסהו‬15.25) Camb. Add. 1846 has ‫נסיהו‬, again differing from all other SP
mss: this time the reason is probably phonetic, due to the pronunciation of the
suffix as -ēhu (cf. similar variants at Gen. 41.52 in von Gall’s apparatus; GSH
§55bαγ, 105a). TgJ,F(VN),Nmg add ‘with the tenth test’ (cf. Num. 14.22): if, as
the place of its citation of Num. 14.22 (before that of Exod. 17.7) suggests,
PRE 44 [345] is referring to the Marah episode, it too makes the subject of
29
The identification is made more explicitly in the Samaritan Asaṭīr (Elfick,
‘The Staff of Moses’, p. 53), which is now dated to the mediaeval period.
15.22-27 421
‫ נסהו‬Israel;30 TgF(P) has instead ‘and he stood up to his test’ (as said also of
Abraham in Aboth 5.4), which will presuppose the general view that God is
the subject (e.g. MRI [Lauterbach 2, p. 94]). In 4QpalExm ‫ נסהו‬is followed
closely by an enlarged waw, which in view of its practice elsewhere probably
indicates not only a division in the text (which is not paralleled in MT or SP)
but that the edge of the fragment is close to the end of a line (as does the waw
at the end of v. 26: cf. DJD IX, pp. 58-61, 91).
‫( ויאמר‬15.26) Sy wʾmr lh adds the addressee as e.g. in 10.29 and 12.31.
‫( שׁמוע תשׁמע‬15.26) TgO,J use the verb ‫קבל‬, which brings out the intended
sense ‘obey’ more clearly (cf. 16.20 and the note there). TgN,F(P) have second
person pl. forms here and throughout the verse.
‫( לקול יהוה‬15.26) A Genizah ms. cited in BHS has ‫בקול יהוה‬, which is the
more common idiom for expressing obedience, but for that very reason it is
likely to be secondary. TgO,J substitute ‘the Memra’ for ‫( קול‬as they do at 5.2)
and TgN,F(P) add it after it.
‫( הישׁר‬15.26) LXX τὰ ἀρεστά, ‘what is pleasing’ (cf. ‫ שׁפיר‬in TgN, Sy),
appears a little imprecise (unlike TgO,J ‫ כשׁר‬and Vulg rectum), but it is LXX’s
regular equivalent for ‫ ישׁר‬in the Pentateuch (Wevers, Notes, p. 240) and may
have been chosen because of its alternative meaning ‘acceptable, approved’
in secular Greek (LSJ, p. 238).
‫( בעיניו‬15.26) 4Q365 appears to read ‫בעינו‬, but this would only be a
phonetic variant (cf. Qimron, p. 59). All the Vss (not only the Tgg) avoid a
reference to Yahweh’s ‘eyes’ by a periphrasis.
‫( למצותיו‬15.26) TgN has no equivalent to the suffix, perhaps a simple
error, or alternatively a reflection of colloquial reference to ‘the command-
ments’ (cf. above on TgJ in v. 22). There is room in a lacuna in 4Q365 for an
additional word or two after this word, perhaps ‫ ולתורותיו‬or ‫ולמשׁפטיו‬.
‫( כל־חקיו‬15.26) 4Q365 prefixes ‫ את‬and uses the pl. of ‫ חקה‬rather than ‫חק‬:
the latter corresponds to the preference of the Temple Scroll for the fem. form
in the pl. (on the distribution in BH see BDB, p. 350).
‫( כל־מחלה‬15.26) The SP ms. Camb. Add. 1846 prefixes waw, which
would make an idiomatic transition to the apodosis. But it is probably second-
ary, and perhaps due to dittography from the previous word.
‫( במצרים‬15.26) Most of the Vss understood ‘the Egyptians’ to be meant,
but Vulg in Aegypto and ms. 5b1 of Sy bmṣryn point to the location ‘in Egypt’.
‫( עליך‬15.26) TgJ has an addition warning of punishment for disobedience
but promising mercy for repentance, probably deduced from the final words
of the verse (cf. MRI [Lauterbach 2, pp. 96-97]).
30
This need not mean that TgJ is based on PRE here, as the ‘ten times that
Israel tested God’ was a well known theme (cf. Aboth 5.7): for more general
questioning of the common view that TgJ is dependent on PRE see R. Hayward,
‘Pirqe de Rabbi Eliezer and Targum Pseudo-Jonathan’, JJS 42 (1991), pp. 215-46.
422 EXODUS 1–18
‫( רפאך‬15.26) Both 4QpalExm (by an enlarged waw) and 4Q365 (by a

vacat) indicate the division here which is also regular in MT and SP mss.
TgF(PVN),Nmg add ‘by my Memra’, identifying God’s ‘word’ as the means (not
the intermediary) by which he heals (cf. Hayward, Divine Name, p. 119, and
the cautionary comments of Chester, Divine Revelation, pp. 293-313 [esp.
311-13]). MRI (Lauterbach 2, p. 96) cites Prov. 3.8 and 4.22 in support of the
Torah being the source of life and health.
‫( אילמה‬15.27) SP omits the directional ending: although SP has no
aversion to this ending as such, it more often omits it after the m.pl./dual
ending (cf. 7.15; 8.16; 9.8: GSH §148d). There is no Qumran evidence in this
case.
‫( ושׁם‬15.27) So also 4QpalExm (‫ )]שׁם‬and most of the Vss. SP, TgJ and three
LXX miniscules of the b family support ‫ובאילים‬, most likely by independent
assimilation to Num. 33.9.
‫( עינת מים‬15.27) TgJ,N,F observe that the number ‘twelve’ meant one for
each tribe (cf. Philo, VM 1.189; MRI [Lauterbach 2, p. 98]).
‫( ושׁבעים תמרים‬15.27) For ‫ תמרים‬LXX has στελέχη φοινίκων, lit. ‘trunks
of date palm-trees’ as at Num. 33.9 and also Job 29.18 (sing.) and Sir. 50.12,
where the Heb. is different: the intention must be to indicate which meaning
of φοίνιξ is intended. This did not stop Ezek.Trag. (Exag. 254-69) from
adding a description of the Phoenix-bird to his account of the scene at Elim
(see on this the very full note of Jacobson, Exagoge, pp. 152-64). A corrector
of LXXF noted that ‘the Hebrew’ had φοίνικες alone, but (although there
is room for doubt) Ezekiel probably lived too early to have used such a
Hebraising revision of LXX. TgO,J,F and Sy used just ‫( דקלים‬cf. Vulg palmae)
but TgN follows it with ‫דתמרים‬, presumably to make clear that the Heb. refers
to a date-palm. TgJ takes up MRI’s observation (cf. Philo, VM 1.189; De Fuga
186) that the number ‘seventy’ corresponds to the number of ‘the elders’ of
Israel (Lauterbach 2, p. 98: cf. Exod. 24.1, 9; Num. 11.16, 24-25), and the Pal.
Tgg make a further comparison with the number of the later Sanhedrin, as
TgF(VN) also does at Num. 33.9. This brief account was also represented in one
of the synagogue mosaics from the fifth–sixth cent. A.D. found at Huqoq in
Lower Galilee. In the summer of 2019 it was announced that a scene showing
the collection of dates from a row of palm-trees is accompanied by a caption
formed by the first two words of Exod. 15.27. I am grateful to Professor
G. Mattingly for drawing this discovery to my attention. On earlier mosaic
discoveries at the site see J. Magness et al., ‘The Huqoq Excavation Project:
2014–17 Interim Report’, BASOR 380 (2018), pp. 61-131 (98).
C h ap t er 1 6 . 1 - 3 6
M an n a an d Q u ai l s
i n t h e Wi ld er n es s of Sin
The boundaries of this long unit, the longest in Exodus 1–20

except for 3.1–4.1-17, are firmly fixed by the subject-matter, from
the location of the (first) divine provision of food in v. 1 to the
explanatory parenthesis in v. 36. The isolated verse about Elim in
15.27, which was traditionally linked to ch. 16 (see the introduction
to 15.22-27), relates to a separate stage on the wilderness journey
and has no material connection with the following narrative. 17.1
marks the next stage(s) and leads into a new episode, as both the
Masoretic mss and the two or three Qumran mss that are available
recognise (see the introduction to 17.1-7). The narrative mostly
progresses in a logical way, with the following main sub-divisions:
(i) itinerary-note marking the arrival in the Wilderness of Sin and
its date (v. 1); (ii) the complaint of the Israelites about lack of
food (vv. 2-3); (iii) the promise of Yahweh to provide food and
his purpose in doing so (vv. 4-5); (iv) the responses of Moses and
Aaron (vv. 6-9) and then of Yahweh (vv. 10-12) to the Israelites’
complaint; (v) the provision of two kinds of food (vv. 13-15); (vi)
Moses’ instructions about the daily collection of manna (vv. 16-21);
(vii) the special arrangements for the Sabbath day (vv. 22-30),
with Yahweh’s reaction to the disobedience of some Israelites
(vv. 27-29); (viii) four concluding notes (vv. 31-36). Only the
beginning of (iii) is marked by a division in the Masoretic mss
(and also in 4QpalExl), which is for the same reason that they also
have divisions before vv. 11 and 28 (in the latter case anticipated
by 4QpalExl), namely that at these points a divine speech begins.
In addition the Leningrad codex (alone) has a division before v. 6
and 4QpalExm, less naturally, has one (within a line) before v. 33.
As indicated more fully in the Explanatory Notes, there are
features within the chapter which have suggested that it was not
originally composed as a single unit.1 For example there are dupli-
1
Houtman, who does not provide full source-critical analyses, clearly recog-
nises this (pp. 323, 350-51).
424 EXODUS 1–18
cations between vv. 4 and 11-12, between vv. 14-15 and v. 31 and
within v. 35; v. 21 sits awkwardly in its present position; and the
Priestly language and concerns which are strongly present in much
of the chapter are absent in other parts of it (vv. 4-5, 13-15, 26-31).
Since at least the middle of the nineteenth century two main ‘hands’ have
been identified at work in its composition, along with minor additions here
and there (for a detailed review of scholarship up to 1974 see Maiberger, Das
Manna, pp. 33-86). At first the two ‘hands’ were seen as the authors of two
separate source-documents, corresponding to the later J and P. Knobel took
this view and already in 1869 E. Schrader had defined their contents in a way
that, broadly speaking, has remained influential until the present day, with
vv. 4-5, 27-31 and 36 attributed to J and the remainder to P.2 Their argument
was based largely on affinities of the language used to other passages, but
was reinforced by the instances of duplication in the chapter. Closer study
led Schrader’s followers to attribute both less and more than this ‘core’ to
J (in the latter case, one may conjecture, also the desire to recover a fuller
J version). In fact the shortage of conclusive data (and perhaps a relentless
quest for greater precision) has led to an unusual divergence of detail even
in this group of commentators, let alone between it and the other two groups
to be treated below. Baentsch took up the suggestion already made by others
that vv. 4bβ and 28 were Deuteronomistic additions and has been widely
followed in this (not, however, by Beer/Galling, Hyatt, Childs, Propp and
Baden). Only a minority of scholars have continued to regard v. 31 as from J
(Gressmann, Noth, Coats [Rebellion, pp. 83-96], Fritz, Childs, W.H. Schmidt,
Propp, Dozeman) and some, following the lead of Noth (ÜGP, pp. 18, 32),
excluded v. 27 (Coats, Fritz, W.H. Schmidt, Propp). By contrast (in addition
to some of the ‘loners’ to be considered below) others enlarged the contribu-
tion of J(E) to the present text by including parts at least of vv. 13b-15, 21
and 35 (Wellhausen, Baentsch, Gressmann [not v. 35], Beer, Hyatt [not v. 35],
Childs, Propp, Dozeman [‘non-P’]).3 A few who excluded the other verses did
attribute part of v. 35 to J (Noth, Coats, Fritz, W.H. Schmidt; cf. Dillmann);
Coats added v. 32 (!) and Baden v. 26.
2
Knobel, Exod.-Lev. (1857), pp. 157-58 (cf. Num.-Jos., p. 548), seems to have
been the first to question the unity of the chapter. For Schrader see W.M.L. de
Wette, Lehrbuch der historisch-kritischen Einleitung in die Bibel Alten und Neuen
Testaments (Berlin, 8th ed., 1869), pp. 281, 284. The attribution of v. 36 to J rather
than P has, however, found little later acceptance: it is probably a relic of the old
view that J was later than P.
3
Wellhausen originally (Composition, p. 78) also attributed parts of vv. 16-20
to JE but this part of his analysis received little support and he himself came to
doubt it as well as some of his other earlier attributions (see the Nachtrag on p.
329 of the 3rd ed. [1889]).
16.1-36 425
A second group of scholars, mainly at the beginning and end of the period
under review, found no evidence of a J source here and attributed the whole
chapter to P and later revisers. The lead in advocating this alternative against
a source-critical solution to the problems was taken by Abraham Kuenen in
one of several critical responses to Wellhausen’s Composition that was first
published in Dutch in 1880 (‘Manna en Kwakkelen’, ThT 14, 281-302; Gn.
tr. in Gesammelte Abhandlungen zur biblischen Wissenschaft [Freiburg and
Leipzig, 1894], pp. 276-94). Kuenen reaffirmed his view in the revised edition
of his Historisch-critisch onderzoek in 1885 (ET The Hexateuch [London,
1886], pp. 331-32; Gn. tr. Historisch-kritische Einleitung 1.1 [Leipzig, 1887],
p. 317), but in 1889 Wellhausen politely but firmly rejected Kuenen’s alter-
native approach (Composition3, pp. 325-29). Kuenen’s argument is initially
traditio-historical, arguing that the negative view of the manna in Num.
11.6(-9) and 21.5 (and according to him also in Deut. 8.3 [and 16?] and 29.5)
must be more original than its, to him, idealised description in Exodus 16. In
any case the Sabbath observance insisted on in vv. 27-30 as well as vv. 22-
26 was impossible in JE and must be an addition made when the narrative
was located after the law-giving at Sinai, like vv. 32-35. The language of the
chapter was predominantly that of P (though Kuenen acknowledged that this
was not the case in vv. 4b, 15a and 28-29), and its purpose was to give the
provision of manna the positive evaluation and prominence which the older
tradition in Numbers 11 had denied it. The disorder in the Priestly text in vv.
6-12 was due to the clumsy insertion of vv. 4-5: vv. 11-12 had originally stood
at the beginning of this section.
Two years after Kuenen’s original article the young Adolf Jülicher, in his
long study of the sources of Exodus chs. 7–24 (‘Die Quellen’), came to very
similar conclusions to Kuenen about Exodus 16, though partly with different
arguments (pp. 279-94). He too found no place for JE in the chapter, attrib-
uting most of the chapter to P and (more significantly in the light of what was
to come later) vv. 4-5, 6-7, 20, 27-30 and apparently 32-34 to a late Deuter-
onomistic redactor (pp. 288-90). Wellhausen thought highly of Jülicher and
remained on friendly terms with him from 1880 to near the end of his life (see
the correspondence in R. Smend [ed.], Julius Wellhausen: Briefe [Tübingen,
2013], esp. pp. 78, 134-35, 319-20, 651), even though he would not have
agreed with his analysis of this chapter any more than with Kuenen’s.
The torch remained with Wellhausen and his successors for almost a
century: apart from the special case of B.D. Eerdmans,4 only some detailed
observations of Kuenen and Jülicher found their way into the (varied)
mainstream analyses of the chapter. But in 1974 E. Ruprecht chose Exodus 16
as the place to say his ‘Farewell to the Yahwist’ (‘Stellung und Bedeutung’).
To be sure, the Yahwist for him means the minimal ‘remains’ found in the
4
On Eerdmans see Houtman, Pentateuch, pp. 173-78; his conclusions about
the present chapter are summarised by Maiberger, Das Manna, pp. 52-54.
426 EXODUS 1–18
present text by Noth, Coats and Fritz (see above), which it was not difficult
to describe as fragments between which ‘die Phantasie gewaltige Lücken
ausfüllen müsste’ (p. 271). Ruprecht’s chosen method insisted that form-
critical considerations must take precedence over Literarkritik, a popular if
one-sided slogan of the time and one which it is not immediately easy to see
as the basis for Ruprecht’s analysis. He chooses to begin, not at the beginning
of the chapter but in the middle, with the Priestly section about the provision
of manna for the Sabbath (vv. 22-26), and then asks what it ‘needs’ (and does
not need) to make a complete narrative. His conclusion is that it needs v. 30
(and 35a?) as a conclusion, vv. 16-21 (sic) as the normal pattern for days other
than the Sabbath, and vv. 2-3, 6-7(8), 9-15 to indicate the people’s need and
its satisfaction in general terms. Everything else is dispensable and therefore
secondary, and vv. 4-5, 28-29 and 31-32 constitute a Deuteronomistic layer of
expansion overlaid on the Priestly narrative, with vv. 33-34 and 35b-36 being
even later additions.
This analysis of the chapter received some early support (Fritz, Tempel
und Zelt [WMANT 47; Neukirchen, 1977], p. 2 n. 10; L. Perlitt, ‘Wovon der
Mensch lebt (Dtn 8,3b)’, in J. Jeremias and L. Perlitt [eds.], Die Botschaft
und die Boten [FS H.W. Wolff; Neukirchen, 1981], pp. 403-26 [408-409]),
but it was rejected by W.H. Schmidt (Exodus, Sinai und Mose, pp. 97-99).
The exclusion of the Yahwist is shared by Maiberger’s much more methodical
work (Das Manna, pp. 87-142, esp. 134-42). But his original P narrative is
much more limited in scope (vv. 1-3, 6-7, 9-15, 21, 31, 35a) than Ruprecht’s
and he regards the Sabbath section in vv. 22-26 as being not the centre of the
story but the first of three layers of expansion (of which vv. 4-5 and 27-30 are
the last). Blum does not go into detail, but agrees that vv. 4 and 28 are part
of a post-Priestly redactional layer, while allowing the possibility that some
older material might have been used in such additions (Studien, pp. 158, 361,
378). More recently L. Schmidt’s two studies have developed an account of
the chapter in which v. 1aα is the only old element and the original Priestly
narrative comprises vv. 1*, 2-3, 9-15 (except for 14bβ), most of vv. 21-26
and v. 35a, with vv. 4-5, 27, 29-30 being from the ‘Pentateuch redactor’
(Studien zur Priesterschrift, pp. 36-45; ‘Die Priesterschrift in Exodus 16’,
ZAW 119 [2007], pp. 483-98). Albertz has a broadly similar view but manages
also to accommodate vv. 6-7, 16-20, 27 and 30 within the earliest Priestly
composition: vv. 4-5 and 28-29 are taken to be insertions by the same ‘Malak-
redaction’ that was responsible for 15.25b-26 (pp. 262-65). This is not so far
from Ruprecht’s proposal.
If this second group of analyses makes a Priestly narrative the main,
indeed the only independent contribution to the text, several of the very
varied third group, by contrast, expand both the extent and the importance
of the non-Priestly narrative within the source-materials available to the
redactor(s). But this is not true of the earliest of them: Carpenter/Harford-
Battersby, followed by McNeile, were close to Kuenen but found fragments
16.1-36 427
of an E account in vv. 4 and 15, while Holzinger added v. 31 and perhaps

(with Wellhausen) vv. 19-20 and 21b to the remnants of JE (which he did not
attempt to divide). Smend (pp. 148-51) and Eissfeldt were the first to greatly
expand the non-Priestly contribution (divided between their two ‘Yahwist’
sources), because they saw the provision of two kinds of food as a defining
characteristic of the Priestly narrative and therefore confined it to parts of
vv. 2-14. Rudolph (pp. 34-36) accepted Kuenen’s approach but turned it on
its head by making the J narrative the only continuous account (vv. 1a, 2*, 3,
4abα, 13b-17, 21, 31), leaving a lot of secondary material (much of it Priestly
in character). This is not too different from the more recent analyses of Levin
and Van Seters. The former (pp. 352-55: cf. Kratz, Komposition, pp. 246-47)
finds no independent Priestly narrative in this chapter (although he generally
does elsewhere in Exodus) and attributes sections with a ‘Priestly’ character
to several very late layers of redactional expansion. For his (exilic) J, which
includes an older form of the narrative, there remain vv. 1a, 4a, 13b-15, 21
and 31: v. 4b and all references to the Sabbath are among the late redac-
tional additions. Van Seters’s account (pp. 181-91) at least has the merits of
simplicity, with all the text allocated either to an original J narrative or to a
single Priestly reworking of it. To provide his J (also exilic) with an appro-
priate introduction, Van Seters ‘borrows’ vv. 1-3 (apart from their Priestly
overlay) and 6-7 from what is ascribed to P by almost all other critics; for the
rest it includes vv. 4-5, 13b-15, 21, 27-31 and 35a. The remainder is P. The
exposition of the ‘murmuring theme’ in vv. 2-3 and 6-7 is compared to other
passages in J (pp. 185, 230, 368-69: but the case for regarding Num. 14.2-3
and 16.3-4 as J rather than P is very disputable) and the exilic date for J is held
to follow from its claimed dependence on Deuteronomy 8 and Ezekiel 20.
D. Frankel’s study is part of a larger and in many ways valuable examina-
tion of all the ‘murmuring’ episodes in the Priestly source (The Murmuring
Stories: see pp. 63-117). Here parallel accounts from J and P are envisaged,
and Frankel is particularly critical of Van Seters’s approach to the non-Priestly
(‘J’) text as a product of the exilic period (pp. 57-61): for him it is substan-
tially pre-exilic in origin. In this he is similar to the source-critics examined
above, but he stands apart from them especially in his treatment of the Priestly
material, which is in part closer to Kuenen and his followers and in part distinc-
tively his own. He begins by arguing that what most have seen as the heart of
the Priestly account in vv. 10b-14a (and also associated material in vv. 3aβ, 7a
and 8) is in fact one of the latest elements in the chapter. He sees a contradic-
tion between vv. 9-10a, where the presence of the Tabernacle is presupposed
(cf. ‘Come near before Yahweh’), and the appearance of Yahweh’s glory in
v. 10b in ‘the wilderness’, not as later in the Tabernacle. The latter fits the
story’s present position before the revelation at Sinai, but this is for him a
(much) later stage in its development than its existence as an independent
narrative set in the vicinity of the Tabernacle. It is this which formed the
‘Early Priestly Story’, not yet part of the Priestly source, which comprised vv.
428 EXODUS 1–18
2, 9-10a, 14b-15a, 31, 33, 34-35a, 36 (pp. 90, 109). Unlike the non-Priestly
account (vv. 4abα, 5, 21, 27-28a, 29-30: cf. pp. 73-81) this included nothing
about the institution of the Sabbath, it was a simple story about the provision
of manna to sustain the people. As such it reflects the oldest surviving form
of the story and it must have a pre-exilic origin. A Priestly editor combined
this with the non-Priestly version, still as an independent story (here Frankel
follows Cassuto, pp. 188-90), adding the Sabbath and other legislation in
vv. 16-20 and 22-26; and finally, with further additions, it was incorporated
into its present position in the Pentateuchal narrative (pp. 96-117). There
seems to be no place here (or in Frankel’s treatment of the other Priestly
murmuring stories) for a continuous independent Priestly narrative: by the
time murmuring stories are incorporated into a comprehensive account of
Israel’s origins, they have already been combined with their non-Priestly
counterparts. Such wider issues about the composition of the Pentateuch as a
whole evidently lay outside the scope of Frankel’s study, but its overall tenor
again justifies his inclusion among the ‘loners’ of our survey.
In the face of such widespread and continuing disagreement

about the literary origins of the chapter it is necessary to be cautious
in one’s conclusions about it and to some extent to be content with
broad results which have extensive support. It is often where details
in the text are made to bear excessive weight and where attempts
are made to define too closely the stages through which the text has
passed that disagreements arise. Sometimes, especially among what
we have called the ‘loners’, an idiosyncratic approach to the Penta-
teuch as a whole may distort judgements about this particular text.
It is, first of all, very widely agreed that much of the chapter comes
from the Priestly tradition in a broad sense. As is shown in detail in
the Explanatory Notes, distinctively Priestly language and concerns
appear in most sections of the chapter (especially in vv. 1-3, 6-12,
16-20, 22-24, 32-35). Cross-connections with these sections make a
Priestly origin likely for vv. 13a, 25 and 36 too. Within the Priestly
material v. 8, which is an incomplete sentence based on vv. 6-7 and
12 but introduced by its own speech-formula, is clearly a secondary
expansion which disrupts its context. It has often been suggested
that either vv. 16-20 or vv. 22-25(26), or both, and vv. 32-34 are
also additions to an originally shorter Priestly story. But in these
cases the arguments, such as they are, are not decisive. Still less are
they when attempts have been made to detect layers of composi-
tion within these sections. Nevertheless it is possible to imagine the
story without them and we should not rule out the possibility that
one or more of them was not part of the original story.
16.1-36 429
It is equally well recognised that vv. 4-5, 13b-15, 21, 27-30 and
31 do not exhibit distinctively Priestly characteristics, while vv. 4-5
and 27-30 duplicate features which appear in vv. 11-12 and 22-25.
As Wellhausen first observed, v. 27 also seems (especially in the
Hebrew text) to begin an account of the Sabbath day after this stage
in the narrative has already been reached in vv. 24-25. Verses 4-5
and 27-30 are therefore generally seen as a separate element in the
chapter, connected by a common concern with the Sabbath and
Yahweh’s instruction or law(s) and by the use of the non-Priestly
expression ‘the people’ for the Israelites, as well as by Yahweh’s
provision of ‘bread’, which is a constant theme throughout the
chapter. But since the days of Wellhausen and Kuenen it has been
debated whether this ‘separate element’ comprises extracts from
a complete parallel version of the manna story or additions made
by a redactor to introduce his own concerns into the main Priestly
account. The strongly theological character of these verses certainly
makes the latter explanation a serious possibility, and the references
to law(s) have led many to see here the activity of a specifically
Deuteronomistic redactor.5
But before attempting, at least provisionally, to resolve this issue
it is important to consider whether vv. 4-5 and 27-30 are the only
verses in the chapter which belong to this ‘separate element’. A
number of scholars have associated vv. 13b-15 with it (see above)
and from a linguistic point of view there is nothing against this. On
the other hand it is difficult to separate v. 13b (and 14) from v. 13a,
which must be from P since only its account speaks about meat as
well as bread, and some mention of the coming of manna is needed
to prepare the way for the Priestly instructions in v. 16. Certainty
is impossible, but it is most likely that vv. 13b-15 are from P. It has
also often been suggested that the first half of the repetitious v. 35 is
non-Priestly; but in the Explanatory Note we take the view (shared
by a number of others) that the whole verse is Priestly, displaying
one of its typical stylistic characteristics. On the other hand, if
vv. 13b-15 are Priestly, it would be strange to have the very similar
statements of v. 31 in the same source and v. 31 might well be the
5
Wellhausen already noted that ‘Der Ton in v. 27-30 erinnert an das
Deuteronomium’, while continuing to maintain that these verses belonged to JE
(Composition, pp. 79, 329); Jülicher unambiguously spoke of ‘Rd’ (‘Die Quellen’,
288-89).
430 EXODUS 1–18
original conclusion of the ‘separate element’ and a further doublet.

More important, v. 21 stands out in its Priestly context for its repeti-
tiveness and its inconsistency with v. 23 (see the Explanatory Note)
and it too should be attributed to the non-Priestly text, as many
commentators have seen. It is also likely that v. 26 is not the conclu-
sion of the Priestly Sabbath legislation but part of the non-Priestly
account (see the Explanatory Note): nothing in its wording is
against this. Since it would follow v. 5 perfectly, it is likely that v. 21
originally stood after it: its present position is due to the compiler’s
understandable wish to keep all text about the weekday provision of
manna together before proceeding to the Sabbath.
There is then, even without vv. 13b-15, more to be reckoned with
when it comes to determining whether the ‘separate element’ is from
a parallel account or the addition of a redactor. To use an expression
favoured by supporters of the latter view there are more ‘fragments’
than they usually recognise (perhaps because they have picked out
the verses that are most amenable to a redactional explanation)
and taken together they comprise a virtually complete alternative
version of the story (vv. 4-5, 26, 21, 27-30, 31), which duplicates
several features of the Priestly version. Already this makes the case
for an extract from a source stronger, perhaps even compelling. In
addition the quails story in Numbers 11, in which there is no sign
of P, takes its departure from the people’s dissatisfaction with their
monotonous diet of manna (vv. 4-9).6 This implies knowledge of
an account of the provision of manna in what is probably an ‘old’
narrative and it would be surprising if some trace of such an account
were not preserved in the wilderness narrative, especially as there
is a further passing reference to manna in Num. 21.5 (not to speak
of Deuteronomy 8, where familiarity with such a narrative is taken
for granted).7
6
The crucial verses are vv. 4-6 and they are clearly presupposed in vv. 10-15.
Verses 7-9 are a parenthesis and are related in some way to both non-Priestly (v. 7:
cf. Exod. 16.31) and Priestly sections (vv. 8-9: cf. Exod. 16.13-14, 23) of Exodus
16. But even if the whole of this parenthesis is a late addition to Num. 11 based on
the combined narrative of Exodus 16, that still leaves Num. 11.6 as a reference to
manna in the older narrative.
7
See further Wellhausen’s closing response to Kuenen’s theory in Composi-
tion3, pp. 328-29, which still carries considerable weight.
16.1-36 431
It is more difficult to be precise about the literary origins of the

non-Priestly manna story. Some parallels of vocabulary with the J
source have been pointed out (e.g. Fritz, Israel in der Wüste, p. 9
n. 3; Wellhausen, Composition3, p. 329, more cautiously spoke of
them as ‘jehowistisch’, i.e. with JE), but they are hardly sufficient
to exclude an Elohistic origin for these verses.8 The latter receives
strong support from the key theological term ‘test’ in v. 4, as in
15.25b (for discussion see the introduction to 15.22-27), but it
occurs in a clause which has been widely suspected of being (with
v. 28) a Deuteronomistic addition. This suspicion arises from the
use of the words ‘law(s)’ (Heb. tôrāh/tôrôt) and ‘commandments’
(miṣwôt), which are certainly common in Deuteronomy. In v. 4
tôrāh can be translated ‘instruction’, which fits the specific and
isolated guidance which is given in v. 26. But the plural forms in v.
28 can hardly be explained in this way, even if (see the Explanatory
Note) the language is hyperbolic. Still, in view of the occurrence of
such legal language in Hos. 8.12 (plural) and the north Israelite Ps.
81.4-5, it should not be assumed that such expressions are impos-
sible before Deuteronomy.9
The chapter thus comprises mainly (see below on v. 1) portions
from JE (more likely E than J) and from P. A summary of each
account, so far as it is preserved, will clarify the differences between
them and the way in which the literary form of the tradition devel-
oped (the latter will have been more complex if the Priestly account
was built up in several stages). In JE (vv. 4-5, 26, 21, 27-30, 31)
the promise of Yahweh to provide bread in the wilderness may
originally have been preceded by a reference to the people’s hunger
(or even a complaint), but if so the compiler preferred the Priestly
introduction in vv. 2-3. The provision is introduced as a test of
obedience (v. 4b) and includes a double provision on the sixth day
(v. 5). This receives its explanation in the instruction in v. 26, which
8
The metaphorical use of ‘rain down’ (Heb. māṭar Hiphil) in Exod. 9.18, 23,
which is usually attributed to J, occurs in sections which are from E according to
our analysis (see the introduction to 9.13-35); and ‘the people’ as a designation for
Israel is found in passages commonly attributed to both J and E.
9
The more so when the existence of divinely authorised legal collections like
the Book of the Covenant (cf. 20.22-26), the Decalogue and less certainly Exod.
34.10-27 in pre-exilic times is taken into account.
432 EXODUS 1–18
refers to the seventh day as ‘a sabbath’, when no ‘bread’ will be

provided. The people gather it each morning, before it melts in the
sun (v. 21). On the seventh day, despite the warning, some of the
people go out to look for more, but in vain (v. 27). Their disregard
for Yahweh’s instruction and extra provision leads him (or, if v. 28
is an addition, perhaps Moses) to rebuke the people and remind
them of his twofold gift, of ‘bread’ and a day of rest (vv. 28-29). The
people rest and give the ‘bread’ the name ‘manna’, and its appear-
ance and taste are briefly described (v. 31).
The original Priestly account included vv. 2-3, 6-7, 9-20, 22-25,
32-36, to which v. 1b would make an apt introduction (see below).
The people (referred to as ‘the whole congregation of the Israelites’)
criticise Moses and Aaron for failing to provide them with food and
even for bringing them out from what they regard as a comfortable
life in Egypt (vv. 2-3). Moses and Aaron reply that they will soon
know that it was not they but Yahweh who brought them out of
Egypt and they will ‘see his glory’; for it is really against him that
they have complained, and he has heard it (vv. 6-7). No assurance
is given at this stage that Yahweh’s appearance will do them good
rather than harm (v. 8 is a later addition) and the summons to meet
Yahweh (v. 9) leaves the outcome open. Only when Yahweh’s glory
is seen outside the camp and he speaks to them through Moses is it
made clear that he is not angry with them but will provide them with
special desert food (vv. 10-12), which comes in the form of quails
and a strange substance which the people do not recognise and call
mān, ‘manna’ (vv. 13-15a). This, Moses says, is the ‘bread’ which
Yahweh promised in v. 12 (v. 15b). Moses then gives instructions
for the family heads to collect each day a set amount (defined by an
otherwise unknown measure) for each member of their households
(vv. 16-18). None of it is to be left overnight; when some Israel-
ites do so it becomes inedible and Moses indicates his disapproval;
nothing more, the point has been made (vv. 19-20). On the sixth day,
evidently without prior warning, the family heads find that there is
sufficient for them to collect twice as much as on other days. When
their tribal leaders report this to Moses, he tells them that Yahweh
has commanded that the following day should be a day of rest and
that what remains after they have prepared enough for the sixth day
should be kept until the next morning (vv. 22-23). When they do so,
it does not become inedible as before and Moses instructs them to
16.1-36 433
eat it, because nothing more will be provided that day (vv. 24-25).
There follows a further divine command, mediated through Moses,
that a day’s portion of manna should be preserved for future genera-
tions to see (v. 32). Moses instructs Aaron to see to this and to put it
in a sacred place (‘before Yahweh’) and Aaron places it in the most
holy place of the Tabernacle (vv. 33-34). The account ends with
a report that this provision continued until the Israelites reached
the land of Canaan and an explanation of the mysterious ‘omer’
measure (vv. 35-36).
Both accounts combine the provision of food in the desert with
the first revelation of the Sabbath to Israel. The murmuring motif is
much less prominent in this narrative than in those which surround
it: it is not present in JE and it disappears from P after v. 12 (cf.
Coats, Rebellion, p. 88; curiously he then concludes [p. 96] that
‘This narrative is dominated by the murmuring motif’!). Even in
P the focus is more on the side-motif of who is responsible for the
Exodus and the leadership claims of Moses and Aaron (vv. 3, 6-7:
Coats, ibid., pp. 89-93). In JE the episode is presented above all as
a test of the people’s obedience, which some of them fail (v. 27).
In P the account is much longer and more detailed. This is partly
due to the extended treatment of the people’s initial complaint,
which is introduced in vv. 2-3 and referred to further in vv. 6-7 and
9-12. The initial uncertainty about how Yahweh will respond to it
prepares for three further Priestly narratives which are structured in
a similar way: the episode of the spies (Num. 13–14*), the rebel-
lion about the extent of the priestly hierarchy (Num. 16–17*) and
the complaint about the lack of water at Meribah (Num. 20.1-13),
all of which end with a painful outcome, either for the people or for
Moses and Aaron.10
10
This pattern seems first to have been identified by R. Rendtorff in 1961 and
was more fully explored by C. Westermann in 1971 (for references see Blum,
Studien, p. 267 n. 145). Childs (pp. 279-80) used the first three examples to
provide an initial rebuttal of the common view that the narrative sequence in vv.
6-12 was not original and needed to be amended either by rearrangement or by
the omission of vv. 6-7(8); see further L. Schmidt, Studien zur Priesterschrift, pp.
35-206, and the idiosyncratic but valuable treatment in Frankel, The Murmuring
Stories.
434 EXODUS 1–18
Five more expansions of the basic themes of food provision

and the Sabbath are also introduced in the Priestly version. First,
a parallel provision of meat, in the form of flocks of quails, that
appears later in the non-Priestly narrative (Num. 11), is anticipated
and described in the first half of the narrative (vv. 3, 6[, 8], 12,
13). Secondly, detailed instructions for the gathering of manna on
weekdays are given and (for the most part) fulfilled (vv. 16-20).
Thirdly, the procedure for preparing and eating food at the end of
the week is carefully spelt out, with emphasis on the consultation of
Moses and the people’s obedience to Yahweh’s command (vv. 22-
25). Fourthly, arrangements are made and carried out for the pres-
ervation of a day’s portion of manna in a holy place (vv. 32-34).
Finally, the provision of manna is said to have continued through-
out the wilderness journey and an explanation of the measure used
for it is given (vv. 35-36). Most of these additions are concerned
only with the manna, which therefore remains (as it was in the non-
Priestly version: cf. Deut. 8) the dominant theme of the narrative.
The provision of quail, which has been stripped of the unpleasant
consequences which it has in Num. 11, seems to be thought of as no
more than a short-term addition (it is significantly not mentioned in
v. 35), as indeed one might readily deduce from Numbers 11.
Not surprisingly, even at a time when the full tabernacle cultus
has not yet been set up, language and other features which belong
to it frequently appear (vv. 7, 9-10, 12, 19, 33-34: see the Explana-
tory Notes). It is these features especially (but also the references
to the Sabbath) which have repeatedly led to the suggestion that
the Priestly version of the story (or even the whole chapter) was
originally placed after the Israelites’ departure from Mount Sinai
(so first de Wette apparently [Beiträge zur Einleitung in das AT,
2 (Halle, 1807), pp. 222-24]: a recent advocate of this view is J.S.
Baden, ‘The Original Place of the Priestly Manna Story in Exodus
16’, ZAW 122 [2010], pp. 491-504, where reference is made to
others who have adopted it). This is probably unnecessary: the most
blatant ‘anachronism’ (in v. 34) can be seen as a parenthetical antic-
ipation of what happened later, like v. 35, while the ‘discovery’ of
the Sabbath before Sinai is not incompatible with its inclusion later
in the Decalogue and elsewhere. In other cases the Priestly writers
must have envisaged some form of the ‘accompanying presence’
of Yahweh with the Israelites on their journey (as the non-Priestly
narrative did in 13.21-22; 14.19-20, 24), even before the construc-
tion of the Tabernacle.
16.1-36 435
Verse 1, the itinerary-note which with 17.1 frames the whole,

requires separate discussion. It bears clear signs of Priestly author-
ship (‘the whole congregation [Heb. ʿēdāh]; the date-formula in v.
1b) and it has commonly been attributed in its entirety to P. But the
‘join’ to the similar verse about Elim in 15.27, which has no Priestly
characteristics, raises doubts and Eissfeldt reasonably saw it as an
older non-Priestly verse which had been overlaid with Priestly
features when the Priestly manna story was inserted (Hexateuch-
synopse, pp. 37, 43-44, 139*; cf. Rudolph, pp. 34-36, 275; Levin,
p. 352). The underlying itinerary-note was probably taken, with the
rest of the ‘string’ to which it belongs (15.27; 17.1*; 19.2a), from
a fuller itinerary by a redactor with a Deuteronomistic perspective
on the wilderness journey (see the Excursus on ‘The Wilderness
Itinerary’ in the introduction to 12.28-42, 50-51; and my ‘The
Wilderness itineraries and the Composition of the Pentateuch’).
Verse 1b (the date-formula) may well have been the original intro-
duction to the Priestly story. Recently L. Schmidt has proposed that
only the words ‘They departed from Elim’ are from the older narra-
tive and that (since he does not think there ever was a non-Priestly
manna story at this point) they were originally joined to the arrival
notice ‘and they camped at Rephidim’ in 17.1 (‘Die Priesterschrift’,
pp. 484-86; cf. Albertz, p. 263). The second half of the itinerary-
note and the rest of the verse (and also 17.1a) were then added to
provide a location for the Priestly manna story. There are several
reasons for rejecting this analysis. Schmidt’s linguistic observation
about the postponement of the subject in 16.1a overlooks an estab-
lished feature of narrative style (see Note a on the translation); the
fragmentation of the two itinerary-notes is artificial and improb-
able (‘the Wilderness of Sin’ is not mentioned in the P story, only
‘the wilderness’ in vv. 2 and 10); and we have provided reasons
above for finding a non-Priestly manna story in the chapter to which
the two itinerary-notes (in full) could have been attached by the
redactor.
The occurrence of manna (and indeed the quails introduced at
this point by P) can plausibly be associated with natural phenomena
which have been observed in the Sinai peninsula in both ancient and
modern times (see the Explanatory Note on vv. 13-15). No doubt,
as with the features that lie behind the narratives in 15.22-27 and
17.1-7, stories about them told by travellers circulated in the neigh-
bouring regions from early times (cf. Gressmann, Mose, pp. 124-45;
Noth, ÜGP, pp. 129-31) and already at this stage they would very
436 EXODUS 1–18
likely have been seen as a provision by their gods for those fortunate
enough to be in the right place at the right time. Even the earliest
traditions that began to be formed around the memories of Israel’s
origins (or more precisely the origins of those ‘proto-Israelites’ who
either were in Egypt or visited Sinai as the mountain sanctuary of
their god Yahweh) may well have included such stories. In their
extant form the oldest written accounts (JE here; the earliest portions
of Num. 11) include didactic elements as well, whose antiquity is
impossible to determine. The observance of the Sabbath as a day of
rest and worship is attested as early as the eighth century B.C. (Isa.
1.13; Hos. 2.13; Amos 8.5) and does not seem to be a new feature
then. An aetiology of its origin could well have appeared in Israel’s
origin traditions by this time: it is not necessary to associate it only
with a time when its neglect was criticised (as in Ezek. 20.21 etc.)
or when it had become especially prominent as a religious institu-
tion in Israel (as in Isa. 56.3-8; cf. Gen. 2.1-3).
The provision of food to Israel’s ancestors in the wilderness
was another of the marks of Yahweh’s care for them in the past (cf.
Deut. 8.3, 16; Pss. 78.23-28; 105.40) which must have encouraged
the belief that he (and not Baal) was also the provider of food for
them in a more fertile place (cf. Deut. 8.17-20). Both accounts in
Exodus 16 speak of miraculously abundant provision (‘rain’ in v. 3;
‘in abundance’ in vv. 8 and 12; for the quails even more so in Num.
11.31-32). In grace Yahweh gives even to those who deny him (v.
3), but he expects his people in need to ‘draw near’ to him (v. 9: cf.
2.23-25). The special provision for the Sabbath (vv. 5, 22, 29) is an
assurance that abstinence from work on it will not lead to hunger.
Each account lays emphasis on one side of the practice. For JE it is
above all a gift from Yahweh to the people as much as the manna
(v. 29), even when it is a commandment whose observance Yahweh
requires (v. 28); in P it is a holy day for Yahweh as well as a time
of rest (vv. 23, 25). In the combined account it is clearly both. The
Sabbath ‘was made’, appointed by God, but it was ‘made for human
beings’ (Mark 2.27), here at least in the first place for Israel (‘you’).
1 [All the congregationa of] [the Israelites set out from Elim and
came to the Wilderness of Sin, which is between Elim and Sinai,]
on the fifteenth day of the second month of/after their departureb
from the land of Egypt. 2 All the congregation of the Israelites
complainedc against Moses and Aaron in the wilderness. 3 The
16.1-36 437
Israelites said to them, ‘If only we had diedd by Yahweh’s hand in

the land of Egypt when we sat by the pot of meat and ate bread
in abundance! For you (pl.) have brought us out to this wilder-
ness to kill this whole assemblye with hunger.’f 4 Yahweh said to
Moses, ‘I tell you, I am going to rain downg food(?) for you from
heaven and the people shall go out and gather a day’s portion
each day, so that I may test them, to see whether they will liveh by
my instruction or not. 5 On the sixth day, when they prepare what
they bring in, it will be double what they gather daily.’ 6 Moses
and Aaron said i to all the Israelites, ‘When evening comes you
will knowj that it was Yahweh who brought you outk of the land
of Egypt, 7 and when morning comes you will seel the glory of
Yahweh, because he has heard your complaintsm against Yahweh –
for we, what are wen that you complaino against us?’ 8 [Moses
said, ‘When Yahweh gives you meat to eat in the evening and
bread in abundance in the morning, because Yahweh has heard
your complaints with which you are complaining against him –
for we, what are we?p Your complaints are not against us butq
against Yahweh.’] 9 Moses said to Aaron, ‘Say to all the congre-
gation of the Israelites, Come near before Yahweh, for he has
heard your complaints’. 10 When Aaron spoke to all the congre-
gation of the Israelites, they looked towards the wilderness, and
there the glory of Yahweh appeared in the/a cloud.r 11 Yahweh
spoke to Moses as follows: 12 ‘I have heard the complaints of the
Israelites. Speak to them as follows. Between the two eveningss
you shall eat meat and in the morning you shall have abundance
of bread and so you will know that I am Yahweh your God.
13 In the evening quailst came up and covered the camp. In the
morning there was a layer of u dew around the camp, 14 and after
the layer of dew had gone up, there on the surface of the wilder-
ness was a powdery dustv, fine like hoar-frost on the ground.
15 The Israelites saw it and said to each other, ‘It is man(na)’w,
because they did not know what it was. Moses said to them, ‘It
is the breadx which Yahweh has given you to eat. 16 This is what
Yahweh has commanded you: Each of you gather of it what he
can eat, an omer per head of the number of your persons – you
shall each take for those in your tent.’ 17 The Israelites did this,
one gathered more and another less. 18 They measured it by the
omery: the one who gathered more did not have too muchz and
the one who gathered less did not have too littlez. Each gathered
according to what he could eat. 19 Moses said to them, ‘No
one must let any of it remain until the (following) morning’.
20 But someaa did not hearken to Moses and let part of it remain
438 EXODUS 1–18
till the morning, and it became full of wormsbb and stank, and
Moses was angry with them. 21 They gathered it every morning,
each according to what he could eat, but when the sun became
hot it would meltcc. 22 On the sixth day they gathered double the
amount of breaddd, two omersee per person, and all the leaders
of the congregation came and told Moses. 23 He said to them,
‘It is whatff Yahweh has spoken of: tomorrow is a day of restgg,
a holy sabbath for Yahweh. Bakehh whatff you want to bakeii and
boiljj whatff you want to boilii, and putkk all the surplus aside
for keeping until the morning. 24 So they put it aside till the
morning, as Moses had commanded, and it made no stink and
there were no worms in it. 25 Moses said, ‘Eat it today, for
today is a sabbath for Yahweh: today you will not find it outside.
26 Six days you shall gather it, but on the seventh day is a
sabbath: there will be none on it’. 27 On the seventh day some of
the peoplell went out to gather, but they found none. 28 Yahweh
said to Moses, ‘How long do you refusemm to keep my command-
ments and instructions? 29 See that Yahweh has givennn you the
sabbath: that is why he gives you bread for two days on the sixth
day. Stay each of you where you are; no one shall go out from his
place on the seventh day.’ 30 So the people rested on the seventh
day. 31 The house of Israel called its name man(na): and it wasoo
like coriander seed, white, and its taste was like a waferpp with
honey. 32 Moses said: ‘This is what Yahweh has commanded:
an omer-full of it is for keeping throughout your generations, so
that they may see the bread with which I fed you in the wilder-
ness when I brought you out from the land of Egypt’. 33 Moses
said to Aaron, ‘Take a jarqq and put an omer-full of man(na)
there and placerr it before Yahweh for keeping throughout your
generations’. 34 As Yahweh commanded Mosesss, Aaron placed
ittt in front of the decreeuu for keeping. 35 The Israelites atevv the
man(na) for forty years, until they came to an inhabited landww;
it was the man(na) that they atexx until they came to the edge of
the land of Canaan. 36 The omeryy is a tenth part of an ephahyy.

a. Heb. ‫עדת‬, as in vv. 2, 9-10, 22: see Note b on the translation of 12.1-20
and the Explanatory Note on 12.3-5. The subject of the two verbs is the same
and Houtman (p. 328) has noted a number of other cases where the subject is
placed after the second verb (Lev. 1.1; Num. 20.22; Josh. 3.1; 2 Kgs 19.36).
16.1-36 439
While this could sometimes be due to the process of literary composition from
separate documents (as is probable in Num. 20.22), such an explanation is not
always likely and it seems that, at least with closely connected verbs, Hebrew
style allowed the postponement of the subject (see also v. 20 below).
b. Heb. ‫לצאתם‬. For similar expressions cf. 19.1; Num. 1.1; 9.1; 33.38;
1 Kgs 6.1: especially the last of these suggests that the sense of ‫ ל‬here is
‘after’, rather than ‘of’ as in other uses of ‫ ל‬in dates, such as in ‫ לחדשׁ‬just
before.
c. Heb. ‫וילינו‬, from ‫לון‬, on which see Note f on the translation of 15.22-27;
see also Text and Versions here.
d. Heb. ‫מי־יתן מותנו‬, lit. ‘Who will give, i.e. grant, our dying [inf. constr.]?’,
a common formula to express a wish more forcefully than by the jussive etc.
(GK §151a-d; JM §164d); more straightforwardly and probably originally of
a desire for a future benefit but here and in 2 Sam. 19.1 (there also with an inf.
constr.) of a wish relating to the past (expressed by ‫ לו‬in Num. 14.2).
e. Heb. ‫הקהל‬. On its difference in sense and range in use from ‫ עדה‬see Note
l on the translation of 12.1-20 and the Explanatory Note on 12.6-7. ‫ קהל‬is very
rare in the non-Priestly portions of Genesis–Numbers (Gen. 49.6; Num. 22.4;
the related verb in Exod. 32.1: cf. TWAT 6, 1208-209 = TDOT 12, pp. 549-50).
f. Heb. ‫ ָבּרעב‬. The def. art. (cf. ‫ ַבּצמא‬in 17.3) is presumably due to the treat-
ment of ‫ רעב‬as a kind of abstract (‘to represent whole classes of attributes or
states’: GK §126n), which is quite common but by no means universal (cf.
Gen. 12.10; 26.1 etc.).
g. Heb. ‫הנני ממטיר‬. On the use of ‫ הנה‬to ‘reinforce affirmation’ and to draw
attention in a non-visual way see JM §164a and Muraoka, Emphatic Words,
p. 138 (cf. 4.23; 7.17, 27; 8.17; 10.4; 14.17), and on the predicative part. after
‫ הנה‬to represent action in the imminent future GK §116p; JM §121e.
h. Heb. ‫הילך‬, lit. ‘walk’, in a very common metaphorical use of the verb
(BDB, pp. 234-35).
i. Heb. ‫ויאמר משׁה ואהרן‬. Before a compound subject the verb is often in the
sing. (GK §146f): so with Moses and Aaron in 4.29; 7.6, 10; 8.8; 10.3; 24.9.
j. Heb. ‫ערב וידעתם‬. ‫ ערב‬is adverbial and as such is followed by the perfect
consecutive as if it were a full temporal clause (GK §118i, 112oo; waw of the
apodosis in JM §176g).
k. Heb. ‫יהוה הוציא‬, with the prefixed subject indicating emphasis upon it and
the contrast with ‫ הוצאתם‬in v. 3.
l. Heb. ‫ובקר וראיתם‬. See Note j above.
m. Heb. ‫את־תלנּתיכם‬.
ֻ For the defective writing followed by gemination
(as also in vv. 8 [2x] and 9; Num. 14.27; 17.20) see BL §24i, 26i′, 61tη: the
etymologically correct full spelling occurs only in v. 12 and Num. 17.25.
n. Heb. ‫ נחנו מה‬is well represented by LXX ἡμεῖς δὲ τί ἐσμεν; ‫ מה‬almost
invariably comes first in its sentence, and its displacement gives special
emphasis to the word that precedes it. The latter is not really a casus pendens
440 EXODUS 1–18
(as in Num. 16.11: Andersen, Verbless Clause, p. 39), even if it is convenient

to represent it in that way in a written translation. ‫ נחנו‬also appears in v. 8 but
elsewhere only in Gen. 42.11, Num. 32.32, Lam. 3.42 and AHI 1.004.10-11.
The longer form )‫ אנחנ(ו‬is standard in early North-West Semitic (DNWSI,
p. 81: no attestations in Ug.), so ‫ נחנו‬was perhaps a colloquial shortening of
this, the agreement with the presumed Proto-Semitic *na/iḫnw (Moscati,
p. 105) being coincidental.
o. Heb. (Qere) ‫ת ִלּינוּ‬.ַ MT’s vocalisation of the Hiphil imperf. here (and in
Num. 14.36Q; 16.11Q) and of the Hiphil part. in v. 8 (and Num. 14.27 [2x];
17.20) is based on the Aram. spelling of some forms of Double Ayin verbs and
may (cf. ‫ נוח‬Hiphil: see Note kk below) be designed to avoid confusion with
forms of ‫ לין‬I = ‘lodge, pass the night’ (GK §72ee). But in 17.3 the standard
form is used. The imperfect may bear an iterative sense, for which the part.
is used in the secondary expansion in v. 8, but a modal interpretation is also
possible (cf. Joosten, Verbal System, pp. 78, 278-79).
p. Heb. ‫ונחנו מה‬. See Note n above.
q. Heb. ‫כי‬. Here in the sense ‘but’, as often after a negative (BDB, p. 474).
r. Heb. ‫בענן‬. If this verse originated in a separate Priestly document, the def.
art. would have been used there (if it was) in an ‘anticipatory’ sense (cf. LXX
ἐν νεφέλῃ), since the theophanic cloud had not previously been mentioned
(cf. GK §126q). But in the extant text, in which the references in 13.21; 14.24
(cf. vv. 19-20) have preceded, ‘in the cloud’ is preferable.
s. Heb. ‫בין הערבים‬, as in the Passover legislation in 12.6: see Note m on the
t. Heb. ‫השׂלו‬. Older editions of MT (e.g. Letteris) record a Qere ‫השׂליו‬, which
is found in some mss: it (or ‫ )הסליו‬is the spelling in later Heb. and Jewish
Aram. (cf. Jastrow, p. 994). 4QNumb is the only Qumran evidence (at 11.32)
and it reads ‫השלו‬. This must be original: the SP spelling (‫ )השלוי‬resembles that
in Ar. and some forms of Aram. (cf. TgN, Sy) and is presumably secondary.
Since quail have not previously been mentioned, the def. art. will be anticipa-
tory (GK §126q), like that in ‫ שׁכבת הטל‬later in the verse. ‫ שׂלו‬generally appears
as a collective sing. (cf. Num. 11.32; Ps. 105.40), and the single occurrence of
a pl. ‫ שׂלוים‬in Num. 11.31 may be corrupt (cf. BHS).
u. Heb. ‫שׁכבת‬, constr. st. of ‫שׁכבה‬, which occurs outside this passage in BH
(6x in Leviticus, once in Numbers) only with ‫ זרע‬in the senses ‘lying’, then
‘fluid’ (cf. ‫ ְשׁכ ֶֹבת‬4x). Some of the Vss (see Text and Versions) apparently saw
it as a part. or other verbal form, with ‫ הטל‬as the subject, but this is impossible
because ‫ טל‬is masc. The influence of this view can still be seen in Luther,
Tyndale, AV and even RV; but by this time the association with the usual
sense of ‫שׁכב‬, ‘lie’, was established (cf. Gesenius, Thesaurus, pp. 1402-403,
and the later lexica). It is, however, unusual for this verb to be used of a thing
(likewise the cognate nouns apparently). Job 38.37 is an example, though
some have preferred to associate ‫ שׁכב‬there (and even some of the nominal
forms) with a homonym related to Ar. sakaba, ‘pour out’ (cf. Barr, Compara-
tive Philology, p. 137; HAL, pp. 1377-79).
16.1-36 441
v. Heb. ‫דק מחספס‬. ‫ דק‬is usually an adj. (as in its second occurrence in this
verse), but here a noun is required, as in Isa. 40.15. On the (secondary) variant
‫ כחספס‬for ‫ מחספס‬see Text and Versions. Morphologically the word is unique in
BH. It is generally explained from the root ‫( חסף‬which is probably correct: see
below), with a curious kind of reduplication in which not the final stem-letter
or two was repeated at the end (as e.g. in ‫ אמלל‬and ‫)סחרחר‬, but just the middle
letter of the stem: so BDB, p. 341, GK §55k, Bergsträsser §20e note a, HAL,
p. 325, Ges18, p. 378, DCH 3, p. 284, with a variety of unparalleled proposals
for the process behind it. Rather than speculate in the dark it is better to follow
the evidence of other Semitic languages which have verb-forms in which an
s is added to the end of the stem even if it does not already contain one (cf.
Moscati, p. 131, for Ar.; Segal, Grammar, p. 56, for MH; perhaps Jastrow,
p. 175, for JewAram [‫)]בלעס‬. In Ar. such forms have an intensive meaning.
Interpretations of the word have been many and various (see the review
by Maiberger, Das Manna 1, pp. 309-22). The traditional ‘round thing’ of AV
and RV (cf. Luther, Tyndale) goes back to the medieval rabbis and ultimately
to Saadya Gaon, but it has no etymological basis and seems to be a guess on
the basis of the comparison with coriander seed in v. 31. A number of the Vss
and Rashi render ‘peeled’ (see Text and Versions), evidently connecting the
form with the root ‫ חשׂף‬as it is used in Gen. 30.37 of the peeling off of bark
from a stick: in post-biblical Heb. this word was spelt ‫( חסף‬Jastrow, p. 489).
The majority of modern interpreters have followed this approach but preferred
a slightly different meaning of the root which is attested in Aram., Ar. and
Eth., ‘be scaly’ (so Gesenius, Thesaurus, p. 504, BDB, p. 341, Maiberger,
p. 315, RSV, NEB, NJPS, NRSV). HAL, p. 325, and Ges18, p. 378, devel-
oped a suggestion of Michaelis based on Ar. ḫašafa = ‘be frozen, crackle’
and suggested ‘knisternd’ or (Ges18) ‘krystallisiert’ (cf. Noth, pp. 103, 107:
the ET has ‘flake-like’ [pp. 130, 134]; Propp, pp. 595-96). The problem with
both these theories is that they ignore the fact that ‫ מחספס‬is spelt with samekh,
whereas a verb with the meanings suggested would, according to the normal
correspondences of consonants between Semitic languages, be spelt with śîn
in BH.11 A different approach may therefore be preferable. In Jewish Aramaic
‫ חסף‬is used in the Pael for ‘to pound’, e.g. of grain or pepper (Jastrow,
p. 489: not in CAL) and it is presumably the basis for Vulg pilo tunsum. This
verb would be cognate with Phoen. ‫‘ = חסף‬break’ in KAI 1.2 (DNWSI, p. 393:
cf. the idiom with Ug. tbr in KTU 1.6.6.29) and a passive part. would mean
‘pounded, powdery’, a meaning that fits very well with the preceding ‫דק‬,
‘dust’ (so already H. Vincent, ‘Les fouilles de Byblos’, RB 34 [1925], 161-93
[186-87], and JB).
11
Within BH (esp. LBH) there is already some alternation between ‫ ס‬and ‫שׂ‬
(cf. Ezra 4.5; Eccl. 1.17: with ‫ חשׂף‬in Sir. 42.1), but it remains exceptional at this
stage. Ug. ḥsp, cited by Cassuto (p. 195) in support of the meaning ‘revealing
(itself)’, does not mean this, but ‘draw, collect [a liquid]’: DULAT, p. 373.
442 EXODUS 1–18
w. Heb. ‫מן הוא‬. ‫ ָמן‬is clearly the Heb. word for ‘manna’ (the English word
is derived from the Aram. equivalent) – it is so used in a number of other
OT passages (and cf. vv. 31, 33, 35) – and it must be related to Ar. mannu,
‘honeydew’, the name for various natural juices that are exuded by trees in
the Sinai desert and elsewhere, especially in Kurdistan: see the Excursus
in the Explanatory Note on vv. 13-15. But the following words (cf. ‫)מה הוא‬
seem to imply that it was given this name because ‫ ָמן‬meant ‘what?’ (see Text
and Versions for the readings of LXX, Vulg and Sy). The element mn is not
otherwise known in Heb. as an interrogative (despite what Josephus says
in AJ 3.32), but it does so appear in several other Semitic languages (BDB,
p. 577, cited Ar., Aram. [for BibAram. see p. 1100], Mand. and Eth., to which
ESA [Moscati, p. 115], Amorite [H.B. Huffmon, Amorite Personal Names
in the Mari Texts (Baltimore, 1965), p. 231] and Akkadian [AHw, pp. 603,
655-56] must now be added), though almost invariably to mean ‘who?’ The
only instances of ‘what?’ are mān in late Syriac, where it is a contraction from
māʾ dēn, and mīnu in Akk.12 The first of these cannot have been known to the
biblical writers and the latter does not correspond exactly. There is also a little
evidence that mān was an alternative for ‘what?’ in Canaanite (cf. Cassuto,
p. 196; Noth, p. 107, ET, p. 135: some instances in Ug. [Sivan, Grammar,
p. 59; DULAT, p. 560] and several in Amarna Canaanite [e.g. EA 286.5: cf.
Rainey, Canaanite in the Amarna Tablets 1, pp. 111-13]), and mainly on this
basis HAL, pp. 564-65, and Ges18, p. 692, maintain the possibility that such
a word also existed in BH. However, an etymology is not a strong basis for
such an inference when imprecise etymologies are quite widespread in the
OT (cf. Y. Zakovitch, ‘A Study of Precise and Partial Derivation in Biblical
Etymology’, JSOT 5 [1980], pp. 31-50), and other explanations are equally if
not more likely. Possibly the writer did not know any explanation of the word
(as seems to be the case in v. 31); possibly the closeness of sound between
‫ ָמן‬and ‫( ָמה‬or even ‫)מה־נָּ א‬
ַ was enough to suggest that the name was given on
this basis (for further discussion along these lines see H. Schult, ‘Mān hūʾ und
mah-hūʾ in Exodus 16,15’, DBAT 1 [1972], pp. 1-9; Maiberger, Das Manna,
pp. 267-79).
x. Heb. ‫הלחם‬. The narrative has already used ‫ לחם‬of the ‘bread’ which the
Israelites had eaten in Egypt (v. 3) and of what Yahweh promised to provide
in the wilderness (vv. 4, 8, 12). In v. 4 it might have the general meaning
‘food’, as it does occasionally elsewhere (BDB, p. 537, 2 and 3): this may
be its most ancient meaning, as in Ar. laḥmun means ‘meat’, with a different
specialisation of the meaning from Phoen., Heb. and Aram.13 But in vv. 3, 8
12
For Aram. CAL s.v. mn, mnʾ (‘what?’) says: ‘The form is extremely
common in Syriac but relatively rare in Jewish dialects…’
13
In Ug. both specialisations may occur, as well as this more general sense:
lḥm can be baked (ʾpy: KTU 1.14.2.30 par.), but sometimes it seems to mean
‘meat’ (cf. DULAT, pp. 496-97).
16.1-36 443
and 12 it has its usual sense ‘bread’ in contrast to ‘meat’ (‫)בשׂר‬, and here ‫לחם‬
refers specifically to the manna and not to the quails (vv. 13-15). The same
will be true in vv. 22, 29 and 32, and the use of ‫ אפה‬in v. 23 suggests that it
was being made into a kind of bread. Yahweh has promised ‘bread’ and the
manna is the fulfilment of this promise and so is designated by the same word,
even though it did not look or taste like normal bread, as the descriptions in
vv. 14 and 31 make clear: it was more like flour or grain. The fact that it was
baked (v. 23) gave further justification for describing the giving of it as (in its
product) a gift of bread.14
y. Heb. ‫בעמר‬. Here (and also more clearly in vv. 32-33) the ‫ עמר‬is a vessel
that holds the quantity that is elsewhere denoted by this word (as in vv. 16,
22, 36: for its size see the Explanatory Note on v. 36). For the ambiguity
compare ‫איפה‬.
z. The Hiphils are best understood as ‘internally causative’ (GK §53d).
aa. ‫ אנשׁים‬is the subject of both verbs: cf. v. 1 and Houtman, p. 328.
bb. Heb. ‫וירם תולעים‬. ‫ וירם‬is not from ‫רום‬/‫ רמם‬I = ‘be high’, but from ‫רמם‬
II = ‘become wormy’ (cf. Rashi), which only occurs here and may be a
denominative from ‫( ִר ָמּה‬v. 24).15 On the form see GK §67n. ‫ תולעים‬seems
hardly necessary, but is probably used after the analogy of verbs like ‫מלא‬, ‘be
full of’ (see GK §117z; JM §125d: the explanation given in GK §121d note
is unnecessarily complex). For a possible wider sense of ‫ תולע‬etc. see IDB 4,
p. 878, and TgO here.
cc. The perfect consecutives ‫ וחם‬and ‫ ונמס‬are iterative in meaning (cf. GK
§112f-g).
dd. Heb. ‫לחם משׁנה‬. Cf. Gen. 43.12: like regular numerals (GK §134a-c, g)
‫ משׁנה‬may either precede or follow its noun.
ee. Heb. ‫שׁני העמר‬, lit. ‘the two omers’, with reference back to vv. 5 and 16:
see also Note yy below.
ff. Heb. ‫ אשׁר‬is used three times in this verse to introduce an ‘independent’
relative clause (GK §138e): on the first occasion it produces an entirely
acceptable equivalent to ‫ הדבר אשׁר‬in vv. 16 and 32, and as there it introduces
a new revelation and command. Similar formulae do sometimes refer back to
an earlier statement in a narrative (e.g. Gen. 42.14; 2 Kgs 9.36), but in Lev.
10.3(P) the same phrase is used to open a new explanatory revelation.
gg. Heb. ‫שׁבתון‬, ‘rest’, an abstract noun formed from the verbal root ‫שׁבת‬
(cf. ‫ צמאון‬from ‫צמא‬, and other exx. in BL §498b-d), which then designates a
‘period of rest’. Most often it occurs in the phrase ‫שׁבת־שׁבתון‬, underlining the
key feature of the Sabbath day (so in 31.15; 35.2), but the phrase was then
also applied to the Day of Atonement (Lev. 16.31; 23.32) and (like ‫ )שׁבת‬to
the sabbatical year (Lev. 25.4). ‫ שׁבתון‬was used alone for rest days at New Year
14
Vulg generally used panis (sing. or pl.), but has cibus (pl.) in vv. 22 and 29
(cf. ‫ מזון‬in TgNmg here and elsewhere in the ch.).
15
BDB (esp. p. 93) follows the more general sense in Ar. ‘grow rotten, decay’.
444 EXODUS 1–18
and at the beginning and end of the Feast of Tabernacles (Lev. 23.24, 39), and
also to designate the sabbatical year as a ‘year of rest’ for the land (Lev. 25.5).
All these other occurrences are in P and H contexts. Only here does it precede
‫שׁבת‬, to provide an explanation in advance for the religious term in apposi-
tion which, while certainly not new to its first readers, had not previously
appeared in the Pentateuchal narrative (in Gen. 2.2-3 only the related verb
is used).
hh. Heb. ‫אפוּ‬,ֵ with the occasional lengthening of the first syllable where it
begins with aleph (GK §23h, 76d), as more commonly in Aram. (Rosenthal,
p. 47; Stevenson, p. 13). The construction here and in the following words is
a variant of the idem per idem pattern found elsewhere in Exodus: see Note k
on the translation of 4.10-17.
ii. Heb. ‫תאפו‬. For the desiderative use of the imperfect see JM §113n and
Joosten, Verbal System, p. 275.
jj. On the meanings of ‫ בשׁל‬see Note t on the translation of 12.1-20. Along-
side ‫ אפה‬it is more likely that it means specifically ‘boil’ than ‘cook’ more
generally. It could refer to the cooking of the quail meat, but Num. 11.8 says
that the manna could be boiled (‫)בשׁל‬, so it is not certain that the quail are in
mind here.
kk. Heb. ‫ ַהנִּ יחוּ‬is (like ‫ ויניחו‬in v. 24: see also vv. 33-34) another example
of the ‘Aramaising’ pronunciation of some Hiphil forms of Ayin Waw/Yodh
verbs (see Note o above) and in this case the variation from the normal form
again corresponds to a difference of meaning: ‘put, place’ rather than ‘give
rest (to)’ (for the latter see 17.11 and 33.14).
ll. Heb. ‫מן־העם‬. For the partitive ‫מן‬, even without a preceding numeral or
noun, see BDB, p. 580.
mm. Heb. ‫מאנתם‬. On the use of the perfect see Note i on the translation of
10.1-20.
nn. The subject precedes the verb, which usually gives it special emphasis:
the point could be that the practice is not merely of human origin. But
Muraoka has noted that when the subject is God or his representative this
alone sometimes seems sufficient reason for placing it first (Emphatic Words,
p. 35; cf. JM §155ne).
oo. Heb. ‫והוא‬, so that the sentence structure (noun clause linked to verbal
clause by waw) is like that of many circumstantial clauses (JM §159). But
there seems to be no special nuance in the connection here, only the attach-
ment of a (further) description of manna.
pp. Heb. ‫כצפיחת‬. The word is a hapax in BH, and the only similar word is
‫צ ַפּ ַחת‬,ַ ‘(water-)jug’, in some places evidently a portable one (1 Sam. 26.11, 12,
16; 1 Kgs 19.6), so perhaps ‘flask’ and typified by the flat so-called pilgrim
flasks found at Iron Age sites (cf. Amiran, Ancient Pottery, p. 276; BRL2,
p. 184). The cognates listed in the lexica suggest the idea of flatness, which
would support this. In MH a similar word means a thick ‘batter’ that could be
poured on to a flat surface (M.Maksh. 5.9: cf. Jastrow, p. 1296).
16.1-36 445
qq. Heb. ‫ צנצנת‬is another hapax in BH. The later interpretations cited by
Jastrow, p. 1293, are only guesses and modern lexicography has not been
any more successful. The Vss suggest ‘jar’ (see Text and Versions), which
is followed by the medieval rabbis. The addition of ‫ אחת‬is an instance of
the occasional use of the numeral as an indefinite article, which is generally
lacking in Heb. (GK §125b; BDB, p. 25, deriving it from the ‘semi-definite’
use in 1 Sam. 1.1 etc.: for a comparable use of (‫ חד)ה‬in Aram. see BDB,
p. 1079; DNWSI, pp. 33-34). According to JM §137u this usage is found
‘especially in the books of Judges, Samuel and Kings’, and they cite no
instances from the Pentateuch. In many apparent cases in the Priestly laws
a different idiom is involved and the ‘unnecessary’ instances in lists such
as Lev. 16.5; 23.18 may reflect the influence of accounting procedures. But
Exod. 29.1, 3 (cf. Houtman 1, p. 60); Num. 13.23; 15.27 are plausible parallels
in the Pentateuch, and the usage does appear in exilic and post-exilic literature
(cf. Ezek. 8.8; Dan. 8.3; 10.5), possibly as an Aramaism.
rr. Heb. ‫והנח‬. On the ‘Aramaising’ form of the Hiphil and its meaning (as
also in v. 34) see Note kk above.
ss. Heb. ‫ אל־משׁה‬is unexpected after ‫צוה‬, which generally takes a direct
object (38 times out of 39 elsewhere in Exodus according to Sanderson,
Exodus Scroll, p. 58, not including the 12 cases where the object is a
pronominal suffix). There are only two other examples with ‫( אל‬2 Sam. 17.23;
1 Kgs 11.10) and a few more with ‫ל‬, including Exod. 1.22. The variation was
probably due to assimilation to the construction with verbs such as ‫ אמר‬and
‫( דבר‬cf. v. 33). On a variant reading here and a proposal to emend the text see
Text and Versions.
tt. Heb. ‫ויניחהו‬, with waw of the apodosis after the preceding subordinate
clause, here (unusually) a comparative clause as in Num. 1.19 (JM §174b,
176a-c). Sy, Vulg and probably LXX did not recognise this and so linked v.
34a to v. 33, but a third person reference to Moses would be most unusual in
his own words (Sy and one ms. of OL significantly read ‘[to] me’ instead).
uu. Heb. ‫לפני העדת‬. For this phrase cf. 30.36; Num. 17.19, 25. In the
Hexateuch ‫ עדות‬in the sing. is found almost exclusively in Priestly contexts,
where it occurs 34 times (the exceptions are Exod. 32.15 and Josh. 4.16);
on the few occurrences elsewhere in the OT see the Explanatory Note. In P
it consisted of the two ‘tablets’ of the Law (31.18; 34.29; cf. 32.15), which
were placed inside the ark (25.16, 21; 40.20); it was consequently known
as ‫( ארון־העדות‬25.22 etc.). The ancient translation of ‫ עדות‬as ‘testimony’ (cf.
Text and Versions), based on a presumed association with ‫‘ = ֵעד‬witness’, has
remained surprisingly popular until the present day (cf. RSV, JB, NEB, NIV,
REB; cf. BDB, p. 730, HAL, p. 747, Ges18, p. 925, DCH 6, pp. 278-80). But
a ‘testimony’ or ‘witness’ is a remarkable word to use of Yahweh’s laws, and
even more so for the pl. form which occurs, for example, frequently in Psalm
119. In fact two of the recent lexica (HAL and Ges18) have abandoned it in
favour of ‘command, law’ for the pl. form and more appropriate translations
446 EXODUS 1–18
have been adopted for the sing. here in NJPS (‘pact’) and NRSV (‘decree’:
cf. Ges18’s mediating ‘Verordnung, Gesetz als Zeugnis des Rechtswillens
Gottes, priesterl.[icher] Ausdr.[uck] f.[ür] die Sinaigesetzgebung’). There is
a good justification for these newer renderings in the meanings of the under-
lying verb, ‫ עוד‬Hiph. Although it is used a few times of ‘bearing witness’
(1 Kgs 21.10, 13; Mal. 2.14; Job 29.11) or ‘calling as a witness’ (Deut. 4.26;
30.19; 31.28; Isa. 8.2; Jer. 32.10, 25, 44), it occurs over twice as often in more
general senses which have nothing to do with ‘witnessing’. It can mean ‘warn’
(e.g. Deut. 8.19), ‘teach’ (1 Sam. 8.9; Ps. 50.7) and especially ‘command’
(Exod. 19.23; Deut. 32.46; 2 Kgs 17.13, 15; Jer. 11.7; 42.19; Ps. 81.9; Neh.
9.34), in other words any kind of solemn declaration, but especially it seems
the words of a prophet. It should therefore not be regarded as a denominative
of ‫עד‬,ֵ except perhaps for the technical uses noted above. T. Veijola has made
a very convincing case for dissociating the other occurrences from the idea
of witness (‘Zur Ableitung und Bedeutung von hēʿīd im Hebräischen’, UF
8 [1976], pp. 343-51: cf. TWAT 5, 1107-30 = TDOT 10, pp. 495-516), but
he perhaps ties his argument too closely to the derivation of Heb. ‫ עדות‬from
Akk. adê and OAram. ʿdn (both pl. words), which mean ‘treaty stipulations’
and then ‘treaty’, so that ‘command’ would be the primary sense. In any case
‫ העדות‬in P clearly refers to the tablets of the law (cf. above) regarded as the
‘decree’ of Yahweh in the singular.16 On the earlier background to this use of
the expression (and also the anachronism that seems to be involved in its use
here) see the Explanatory Note.
vv. The subject precedes the verb, because this is not the continuation of
the narrative stream (which is in 17.1), but a parenthesis that refers to the
whole wilderness journey (cf. 12.38, and more generally JM §159f).
ww. Heb. ‫ארץ נושׁבת‬. Niphal participles sometimes convey a potential
rather than an actual sense (cf. GK §116e; JM §121i), so ‫ נושׁבת‬could mean
‘habitable’, just as ‫ נורא‬in 15.11 means ‘fearsome, to be feared’. But the other
occurrences of ‫ נושׁב‬all seem to be straightforward passives.
xx. The object unusually precedes the verb, either to create a partial
chiasmus between the two almost synonymous parts of the verse (JM §155ng)
or to emphasise that the manna was all that the people ate during the journey
(JM §155o).
yy. The use of the def. art. with ‫ עמר‬may refer back to its use earlier in
the chapter, but it seems from Ezek. 45.11-12 that at least the better-known
measures were regularly determined by the article in such general statements,
which will explain ‫ האיפה‬here: cf. the criterion of ‘identifiability by the hearer’
proposed by P. Bekins, ‘Non-Prototypical Uses’.
16
NJPS ‘pact’ is based on the treaty connotations of the Akkadian and Aramaic
cognates, but this is less appropriate to P’s covenant theology, which does not
regard the commandments as the conditions of a covenant as Deuteronomy does.
16.1-36 447
Explanatory Notes
1. The long narrative that is to follow is first located, in the Wilder-
ness of Sin, by an itinerary-note of the typical two-part form, with
a double expansion in v. 1aγb. Elim, the point of departure, makes
a connection with the isolated itinerary-note in 15.27.17 16.1 forms
part of an ‘itinerary-chain’ which also includes 17.1 and (part of)
19.1-2. The same sequence, with some additional names, appears in
the extensive itinerary in Num. 33.1-49 (cf. vv. 9-15), from which
it may have been extracted (see the Excursus on ‘The Wilderness
Itinerary’ in the introduction to 12.28-42, 50-51). The presence of
Priestly features here (‘congregation’ and the date-formula, on which
see respectively the Note on vv. 2-3 and below) and in 17.1 (see the
note there) has led many commentators to attribute the ‘chain’ to
P, but in 19.1-2 the ‘two-part formula’ (v. 2a) is clearly separate
from and duplicated by the Priestly note of arrival at Sinai in v. 1.
The Priestly features here (and in 17.1) are therefore secondary and
the result of assimilation to the dominant Priestly character of the
following narrative, as Eissfeldt already saw (Hexateuchsynopse,
pp. 139-46*; cf. my ‘The Wilderness Itineraries and the Composi-
tion of the Old Testament’, pp. 2-3). The original wording, ‘(They)
departed from Elim and came into the Wilderness of Sin, which
is between Elim and Sinai’, will have been added by a redactor
originally to introduce the older non-Priestly manna story of which
sections survive in vv. 4-5 and elsewhere in the chapter (see also
the discussion of 15.27 in the previous section). The words ‘which
is between Elim and Sinai’ are not present in Num. 33.11, where
the geographical situation is clearer, but were perhaps needed here
to avoid the possible misconception (which has recurred in modern
times: e.g. Noth, p. 106, ET, p. 133; Propp, p. 592) that Sin and
Sinai were two names for the same place: they do differ by only a
single final letter and may indeed be connected in some way. The
date-formula at the end of the verse is part of a sequence of dates in
the Priestly narrative (cf. 12.40-42 with v. 6; 19.1; 40.1, 17; Num.
1.1; 10.11) and may originally have been the beginning of its story
of Yahweh’s feeding of his people in the wilderness, which now
begins in v. 2. LXX in fact connects the date-formula grammatically
17
As a result the standard (and ancient) division of the Heb. text includes
15.27 with the first part of ch. 16 (see Text and Versions on ‫ רפאך‬in 15.26).
448 EXODUS 1–18
with the ‘complaining’ in v. 2 (see Text and Versions), and ‘in the
wilderness’ there may have been the only geographical location
originally given for the Priestly version of the story. The reason
for giving such a precise date here may have been, as early Jewish
exegesis suggested (see Text and Versions), to fix the time when
the (unleavened) bread baked in Egypt ran out and was replaced
by the manna. But it may also have been to mark the date when,
as the following narrative in both its versions envisages, Israel was
believed to have begun to observe the Sabbath for the first time (cf.
vv. 22-30).
The Wilderness of Sin is mentioned only here, in 17.1 and in
Num. 33.11-12: there is no useful evidence for its location outside
the Bible.18 The name is unlikely to have anything to do with Sin in
Ezek. 30.15-16 (though the Heb. spelling is the same), perhaps the
older name for Pelusium on the Mediterrranean coast of Egypt (but
the text may be corrupt), or with Sîn the Mesopotamian moon-god,
for which the West-Semitic equivalent often begins with Sh- (see
DDD, 1480-81): there is no evidence in either case for a connec-
tion with the region that is likely to be meant. A location for the
Wilderness of Sin can only be conjectured on the basis of its place
in the biblical wilderness itinerary and a choice between the vari-
ous candidates for the identification of Mount Sinai (which is so
named for the first time in Exodus here: cf. 19.1-2, 18, 20, 23). Even
if the traditional view that the latter was in the south of the Sinai
peninsula is accepted (e.g. my Way of the Wilderness, pp. 63-69),
there remain two competing possibilities, as this part of the itiner-
ary can be mapped on a ‘northern’ or a ‘southern’ route. According
to the former view, favoured by Abel (Géographie 1, pp. 435-36; 2,
pp. 210-13: cf. Knobel, pp. 163-64, and M.-J. Lagrange, ‘L’itineraire
des Israélites du pays de Gessen aux bords du Jourdain’, RB 9 [1900],
pp. 63-86, 273-87, 443-49 [83-86]), the Wilderness of Sin was
inland, at Debbet er-Ramleh, close to Serabit el-Khadem; according
to the latter, which had early support from Egeria/Peter the Deacon
(Wilkinson, Egeria’s Travels, p. 208) and continued to be advo-
cated in the nineteenth century (Robinson, Biblical Researches 1,
pp. 106-107; Stanley, Sinai, pp. 37-39, 69; Palmer, Desert 1, p. 275),
it was on the coastal plain at el-Marḫa close to the mouth of Wadi
18
The equation with Alush (Num. 33.13-14) in Bereshit Rabbah 48 and
elsewhere is based purely on etymology and is of no geographical value.
16.1-36 449
Sidri.19 For identifications based on different presuppositions about

the location of Mount Sinai see e.g. J. Koenig, ‘Le Sinai montagne
de feu dans un desert de ténèbres’, RHR 167 (1965), pp. 129-55
[153-54] (a general name for the volcanic ḫarrat of N.W. Arabia),
and Harel, Masaʿey Sinay, pp. 205-206 (south of Uyun Musa near
the entrance to Wadi Sudr and Jebel Sin Bisher).
2-3. Unlike 15.22-23 and 17.1 this story of divine provision
begins with the people’s complaint: the reason for it is only stated
in their words at the end of v. 3, ‘with hunger’. The complaint
itself becomes a prominent feature of the first part of the story, not
immediately, but in vv. 7-9 and 12 (on the translation ‘complained’
see Note f on the translation of 15.22-27). The people are referred to
as ‘(all) the congregation [ʿēdāh]’ of the Israelites (cf. vv. 9, 10, 22),
a designation that is restricted to the Priestly strand of the Penta-
teuchal narrative (from Exod. 12.3 onwards: see the Explanatory
Note there) and passages influenced by it, including 16.1 and 17.1.
The inclusion of Aaron with Moses as an object of their complaint
reflects the development of his role as a leader of the people in
the Priestly narrative, which continues to be prominent later in the
story (vv. 6, 9-10, 33-34). The complaint itself takes up a tradi-
tional motif, which is found also in the non-Priestly account(s): by
comparison with life in the wilderness the time of slavery in Egypt
is recalled as one of relative comfort and safety, and the people’s
leaders are accused of being responsible for their present peril (in
non-P cf. 14.11-12; 17.3; Num. 11.4-6, 18, 20; 14.3-4; 16.12-13;
20.5; elsewhere in P Num. 14.2; 20.3b-4). The responsibility of
the leaders is highlighted by the words ‘you have brought us out
(sc. from Egypt)’ and it is no accident that this makes the accusa-
tion substitute their role for that which both older tradition and
the Priestly narrative itself (cf. 6.6-7; 7.4-5; 12.42, 51) had given
to Yahweh: hence the responses to this part of the complaint in
vv. 7-8. The contrast with Egypt speaks of two kinds of food, meat
and bread, and this too will be taken up in the two-fold provi-
sion of food later in the story (vv. 8, 12, 13-14). Sitting ‘by the
pot of meat’ no doubt reflects common eating practice at the time
of writing: the meat was boiled on a fire and the family or larger
group used a utensil to take pieces from the cooking pot. Such pots
19
Bodenheimer (Ergebnisse, p. 83) located it a little to the south, near Wadi
Feiran.
450 EXODUS 1–18
are well known from archaeological excavations (see e.g. Amiran,

Ancient Pottery, plates 42, 75, 76) and are mentioned in biblical
narratives (e.g. 1 Sam. 2.12-17; 2 Kgs 4.38-41). According to the
stricter Priestly laws (Lev. 17.2-4) meat was to be eaten as part
of a sacrificial ritual (Deut. 12.15-27 is less restrictive), but this
is not always mentioned in the narratives and may not have been
a universal practice (cf. King and Stager, Life in Biblical Israel,
pp. 67-68: further on meat consumption N. Macdonald, What Did
Ancient Israelites Eat? [Grand Rapids, 2008], pp. 61-79)
4-5. Yahweh’s words to Moses make no mention of Israel’s
complaints (unlike the otherwise similar statement in vv. 11-12)
but they do deal with the same issue, namely Israel’s hunger in the
wilderness. In their present position they can be read as showing
a gracious willingness on Yahweh’s part to overlook the people’s
discontent and accusations and to provide for their needs out of
undeserved kindness and faithfulness, and this may well be what
the compiler of the narrative intended. But their language and in
part their content are very different from vv. 2-3 and the later verses
which are more closely connected with them (especially vv. 6-12).
Unlike v. 12 they speak only of a single kind of food – even if Heb.
leḥem is translated ‘food’ here (as we have done with some hesita-
tion), this differs from its specific use for ‘bread’ as opposed to meat
in v. 3 and the rest of the chapter. The purpose of testing Israel’s
willingness to obey Yahweh’s ‘instruction’ (or ‘law’, Heb. tôrāh) is
not alluded to again except in v. 28, which is part of another section
of the story that sits awkwardly in its present position (see the notes
on vv. 22-30), and the parallels to it are in passages which show
none of the Priestly characteristics of vv. 2-3 and the bulk of this
chapter (see the Explanatory Note on 15.25b-26). Verses 4-5 are
therefore generally thought to come from either a different account
of this episode or, less likely, a later redactional reworking of the
Priestly narrative (see further the introduction to this chapter). The
vivid metaphorical use of ‘rain down’ for other kinds of ‘downpour’
sent by God is paralleled in 9.18, 23; Gen. 19.24; ‘the people’ is a
common expression for the Israelites in the non-Priestly narrative
(e.g. 15.24); and the expression ‘a day’s portion each day’ (debar-
yôm beyômô) appears also in 5.13, 19, but not elsewhere in this
chapter, where other expressions are used. The ‘test’ appears from
the sentence division here (which corresponds to the break between
v. 4 and v. 5) to be the daily collection of sufficient ‘food’ for each
16.1-36 451
day; the additional food collected on the sixth day is more of a

bonus, to ensure that the people do not have to work on the Sabbath
(cf. v. 29). If so, then vv. 27-30 represent an unanticipated violation
of the divine plan. The language of ‘food/bread’ from heaven is
recalled in Ps. 105.40 (cf. 78.23-25) and Neh. 9.15 (cf. v. 20): in all
these passages the meaning of leḥem is clearly ‘bread’. For the later
development of this narrative of divine provision in Judaism and
the New Testament (especially John 6) see P. Borgen, Bread from
Heaven (NTSup 10; Leiden, 1965); B.J. Malina, The Palestinian
Manna Tradition (AGSU 7; Leiden, 1968).
6-7. The response of Moses and Aaron (including the naming of
the latter) does take up the Israelites’ complaint, though explicitly
only at the end (v. 7b). But from the beginning they insist that it is
misdirected, for it was not they who had ‘brought you out of the land
of Egypt’, but Yahweh himself (v. 6b): the word-order of the Heb.
indicates that this is an emphatic correction of the people’s words
in v. 3b (see Note k on the translation). Something will happen ‘in
the evening’ to make them ‘know’ this – it is not clear what at this
stage – and ‘in the morning they will see the ‘glory’ (Heb. kābôd)
of Yahweh. All this is ‘because Yahweh has heard your complaints’
which, it is now explicitly stated, are ‘against Yahweh’ himself.
Taken by itself, as the narrator probably intended his readers to do,
this response gives no clear indication that Yahweh will respond
positively to the people’s direct challenge to him, if anything the
opposite is suggested. In the majority of the Exodus narrative and
most recently in 14.4, 18 it is the recalcitrant Egyptians who are
to ‘know’ who Yahweh is when he ‘gets glory’ (Heb. kbd Niphal)
over them by punishing them. The better expectation of what such
‘knowledge’ could mean that was encouraged in 6.7 and 10.2 might
seem to have been nullified by the people’s ungrateful words.
8. But there follow words (now attributed to Moses alone:
perhaps because he alone has heard Yahweh’s promise in v. 4?)
which, while having much in common with vv. 6-7 (including their
threatening finale), remove any suggestion that Yahweh is going
to punish his people for their recalcitrance: evening will bring the
‘meat’ and morning the ‘bread in abundance’ that they recall having
enjoyed in Egypt (v. 3). There is no doubt that v. 8 is modelled on
vv. 6-7 and is intended to specify their meaning in a way that is
confirmed by what happens in vv. 12-14. In fact even grammatically
the verse is an appendage to what Moses and Aaron have said: it is,
452 EXODUS 1–18
as our translation makes clear, an incomplete sentence consisting

only of a temporal clause followed by a parenthesis (the rendering
of the parenthesis as a main clause [Maiberger, Das Manna, pp.
114-15; Houtman, p 334] is forced). This alone makes it likely
that the verse is a later addition to the Priestly narrative. But by
its premature anticipation of the outcome announced by Yahweh in
v. 12 in very similar terms, it clumsily removes the tension which
vv. 9-10 actually sustain (see below). Different authorship also helps
to make better sense of some minor linguistic variations between
vv. 6-7 and v. 8: the fuller expressions for ‘in the evening’ and ‘in
the morning’ and the use of the participle rather than the imperfect
for ‘you complain’, both in a simple sense more ‘correct’.
9-12. Thus far Moses and Aaron have only told the Israelites
what ‘will’ happen and corrected their misguided complaint about
their lack of food. Now, addressing them with their full designation
as ‘the congregation of the Israelites’ used earlier in vv. 1-3, they
call them to come into Yahweh’s presence to receive his response to
their complaints. Aaron speaks to them on Moses’ behalf, as he had
earlier been appointed to do with Pharaoh (7.1-2; for Aaron’s role as
an intermediary cf. also 4.14-16 and Aaron’s performance of acts of
power in 7.8-10 etc.). This is not established as a regular pattern in
the ensuing narrative, although Aaron is sometimes given specific
tasks to fulfil in connection with the sanctuary, as in vv. 33-34 (see
the note there). Here too the gathering of the people ‘before Yahweh’
is a kind of cultic activity (for the phrase see 16.33; 27.21; 28.12
etc.) and Aaron’s later priestly responsibilities for such activity are
evidently anticipated here, even though as yet there has been no
mention of a set place of worship or his appointment as a priest. In
any case, when Yahweh’s ‘glory’ appears here it is not, as later, in
the camp of the Israelites (40.34-35) but outside it in the wilderness
(v. 10), just as it will first settle on Mount Sinai itself before entering
the camp (24.16-18). Yahweh’s ‘glory’ (Heb. kābôd) is the visible
manifestation of his presence in the Priestly narrative, where it is a
concept of central theological importance (cf. THAT 1, 802-812 =
TLOT 1, pp. 596-602; also TWAT 4, 23-40 = TDOT 7, pp. 22-38). It
is described as ‘like a devouring fire’ (24.17), and when it appears
at the time of the ordination of Aaron and his sons to be priests
‘Fire came out from Yahweh and devoured the burnt offering…’
(Lev. 9.23-24; cf. 10.1-2). It is regularly, as here, associated with
a theophanic cloud and when the tabernacle has been completed,
16.1-36 453
it ‘fills’ it so that Moses is no longer able to enter it (40.34-35:

cf. 29.43). Later, on another occasion when the Israelites had
complained about their situation, Yahweh’s glory appeared there
and he spoke to the people in great severity (Num. 14.10, 26-39*; cf.
16.19-20; 17.7-9). Both the expression and the manifestation which
it describes seem to have their roots in older tradition. Indeed paral-
lels to the latter are widespread in the portrayal of deites and kings
in the ancient Near East (cf. TWAT 4, 30-32 = TDOT 7, pp. 29-31).
The term kābôd appears in cultic texts from the pre-exilic temple,
though perhaps not at first with reference to Yahweh’s presence
there (Pss. 19.2; 24.7-10; 29.1-3, 9; 97.6; Isa. 6.3). But by the time
of Ezekiel, to whom the Priestly conception of Yahweh’s kābôd is
most closely related, the majestic ‘glory’ of Yahweh was evidently
seen as the mode in which his presence in the Jerusalem temple was
to be understood (cf. Ps. 63.3 and perhaps Moses’ request in Exod.
33.18, though significantly it is declined), so that its departure
would leave the temple and the city unprotected (cf. Ezek. 1; 8-11;
and by contrast 43.1-5). From a visual point of view the Priestly
description seems to draw on the tradition of the pillar of cloud and
fire which accompanied and protected the Israelites on their journey
through the desert (cf. 13.21-22; 14.19-20, 24). But for the Priestly
writer (see above) the appearance of Yahweh’s kābôd is as capable
of bearing a threatening note as a reassuring one, and perhaps the
more so when, as here, it is seen outside the camp rather than within
it. It is not until Yahweh speaks to Moses in vv. 11-12 that it is finally
made clear that his hearing of the people’s complaints will lead him
to provide for their needs rather than rebuking and bringing judge-
ment upon them as he had upon the Egyptians. In the end the pattern
established earlier in the Priestly account (2.23-25; 6.5: note the
recurrence of ‘(have) heard [šāmaʿ]’ there) continues and Yahweh
passes over the angry words of his people, as he does again in Num.
20.6-11 when they lack water. The provision of meat in the evening
and bread in the morning corresponds to the eating habits of at least
some Israelites (cf. Gressmann, Anfänge, p. 86: however, his refer-
ence to 1 Kgs 17.6 is based on acceptance of the shorter LXX text).
Two other passages (Gen. 18.6-7; 1 Sam. 28.24) suggest that meat
and bread were eaten together as one might expect. In the older
tradition too Yahweh had responded in this way to his people’s cry
for help, even if it was couched in words of complaint (14.10-12;
15.24-25a; 17.1-6: perhaps also here in vv. 4-5, though any such
454 EXODUS 1–18
beginning of the older version has not been preserved). Two further
features of Yahweh’s response here display its Priestly origin. The
phrase ‘between the two evenings’ (i.e. at twilight) specifies more
precisely than ‘in the evening’ the occasion of feeding with ‘meat’
and coincides with the timing of the slaughter of the Passover victim
and the evening offering in the tabernacle ritual (12.6 [see the
Explanatory Note there]; 29.39, 41): perhaps both are deliberately
recalled here. That Israel ‘will know that I am Yahweh your God’
when their need for food is met also picks up a Priestly formula,
one that has already been used in the message of deliverance from
Egypt which Moses was given for the Israelites (6.7): it will recur
in 29.46, where it indicates that such knowledge is to be confirmed
again when Yahweh, or rather his ‘glory’, takes up residence in the
tabernacle.20 The ‘recognition-formula’ was apparently taken over
by P from Ezekiel, where it occurs very frequently, and adapted
to the specific contours of the Exodus narrative.21 Earlier in this
chapter Moses and Aaron have already used it in their response to
the Israelites’ complaints (v. 6). Some commentators have seen this
as a sign that vv. 6-7 are a later elaboration of the original Priestly
narrative or originally followed v. 12 (see the introduction to the
chapter), but the differences between vv. 6-7 and v. 12 justify seeing
them in their present sequence as parts of a unified, developing
narrative. The words of Moses and Aaron do not reveal whether the
people’s coming recognition of Yahweh as the God of the Exodus
will mean judgement (as it did for the Egyptians) or the assurance
that Yahweh will provide for their needs. Now in v. 12 it is made
clear that it is the latter and the inclusion this time of ‘your God’
20
A similar expression occurs in 31.13, with ‘who sanctify you’ in place of
‘your God’, wording which corresponds very closely to Ezek. 20.12 (cf. v. 20,
also Lev. 20.8; 21.8, 15, 23; 22.9, 16, 32). Exod. 31.12-17, which expounds the
meaning of the Sabbath day, is commonly regarded as a secondary expansion of
the original Priestly account of the plans for the tabernacle (e.g. Noth, p. 192, ET,
p. 234).
21
For a full discussion of it see the studies of W. Zimmerli, ‘Ich bin Jahwe’,
‘Erkenntnis Gottes nach dem Buch Ezekiel’, and ‘Das Wort des göttlichen Selb-
sterweises (Erweiswort), eine prophetische Gattung’, reprinted in his Gesammelte
Aufsätze, 1 (Munich, 1963), pp. 11-40, 41-119, 120-32; ET in I Am Yahweh (ed. W.
Brueggemann: Atlanta, 1982), pp. 1-28, 29-98, 99-110. For a different view of the
ultimate origin of the ‘self-identification formula’ see J.W. Hilber, Cultic Prophecy
in the Psalms (BZAW 352; Berlin, 2005), pp. 132-33.
16.1-36 455
underlines the basis for this in the covenant relationship that was
central to 6.7 (and 6.2-8 as a whole). On the other hand, the fact that
v. 12 tells Moses to do exactly what he has already done according
to v. 8 adds weight to the arguments noted in the comment there for
that verse being a clumsy and secondary elaboration of vv. 6-7.
13-15. The narrative continues from v. 12 as expected, with
the provision of two kinds of food. The ‘meat’ and the ‘bread’
that had been promised turn out to correspond (to some extent
at least) to phenomena that have continued to be observed in the
Sinai peninsula. The meat comes in the form of quail, probably
Coturnix coturnix, the common quail, whose migratory route from
Europe and Western Asia to and from Africa took many of them, in
autumn and spring respectively, across Egypt and the Sinai penin-
sula (Cansdale, Animals of Bible Lands, pp. 167-68: cf. Jos., AJ
3.25 [in the Ἀράβιος κολπός, apparently the Gulf of Suez]). Quail
(pʿrt) were in fact popular in ancient Egypt as food, to judge from
surviving grave-reliefs: the favoured time for catching them was
the harvest-month of Paophi (LexAeg 6, 1094-95). A rare example
of a tomb-painting showing the use of a net to catch them appeared
in the tomb-chapel of Neb-Amun (cf. R. Parkinson, The Painted
Tomb-Chapel of Nebamun [London, 2008]), pp. 35, 118). Alfred
Kaiser, who spent many years living in the Sinai peninsula, reported
that quail congregated on the Mediterranean (northern) coast, but
for the southern part of their journeys they generally followed either
a south-westerly path (through the Nile valley) or a south-easterly
one (via Transjordan and N.W. Arabia) to maximise the food
supply: only much smaller numbers were observed in the central
and southern parts of the peninsula (‘Neue naturwissenschaftliche
Forschungen auf der Sinai-Halbinsel [besonders zur Mannafrage]’,
ZDPV 53 [1930], pp. 63-75 [72]; cf. C.S. Jarvis, ‘The Forty Years’
Wanderings of the Israelites’, PEQ 70 [1938], pp. 25-40 [30-31],
and Yesterday and Today in Sinai, pp. 183-84, 258-64). Josephus
(see above), however, seems to have known of larger numbers in
the Gulf of Suez. They are mentioned only briefly here (v. 13a; but
see the Note on v. 23); a fuller account of them is given in Num.
11.31-33 and, without the name of the species, Ps. 78.26-31 (more
briefly in Ps. 105.40). They are not included in the lists of ‘unclean’
birds in Lev. 11.13-19 and Deut. 14.11-20, which name only those
forbidden as food. It is not clear whether the writer here regards
them as a provision for a single occasion (as e.g. Houtman deduces
456 EXODUS 1–18
from their absence in the second half of the chapter [p. 321]) or
as one which continued. The issue is complicated by the fact that
Numbers 11 seems to regard them as a new, unprecedented provi-
sion to add variety to a previously limited diet (cf. vv. 4-6), but the
contradiction is probably due to the different ordering of events in
the Priestly and non-Priestly accounts (see the introduction to this
chapter).
The ‘bread’ is described much more fully, in what seems to be a
unified account (vv. 13b-15) which is partly paralleled in v. 31 and
more fully in Num. 11.7-9. Unlike the quail, presumably, this other
kind of nourishment was unfamiliar (v. 15: cf. Deut. 8.3, 16), no
doubt to most readers of the narrative as well as to their ancestors.
So these verses explain its origin, describe its appearance (on the
translation ‘powdery’ and other views see Note v on the transla-
tion) and give it a name (see Note w), though it continues to be
referred to as ‘bread’ (vv. 15b, 22, 29, 32) until near the end of the
chapter (vv. 31, 33, 35). From ancient times it has been compared
to a phenomenon which was known to occur in the Sinai peninsula
and elsewhere (Jos., AJ 3.31; Origen on Num. 11.6; Ambrose, Ep.
64; Anon.Plac., Itinerarium 39 [CCSL 179, p. 149]).
Excursus on Observations and Studies

of Phenomena Comparable to the Biblical Manna22
The early sources do not go into any detail about the origin of the ‘manna’
of their own time, except in some cases to say that it ‘came down’ from
heaven or the sky like rain or dew. But it is likely that they are referring to
what was later recognised as a product from the tamarisk trees which are still
widespread in the Sinai peninsula (see below). In the Middle Ages comparison
with phenomena specific to the peninsula was displaced by references to
a product of the al-ḥāj bush, which grows in dry areas across much of the
Middle East.23 In Armenia, Mesopotamia and especially in parts of Persia
it exudes a sweet, syrupy substance (taranjabin in Persian), which has long
been collected for medical use and as a source of sugar. This is probably
what Georgius Syncellus (c. A.D. 800) had in mind when he wrote of τὸ
μάννα κομισθὲν ἐκ τῆς Παρθικῆς which he had seen and tasted (Ecloga
22
See further and for fuller references the very detailed study of Maiberger,
Das Manna, pp. 325-438; more briefly IDB, 3, 259-60.
23
Cf. Maiberger, Das Manna, pp. 336-50.
16.1-36 457
Chronographica 244 [ed. Mosshammer]), and it is named as such in one

explanation cited in the Tafsīr of aṭ-Ṭabarī (c. 900) on Sura 2.57 of the Koran
(other explanations mention only ‘trees’) and in Ibn Ezra’s vigorous refuta-
tion of the late ninth-century Jewish rationalist Ḥī wī al-Balḫī (who came
from what is now northern Afghanistan) in his longer commentary on Exodus
(Rottzoll 1, p. 446).
In the fifteenth century the biblical manna was sometimes (e.g. by Pico
della Mirandola) equated with a juice harvested from a kind of ash-tree
(Fraxinus Ornus) in Calabria and Sicily. But with the renewal of travel to
the Sinai peninsula, beginning with Johannes Tucher and Sebald Rieter in
1479, attention was drawn again to the ‘manna’ that could still be found
there, though it was not (except for the isolated and only much later published
account of Samuel Kiechel [1588]) thought to be a product of a particular
type of tree until the late eighteenth century and only U.J. Seetzen in 1807
and J.L. Burckhardt in 1812 could conclusively establish that the tamarisk
(gallica) was its source. This species is widely attested in North Africa and
the Middle East, but it only produces ‘manna’ in certain parts of the southern
Sinai peninsula and of Persia.24 There seems to have been a suggestion even
before this that an insect may have played a part in its production, but this
was only confirmed in the 1820s by the researches of C.G. Ehrenberg, who
observed numbers of lice of a species which he named Coccus manniparus
(it is now known as Trabutina mannipara) on the branches of the tamarisks.
But he mistakenly assumed that, as with similar phenomena on other trees,
they simply bit into the bark of the tree and so released the juice which turned
into ‘manna’. It was only a hundred years later that closer study by A. Kaiser
and F.S. Bodenheimer established that the lice consumed the juices and
then excreted a liquid rich in carbohydrates which they did not need, from
which the ‘manna’ was formed. Bodenheimer also discovered another insect
(Naiacoccus serpentinus) which was involved in the same process.25 It should
be noted that this phenomenon has only been observed in the mountainous
region of the peninsula and in some wadis that run westwards to the Gulf
of Suez and that it does not occur throughout the year or in every year. It
is clearly related to the availability of water, either on the surface or under-
ground, and the rainfall varies from year to year. The quantity produced is of
course not large and it forms only a small part of the diet of the Bedouin who
24
For a list of the places where it has been found in Sinai see Maiberger, Das
Manna, pp. 389-90.
25
See Kaiser, ‘Neue naturwissenschaftliche Forschungen’ (an appreciative
review of Bodenheimer’s Ergebnisse which followed an earlier publication of
his own in 1924); Bodenheimer, Ergebnisse der Sinai-Expedition 1927 (Leipzig,
1929), pp. 45-89; id. ‘The Manna of Sinai’, BA 10 (1947), pp. 2-6; and the
summary in Maiberger, Das Manna, pp. 401-403.
458 EXODUS 1–18
collect it. It is possible that larger stands of tamarisks existed in earlier times,
before many were cut down for their wood, but there is no reason to suppose
that climatic changes have occurred in the past 5,000 years which would have
made the ‘manna’ available over a wider area.
In the meantime a rival explanation had been put forward and attracted
considerable support, because it could offer a closer parallel to some aspects
of the biblical narrative.26 The latter affirmed that the manna could be ground
and boiled (Num. 11.8; cf. Exod. 16.23), neither of which is possible with
tamarisk manna; and newer studies of the latter had made clear that it was
in no sense ‘rained’ from heaven or connected with dew (Exod. 16.4, 13-14;
Num. 11.9). All these features were, however, matched in a lichen found in
south-east Europe, south-west Asia and North Africa, about which a growing
number of reports reached Europe in the first half of the nineteenth century
(Lecanora esculenta, L. desertorum, L. Jussifii). A ‘manna-rain’ was first
reported in 1824. A number of popular writers seized on this parallel, but
it seems to have gained little support from biblical scholars, although both
P. Haupt and R. Meyer found a place for it in their accounts of the biblical
descriptions of manna.27 The fatal objection to this theory was that this lichen
has never been found in Palestine, Sinai or Egypt.
In the Exodus narrative the origin of the ‘bread’ is associated

with dew, but not identified with it (similarly in Num. 11.9). The
account might be read as saying that when the dew evaporated the
‘bread’ was left where it had been. Here it is emphatically said
to lie ‘on the surface of the ground’, thus resembling a layer of
frost, and perhaps in colour also (cf. v. 31). It is ‘fine’ or ‘thin’
(Heb. daq can have both meanings). There is no mention of any
association with trees here (though the comparison with bedōlaḥ,
probably ‘bdellium’ or ‘gum resin’, in Num. 11.7 could imply one).
There is debate about the translation of the Israelites’ response in
v. 15 (see further Note w on the translation). Probably the majority
of modern scholars follow the view already taken in some of the
26
It was not the only rival to the two most persuasive natural explanations of
the biblical manna. In addition to the ‘ash-manna’ mentioned above, numerous
exudations from other trees have been observed and referred to as ‘manna’, such
as the Kurdish ‘oak-manna’ and others noted by Maiberger, Das Manna, pp. 325-
28, 419-21. It is of course not necessarily the case that the use of the name is meant
to equate such phenomena with the biblical manna: it may simply reflect their
similarity to descriptions of it.
27
Cf. Haupt, ‘Biblical Studies’, AJP 43 (1922), pp. 238-49 (247-49); TWNT
4, 466-70 = TDNT 4, pp. 462-66.
16.1-36 459
ancient translations that they asked a question: ‘What is it?’, as the

explanation that follows might seem at first to suggest. Moses then
answers the question at the end of the verse. But the word translated
‘What?’ is not the regular Hebrew word for this (mah): it is mān,
which is elsewhere in Hebrew, including vv. 31, 33 and 35 below,
always the word for manna. So it is more natural to understand it
in that way here and to translate: ‘It is manna’. The Israelites are
then thought of as coining a new name for something they have
not seen before. The suggestion that mān was originally an alterna-
tive word for ‘What?’ in Hebrew lacks any support in the language
itself, whatever one may make of the possible parallels in related
languages. Its true etymology remains uncertain (see Maiberger,
Das Manna, pp. 280-308, for a review of earlier proposals and
the conclusion that it was originally used, like Arabic mannu, for
a ‘thin’ deposit on various trees and plants that was believed, like
the dew, to come down from heaven). The author of v. 15 probably
saw sufficient similarity between this word and the regular Hebrew
expression for ‘What?’ to suggest why the name came to be given.
Such etymological speculation, which is often not based on precise
linguistic equivalence, is common in the biblical explanation of
names (e.g. 2.10: see the Explanatory Note there).
16-21. In these verses the daily gathering of the manna is precisely
regulated by Moses’ instructions in vv. 16 and 19 and, while obedi-
ence to them is rewarded (vv. 17-18), disobedience leads to spoiling
of the divine gift and an angry response from Moses (v. 19). This
corresponds closely to a recurring pattern of the Priestly corpus (e.g.
39.42-43; 40.16-34; Lev. 10.1-3, 16-18; 24.10-23; Num. 15.32-36).
The instructions and the account of their fulfilment also correspond
in some details to the Priestly regulations for Passover in ch. 12 (as
earlier in v. 12: compare vv. 16 and 18 with 12.4, v. 17 with 12.28,
vv. 19-20 with 12.10) and to the clearly Priestly appendix to this
chapter (compare v. 16 with v. 32; and the references to the omer).
There is some tension between v. 21 and the rest of the section: the
first part of the verse (in which ‘according to what he could eat’ is
expressed differently from vv. 16 and 18 and also 12.4) duplicates
the end of v. 18 and would come more naturally before than after
vv. 19-20, while the statement that the manna melted in the sun
seems incompatible with v. 19. So v. 21 may be from a different
(older?) version of the story, like vv. 4-5 (see further Frankel, The
Murmuring Stories, pp. 78-81).
460 EXODUS 1–18
22-30. The special arrangements for the Sabbath day’s provi-

sion are described in greater detail, which befits the importance of
Sabbath observance, but there are signs that they may be drawn
from two different accounts of the episode. In vv. 22-25 features of
the Priestly corpus are again strongly present: in v. 22 the omer and
‘the leaders of the congregation’ (neśîʾê hāʿēdāh: here mentioned
for the first time, next in Num. 1.16 and with this designation in
Num. 16.2); in v. 23 the expressions ‘a day of rest’ (see Note gg
on the translation), ‘a holy Sabbath for Yahweh’ (cf. 31.14-15;
Lev. 23.3) and ‘for keeping’ (cf. vv. 32-34); in v. 24 ‘as Moses had
commanded’ (Lev. 9.5, 21), while the reference to lack of worms
and stink is clearly meant to contrast with, and presupposes, v. 20;
in vv. 23 and 25 Moses continues to guide the people’s behaviour
as in vv. 16 and 19. But from v. 26 on such features are lacking
until v. 32 and ‘the people’ (Heb. hāʿām) becomes the designa-
tion for the Israelites (vv. 27, 30), as in v. 4 and more generally
in the non-Priestly sections of Exodus (cf. 14.21; 15.24). Both the
special provision on the sixth day (v. 29) and, more distinctively, the
mention of Yahweh’s ‘instruction(s)’ are anticipated in vv. 4-5. The
beginning of Yahweh’s rebuke of the Israelites is almost identical in
form to that addressed to Pharaoh in 10.3, a non-Priestly passage.28
The precise demarcation line between the two versions is a little
uncertain. Verse 26 could be (as most analyses conclude) the end
of the Priestly account, summing up its teaching in a form which
resembles the Sabbath commandment in the Decalogue, which is
clearly in the background of this whole section of the chapter. The
attachment of it to Moses’ words in v. 25, with the same second
person plural (‘you’) address to the people, would support this. On
the other hand, the wording is not particularly close to 31.15(P)
and v. 26 (the end of which duplicates the end of v. 25) might be
the beginning of the non-Priestly account: without it v. 27 begins
abruptly and it is not clear what the ‘commandments and instruc-
tions’ in v. 28 are. On balance there is much to be said for this
alternative, and then v. 26 could originally have been the end of
Yahweh’s words in vv. 4-5, inaugurating the Sabbath for the first
28
As Wellhausen observed (Composition, p. 78), it is also strange that v. 27 has
a new introduction to the Sabbath day (this is clearer in the Heb.: cf. v. 22) when
vv. 24-25 have already brought the narrative to that point.
16.1-36 461
time.29 The join is smooth as far as the content is concerned and the
use of the second person to refer to the people (instead of the third
person in vv. 4-5) could be an adaptation to fit its present context.
But the first part of v. 4 has ‘for you’, so a reversion to the second
person in the original source is also possible.
According to this analysis, what was probably the older version
of this section led on directly from Yahweh’s announcement to
Moses in vv. 4-5. The logical position for v. 21 in the underlying
source would be after v. 26: its present placing would be due to the
redactor’s careful organisation of the material before him. Origi-
nally it prepared the way for v. 27, which recounts the violation
of the Sabbath by a (small?) group of Israelites. Yahweh’s rebuke,
like his promise in vv. 4-5 (and the warning in v. 26), is mediated
through Moses and probably speaks of ‘commandments and instruc-
tions’ in the plural for greater rhetorical effect: it is clearly only
breach of the Sabbath that is in view. The reminder of what Yahweh
requires (v. 29) keeps to the wilderness situation in its wording and
need not in the mind of the writer have had in view the later rulings
against travel on the Sabbath: it simply uses the narrative context to
reinforce the ban on work on the Sabbath and emphasises that this
is as much a ‘gift’ as the provision of the manna. Appropriately v.
30 speaks positively, not negatively, about what the people do: they
‘rested’.30
The Priestly version, at least in what has been preserved,
contents itself with commands about the observance of the Sabbath
day (though v. 24 in effect indicates the people’s obedience). The
Sabbath is initially introduced to the people, not by a word from
Yahweh, but by the double provision of manna which they find on
the sixth day (v. 22). It is only when their tribal leaders report this
to Moses that he informs them of what Yahweh has declared (to
him, evidently): the seventh day is to be a day of rest dedicated
29
So already Baden, ‘The Priestly Manna Story’, pp. 492-93.
30
A problem for the unity of vv. 28-29 is sometimes seen in the fact that v. 28
is a word of Yahweh, while the use of the third person of him in v. 29 would fit
more easily with it being a word of Moses. But the emphasis placed on the subject
by the word-order may have led to the grammatical shift in v. 29. See also the
parallels for such a change cited by Baden, ‘The Priestly Manna Story’, 494 n. 13.
Of course if v. 28 is an addition, as many think, this would remove the problem
altogether.
462 EXODUS 1–18
(‘holy’) to Yahweh and henceforth to be known as the Sabbath.31

In the present text, where the two accounts have been combined, it
is possible to see Moses as referring back to what he has been told
in v. 5 (though that did not say anything about the sanctification
of the seventh day); in the independent Priestly version he was
explaining an unexpected event by passing on to the people a new
word of Yahweh (as in vv. 16 and 32), which also contained his
instructions for what to do (on the linguistic issue cf. Lev. 10.3 and
Note ff on the translation). Careful preparation is to be made on the
day before (here ‘bake’ and ‘boil’ may have the normal means of
cooking bread and meat in view, though Num. 11.8 represents both
as methods used in the preparation of the manna itself) and then the
food prepared may be eaten on the Sabbath (vv. 23, 25). Cooking is
evidently regarded as work and so to be avoided on the Sabbath as
in the Mishnah; it required the lighting of a fire, which was prohib-
ited according to 35.3.32
31. It is surprising to find this further statement about the naming
of the manna and its description, as though nothing similar had
appeared earlier (in vv. 14-15), and it presumably belongs to a
different version of the story.33 Although some scholars have attrib-
uted vv. 13-15 or part of them to the older (‘J’) narrative (see the
introduction to this chapter), they fit smoothly into the surrounding
Priestly context and something like them is needed to bridge the gap
between v. 12 and v. 16. It is therefore likely (and there is nothing
against it) that v. 31 belonged to the non-Priestly version and in
fact formed (part of, perhaps: see below on v. 35) its conclusion,
following directly on from v. 30.34 ‘The house of Israel’ is a relatively
31
A (probably later) Priestly law for the longer-term observance of the Sabbath
appears in 31.12-17.
32
Frankel (p. 106), following Weinfeld, denies this and takes leaving the
manna for the Sabbath uncooked overnight as a test of the people’s obedience
(cf. v. 4). But this depends on his view that vv. 22-26 come from the editor who
combined the P and non-P stories together.
33
This is often denied (e.g. by Baden, ‘The Priestly Manna Story’, p. 495
n. 20), but it remains the most plausible explanation, especially (but not only) if,
as we have done, the people’s words in v. 15 are translated as a statement about
the manna.
34
The objection that ‘its name’ then has nothing to refer back to (so recently
Frankel, p. 74; L. Schmidt, ‘Priesterschrift’, p. 493) is pedantic: ‘it’ throughout the
chapter has been the manna, and ‘bread’ in v. 29 has been a recent reminder of this.
16.1-36 463
uncommon expression in the Pentateuch and elsewhere occurs only

in P and H (40.38; Lev. 10.6; 17.3, 8, 10; 22.28; Num. 20.29). But it
is much more common in the rest of the Old Testament, especially
in the historical books and in prophecy (where 83 of its 147 occur-
rences are in Ezekiel), and it can refer either to the northern kingdom
(as in 1 Kgs 12.21) or to the whole people (e.g. 2 Sam. 6.5; 16.3). It
therefore need not point to a Priestly origin for the verse and it may
perhaps have been deliberately chosen to attribute the use of the
name ‘manna’, according to this writer, to people of his own time,
rather than to the wilderness generation as in v. 15. The description
of the manna given here is different from that in v. 14 and corre-
sponds closely to features of the ‘tamarisk manna’ which is still
found in the Sinai peninsula (see the Excursus on vv. 13-15).35 As
was already recognised by Rashi and other Jewish commentators,
‘like coriander seed’ and ‘white’ are two separate descriptions, since
the coriander seed is black or grey (cf. IDB 1, pp. 681-82; 2, p. 289;
Maiberger, Das Manna, p. 179) and the manna is compared to it
because of its size and shape, not its colour. Since the statement in
v. 21 that the manna melted in the sun also corresponds to observed
reality, the older narrative as a whole displays a closer acquaintance
than v. 14 (and vv. 20 and 23) with the actual phenomena.36
32-34. Finally, before what may be called an appendix to the
narrative (vv. 35-36), by Yahweh’s command an omer of manna
is set aside as a witness to future generations of Yahweh’s provi-
sion for their ancestors’ needs. The three verses make a connected
sequence: Moses announces a divine command addressed to the
people as a whole (v. 32), he tells Aaron to take the necessary steps
(v. 33) and Aaron fulfils the instruction (v. 34). Inevitably there is a
good deal of repetition between these verses, but the writer varies
the wording to allow different aspects of the memorial to be intro-
duced at each stage: the purpose in v. 32, the container in v. 33 and
35
Cf. Bodenheimer, Ergebnisse, pp. 86-87. Alfred Kaiser denied the similarity
to coriander seed (‘Neue naturwissenschaftliche Forschungen’, 72), but he may
have been attaching too much importance to the difference in colour.
36
Cf. Maiberger, Das Manna, p. 403. Num. 11.7-8 seems to be a later and
artificial addition based on the combination of the two accounts in Exod. 16, and it
has a different view of the taste of the manna. But the comparison with bdellium is
a new feature which matches both the texture and, at times, the colour of tamarisk
manna.
464 EXODUS 1–18
the precise location in v. 34. The repeated elements – the quantity,

the long-term intention, Aaron’s role (all mentioned twice) and the
‘keeping’ of the manna (thrice) emphasise the key components of
the process. Several features link these verses to Priestly sections
earlier in the chapter (the prominence of Aaron [cf. vv. 2, 6, 9-10:
here especially because of the place where the manna is to be kept],
the wording at the beginning of v. 32 [cf. v. 16], the omer measure
[vv. 16, 18, 22], ‘for keeping’ [v. 23]) or elsewhere (‘throughout
your generations’ [cf. 12.14, 17, 42; 29.42 etc.], ‘the decree’ [25.16,
21-22 etc.]). There is a particularly close parallel with the placing of
Aaron’s staff ‘before the decree’ in Num. 17.25-26 as a warning to
anyone who would challenge his family’s priestly prerogatives, and
these verses are sometimes ascribed to a later layer of P as a result
(so Albertz, pp. 276-77). But here such concerns about the priestly
hierarchy are only hinted at and the primary purpose is, as with
Passover and Unleavened Bread, to provide a material reminder of
Yahweh’s care for his people at the time of the Exodus (v. 32; cf.
12.14, 42).
In v. 34 ‘before the decree’ is probably the best translation of
Heb. lipnê hāʿēdūt (for full discussion see Note uu on the trans-
lation). The older translation ‘before the testimony’ is certainly
wrong, as the idea of ‘witness’ can only be artificially introduced
in the occurrences of ʿēdūt. The related verb has a wider meaning
of ‘solemnly declare’ in a variety of contexts, including the word of
God through a prophet (Deut. 8.19; Ps. 81.9; 2 Kgs 17.13, 15). The
renderings ‘pact’ (NJPS) and ‘covenant’ (NRSV) here are certainly
an improvement, but probably keep too close to the meanings of
cognate words in Aramaic and Akkadian. A better guide is the use
of the plural in Deuteronomy (4.45; 6.17, 20) and elsewhere, where
it is clearly an alternative word for the laws in that book and NJPS
has ‘exhortations’ and NRSV has ‘decrees’. Elsewhere in the Penta-
teuch (always in P except for 32.15) the singular form is used for
the tablets of the law which were kept in the ark (cf. 25.16, 21;
31.18; 34.29). Neither the ark nor the tablets of the law have been
mentioned hitherto, but in P they were to be kept in the holy of
holies of the Tabernacle whose construction is described in Exodus
25–31 and 35–40. The expression ‘before the decree’ (hāʿēdūt) is
also used for the placing of the holy incense (30.36) and the staffs
of Aaron and other tribal leaders (Num. 17.19, 25) in the holy of
holies.
16.1-36 465
Thus the expression presupposes the Priestly Sinai narrative and

in particular the construction of the Tabernacle later in Exodus.
According to some (most recently Baden in ‘The Priestly Manna
Story’) this is because the Priestly manna story was originally
placed somewhere in the middle of the book of Numbers, for which
other arguments are brought forward, including the anticipation of
the arrival in Canaan in the next verse (see further the introduction
to this chapter). Such a drastic solution is probably unnecessary, as
the Priestly author (or redactor, if vv. 32-34 are a secondary addition
to P) may have attached his account of what happened later to the
main account of the manna here.
It remains to consider the possible origin of the Priestly use of
ʿēdūt to refer to the tablets of the law and the reason for its choice.
As already noted, the plural of the same (or according to some,
a similar) word is used in Deuteronomy for the individual laws
included there. Since it always appears there with other expres-
sions, it is possible that it was intended to designate the Decalogue
in particular, which appears again in Deut. 5.6-18(21). But there is
no proof of this and elsewhere Deuteronomy refers to the Decalogue
only as ‘the (ten) words’ (Heb. debārîm: 5.22; 9.10; 10.2, 4: cf. Exod.
20.1; 34.1, 27(?), 28). So the use of ʿēdūt for the Decalogue may be
an innovation of P as well as the choice of the singular to emphasise
the unity and special character of these commands, even if the word
itself was taken over from earlier theological usage. There are also
a few other possible traces of the earlier use of ʿēdūt in the singular
which may be relevant. The most securely dated is the reference
to the ʿēdūt that was given to King Joash by Jehoiada the priest at
his coronation (2 Kgs 11.12 par. 2 Chr. 23.11: the former tendency
to emend this word to one for a piece of jewellery seems to have
been abandoned since von Rad’s 1947 essay on ‘Das judäische
Königsritual’, TLZ 72 (1947), pp. 211-16, ET in his The Problem
of the Hexateuch and Other Essays (Edinburgh, 1966), pp. 222-31;
cf. HAL, Ges18, G.H. Jones, 1 and 2 Kings [NCB; London, 1984),
p. 481). It would appear that it is only mentioned here because of
the exceptional circumstances and that it was a regular part of the
coronation ceremony. The contents of the document are unclear. In
Ps. 78.5 ʿēdūt is used of the laws given to Israel by Yahweh (cf. the
pl. in v. 56 and again in Ps. 99.7). In Pss. 81.6 and 122.4, both of
which may well be pre-exilic, ʿēdūt refers to a regulation requiring
attendance at a festival, the former probably at a northern shrine (cf.
466 EXODUS 1–18
‘Joseph’: perhaps Bethel) and the latter in Jerusalem. Psalm 93.5

uses the plural of decrees associated with Yahweh’s temple. Any
or all of these models may have been drawn upon by Deuteronomy
and P, with the first and last being most likely in P because of their
connections with the temple and the priesthood.
Why did P adopt this expression for such a central place in its
presentation of the Sinai narrative? It is likely to be connected with
P’s concept of covenant and the significance of the revelation at
Sinai for P’s theology. Unlike the conditional covenant theology
of Deuteronomy, P uses the expression ‘covenant’ (Heb. berît) for
unconditional, irrevocable commitments of God to mankind and
to his people, made with Noah (Gen. 9), Abraham (Gen. 17) and
Phinehas the grandson of Aaron (Num. 25.10-13, probably from
a later stratum of P). There is no mention in P of a covenant being
made at Sinai as in Deuteronomy 5 (and indeed in Exod. 24.3-8):
for P Sinai is primarily the place where Israel’s cult and priesthood
were instituted. But evidently P did not wish, or did not feel able, to
deny that the ten commandments had been given to Israel at Sinai.
What was needed was an expression to describe them (and the ark
in which they were placed) which did not echo the Deuteronomic
language about a covenant at Sinai (or Horeb as it calls it), since the
covenant which now (again?) defined Israel’s standing before God
was the covenant with Abraham (and Isaac and Jacob), which P
mentions as such in Exod. 2.24 and 6.4-5. P evidently regarded the
commandments, like the revelation about worship, as a declaration
of Yahweh’s will which was not part of a conditional covenant and
the term ʿēdūt, in the sense of ‘decree’, was ideally suited to this
purpose.
35. Here again, as in the previous verse, the narrator jumps
forward to take in what was still in the future at this stage of
the story, this time in a parenthesis (the first of two with which
the chapter concludes) which makes no secret of its longer-term
perspective. The provision of manna throughout the forty years of
the wilderness journey also forms part of the parenesis in Deut.
8.2-3. Joshua 5.11-12 specifies the encampment at Gilgal, just after
the crossing of the river Jordan, as the occasion when the Israelites
first ate the crops of Canaan and the manna ceased. This conforms
to the view of the eastern frontier of Canaan which is found in Ezek.
47.18 and in some other passages which exhibit connections with
the Priestly literature (e.g. Num. 34.10-12; 35.14; Josh. 22.9, 24-25,
16.1-36 467
32): from the Sea of Galilee to the Dead Sea the boundary follows
the Jordan river and the regions in Transjordan occupied by some of
the Israelite tribes are not in Canaan. This appears to be an ancient
conception, since it corresponds (so far as the evidence goes) to the
use of the term ‘Canaan’ in second-millennium sources such as the
Amarna letters and in earlier Old Testament passages (cf. Aharoni,
The Land of the Bible, pp. 67-77; A.R. Millard, ‘The Canaanites’,
in POTT, pp. 29-33; TWAT 4, 231-38 = TDOT 7, pp. 217-24). The
vocabulary of v. 35 is not distinctively Priestly, though ‘inhabited’
(Heb. nôšāb) is attested elsewhere only (three times) in Ezekiel.
The repetitiveness of the verse has led some scholars to divide it
between the two main sources found in the chapter (see the introduc-
tion), but the chiastic inversion between the two halves of the verse
corresponds to a frequent stylistic feature of biblical Hebrew, and
especially the Priestly source (cf. Gen. 1.5, 27; Exod. 27.44; Lev.
10.13-14; 16.33: see M. Paran, Forms of the Priestly Style, pp. 49-
136 [summary pp. viii-xi]; McEvenue, Narrative Style, pp. 43,
51-52). So the verse should not be divided and it may well belong
to P.
The statement here and in the other passages mentioned above
that manna was gathered throughout the wilderness journey does
not correspond to observations of the ‘tamarisk manna’, which are
limited to the southern part of the Sinai peninsula and to a short
period in each year. If, as seems likely, that is what lies behind this
narrative, the claim made in this verse (like the quantity of manna
involved and the weekly interruption in its provision) is an indica-
tion of the extent to which the narrative has developed from the
natural phenomenon on which it is based.
36. This second parenthesis, which belatedly explains the meas-
ure ‘omer’ that has appeared several times in Priestly sections of the
chapter, must be part of (or more likely an addition to) the Priestly
account. In its form it resembles Ezek. 45.11-12, but its purpose
seems to be different. The Ezekiel passage may well be introducing
a reform of the system of weights and measures (for archaeologi-
cal evidence on the different pre-exilic relation between gerah and
shekel see R. Kletter, Economic Keystones [JSOTSup 276; Sheffield,
1998], esp. pp. 80-84), but here an otherwise unknown (and possi-
bly invented) measure is being explained. Elsewhere the tenth part
of an ephah is called an ʿiśśārôn, ‘a tenth’, from the numeral ten,
ʿāśār (29.40 and 32 other occurrences, all in P), or the expression
468 EXODUS 1–18
employed here, ‘a tenth of an ephah’ (ʿaśîrît hāʾêpāh), was used

(Lev. 5.11; 6.13; Num. 5.15; 28.5). The ephah was a well-known
measure for grain (e.g. Deut. 25.14-15; Amos 8.5). Its equivalent
in modern terms has been variously calculated, as the evidence is
complex and ancient norms may well have varied: BRL2, p. 205,
suggests a little under 20 litres, while ABD 6, pp. 903-904, prefers a
lower figure, closer to 10 litres.37
Text and Versions

‫( בני־ישׂראל‬16.1) TgJ has simply ‫דישׂראל‬, probably to conform to ‫סבייא‬
‫ דישׂראל‬in its addition at the end of 15.27.
‫( בחמשׁה עשׂר יום‬16.1) LXX τῇ δὲ πεντεκαιδεκάτῃ ἡμέρᾳ separates
the date-formula from the note of arrival (as if a waw had preceded the date)
and links it instead to the complaint in v. 2 (where the initial waw is ignored
or treated as waw of the apodosis). This appears to presuppose the tradition
recorded in Jos., AJ 3.11 and in MRI (Lauterbach 2, pp. 99-100; cf. TgJ on
12.39 and 16.2) that the bread baked when leaving Egypt ran out on the 15th
of Iyyar.
‫( לחדשׁ השׁני‬16.1) TgJ adds that it was the month of Iyyar, just as it had
inserted the name ‘Nisan’ in 12.8.
‫( לצאתם‬16.1) LXX (ἐξεληλυθότων αὐτῶν) and Vulg explicitly under-
stand ‫ ל‬to mean ‘after’ here, whereas TgNmg and Sy point to ‘during’ (see
Note b on the translation). Sy replaces the suffix with ‘the Israelites’, perhaps
because it too, like LXX, took the date-formula with v. 2.
‫( מארץ מצרים‬16.1) Sy has just ‘from Egypt’, in line with MT’s reading
elsewhere, most recently in 14.11 (cf. vv. 6 and 32 below).
‫( וילינו‬16.2) The Kethibh has the Hiphil form (cf. the part. ‫ מלינים‬in v. 8),
but the Qere ‫ וַ יִּ לּוֹנוּ‬is the Niphal as in 15.24, which SP and 4QpalExl ‫ וילנו‬were
probably also intended to represent (plene forms are preferred for the third
person pl. Hiphil according to GSH §82cβ). With so much variation (see also
v. 7), which could be due in part to the similarity of waw and yodh in early
forms of the square script, it is difficult to be confident of the original read-
ing. But there may be something to be said for the view that ch. 16 originally
used the Hiphil throughout and that the Niphal forms were assimilated to
37
Earlier estimates were even more divergent, if more confident and precise:
Gressmann, Anfänge, p. 82 n. 6, was in no doubt that an omer was 3.64 litres, so
that the ephah would be 36.4 litres; while McNeile (p. 100) reckoned the ephah
as 65 imperial pints in biblical times, but later under Greek influence it was 71.28
pints.
16.1-36 469
15.24. Here, as in 15.24, LXX and Vulg (but this time not SP) again have the
easier and therefore secondary sing. form. The Vss mainly render with the
same verbs as in 15.24 but Sy shifts to wrṭnw, which has the same meaning
as wʾtrʿmw had there but is less ambiguous, and it maintains this preference
in vv. 7-8. TgJ prefixes a statement that the bread brought from Egypt had run
out (cf. the note on ‫ בחמשׁה עשׂר יום‬in v. 1).
‫( עדת‬16.2) TgJ has no equivalent, like Heb. in v. 3. By contrast TgN has
‫ עדת‬as well as its standard translation ‫כנישׁתא‬, a very obvious case of the intru-
sion of MT language into TgN.
‫( במדבר‬16.2) LXX has no equivalent (one is supplied in Symm, Theod
and the O-text): since the expression is scarcely necessary after v. 1, this might
be the original text, with the other witnesses pedantically adding ‫ במדבר‬to
prepare for ‘to this wilderness’ in v. 3.
‫( מי יתן מותנו‬16.3) The Vss render with the appropriate idiom for express-
ing a strong wish in their own languages. TgN again adds the MT wording to
its translation (cf. above on ‫ עדת‬in v. 2), including ‫ מותנו‬in place of an Aram.
equivalent.
‫( ביד־יהוה‬16.3) Tgg avoid the anthropomorphism as in the other three
instances in the Pentateuch (9.3; Num. 11.23; Deut. 2.15), using ‘before
the Lord’ (TgO,N) or ‘by the Memra of the Lord’ (TgJ). LXX substitutes
πληγέντες ὑπὸ κυρίου here, but renders literally in the other passages, so it
is only clarifying the meaning, as also in 24.11 (cf. Fritsch, p. 14).
‫( סיר הבשׂר…לחם‬16.3) LXX and Vulg use pl. forms to match the reality
more closely, but Tgg and Sy do so only with ‫סיר‬.38
‫( לשׂבע‬16.3) Tgg and Sy all paraphrase slightly by attaching the part. of
the related verb in the pl.
‫( כי‬16.3) Vulg cur intensifies the accusation by borrowing the language
of 17.3.
‫( אל־המדבר הזה‬16.3) TgO has no equivalent to ‫הזה‬, perhaps for stylistic
reasons (cf. below on Vulg).
‫( את־כל־הקהל הזה‬16.3) Vulg has no equivalent to ‫הזה‬, and Sy (except for
5b1) has ‘of the Israelites’, both probably to avoid the excessive repetition of
the demonstrative. LXX has συναγωγήν (its usual word for ‫ )עדה‬for ‫ קהל‬here,
as it does quite frequently (e.g. Lev. 4.13-14), lending support to J. Barr’s
contention that LXX does not see a fundamental difference between the two
Heb. words (cf. Semantics, pp. 119-29). It remains striking, however, that
ἐκκλησία is used regularly outside the Pentateuch for ‫ קהל‬but never for ‫עדה‬.
After v. 3 4QpalExl has a vacat, which indicates the antiquity of the
division at this point which is found in MT and SP mss.
38
Muraoka (Lexicon, p. 68) gives ‘food’ rather than ‘loaves’ as the meaning
of the pl. of ἄρτος: perhaps this was a device of the translators to distinguish the
broader and narrower meanings of ‫לחם‬.
470 EXODUS 1–18
‫( יהוה‬16.4) TgNmg as often prefixes ‘the Memra of’.

‫( ממטיר‬16.4) On LXX ὕω see Text and Versions on 9.18. Vulg pluam
(alone) indicates the (imminent) future reference that is intended. Tgg ‫מחית‬,
‘send down’, eliminates the metaphor, but they use the same verb even in Gen.
7.4, where actual rain is involved.
‫( לחם‬16.4) LXX ἄρτους and Vulg panes have the pl. as in v. 3 and their
renderings have a curiously literal character (cf. vv. 7-8).
‫( מן־השׁמים‬16.4) TgJ adds ‘which has been reserved for you from the
beginning’, recalling the tradition that manna was one of the special items
created on the eve of the first Sabbath day (cf. MRI [Lauterbach 2, pp. 124-
25]; TgJ on Gen. 2.2; Num. 22.28).
‫( ויצא‬16.4) Tgg and Sy have the pl. form, conforming to the verb that
follows.
‫( ולקטו‬16.4) Vulg colligat follows the stricter Latin rules of concord.
‫( דבר־יום ביומו‬16.4) Tgg and Sy follow MT closely, though TgN ‫סכום‬,
‘amount’, and Sy mʾkwltʾ, ‘food’, aptly use more specific alternatives for ‫דבר‬.
LXX’s paraphrase, though less free than in 5.13, 19 (and Vulg here), was
harmonised with MT by the addition of ῥῆμα and αὐτῆς in Aq and the O-text.
‫( אנסנו‬16.4) So also 4QpalExm and nearly all SP mss (Camb. 1846 reads
‫אנסינו‬, which must be a secondary phonetic variant); LXX πειράσω αὐτούς
(cf. Tgg, Sy) will be an adaptation to its pl. variant for ‫( הילך‬see the next note).
TgNmg has second person pl. pronouns for the people both here and there,
perhaps reflecting liturgical use.
‫( הילך‬16.4) Most SP mss read the pl. ‫הילכו‬, breaking the concord with the
suffix of ‫ אנסנו‬but agreeing with ‫ולקטו‬, the last verb of which the people were
the subject. LXX, Tgg and Sy also have pl. forms here, but Vulg agrees with
MT’s sing. There is no Qumran evidence. The difficilior lectio here is ‫הילכו‬
and it is probably original, with MT having assimilated it to the suffix of ‫אנסנו‬.
‫( בתורתי‬16.4) The Vss all render ‫ תורה‬here with words meaning ‘law’: it
is only in recent times that EVV have departed from this (NJPS, NEB, NIV,
REB, NRSV). A number of SP mss (including one of Crown’s and Camb.
1846) support a pl. reading of the noun, as do TgJ,N and Sy. But this may be
due to no more than the influence of the pl. form of this word in v. 28.
‫( והכינו‬16.5) Vulg parent (cf. sit) in the subjunctive implies the subordina-
tion of v. 5 to ut in v. 4 – a typical piece of ‘tidying up’ the looser syntax of
Heb. (and OL). TgN ‫ויצנעון‬, ‘they shall set aside’, seems to be thinking only of
the Sabbath portion: the mg, like the other Vss, has ‘prepare’.
‫( יביאו‬16.5) Some SP mss (inc. Sadaqa) read ‫הביאו‬, with a similar concern
to mark the ‘future perfect’ to SP’s ‫ ברכך‬in Deut. 16.10 (cf. GSH §172a).
TgNmg substitutes ‫ ילקטו‬from v. 4; TgJ has a long addition incorporating the
rules about the erub from B.Erub. 71a-b (MRI [Lauterbach 2, p. 104] refers
to this more briefly).
‫( משׁנה על‬16.5) LXX has just διπλοῦν, which to make a grammatical
connection with what follows should probably be parsed as a neuter part.,
‘doubling’: the other Vss render freely with idiomatic equivalents.
16.1-36 471
‫( יום יום‬16.5) Many SP mss (inc. all of Crown’s and Camb. 1846) read
‫( יום ויום‬cf. SP at Gen. 39.10), a variation found in late BH (Esth. 2.11; 3.4)
and at Qumran (11QT 15.1, 5; 4Q210[Aram.] fr.1 3.4-5), with a distributive
use of waw that is ‘late’ and post-biblical (BDB, p. 253), cf. TgJ here: TgO, Aq
and Theod reproduce MT, which must be original. The other Vss render freely,
either idiomatically (Symm, Sy, Vulg) or awkwardly due to repetition of their
equivalents for ‫ דבר־יום ביומו‬in v. 4 (LXX [in part], TgN).
‫( ויאמר‬16.6) So also SP; among the Vss only Vulg has a pl. verb.
‫( אל־כל־בני ישׂראל‬16.6) LXX πρὸς πᾶσαν συναγωγὴν υἱῶν Ἰσραήλ
follows vv. 1-2, where ‫ כל‬also appears.
‫( ערב‬16.6) The early Sy ms. 5b1 adds bhdʾ, ‘this’, a rare case of its
varying from MT.
‫( אתכם‬16.6) TgJ,N add ‫פריקין‬, ‘redeemed’, as often, most recently in 13.14
(see Text and Versions).
‫( מארץ מצרים‬16.6) Sy has no equivalent to ‫ארץ‬, as in v. 1 (see the note).
‫( וראיתם‬16.7) TgJ has ‘(the glory of the Shekinah of the Lord) shall be
revealed against you’, using the same expression as it (and TgO,N) has in v. 10
for ‫( נראה‬cf. Chester, Divine Revelation, pp. 80-81). TgN also introduces the
Shekinah here, but TgO follows MT.
‫( בשׁמעו‬16.7) LXX omitted an equivalent to the suffix because the refer-
ence to God was clear enough from the context (Wevers, Notes, p. 246), but
αὐτόν was added by the Three and the O-text and κύριον (from v. 8) in many
other mss. TgO,J and Sy turn the expression into a passive form; likewise in
TgN with the m. sing. part. ‫ ושׁמיע‬but the causal connection is ignored.39
‫( את־תלנתיכם‬16.7) LXX and Sy have sing. forms of the noun here and
in vv. 8, 9 and 12, Vulg (perhaps for the sake of variety) only here and in
vv. 8b and 9: only in v. 12 is the consonantal text of MT truly ambiguous, but
a Vorlage reading ‫ תלנתכם‬may have existed and even have given the original
reading. The Tgg evidently knew the pl. reading.
‫( על־יהוה‬16.7) Tgg and Sy render ‫ על‬with ‘before’; LXX ἐπὶ τῷ θεῷ (cf.
vv. 8-9) perhaps stresses the human–divine contrast (Wevers, ibid.).
‫( ונחנו מה‬16.7) SP ‫ אנחנו‬normalises the form of the personal pronoun.
TgJ,N,F, as well as (like Sy) repeating ‘we’, add ‫חשׁיבין‬, ‘reckoned’, presumably
to exclude even the idea that Moses and Aaron might be divine.
‫( תלונו‬16.7) The Niphal of the Kethibh is probably intended by all mss of
SP except one, while 4QpalExl’s damaged text seems to have the Hiphil of the
Qere. On the variation in general see the note on v. 2.
‫( ויאמר משׁה‬16.8) Sy interposes, as elsewhere, lhwn, ‘to them’.
‫( בתת יהוה‬16.8) Most of the Vss leave the sentence incomplete as a
temporal clause, but Vulg dabit and TgJ’s prefixing of ‘Then you will know’
(from v. 7) smoothen the connection. TgJ departs from the straightforward
meaning of ‫ בתת‬in rendering with ‫( בדיזמן‬AramB: ‘prepares’).
39
‫ ית‬suggests an active verb, but it seems to be used (with the subj.!) even in
passive renderings.
472 EXODUS 1–18
‫( בשׁמע יהוה‬16.8) Tgg again give a passive rendering, but Sy has the
active of the Heb. this time. TgNmg preserves an unusual use of ‫ גלי קדם‬for
God hearing.
‫( את־תלנתיכם‬16.8) Vulg has a pl. noun here but later in the verse it reverts
to the sing. exhibited elsewhere by LXX and Sy, while Sy paraphrases by
repeating the verbal expression used just before.
‫( עליו‬16.8) LXX καθ’ἡμῶν ‘corrects’ the Heb. to agree with the narrative
in v. 2 (and the wording of the end of v. 7).
‫( ונחנו מה לא עלינו‬16.8) LXX and Vulg both tighten the logic of Moses’
words here by inserting ‘For’, though at different points.
‫( על־יהוה‬16.8) LXX again uses θεός for the divine name here, as in
v. 7 (see the note there) and again in v. 9, but earlier in the verse it twice has
κύριος. Tgg created distance between the murmurings and the divine name,
either by inserting ‘the Memra of’ (TgO,J) or by replacing ‫ על‬with ‘before’
(TgN, like all the Tgg in v. 7): notably none of them felt the same scruple over
the suffix referring to God in ‫עליו‬.
‫( קרבו‬16.9) So all the Vss; in AramB (p. 71) TgN ‫ קרבו‬is rendered ‘Offer
sacrifice’, which the Pael can mean, but there is no reason to introduce this
idea here (cf. MRI [Lauterbach 2, p. 106]).
‫( שׁמע‬16.9) Tgg (but not Sy) again render as a passive with ‘before him/
the Lord’.
‫( תלנתיכם‬16.9) LXX, Vulg and Sy again have the sing. noun.
‫( בני־ישׂראל‬16.10) TgJ has just ‫ישׂראל‬, as in v. 1.
‫( ויפנו‬16.10) LXX ἐπεστράφησαν (cf. Theod.) follows the sense ‘turn’
for ‫פנה‬, like Tgg and Sy; Aq ἔνευσαν (cf. Symm) uses its regular equivalent
for ‫פנה‬, the change being not so much a change of meaning as to distinguish
between renderings of ‫( שׁוב‬which ἐπιστρέφω most often translates) and ‫פנה‬.
Vulg respexerunt plausibly prefers the sense ‘look’ (cf. the following ‫ והנה‬and
Lev. 26.9; Num. 12.10).
‫( והנה‬16.10) LXX has simply καί; the Three and the O-text add ἰδού.
‫( כבוד יהוה‬16.10) TgJ,N interpose ‘the Shekinah of’ as in v. 7 and else-
where: see the note on v. 7.
‫( נראה‬16.10) All the Tgg render with ‫אתגלי‬: for the equivalence see
Chester, Divine Revelation, pp. 31-79.
‫( בענן‬16.10) TgJ renders with ‫בענן יקרא‬, a phrase introduced already in
Gen. 2.6 (cf. also Lev. 16.2; Deut. 31.15) and one which reflects the close
association of ‫ כבוד‬and ‫ ענן‬in the Heb. text (cf. 24.16; 40.34-35; Num. 17.7).
‫( וידבר‬16.11) Sy uses ʾmr (rather than the expected mlyl) here and also for
‫ דבר‬in v. 12: cf. 14.1-2 and Text and Versions there. The Vatican ms. of TgO
omits the whole verse, as it does wherever this expression occurs (cf. AramB
6, pp. 36-37).
‫( שׁמעתי‬16.12) Tgg again paraphrase in the passive, as in vv. 7-8, but here
TgO,J also have the m. sing. form ‫שׁמיע‬, as TgN does throughout. TgNmg, as in
16.1-36 473
v. 8, preserves a rendering with ‫גלי‬. TgJ (cf. TgNmg) reads ‘before him’, mistak-
enly following its version of v. 7.
‫( תלונת‬16.12) SP ‫ תלנות‬clearly indicates the pl. form here. LXX and Sy
have the sing. again, but Vulg follows the pl. reading here. TgNmg has ‘your
murmurings’, another error based on v. 7.
‫( בין הערבים‬16.12) The Vss render essentially as they do in 12.6 (see the
comments in Text and Versions there). LXX prefixes τὸ to πρὸς ἑσπέραν,
perhaps to match τὸ πρωί in the next phrase, and Vulg vespere blurs the
precision of the Heb., so avoiding a conflict with ‫ בערב‬in the next verse: Vulg
generally uses ad vesperam/um, the only other exception being 29.39, where
v. 41 provides the more precise expression.
‫( בשׂר…לחם‬16.12) LXX and Vulg again have pl. forms.
‫( בערב‬16.13) LXX ἑσπέρα, apparently the subj. of ἐγένετο, could
suggest a Vorlage ‫( ערב‬for the expression cf. Gen. 1.5 etc.), but SP and all the
other Vss agree with MT: LXX itself might be a paraphrase of the majority
reading. The Three and the O-text read ἐν ἑσπέρᾳ.
‫( השׂלו‬16.13) See Note t on the translation for the variant (and secondary)
spellings of SP and some mss and edd. There is no Qumran evidence here but
at Num. 11.32 4QNumb reads ‫השלו‬. LXX ὀρτυγομήτρα (its regular rendering:
for its meaning see LSJ, p. 1257) and TgJ,Nmg ‘pheasants’ identify the birds as
a related species, for reasons that remain unclear.
‫( שׁכבת הטל‬16.13) The rarity of ‫ שׁכבה‬and the obvious inappropriateness
here of the senses which it has in its other occurrences (on which see Note u
on the translation) led the Vss to a variety of interpretations in this and the
following verse. LXX and Vulg saw ‫ שׁכבת‬as a verbal form; καταπαυομένης
τῆς δρόσου in the former possibly associated it with √‫( שׁבת‬so Barr, Compar-
ative Philology, p. 137 n. 2; cf. Wevers, Notes, p. 249), but perhaps more
likely uses καταπαύω in the sense ‘give rest to’, which it has several times
in LXX (cf. Exod. 10.14; 33.14; Deut. 3.20; 12.10; 25.19), as a tolerable
equivalent to ‫( שׁכב‬cf. BAlex, p. 183, ‘se déposait’, and its note).40 Vulg iacuit
has a more obvious connection with ‫( שׁכב‬like Aq, Symm and Theod in v. 14).
TgO ‫ נחתת‬is evidently based on Num. 11.9 (for the idea cf. 2 Sam. 17.12): the
vocalisation in Sperber’s text suggests a participle, ‘(was) coming down’, but
in v. 14 a constr. noun is required so probably here too, ‘(was) a descent, fall
of (dew)’; similarly Ibn Ezra, citing Job 38.37 and Num. 11.9. The other Tgg
and Sy are more speculative and obscure: TgJ ‫ אנחות‬was apparently meant in
the sense ‘tray, board’ (cf. Jastrow, p. 82) in view of the addition ‘congealed
40
Barr’s view presupposes that the translator read the word as ‫כשׁבת‬: in another
place he cites two further passages where LXX is based on a different ordering
of the consonants from MT (‘Vocalisation and the Analysis of Hebrew among the
Ancient Translators’, in Hebräische Wortforschung [FS W. Baumgartner; VTSup
16; Leiden, 1967], pp. 1-11 [10]).
474 EXODUS 1–18
[cf. AramB 2, p. 208 n. 14], arranged like tables’ (cf. MRI [Lauterbach,
p. 111]); TgN ‫( עננית‬or ‫ות‬-) may mean ‘a little cloud’ (AramB; cf. TgJ in v. 14,
but CAL follows Sokoloff with ‘a covering’); and Sy dymtʾ is rendered ‘a
mist’ or (here) ‘a fall of dew’ by Payne Smith, p. 90, but ‘a type of cloud’ in
CAL. None of this provides any basis for departing from the readily intel-
ligible text of MT (and SP) as it is now understood.
‫( ותעל שׁכבת הטל‬16.14) TgO,N and Sy follow MT, using the same equiva-
lents for ‫ שׁכבת‬as before. LXX* omitted the clause, perhaps thinking that it
added nothing new, but the gap was filled by the Three and in the O-text (which
used Symm’s version: cf. Wevers, Notes, p. 249 n. 10), and also in Vulg: its free
version (cumque operuisset [sc. the dew] superficiem terrae) seems to follow
the rabbinic tradition that the manna came down on top of the dew. The latter is
clearly present in TgJ’s expansion: ‘The clouds went up and brought down the
manna on the tray of dew’ (see further AramB 2, p. 208 nn. 15-16).
‫( והנה על־פני המדבר‬16.14) Vulg apparuit in solitudine is still free, but
closer to the Heb. than in the previous clause.
‫( דק‬16.14)1o While LXX λεπτόν and Vulg minutum represent the meaning
well, they do not retain the association of ‫ דק‬with grinding and crushing as
well as the equivalents in Tgg and Sy.
‫( מחספס‬16.14) 1QEx reads ‫כחספס‬, which might seem to be the Vorlage
for LXX and Vulg, which both have a comparative particle. But LXX ὡσεὶ
κόριον is certainly based on v. 31 (presumably to deal with an unintelligible
word) and Vulg’s quasi (pilo tunsum: see the discussion of this in Note v
on the translation) could have been derived independently from the context.
The reading of 1QEx (which Propp, p. 586, regards as original) could have
a similar origin, helped perhaps by the similarity between mem and kaph in
the Qumran script. The other Vss follow MT’s reading as a passive participle:
TgO ‫ מקלף‬and Sy wmtqlp, ‘peeled, stripped’, derive the Heb. from ‫( חשׂף‬cf.
Gen. 30.37), as Aq and Theod (latinised by BM and Wevers as decorticatum:
λελεπισμένον acc. Wevers, Notes, p. 250 n. 11) and Symm (manifestum: cf.
LXXFb ἀνασυρόμενον) will also have done; TgJ ‫ מסרגל‬and TgN,F ‫ מפספס‬are less
clear, but Salvesen, Symmachus, pp. 96-97, suggests ‘levelled out’ and ‘spread
out’ respectively (cf. Jastrow, p. 1023; CAL, s. vv.: it has ‘arranged in straight
lines’ for ‫מסרגל‬, which is closer to its meaning elsewhere).
‫( דק‬16.14)2o Most of the Vss render as before, but LXX λευκόν, ‘white’,
will have been erroneously imported from v. 31 with its equivalent for ‫מחספס‬
(cf. TgNmg) and Vulg does not render it all, seeing the repetition as otiose or
mistaken. TgO reinforces with ‫דגיר‬, ‘heaped up’ (so AramB), unless the variant
‫כגיר‬, ‘like lime’ (so Rashi), is followed.
‫( ככפר‬16.14) Some SP mss (cf. Tal, Crown, Camb. 1846) vocalise ‫ככופר‬.
πάχνη in LXXFb (and Symm?) is perhaps preferred because of the ambiguity
of LXX πάγος, which can mean ‘rock’.
‫( על־הארץ‬16.14) The ms. of TgJ prefixes ‘which (is)’, attaching the phrase
to ‘hoar frost’ since the manna has already been located earlier in the verse: it
need not presuppose a Vorlage different from MT and SP.
16.1-36 475
‫( ויראו‬16.15) LXX and Vulg subordinate to the following clause and add
an object for stylistic reasons; TgJ interposes ‘and they were amazed’.
‫( מן הוא‬16.15) LXX assumes that the meaning is the same as ‫ מה־הוא‬later
in the verse (cf. Symm), as does Vulg which transliterates the Heb. with man
hu before explaining it, in the way that some Semitic expressions are inter-
preted in the Gospels (e.g. Mk. 6.41). Sy mnʾ (5b1 mnw) probably intended the
same as LXX. TgO ‫( מנא הוא‬cf. TgN) and TgJ ‫ מאן הוא‬are commonly translated
in the same way (cf. AramB), but these Vss use ‫מא‬/‫ מה‬later in the verse for
‘what?’ and they may therefore imply ‘It is man(na)’ here (on the problem see
B. Malina, The Palestinian Manna Tradition, pp. 55-56). This is also a view
recorded in MRI (Lauterbach, p. 114) and it seems to lie behind the attempt at
an etymology in Rashi and Ibn Ezra.
‫( מה־הוא‬16.15) For LXX’s τί ἦν (the imperfect is due to the subordination
to οὐ γὰρ ᾔδεισαν) LXXFb preserves the Hebraising reading τί αὐτό.
‫( הלחם‬16.15) LXX ἄρτος and Vulg panis continue their specific rendering
of ‫לחם‬, though now more precisely in the sing. TgJ adds a further reference to
the primordial creation of manna (cf. v. 4).41
‫( לאכלה‬16.15) LXX φαγεῖν (cf. Vulg) renders as if there was an
infinitive in the Heb., assimilating to v. 8: in Gen. it uses βρωσις for ‫אכלה‬. The
forms used in Tgg and Sy may be more precise, as they are attested as nouns
as well as infinitives.
‫( זה‬16.16) Sy whnʾ makes the connection explicit: the asyndeton of the
other witnesses will be more original.
‫( אישׁ לפי אכלו‬16.16) LXX ἕκαστος εἰς τοὺς καθήκοντας (cf. v. 18),
‘each for his associates’, anticipates the clear specification at the end of the
verse (contrast LXX’s precise rendering in 12.4). καθήκων in this sense
is most unusual and perhaps an error of the translator for the classical
προσήκων. The Three follow MT (on the specific replacement of ἕκαστος
by ἀνήρ in Theod and Aq see O’Connell, Theodotionic Revision, pp. 275-78);
Vulg’s paraphrase omits any equivalent to the suffix of ‫אכלו‬, so avoiding any
conflict with the rest of the verse.
‫( עמר‬16.16) LXX and Vulg transliterate (γόμορ; gomor) and Tgg repro-
duce the Heb. word; Sy kylʾ, ‘a measure’, provides clarification without being
specific.
‫( לגלגלת‬16.16) TgN ‫ לכל גלגלת‬brings out the sense of the Heb. idiom.
‫( לאשׁר באהלו‬16.16) LXX σὺν τοῖς συσκηνίοις ὑμῶν, ‘with your tent-
mates’, is probably a free (and not entirely accurate) paraphrase of the Heb.;
the O-text has τοῖς ἐν in place of σύν and Vulg quae habitant in tabernaculo
41
G. Vermes, ‘ “He is the Bread”: Targum Neofiti Exodus 16:15’, in E.E. Ellis
and M. Wilcox (eds.), Neotestamentica et Semitica (FS M. Black; Edinburgh,
1969), pp. 256-63, took TgN’s ‫ הוא לחמא‬as a reference to Moses himself because
the ms. reads ‫ משׁה‬for the preceding ‫ מה־הוא‬of MT. But this is probably a scribal
error (cf. the editio princeps ad loc. and p. 60*; Childs, p. 274).
476 EXODUS 1–18
takes the correction further: TgO and Sy follow MT, on which TgJ,N are
evidently also based.
‫( וילקטו‬16.17) TgJ supplies an object by adding ‘the manna’.
‫( המרבה והממעיט‬16.17) Tgg (and in effect LXX) render literally; Symm
ὁ μέν… ὁ δέ… (cf. Wevers, Notes, p. 252 n. 14) improved the Greek style,
followed by Vulg alius…alius (cf. Sy). TgNmg records a version which added
‫שׁבטא‬, ‘the tribe’, to each expression (also in v. 18), presumably thinking of the
larger and smaller numbers for each tribe in the census lists (for a different
kind of specification see MRI [Lauterbach, p. 115]).
‫( וימדו‬16.18) LXX* probably rendered precisely by the indicative καὶ
ἐμέτρησαν. Rahlfs καὶ μετρήσαντες followed the reading of A(c), B and
some miniscules, but it does not connect with what follows and it is likely
to be a secondary variant influenced by the wording of v. 17 (Wevers, Notes,
p. 252; THGE, p. 219). In any case the underlying Heb. would be the same
(= MT, SP). Symm and Sy add ‘it’ to provide an object.
‫( בעמר‬16.18) LXX again transliterates, but Vulg ad mensuram gomor
recognises that a container is now involved: cf. perhaps TgJ’s addition of
‘from the measure’ in the next two clauses.
‫( ולא העדיף‬16.18) LXX and TgJ have no ‘and’: this could be the original
reading, as the conjunction is more often added than omitted. LXXO conforms
to MT as usual.42 A few SP mss, including Camb. Add. 1846 and one or two
other early mss, read ‫ עדיף‬here and ‫ חסיר‬in the next clause: the adjectives may
have been substituted for the Hiphils of the standard text as a typical instance
of SP simplifying an unusual expression. Unfortunately 4QpalExl preserves
only the end of ‫ )ה(עדיף‬and so is unable to show whether this was an early
variant.
‫( לפי אכלו‬16.18) LXX rendered as in v. 16 but with the addition of
παρ’ἑαυτῷ to reflect the suffix of ‫( אכלו‬presumably read as equivalent to
‫אכליו‬, ‘those who ate with him’): in v. 16 παρ’ἑαυτῷ was added in the O-text
alone for the same reason.
‫( לקטו‬16.18) Vulg congregarunt uses a verb which almost always applies
to people or animals; its stylistic concern not to repeat the same equivalent for
a Heb. word is also very evident in the two preceding clauses.
‫( ויאמר‬16.19) Most mss of Sy have just ʾmr, but 5b1 as usual keeps to MT
and preserves the waw.
‫( ולא שׁמעו‬16.20) TgO,J ‫ ולא קבילו‬as usual bring out the contextual sense
‘(did not) obey’; the other Vss use their standard renderings of ‫שׁמע‬.
‫( ויותרו‬16.20) Vulg’s dimiserunt should mean ‘let go’ rather than ‘leave,
keep’, and OL had used relinquo here as in v. 19: Vulg again seems to be
seeking variety, but less successfully than in v. 18.
42
The reading καὶ οὐχ εὗρον πλέον attributed to Symm in Syh does not make
sense and εὗρεν should be read instead.
16.1-36 477
‫( אנשׁים‬16.20) TgJ (cf. TgNmg) identified them as Dathan and Abiram, of

whom ‫ אנשׁים‬is used in Num. 16.26, and added ‘sinful’, its rendering for ‫רשׁעים‬
there (cf. the midrashic sources in AramB 2, p. 208 n. 22, and TgJ’s additions
in 2.13 and 14.3). MRI (Lauterbach 2, p. 116) instead gives the general and apt
explanation that the ‘men’ were ‘lacking in faith’ and ‘not good’.
‫( ממנו‬16.20) Vulg oddly has (quidam) ex eis, presumably taking the suffix
collectively (though it correctly had ex eo in v. 19!).
‫( וירם‬16.20) The closest rendering to the modern understanding of this
hapax legomenon (see Note bb on the translation) is TgN’s ‫ועבדו‬, which seems
to have recognised the tautology. LXX ἐξέζεσεν, ‘broke out in’ (idiomatic
Greek for a worm infestation: cf. Hdt. 4.205), Symm ἀνέβρασεν, ‘boiled up
with’, and the ‘swarmed’ of TgO,J and Sy are probably all (as more clearly the
‫ רמרם‬of TgF,Nmg) based on a free understanding of the form as derived from
‫‘ = רום‬be high’.
‫( תולעים‬16.20) Most Vss have a word for ‘worms’ here, but TgO has the
more general ‫רחשׁא‬, ‘creeping things’.
‫( ויבאשׁ‬16.20) LXX ἐπώζεσεν preserves the connotation of ‘smell’ which
is characteristic of BH ‫באשׁ‬, but the Aram. Vss use words which reflect the
wider sense, which it has in Aram., of ‘be bad, decay’: hence no doubt Vulg
computruit.
‫( ויקצף‬16.20) LXX ἐπικράνθη (only elsewhere for ‫ קצף‬at Jer. 37[MT].15)
surprisingly departs from its normal association of ‫ קצף‬with ὀργή and θυμός:
perhaps the translator thought that ‘vexation’ was more appropriate than
‘anger’ to the ‘humble’ (πραΰς) Moses of Num. 12.3 (cf. the ideal in Jos., AJ
4.328-29).
‫( אתו בבקר בבקר‬16.21) Vulg abbreviates, with no equivalent to ‫ אתו‬and
mane only once (but its imperfect verb is another way of indicating the
regularity). TgJ rephrases the temporal expression and makes additions to
locate the melting of the manna at ‘the fourth hour’ (cf. MRI [Lauterbach 2,
p. 117]).
‫( אישׁ כפי אכלו‬16.21) LXX ἕκαστος τὸ καθῆκον αὐτῷ is closer to the
sense than in vv. 16 and 18 but still omits any reference to eating. SP reads
‫ לפי‬as in vv. 16 and 18, a clear case of assimilation, and a Geniza ms. (cf.
BHS) and TgO,J do likewise; LXX, Vulg and Sy are too free for any conclu-
sion about their Vorlagen to be drawn. Aq and Symm appear to presuppose
‫כפי‬, but Theod εἰς τοὺς ἐσθίοντας αὐτό points to ‫לפי‬, with a rendering of
‫ אכלו‬which, though more literal, oddly analyses it like LXX earlier as a plural
part. with suffix.
‫( וחם‬16.21) SP ‫ וחמה‬adjusts the ending to fit the usual fem. gender of
‫שׁמשׁ‬. There is no Qumran evidence to determine the antiquity of the change.
TgO, like LXX and Vulg here and elsewhere, creates a subordinate clause out
of the Heb. parataxis and also prefixes ‘what remained of it (sc. the manna) on
the surface of the open country’ to provide a precise subject for ‫ונמס‬. TgO,J,N all
add ‘upon it’ to make explicit the connection with the manna. TgN curiously
has ‘rose’ for ‫( חם‬and its mg even more curiously ‘set’!).
478 EXODUS 1–18
‫( ונמס‬16.21) TgJ has a long addition about what happened to the molten
manna, as in MRI (Lauterbach 2, p. 117), except that the Israelites are the
beneficiaries rather than Gentiles.
‫( לחם‬16.22) LXX τὰ δέοντα, a Koine expression for the ‘necessities’ of
life, i.e. food (cf. 21.10; 1 Kgs 5.2; Prov. 30.8; Tob. 5.15; 2 Macc. 13.20: for
pars. cf. LSJ, p. 379; BAG, p. 171 [s.v. 5] with evidence from papyri). Aq,
Symm and Vulg follow MT, the last adding id est to connect with the next
phrase.
‫( לאחד‬16.22) TgJ,N (cf. Nmg) expand to make clear that all humans are meant.
‫( נשׂיאי‬16.22) Most of the Vss use general words for ‘leaders, officers’, but
Sy qšyšy equates them specifically with the ‘elders’ (for this word cf. 10.9;
17.5; 19.7: more often Sy uses sb for ‘elder’).
‫( העדה‬16.22) TgNmg prefixes ‫עם‬, the other word used for the community
in the chapter (cf. its additions in vv. 10 and 35).
‫( ויאמר‬16.23) LXXB carelessly has κύριος as the subject (cf. Wevers,
Notes, p. 254; but its addition of οὐ before τοῦτο is a stylistic ‘improvement’
intending a question: cf. BDF §440 and the question mark in Swete ad loc.);
almost all other LXX witnesses have ‘Moses’, like TgJ,N and Sy. 4QpalExl has
no addition here and there is probably insufficient space for one in the lacuna
after ‫( אלהם‬DJD IX, p. 37). The shorter text of MT and SP is certainly superior.
‫( הוא‬16.23) LXX (τοῦτο τὸ ῥῆμά) and TgN expand to agree with the
fuller expressions in vv. 16 and 32. Again the editor of 4QpalExl (loc. cit.)
considers that there would not have been space for the additional word in the
lacuna.
‫( יהוה‬16.23) TgNmg as often prefixes ‘the Memra of’; TgJ’s addition of
‫עבדתון‬, ‘you have done’, wrongly identifies what Yahweh has spoken of with
the people’s action rather than his provision (cf. v. 5).
‫( שׁבתון שׁבת־קדשׁ‬16.23) The order of words in the Heb. is unique here
(see Note gg on the translation) and only Vulg (requies sabbati sanctificata)
preserves it, though with ‫ שׁבתון‬apparently in the constr. st.: the rest follow the
usual order, with ‫ שׁבת‬before ‫שׁבתון‬. This is particularly clear in LXX and Sy,
where non-cognate words (ἀνάπαυσις and nyḥʾ) render ‫שׁבתון‬, but in TgO,J the
initial ‫ שׁבא‬is a form of the word for ‘sabbath’ (cf. Jastrow, p. 1509; CAL) and
TgN ‫ שבת שבתא קדש‬should probably be understood in the same way, as this is
its regular rendering of ‫ שׁבת שׁבתון‬elsewhere. The unusual wording of MT and
SP will have been assimilated in these Vss to the familiar expression found
elsewhere: its distinctiveness is appropriate here, as the Sabbath is being
mentioned for the first time by name and needs explanation.
‫( ליהוה‬16.23) Tgg as usual have ‫ קדם‬in place of ‫ל‬.
‫( את־אשׁר תאפו אפו ואת אשׁר־תבשׁלו בשׁלו‬16.23) LXX, Tgg and Sy render the
Heb. precisely, with the Aram. Vss employing the cognate verb in each case
and only TgJ adding ‘tomorrow’ and ‘today’ to insist that food to be eaten on
the Sabbath must be cooked on the day before (cf. the next note). Vulg curi-
ously renders the first clause in a very general way (quodcumque operandum
16.1-36 479
est facite) and the second in a way that that loses the specificity of the Heb.
here (as already in OL): quae coquenda sunt coquite. From its use of coquere
to render both ‫( בשׁל‬in 12.9) and ‫( אפה‬in 12.39) it seems to have been ignorant
of (or uninterested in) the distinction between them: Latin certainly had the
means to express it if required.
‫( ואת כל־העדף‬16.23) TgJ adds in the light of post-biblical interpretation
of the Sabbath law (as e.g. in M.Shab. 1.10; 3.1-3; cf. MRI [Lauterbach 2,
p. 118]: on evidence for the development of the regulations see de Vaux, Insti-
tutions 2, pp. 380-82, ET pp. 482-83; ABD 5, pp. 853-54) ‘from what you eat’,
i.e. what has already been cooked.
‫( לכם‬16.23) LXX has no equivalent: it probably seemed unnecessary, but
Heb. generally specifies the keepers (cf. 12.6).
‫( למשׁמרת‬16.23) LXX εἰς ἀποθήκην, ‘into store’ (as again in v. 32), and
Sy qryrʾ, ‘cool’ (or possibly ‘as dough’ [cf. 12.34]) depart from their usual
renderings in the light of the context.
‫( עד־הבקר‬16.23) SP ‫עד בקר‬, without the article; both MT and SP exhibit
the variation (cf. vv. 19-20, 24), but SP several times assimilates to a
preceding form (so in 14.27 after 14.24, and in 29.34 after 27.21: but not in
16.24!) and may do the same here. 4QpalExl does not preserve the end of the
verse and the Vss provide no decisive evidence (not even LXX εἰς πρωί: cf.
vv. 19-20, 24!).
‫( ויניחו אתו עד־הבקר‬16.24) Vulg feceruntque ita avoids repetition and
abbreviates: LXX and Sy have ‘from it’, i.e. ‘some of it’, for ‫אתו‬, hardly
appropriately and probably drawing on vv. 19-20 (cf. Wevers, Notes, p. 255).
‫( צוה‬16.24) LXX adds αὐτοῖς, which is obelised in Syh and omitted in
part of the O-text. Sy pqd ʾnwn again agrees with LXX.
‫( היתה‬16.24) So also SP, but Sadaqa has ‫יהיה‬, which must be a slip based
on v. 26, either in his ms. or in his copying. Vulg inventus est provides enrich-
ment to the bare statement of the Heb., but at the price of adding a fourth
occurrence of the verb to the three in vv. 25-27.
‫( בו‬16.24) ἐν αὐτῷ (Rahlfs, Wevers) must be the original LXX reading:
LXXB’s ἐν αὐτοῖς can only be an error, perhaps caused by its αὐτοῖς earlier
in the verse. TgN ‫ בגויה‬is a slight imprecision, which its mg corrects.
‫( ויאמר משׁה‬16.25) Sy in this case does not (except for 5b1) represent the
waw, preferring asyndeton for the direct response (cf. vv. 19, 23); as often it
specifies the addressee, adding lhwn.
‫( אכלהו‬16.25) LXX φάγετε, with the object understood: the Three and
the O-text add αὐτό to agree with MT. Possibly the suffix is secondary in MT
and the other witnesses (cf. below on ‫)תמצאהו‬, but both the main Heb. tradi-
tions have it.
‫( היום‬16.25)2o Vulg has no equivalent, but in view of its tendencies to
abbreviate and avoid repetition this need not imply that it was lacking in
Vulg’s Vorlage.
‫( ליהוה‬16.25) Tgg as usual have ‫ קדם‬in place of ‫ ל‬before the divine name.
480 EXODUS 1–18
‫( היום‬16.25)3o LXX* has no equivalent, but the translator may have

thought it otiose: the Three, Syh and one ms. of the O-text add σήμερον.
‫( לא תמצאהו‬16.25) SP ‫ לא תמצאו‬leaves the object unexpressed, as do
LXXAFM and most other mss. The O-text and the Three agree with MT, as
do Tgg and Sy. LXXB, a few miniscules and Vulg turn the expression into
a passive form and both Rahlfs and Wevers regard this as the original LXX
reading: in any case it probably presupposes MT’s suffix. Again, however, the
latter could be secondary (cf. the wording at the end of v. 27).
‫( בשׂדה‬16.25) TgN’s ‫ באפי ברא‬might seem to be an expansion based on
MT’s phrase in v. 14 (‫)על־פני המדבר‬, but its occurrences in 9.19, 21, 22, 25
suggest that it is probably just an idiomatic equivalent.
‫( תלקטהו‬16.26) As in the previous verse, evidence for the suffix is
divided: LXX (apart from the O-text) and Vulg ignore it – both of them also
render the imperfect with an imperative (cf. ‫ אכלהו‬in v. 25) – as does Sy5b1. But
in addition to MT and SP it seems to have been present in 4QpalExl, and it
is translated in Tgg and the other mss of Sy. The agreement of the three Heb.
witnesses is probably decisive.
‫( וביום השׁביעי‬16.26) So Sy5b1, but the remaining Sy mss omit ‫ב‬, no doubt
influenced by the familiar wording of the Decalogue (20.10: both there and
in Deut. 5.14 there is a variant with ‫ב‬, which could have been based on the
present passage).
‫( שׁבת‬16.26) Vulg adds Domino from the previous verse; TgJ ‘which is the
sabbath’ makes the second half of the verse into a single sentence.
‫( לא יהיה־בו‬16.26) The isolation and total pronominalisation of this state-
ment invited clarification, either by a backwards connection (LXXB,mss ὅτι
[on which see Wevers, THGE, p. 244]; Vulg more appositely idcirco) or by
rewording (Vulg non invenietur as in v. 25, but here further from MT) or by
additions: TgJ adds ‘the manna’ and ‘coming down’, TgN ‫מניה‬, ‘any of it’.
LXX*, TgO and Sy (if its hwʾ is a part. equivalent to a future) simply repro-
duced the Heb.
‫( ויהי ביום השׁביעי‬16.27) Vulg venit septima dies is a neat stylistic
paraphrase, but it exposes the duplication in the narrative all the more clearly!
‫( יצאו‬16.27) Vulg et egressi avoids the need for coordination later in the
verse, but it has to add et here to connect with venit.
‫( מן־העם‬16.27) TgJ (cf. TgNmg) adds ‘the wicked’; MRI (Lauterbach 2,
p. 120) as in v. 20 says that they were people of little faith.
‫( ללקט‬16.27) TgJ adds ‘the manna’ again.
After this verse 4QpalExl had both an incomplete line of text and an empty
line (cf. DJD IX, pp. 19-20), corresponding to the division in MT and SP mss.
‫( מאנתם‬16.28) On the versional renderings see Text and Versions on 9.2
and 10.3 and Wevers, Notes, p. 257. Here the present tense intended is well
represented in the forms chosen.
16.1-36 481
‫( לשׁמר‬16.28) LXX εἰσακούειν is a rare equivalent for ‫שׁמר‬, being most

often used for ‫שׁמע‬, as in v. 20: its use there probably suggested it to the trans-
lator here. The other Vss render the Heb. more precisely.
‫( מצותי ותורתי‬16.28) Although the vocalisation of MT makes both words
pl. (cf. TgN, MRI [Lauterbach 2, p. 121], Sy), their consonants could be read
as sing. (likewise the majority of SP mss and, perhaps, 4QpalExl: see DJD IX,
p. 37). There is support for a sing. rendering of ‫ תורתי‬in LXX, Vulg, TgO and
TgNmg (‫ואחוויית אורתי‬, which is a fuller version of TgJ’s reading here: cf. Lev.
14.54; Num. 6.13), but ‫ מצותי‬is taken as a pl.
‫( כי‬16.29) LXX γάρ takes ‫ כי‬as causal, but Vulg quod (cf. vv. 6, 12) and
probably the other Vss intend the sense ‘that’.
‫( השׁבת‬16.29) SP ‘improves’ the grammar by prefixing ‫ ;את‬LXX prefixes
τὴν ἡμέραν ταύτην, ‘this day (as)’, emphasising like ‘today’ in v. 25 that the
words (and especially those in v. 29b) were spoken on the Sabbath day and
refer to it: compare the parenesis in the first comment on the verse in MRI
(Lauterbach 2, p. 122). 4QpalExl does not survive at this point, but there is
insufficient space for LXX’s longer text (DJD IX, p. 37).
‫( הוא‬16.29) Sy (except 5b1 and some later mss) replaces ‫ הוא‬with mryʾ,
presumably for added emphasis: this may also be why LXX unnecessarily
rendered ‫ הוא‬by αὐτός.
‫( נ ֵֹתן‬16.29) LXX ἔδωκεν, Vulg tribuerit (perf. subj.) and perhaps TgN
render as a perfect rather than MT’s part., assimilating to the form earlier in
the verse (Sy yhb may be a part. here).
‫( לחם יומים‬16.29) LXX and Vulg render ‫ לחם‬in the pl., as earlier in the
chapter. For ‫ יומים‬Vulg and TgN have ‘double’, perhaps from v. 22.
‫( תחתיו‬16.29) See Text and Versions on 10.23 for the similar treatment
of the same expression there; LXX* εἰς τοὺς οἴκους (!) ὑμῶν again prefers
a concrete equivalent here, though in other books this is avoided. Many mss
have παρ’ἑαυτῷ here, including some witnesses to the O-text, from which it
may derive.
‫( אל‬16.29) So also SP according to von Gall, but many of his mss read
‫ואל‬, as do those used by Sadaqa, Tal and Crown (two) and Camb. 1846, so this
is the predominant SP reading. But it will be secondary, despite the support of
Sy and a few LXX witnesses, since the addition of waw is a frequent textual
development.
‫( ממקמו‬16.29) TgJ adds ‘to travel beyond 2000 cubits’, reflecting the limit
imposed in rabbinic law (as Rashi notes), just as after its equivalent to ‫תחתיו‬
it has ‘and do not remove anything from “territory” to “territory” beyond 4
cubits’ (cf. MRI [Lauterbach 2, p. 122]; Jub. 2.30; 50.8). Sy’s specific mn trʿ
byth, ‘from the door of his house’, probably reflects a Jewish background (as
might LXX’s οἴκους in the previous clause), although Weitzman finds this
more plausible for the unique (and perhaps older) reading of 5b1, mn tḥwmh,
482 EXODUS 1–18
‘from his limit’, a technical expression of rabbinic law, and attributes the
majority reading to the influence of 12.22 (Syriac Version, p. 159; for MH
‫ תחום‬cf. Jastrow, p. 1660).
‫( השׁביעי‬16.29) SP reads ‫( השׁבת‬cf. some mss of Sy); the Vss support
MT. Both expressions occur in the context, but ‫ השׁביעי‬is regular in the other
temporal expressions nearby: so one might well conclude that SP has simply
repeated ‫ השׁבת‬from earlier in this verse.
‫( וישׁבתו‬16.30) SP has the sing. form to correspond exactly to the subj.
‫ ;העם‬likewise LXX, Vulg and some later Sy mss (but see Text and Versions on
15.24). MT follows the common use of the pl. with collectives (GK §145b-c)
and will be more original. LXX ἐσαββάτισεν, ‘observed the sabbath’ (cf.
Vulg), took ‫ ָשׁ ַבת‬as a denominative here (cf. Lev. 23.32; 25.2: more commonly
so in MH and JAram; cf. MRI [Lauterbach 2, pp. 122-23] and Jastrow,
pp. 1519-20), but not elsewhere in Exodus, even in the very similar verses
23.12 and 34.21: thus LXX marks explicitly the first observance of the new
commandment.43 TgO ‫ ושׁבתו‬may have a similar sense, but the other Tgg and
Sy have ‘rested’, which also appears in some witnesses to TgO, mainly printed
edd.
‫( ויקראו‬16.31) Only Vulg has a sing. verb here, to agree with its rendering
of ‫( בית‬on LXX see the next note).
‫( בית ישׂראל‬16.31) LXX, TgN and Sy have ‘the children of Israel’,
conforming to the regular designation used earlier in vv. 1-3, 9-10, 12, 15, 17.44
SP, 4QpalExl, 4QpalExm and the other Vss agree with MT, which is clearly
original just because it is so unusual in Exodus (see the Explanatory Note).
‫( מן‬16.31) So also SP: LXX and Vulg transliterate the Heb. here, as in
vv. 33 and 35 (so Vulg also in Num. 11 and Ps. 78 [Heb.]), while the Aram.
versions naturally have the longer emphatic forms (from which, through
their transliterated forms in LXX and Vulg elsewhere, the familiar ‘manna’
is derived).45
‫( גד‬16.31) The main ancient Vss (except for TgO: see Maiberger, Das
Manna, p. 180) all have words for ‘coriander’ but SamGk oddly has ὀρύζης,
‘of rice’ (so also SamTgJ ‫כארז קליף‬, similarly at Num. 11.7: cf. Tal, Dictionary
1, p. 60, ‘like peeled rice’).
‫( כצפיחת‬16.31) Most of the Vss plausibly render the hapax legomenon of
the Heb. with words for ‘flour’ or ‘cakes’ (cf. MRI [Lauterbach 2, p. 124]),
which have then tended to influence renderings, ancient and modern, of ‫ לשׁד‬in
43
The ἤργησεν, ‘did nothing’, of Fb will be a ‘normalising’, non-specific
correction: its source is unknown, but cf. TgJ,N and Sy.
44
MRI (Lauterbach 2, p. 123) attributes this reading to the ‘Allegorists’.
45
Walters (p. 170) suggested that LXX’s exceptional adherence to the Heb.
form in vv. 31-35 was designed to recall v. 15, where ‫ מן‬is transliterated in some
of the Vss (but not LXX!).
16.1-36 483
Num. 11.8; but Sy and Symm have ‘honeycomb’ (kkrytʾ ddbšʾ; μελικήριον);
for ‫ צפיחת בדבשׁ‬here Aq apparently replaced ‘honey’ with the ‘oil’ (ἔλαιον) of
Num. 11.8: on both verses see Salvesen, Symmachus, pp. 122-24.
‫( יהוה‬16.32) TgNmg as often prefixes ‘the Memra of’.
‫( מלא‬16.32) The absence of a verb caused SP and most of the Vss some
difficulty, which was resolved in two different ways: either ‫ מלא‬was itself
taken to be an imperative verb, whether sing. (closer to the Heb.: Vulg) or pl.
(conforming to the second pl. forms later in the verse: SP, LXX, TgO), or a
verb was supplied (Sy, TgJ). TgN ‫ מלה‬could be a sing. imperative or the noun
with ‘let (it) be’ understood, as in MT.
‫( ממנו‬16.32) LXX’s τοῦ μάν might be understood as a clarification (so
Wevers, Notes, p. 259: cf. TgJ in vv. 26-27), but one is scarcely needed so
soon after ‫ מן‬in v. 31 and so perhaps LXX had (or thought it had) a Vorlage
which corresponded (most likely as the result of a secondary assimilation) to
the similar phrase in v. 33.46
‫( למשׁמרת‬16.32) On LXX see the note on v. 23.
‫( יראו‬16.32) Vulg noverint, ‘may know’, is hard to explain as a contextual
or paraphrasing rendering and seems most likely to be due to knowledge of
the itacistic LXX variant εἰδῶσιν (on which see Wevers, Notes, p. 259). TgJ
adds ‘the rebellious generations’ as the subject (cf. MRI’s comparison of Jer.
2.31 [Lauterbach 2, p. 126]).
‫( האכלתי אתכם‬16.32) LXX ἐφάγετε ὑμεῖς is probably a free or careless
rendering of the wording presupposed in the other witnesses: the pointless
ὑμεῖς must be based on ‫ אתכם‬or something like it. Fritsch (p. 51) connects it
with other passages where LXX avoids making God the subject (1.21; 23.7;
34.9).
‫( בהוציאי אתכם‬16.32) Again LXX has a free rendering, ὡς ἐξήγαγεν
ὑμᾶς κύριος, which echoes a recurrent formula of ch. 13 as well as the
Decalogue; Vulg educti estis exhibits a tendency to use passive forms for the
sake of Latin style (with this verb cf. Num. 19.3; Josh. 10.24).
‫( מארץ מצרים‬16.32) Sy has no equivalent to ‫ארץ‬, as often in this formula
(see the note on v. 1). After ‫ מצרים‬4QpalExm has a short interval which is not
paralleled in SP or MT, but one may have been present in 4QpalExl (DJD IX,
p. 38).
‫( צנצנת‬16.33) The Vss all have, as the context would suggest, words for
‘jar’ or ‘flask’: LXX (χρυσοῦν, presumably on the analogy of the other golden
objects in the Tabernacle: cf. Heb. 9.4) and TgJ (‫דפחר‬, ‘of clay’: cf. MRI
[Lauterbach 2, p. 125], which deduces this from the need to keep the manna
cool [‫]צנן‬, as in Sy at v. 23) add descriptions of the material used.
46
AramB 2, p. 72 and n. oo, takes TgN ‫ מנה‬to mean ‘(of) manna’ as in v. 31, but
Tg often writes the third person m.s. suffix defectively (e.g. vv. 16 and 19 above:
N
cf. 12.44-45) and may simply be following MT here.

484 EXODUS 1–18
‫( שׁמה‬16.33) SP has ‫שׁם‬, which can also mean ‘thither’ in BH (cf. BDB,
p. 1027), as it sometimes does elsewhere when MT has ‫( שׁמה‬e.g. Gen. 14.10;
23.13), but not always (cf. 10.26), and sometimes the readings are reversed
(Num. 35.6, 11). Neither of the Qumran palaeo-Heb. mss preserves this word
and the versional renderings are inconclusive. Although GSH §148dβ attaches
little significance to the SP variants, since they were pronounced the same, one
might argue that here it is MT which has introduced the more specific form
secondarily.
‫( מלא‬16.33) LXX πλῆρες, ‘full’, i.e. ‘in full’, is presumably the comple-
ment (cf. 9.8, where ‫ מלא‬must be the noun and so also here). Vulg paraphrases.
‫( והנח‬16.33) SP ‫ והניח‬is its regular form for the Hiphil m. sing. imperative,
but its use here suggests that Sam. Heb. did not formally distinguish the two
senses of ‫ נוח‬Hiphil as the Masoretes did (see the note on v. 23). 4QpalExm
agrees with MT’s spelling of the consonants, but its vocalisation is of course
unknown. LXX has καὶ ἀποθήσεις: the future is a relatively rare equivalent
for the Heb. imperative (23 cases in the Pentateuch acc. to the tables in Evans,
Verbal Syntax, pp. 281-96).
‫( יהוה‬16.33) LXX τοῦ θεοῦ as occasionally elsewhere (cf. Lemmelijn, p.
138), but all the other witnesses (inc. 4QpalExm) agree with MT. TgN adds ‘for
a testimony’, summarising v. 32b: the idea is also found in MRI (Lauterbach
2, p. 126).
‫( למשׁמרת‬16.33) LXX now abandons its previous rendering (see the notes
on vv. 23 and 32) in favour of the more appropriate εἰς διατηρησιν (as also
in v. 34). Wevers, Notes, p. 260, offers a possible explanation for the change
at this point.
‫( כאשׁר‬16.34) It is clear that Sy (see the next note) and Vulg (in view of
its ‘and’ [posuitque] after it) took v. 34a as the conclusion of v. 33. Wevers
may well be correct in thinking that LXX did too (cf. Notes, p. 261), since ὃν
τρόπον almost always connects backwards and its και after v. 34a suggests a
break there: the waw apodosis is generally not given an equivalent in Exod.
(see Aejmaleus, ‘The Significance of Clause Connectors in the Syntactical
and Translation-Technical Study of the Septuagint’, in On the Trail, pp. 49-64
[57]). In the Tgg the function of the waw (which they all reproduce) is unclear;
but AramB in each case treats it as waw apodosis. On the likely clause struc-
ture of the Heb. see Note tt on the translation: it is not necessary to suppose
that a clause has been lost at the beginning of the verse (BHS).
‫( אל־משׁה‬16.34) SP, 4QpalExm and a Genizah ms. (BHS) read ‫ את‬for ‫אל‬
(cf. TgJ,N), while TgO ‫ למשׁה‬supports the MT reading; the datives of LXX and
Vulg are inconclusive because the associated verbs regularly take the dative
anyway. Sy pqdny, ‘commanded me’, follows out the logical consequences
of linking v. 34a to the previous verse (see the previous note) and leaves
its Vorlage obscure. Sanderson (Exodus Scroll, p. 58) has argued that ‫את‬
is original, because 38 of the other 39 occurrences of ‫ צוה‬in Exodus which
identify the recipient of the command do so with ‫( את‬the exception is in 1.22).
16.1-36 485
The textual evidence for ‫ את‬is weighty too, but weakened by the possibility
that it is due to ‘normalising’ the grammar (a frequent characteristic of the
tradition to which SP and 4QpalExm belong). The rare construction of MT
(which is occasionally found: see Note ss on the translation) is probably
original here.
‫( ויניחהו‬16.34) Von Gall printed ‫ וינחהו‬as the SP reading, but it is supported
by only three of his mss (not the earliest) and that used by Tal: the remainder
read as in MT (and 4QpalExl and 4QpalExm) and surely preserve the original
SP reading.47
‫( לפני העדת‬16.34) The Vss uniformly (except for Vulg which paraphrases
with in tabernaculo here) render ‫ עדת‬with words for ‘testimony, witness’
(LXX τοῦ μαρτυρίου, Tgg ‫[ סהדותא‬cf. Sy]), as they continue to do in its
many occurrences from 25.16 onwards. This is almost certainly not its true
meaning, which was forgotten at an early date (see Note uu on the transla-
tion). Sy lshdwtʾ ignores the local implications of ‫ לפני‬and takes the word in a
general sense, no doubt to avoid the anachronism which the Heb. seemed to
create (for its positive intention cf. TgN’s addition of the same word in v. 33).
‫( ובני ישׂראל‬16.35) TgNmg prefixes ‫ עם‬as in v. 22 to include the alternative
designation for the Israelites (used most recently in v. 30).
‫( ארבעים שׁנה‬16.35) Most Sy mss agree with MT, but 7a1 and 7a13 prefix
bmdbrʾ (cf. v. 32). TgJ added ‘in the lifetime of Moses’ (see the next note but
one).
‫( ארץ נושׁבת‬16.35) TgN has ‘the land of their dwellings’, a phrase it used
in Deut. 2.12 for ‫ ארץ ירשׁתו‬and earlier in Exod. 6.4 (and Gen.) for ‫ארץ מגורים‬:
the intention is to emphasise that this was to be Israel’s own land.
‫( אכלו‬16.35)2o TgJ adds ‘for forty days after his [sc. Moses’] death’ and
after ‫‘ עד‬they crossed the Jordan and’: cf. MRI (Lauterbach 2, pp. 126-28),
where alternative calculations are also given, and other refs. in AramB 2,
p. 209 n. 31.
‫( אל־קצה ארץ כנען‬16.35) For ‫ קצה‬ms. 5b1 of Sy and some others have a pl.
equivalent, a common Aram. idiom (cf. TgO,J,N and 13.20): 7a1’s sing. lswph
may be a pedantic correction to MT. LXX εἰς μέρος τῆς Φοινίκης avoids
the precise rendering used in 15.15 (and its plus in 12.40), presumably to
give readers a contemporary equivalent (cf. 1.11, perhaps 14.2, 9 and [for this
expression] 6.15 and Josh. 5.1, 12): for the wider extent of Φοινίκη (reflected
in μέρος here) see Strabo, Geog. 2.5.24. Symm corrected to MT, which Vulg
also followed (cf. SP, 4QpalExl, Tgg, Sy).
47
Both Rahlfs and Wevers adopt καὶ ἀπέθετο as the original LXX reading,
rather than καὶ ἀπέθηκεν of B etc., but Wevers’ justification for this (if that is
what it is meant to be: Notes, p. 261) is hardly persuasive: if the active form is
indeed secondary (and it is used twice in LXX Num.), it is best regarded as an
Atticism (cf. the listings of occurrences in LSJ, p. 223).
486 EXODUS 1–18
There was space for 16.36 in the lacuna at the end of a line in 4QpalExl,
but DJD IX, p. 93, notes that there seems not to be room for all of vv. 35-36 in
ll. 5-7 of col. xvii of 4QpalExm, since the enlarged waw of 17.1 appears in the
middle of l. 7. The editors conclude that it is ‘most likely that the parenthetic
statement which is 16.36 was not present’. This is certainly one possibility, but
it is not the only one: the repetitiveness of v. 35 means that an omission could
have occurred there by homoeoteleuton (most likely on ‫ את־המן‬or ‫ )עד־באם‬and
then there would have been room for v. 36. For another (possible) case of
omission by homoeoteleuton in 4QpalExm see 31.13-14 (DJD IX, pp. 122-23).
‫( האיפה‬16.35) Most of the Vss assumed that ‫ האיפה‬would no longer be
intelligible and expressed its meaning in other terms (only Aq and Vulg
transliterated the Heb. word). LXX has τῶν τριῶν μέτρων (cf. Isa. 5.10),
where μέτρον has a specific equivalence (cf. LSJ, p. 1089): in Gen. 18.6 it
corresponds to Heb. ‫סאה‬. Tgg, with ‘a tenth part of three seahs’, agree with
this reckoning (cf. also B.Men. 77a). Sy wkylʾ ḥd mn ʿsrʾ hwʾ lsʾtʾ assumes
that a ‫ סאה‬was equal to an ‫איפה‬, as it does also in Isa. 5.10. Payne Smith, p.
356, gives a much larger modern equivalent for sʾʾ than is usual for Heb. ‫סאה‬,
which (if reliable) would explain the anomaly. But Sy may have been misled
by the fact that LXX often has μέτρον as an imprecise equivalent to ‫איפה‬
when the exact quantity meant is immaterial (cf. Deut. 25.14-15; Prov. 20.10;
Amos 8.5; Zech. 5 passim). MRI (Lauterbach 2, p. 128) gives its equivalent
in different terms.
C h ap t er 1 7 . 1 - 7
W at er fr om th e R ock at R e p hi di m
The beginning and end of the section correspond to the narrative

structure of the text and to corresponding divisions in both MT
and SP (for the beginning 4QpalExl, 4QpalExm and less certainly
4QExc provide early agreement; at the end all three mss ended
v. 7 in mid-line [see Text and Versions]). The section begins with a
note of movement from the location of the last episode and, while
v. 8 places the following episode at the same location as vv. 1-7, the
two sections of the chapter deal with quite different kinds of danger
for the Israelites, lack of water in the former and an enemy attack
in the latter. SP has a further division after v. 4 (so also probably
4QExc and 4QpalExm), which will be due to the opening of divine
speech in v. 5 and is not significant for the separation of major
textual units.
The section is dominated by an alternating sequence of utter-
ances by different speakers, who generally respond to a contrasting
previous utterance: the people and Moses (v. 2); the people (v. 3b);
Moses (v. 4); Yahweh (vv. 5-6a); Moses (implied) and the people (as
if recapitulated) (v. 7). By contrast actions and states of affairs are
in the background (vv. 1, 3a) or recounted in summary form (v. 6b).
Eight short sub-units can be distinguished: (i) itinerary-note marking
the movement to Rephidim, where there was no water (v. 1); (ii) the
people’s ‘dispute’ with Moses and request for water (v. 2a); Moses’
disputatious response (v. 2b); (iii) the people’s complaint against
Moses because they are dying of thirst (v. 3); (iv) Moses’ appeal
to Yahweh for help (v. 4) and Yahweh’s reply to Moses, instructing
him what to do and promising that by his oversight the people will
have water to drink (vv. 5-6a); (v) a summary of Moses’ fulfilment
of Yahweh’s instructions (v. 6b); (vii) Moses names the place in
accordance with (his view of) the people’s initial behaviour (v. 7a);
and (viii) a purported ‘citation’ of the people’s earlier words (v. 7b).
The episode is located in the vicinity of the mountain of God,
both by the sequence of the itinerary (cf. 19.2) and by the mention
of ‘Horeb’ in v. 6, that is, close to the place where Yahweh will meet
488 EXODUS 1–18
with the people and make a covenant with them and they will rebel
against him by making and worshipping a golden ‘calf’ (Exod.
32.1-6).
A very similar episode, but with some notable differences, is
narrated near the end of the wilderness journey (Num. 20.1-13)
and its characteristics need to be kept in mind in any assessment of
Exod. 17.1-7. It comprises the following elements: (i) itinerary-note
recording the people’s arrival at Kadesh in the wilderness of Zin
(v. 1a); (ii) the death and burial of Miriam (v. 1b); (iii) lack of water
(v. 2a: cf. Exod. 17.1b); (iv) the people’s dispute with Moses (and
Aaron) and the charges they bring against them (vv. 2b-5: cf. Exod.
17.2a, 3b); (v) recourse of Moses and Aaron to the Tent of Meeting
(v. 6a: cf. Exod. 17.4); (vi) appearance of Yahweh’s glory to them
and Yahweh’s instruction of them about bringing water from a rock
by speaking to it in the presence of the people (vv. 6b-8: cf. Exod.
17.5-6a); (vii) the action of Moses (and Aaron) in taking ‘the staff’
from ‘before Yahweh’, gathering the people together, challenging
them and striking the rock with the staff which produces water for
the people to drink (vv. 9-11: cf. Exod. 17.2b, 6b); (viii) Yahweh
rebukes Moses and Aaron for their lack of faith and reverence
towards him and declares that neither of them will enter the promised
land (v. 12); (ix) identification of the place (by the narrator) as ‘the
waters of Meribah’, where the people disputed with Yahweh (v. 13:
cf. Exod. 17.7).
The similarities are evident from the cross-references given: most
of the elements of the Exodus story are there. But not all are present,
and some details have been changed or moved and others have been
added. The addition of Aaron alongside Moses, the description of
the people as ‘the congregation’ (hāʿēdāh) and the terminology used
in v. 6 leave no doubt that the passage is of (mainly) Priestly author-
ship, and the incorporation of Moses and Aaron’s disobedience at
the end produces a completely different outcome, which is picked
up in several later passages (Num. 20.23-29; 27.12-14; 33.38-39;
Deut. 32.48-52). Nothing is said here about Massah or about the
‘testing’ of Yahweh by the people. Other passages, in the Psalms,
also handle the story in a partly different way from Exod. 17.1-7 (cf.
Pss. 78.15-16; 105.41; 106.32-33).
But the Exodus passage itself displays some inconsistency in its
own presentation of the episode which has given rise to a variety
of critical assessments of its origins. Verses 1b-2 and 3 look like
17.1-7 489
alternative introductions (as they stand they overlap to a consider-

able extent); the names ‘Horeb’ and ‘Meribah’ (which is elsewhere
associated with Kadesh: cf. Num. 20.13; 27.14; Deut. 32.41; Ezek.
47.19; 48.28) relate to different stages on the journey; and in v.
7 words of the people are quoted which are quite different from
anything that they say earlier in the section.
Apart from v. 1 (the attribution of which is discussed separately below),

Knobel assigned the whole passage to his Rechtsbuch (the later E: cf. Exod.-
Lev., pp. 157-58; Num.-Jos., p. 532). Wellhausen ‘reluctantly’, as he says
himself (Composition, p. 79), recognised the signs of multiple authorship but
could not produce a full analysis and Holzinger got little further. The first to
produce a clear division was Dillmann, who assigned vv. 3-6 to E (his B) and
vv. 2 and 7 to J (his C: similarly, with an unusual correlation with parts of
Num. 20.1-13, C.H. Cornill in his ‘Beiträge zur Pentateuchkritik’, ZAW 11
[1891], pp. 1-34 [20, 33]), and later Smend, Eissfeldt and Beer followed him,
allocating in addition only v. 6aα from the central section to J. But before
long a much more complex subdivision of the passage between J and E,
which distinguished between a ‘Massah story’ and a ‘Meribah story’, became
the norm (cf. Carpenter/Harford-Battersby, Baentsch, Gressmann, McNeile)
and this still found support with Hyatt and, apparently, Childs. Rudolph (pp.
36-37) reverted to the older view of the passage’s unity (cf. Eerdmans) and
so eliminated the E strand, but to deal with the problems he had to transpose
v. 3aα after v. 1bα and ‘in Horeb’ to the end of v. 5, in both cases without
any textual support. Subsequent scholarship has gradually abandoned the
quest for parallel sources in the central verses of the passage, and dealt with
the problems in different ways. Noth thought that v. 3 was from E and the
Massah/testing sections and ‘in Horeb’ from a redactor (pp. 110-12, ET,
pp. 137-40; ÜGP, pp. 32, 39), and explanations for all the non-J material as
variants in tradition or redactional additions became increasingly popular
(Coats [Rebellion, pp. 53-82 (54-63)], Fritz, Ruprecht [‘Stellung und Bedeu-
tung’, pp. 302-304], W.H. Schmidt, Houtman, Schart, Levin, Graupner).
Schart in fact revived the view of Dillmann and Cornill (without the latter’s
diversions into Num. 20) that only vv. 2 and 7 needed to be distinguished from
the rest, in his opinion as an early redactional addition related to the mention
of ‘testing’ and divine presence in Num. 14.11, 22, 39-45 (Mose und Israel,
pp. 167-72: cf. N. Macdonald, ‘Anticipations of Horeb: Exodus 17 as Inner-
Biblical Commentary’, in G. Khan and D. Lipton [eds.], Studies on the Text
and Versions of the Hebrew Bible in Honour of Robert Gordon [VTSup 149;
Leiden, 2012], pp. 7-19 [10 (with n. 7) and 13 (with n. 16)]).
Blum (Studien, pp. 148-52) and Van Seters (Life, pp. 191-98) by contrast
agreed with Rudolph about the unity of the passage, without any need for
transpositions of text or substantial redactional additions. This was partly due
to a better perception of the literary connections within the passage and partly
490 EXODUS 1–18
to the later dates which they gave to the literary works in which the passage
was included (Kd according to Blum, exilic J according to Van Seters). For
one or both of these reasons the majority of recent commentators have taken
a similar view (Propp, Kratz [who dates the section unusually late: Komposi-
tion, pp. 246-47, 302-304], Dozeman, Albertz, Baden [pp. 77-80]).
There is insufficient evidence to trace even fragments of two

once complete parallel narratives behind the present text, but the
problems of the beginning and end of it do require a solution, the
more so as the ‘framing’ verses, vv. 2 and 7, are closely connected
and embody a distinctive theological critique of the people. It is
true that vv. 3-6 reflect a challenge to Moses’ authority which is
responded to in the manner in which water is provided, but this is
best seen as a feature which attracted the aetiological expansion at
its beginning and end, rather than as evidence that the whole passage
is a unity.1 It is certainly possible to see vv. 2 and 7 as a redactional
addition with Schart and Macdonald (the latter less convincingly
identifies the added material at the beginning as vv. 2aβ-3a), related
to Numbers 14 and presumably drawing the two place-names
from other elements of the wilderness tradition. But the ingenious
suggestion of Cornill (see the reference above) that vv. 2 and 7 were
originally part of a similar story located at Kadesh (of which other
fragments survive in Num. 20.1aβ, 3a, 5) can be developed into a
viable alternative in the light of a suggestion made by D. Frankel
(The Priestly Murmuring Stories, pp. 284-87). Frankel rejected
the view of Rudolph (pp. 84-87), Noth (Numeri, pp. 127-29, ET,
pp. 143-47) and L. Schmidt (Studien, pp. 45-52) that elements
from Exodus 17 had been secondarily added to the Priestly story
in Num. 20.1-13, and argued instead that the non-Priestly material
there came from the beginning of the spies story in Numbers 13
(cf. 13.19 and 23 with 20.5), where it would fit well before the
surviving non-Priestly text in v. 17. Exodus 17.2, 7 could origi-
nally have been part of the same introductory section, which at this
stage had nothing to do with a lack of water. As Frankel partly sees
(p. 286), possible confirmation of this can be found in Psalm 95.
1
For aetiological features in narratives as commonly secondary see Herrmann
as cited below in the Explanatory Note on v. 7, p. 48: to the references given there
B.O. Long, The Problem of Etiological Narrative in the Old Testament (BZAW
108; Berlin, 1968), should be added.
17.1-7 491
There Massah and Meribah are also mentioned as a place of rebel-

lion and one which led immediately to the forty-year wandering in
the wilderness (vv. 10-11: cf. Num. 14.20-39; Deut. 1.19-40), so
that the usual association of this psalm with the episode in Exodus
cannot be original. The most likely occasion for the displacement
of Exod. 17.2, 7 and Num. 20.1aβ, 3a, 5 to their present positions
would be when the Priestly and non-Priestly narratives were spliced
together, since it is the redactor’s use of the Priestly introduction
to the spies story (Num. 13.1-17a) which has left no room for the
original non-Priestly version. The puzzling mention of ‘Horeb’ in v.
6 (see the Explanatory Note) may have been added at the same time:
Macdonald (see above) has shown how the keynotes of the text in
its present form cohere together and prepare the way for the Sinai
narrative in chs. 19–34. As for the main narrative here (vv. 3-6), the
mention of Moses’ staff and the specific link to 7.15 in v. 5 point to
its belonging to the J strand of the non-Priestly account.
The itinerary-note in v. 1 is closely linked, both by the toponym
‘the Wilderness of Sin’ and by the presence of Priestly language,
to 16.1, and discussion of its origin has naturally proceeded
along similar lines (see on this the introduction to 16.1-36). Here
all of v. 1abα (i.e all except the note about the absence of water,
which has been associated with [part of] the following narra-
tive) has also been widely regarded as Priestly: so most recently
Blum (Studien, p. 148 n. 198, 277-78) and Graupner (tentatively:
Elohist, pp. 91-92). But the verse probably has an older basis to it
(Davies, ‘The Wilderness Itineraries and the Composition of the
Pentateuch’, pp. 2-3, following Eissfeldt; Levin, p. 357): this may
comprise not only ‘They travelled from the Wilderness of Sin and
camped at Rephidim’ but ‘by stages’ (cf. Gen. 13.3) and ‘where
there was no water for the people to drink’ (cf. Num. 33.14).2
Most likely it is part of a large-scale redactional operation which
provided the non-Priestly wilderness narrative with most of its
detailed geographical structure (see the Excursus on ‘The Wilder-
ness Itinerary’ in the introduction to 12.28-42, 50-51). Whether this
2
For discussion (and rejection) of the recent proposal that only v. 1b is of older
origin and that it was originally joined directly to 16.1aα (L. Schmidt; Dozeman;
Albertz: so also apparently Schart, Mose und Israel, pp. 167, 168 n. 71, 184-85)
see the introduction to 16.1-36.
492 EXODUS 1–18
preceded or followed the insertion of v. 2 (and v. 7) is not certain:

17.1 would lead well into either v. 2 or v. 3.
There has been much discussion, especially in earlier scholar-
ship, about the place of this passage in the history of the wilderness
tradition. Much of it was based on the assumption that at least one
of the names ‘Massah’ and ‘Meribah’ was originally associated with
the discovery of a spring in the vicinity of Kadesh (so especially
Gressmann, who included his treatment of the passage under the
heading ‘Die Quellwunder von Kades’ [Anfänge, pp. 77-80]: see
also Coats and Fritz). But if we are right in regarding vv. 2 and 7 as
a secondary expansion of the original story, the background of the
two main components of the passage must be examined afresh and
separately. On the one hand, vv. 3-6 are essentially a miracle story
(like the two preceding narratives) and specifically one about the
divine provision of water through the agency of Moses, similar to
and possibly ultimately derived from the brief fragment of narrative
and poetry in Num. 21.16-18 which is now attached to an itinerary
referring to a later part of the wilderness journey in Transjordan.
This was no doubt originally a free-floating tradition about the
journey through the wilderness, like Exod. 17.3-6, and if it ever
had a specific location in mind this is no longer accessible to us.
Its general background will lie in phenomena like those noted in
the Explanatory Note on vv. 5-6, which are also celebrated in Deut.
8.15; Pss. 78.15-16, 20; 105.41; 107.9; 114.8. But it now incor-
porates the recurrent motif of ‘complaint’ against Moses (which is
absent in Num. 21.16-18) and the vindication of his authority by his
role in the provision of water.3
The tradition-history of vv. 2 and 7 is more difficult to reconstruct
because there is more evidence to consider and their interrelation-
ships are vague and open to different explanations (see the review
in the Explanatory Note on v. 7). Formally they are an aetiological
3
Cf. Coats, Rebellion, pp. 60-62. Frankel, Murmuring Stories, pp. 298-300,
who draws attention to Num. 21.16-18, is right to point out that the gathering of
the people there to witness the event is more closely followed in the P account
in Num. 20.8-10 than in Exod. 17.5-6. But this need not mean that there was an
‘early Priestly version’ of the episode, as he supposes: the Priestly author could
have drawn this motif, which serves his own purposes in Num. 20.10, directly
from Num. 21.16, just as he probably (see below) drew much of the rest of his
material from Exod. 17.1-7.
17.1-7 493
supplement to the main narrative, with an explanation for each of

the names and an elaboration in the second case (cf. Gen. 4.25;
29.32).4 The two names are brought together only twice elsewhere,
in Deut. 33.8 and Ps. 95.8-9. The latter passage almost certainly
alludes to Exod. 17.7, whether (as most think) in its present context
or, as suggested above, when the verse was part of the non-Priestly
narrative now preserved in Numbers 13–14. In either case it does
not help to locate the origin of this pairing. Deuteronomy 33.8 is
more promising, since it is a part (though not necessarily an original
part) of what seems to be a collection of early poetic sayings about
the individual tribes of Israel. But it differs considerably from Exod.
17.2 and 7 in its focus on the tribe of Levi alone (or even, according
to some, on Moses alone), in the fact that it is Yahweh who is doing
the ‘testing’ and ‘disputing’ rather than Israel, and in the positive
outcome of the ‘test’. The mention of a ‘testing’ at Meribah alone in
Ps. 81.8, probably a north Israelite psalm (cf. ‘Joseph’ in v. 6) and so
relatively early, shares the second and third of these characteristics,
but (like Exod. 15.25; 16.4; 20.20: cf. Gen. 22.1) it speaks of Israel
being ‘tested’, not the Levites or Moses. Both the location and the
interrelationship (if there is one) of these passages are obscure,
but they share what may be an ancient folk-lore motif which has
been applied to Israel and some of its leaders. On its own the name
Massah occurs only in Deut. 6.16 and 9.22, in both of which it is a
place of Israelite rebellion against Yahweh: the latter verse places it
with two place-names that occur in Numbers 11, i.e. in the second
part of the wilderness journey, after the departure from Mount
Sinai/Horeb. Meribah is also a place of rebellion against Yahweh in
Num. 20.1-13(P) and other passages related to it, but these are late
passages and probably dependent on Exod. 17.1-7 in more or less
its present form.
Apart from the place-names, the idea of a challenge to Moses
(and Yahweh) by the people is evident in a number of non-Priestly
wilderness narratives (as also in Deut. 1.26-35) and this could have
led to the revision of an older positive interpretation of the names
Massah and Meribah. The specific expression of this in the form
of Israel ‘testing’ Yahweh is only found elsewhere in Num. 14.22,
4
In view of these parallels there is no need to regard v. 7bβ as an even later
expansion, as some have done (e.g. Herrmann, pp. 51-52).
494 EXODUS 1–18
Deut. 6.16 (with the name Massah) and in Pss. 78.18, 41, 56; 95.9
(with Massah); 106.14, among which Psalm 78 and Numbers 14 are
probably the oldest texts to preserve this tradition.
By the time of Ezekiel, or to be more precise the time of what is
probably one of the latest passages in the book of Ezekiel, there was
a place called ‘the waters of Meribath-kadesh’ on or near what was
then regarded as the southern boundary of ‘the land of Israel’ (Ezek.
47.19; 48.28). This is very probably the area of springs about 50
miles south-west of Beersheba where the name of biblical Kadesh is
preserved in Wadi Qadeis and Ain Qadeis. The name also appears in
Num. 27.14 and Deut. 32.41 and it is presupposed in the aetiology
in Num. 20.13. It is only in these late texts that the name Meribah
is explicitly associated with Kadesh. In what is probably an older
description of the southern boundary in Num. 34.4 and Josh. 15.3
Kadesh-barnea appears in its place.
If Exod. 17.7 was originally positioned somewhere in the older
spies story in Numbers 13–14 (as suggested above), a story which
is located at Kadesh according to Num. 13.26 (cf. Deut. 1.19, 46),
this would take the association back further (and provide an expla-
nation for the toponym in Ezekiel and P), but it is not clear how
far. Even then, the once popular hypothesis that already at the
time of the Exodus there were places called Marah and Meribah
in the vicinity of Kadesh would remain far from being proved.
Equally, the notion that the names originally designated ancient
centres of judicial activity (‘Trial’ and ‘Dispute’) is no more than
a conjecture, whatever one may make of the isolated reference to
‘En-mishpat (“the well of judgement”), that is Kadesh’ in Gen.
14.7.
The narrative, in both its original and expanded forms, combines
two topics which both recur a number of times in the wilderness
narratives: Yahweh’s provision of water in the desert through
the agency of Moses and Israel’s complaints which according to
Moses at least call in question Yahweh’s care for them and his
presence among them. Both of these themes are taken up in the
Psalms and especially in Psalm 78, where Israel’s complaints are
regarded not only as a ‘test’ of Yahweh but as forgetfulness of his
provision in the past (vv. 11, 42). Yet here, once again, as in the two
previous episodes in Exodus, Yahweh shows patience and mercy
and responds positively to the people’s need and (as in 15.25) to
17.1-7 495
Moses’ prayer on their behalf. In the end, it is Yahweh’s faithful

provision which is remembered and expected again by the prophets
of salvation in the book of Isaiah (35.6-7; 43.19; 44.3: cf. 55.1).
1 [The whole congregation of the Israelites travelleda from the

Wilderness of Sin by stagesb according to Yahweh’s command
and camped at Rephidim, where there wasc no water for the
peopled to drink.] 2 [So the people disputed withe Moses and
said, ‘Givef us water so that we may drink!’ Moses said to them,
‘Whyg do you dispute withh me? Whyg do you put Yahweh to the
test?’] 3 The people became thirsty there for water and the people
complained to Moses. They said, ‘Why everi did you bring us up
from Egypt, so as to kill usj, ourj children and ourj animals with
thirst?’ 4 Moses cried to Yahweh, saying ‘What shall I do withk
this people? They are close to stoning mel!’ 5 Yahweh said to
Moses, ‘Pass on ahead of the peoplem, and take with you some
of the elders of Israeln; and your staffo, with which you struckp
the river Nile, take that in your hand and go! 6 I tell youq, I am
going to stand therer in front of you on/by a crags [in Horeb],
and you shall strike the cragt. Then water will come out of it and
the people shall drink.’ Moses did so in the sight of the elders of
Israel. 7 [He called the name of the place Massah and Meribah,
because the people disputed and put Yahweh to the test, by
saying ‘Is Yahweh in our midst or not?]’

a. Heb. ‫ויסעו‬, from ‫נסע‬: for the sense ‘journey, travel’ see 12.37 and Note
o on the translation of 12.28-42, 50-51. Here it is the inclusion of ‫למסעיהם‬
(see the next note) which focuses attention on the journey as distinct from its
beginning (contrast 13.20 and 15.22).
b. Heb. ‫למסעיהם‬. Apart from a few places where it refers to departure (Num.
10.2, 6, 28; perhaps Deut. 10.11), ‫ מסע‬means a ‘journey’ and especially a day’s
journey or a ‘stage’ of a journey, as in 40.36, 38 and Num. 33.1-2. Elsewhere
expressions like ‫ למסעיהם‬are used of a comparatively long journey that would
take more than one day (Gen. 13.3; Num. 10.12) and here it probably also
implies that one or more intermediate stops have been passed over (cf. Num.
33.12-14).
c. Although ‫אין‬, being originally a noun, gives no indication of tense in
itself and most often refers to the present, when it follows a past tense verb it
naturally shares that time-indication (cf. Gen. 5.24; 11.30).
496 EXODUS 1–18
d. Grammatically ‫ העם‬may be regarded as either a nominative or a genitive

(GK §115e).
e. Heb. ‫ ִעם‬is a frequent alternative to the construction of ‫ ריב‬with ‫( ב‬cf.
Gen. 26.20; Num. 20.3): see also Note h and Text and Versions.
f. The pl. form ‫ תנו‬of MT does not make sense and is probably a corruption
of the emphatic form of the sing. imperative ‫תנה‬: see Text and Versions.
g. For ‫‘ = מה‬Why?’ see 14.15 and Note kk on the translation of 14.1-31.
h. Heb. ‫עמדי‬. The longer form ‫( עמד‬used instead of ‫ ִעם‬only with the
first person sing. suffix), which occurs only here in Exodus but 14 times in
Genesis, is generally thought to be related to Ar. ʿinda = ‘beside’.
i. Heb. ‫למה זה‬, as in 2.20 and 5.22 (see Note z on the translation of 2.11-22).
The following inf. may express the result rather than the implied purpose (cf.
Jer. 44.7 and IBHS §36.2.3d).
j. In these three instances Heb. has first person sing. forms (‘me’, ‘my’),
a surprising change from the first person pl. object suffix for the people in
‫ העליתנו‬earlier in the verse, but one that is paralleled in Num. 20.19, where
again the initial pl. is seen to imply (there two) sub-units for which the sing.
pronouns are used. On the understandable simplification in most of the Vss
and one Heb. ms. see Text and Versions. The use of sing. endings for the
people as a whole is in itself not uncommon (cf. 15.25-26). As for ‫ואת־מקנַ י‬, the
noun ‫ מקנה‬is generally found as a (collective) sing. (see Note g on the transla-
tion of 9.1-7), and even some other forms which have the suffixes appropriate
to a pl. are generally taken as sing. (GK §93ss; JM §96Ce), so this one
(elsewhere only in Num. 20.19) is surely sing. too, perhaps by contraction.
k. Heb. ‫לעם‬. ‫ל‬, when construed with ‫עשׂה‬, can have either of its most
common meanings, ‘to, unto’ (cf. AV, RV) and ‘for’, and it is used of doing
both good (cf. 18.9; 20.6) and harm (18.8: cf. Gen. 42.28), as well as more
generally (cf. BDB, pp. 793-94). ‘With’ is the idiomatic (and ambiguous)
English equivalent (cf. also Luther ‘mit’) and appropriate in a question which
is seeking guidance (though in view of the next sentence it probably expresses
exasperation too). Albertz sees its use in questions as generally positive
(pp. 286-87: cf. 33.5), but recognises that the following words suggest a more
negative reaction here.
l. Heb. ‫עוד מעט וסקלני‬, lit. ‘A little more and they will stone me’; the other
uses of such expressions are in prophetic announcements of judgement or a
similar context (Isa. 10.25; 29.17; Jer. 51.33; Hos. 1.4; Ps. 37.10), but it is
only likely that these reflect a wider colloquial usage such as is found here.
For the perfect consecutive following a temporal expression see 16.6-7 and
Note j on the translation of 16.1-36.
m. Heb. ‫ עבר לפני העם‬is to be taken straightforwardly, as e.g. in Gen. 32.17;
33.3, 14.
n. Heb. ‫מזקני ישׂראל‬: for the partitive ‫ מן‬cf. 16.27 and BDB, pp. 580-81.
o. The ‘fronting’ of the object ‫( מטך‬which creates a chiasmus) and the
repetition of the imperative verb ‫ קח‬give added prominence to Moses’ staff
and its earlier appearance in the Exodus narrative (7.15, 17[?], 20: see the
17.1-7 497
Explanatory Note here). On the Heb. word-order see Muraoka, Emphatic

Words, pp. 38-40; JM §155o, oa.
p. Heb. ֹ‫אשׁר הכית בּו‬, with dagesh forte conjunctivum (GK §20f).
q. ‫ הנני‬is followed by a participle in Exod. 3.4; 8.17; 9.18; 10.4; 14.17;
16.4; 34.11 and frequently elsewhere, especially in prophecy: see Note g on
the translation of 16.1-36 on its function here.
r. Heb. ‫ ָשּׁם‬: on the dagesh see Note p above.
s. Heb. ‫על־הצור‬. With ‫( עמד‬and ‫ על )נצב‬can mean ‘by, beside’ as well as
‘on’, and this might be the meaning here. Unless reference is made to a well-
known rock at Horeb, which seems unlikely in the context, the def. art. is
probably ‘anticipatory’ (cf. Luther: ‘auf einem Fels’), as in ‫ הסנה‬in 3.2 (see
also Note o on the translation of 2.11-22). A ‫ צור‬can be a ‘boulder’ (Judg.
6.21; 13.19; 2 Sam. 21.10; Prov. 30.19), but it is more often a much larger
expanse of rock (Exod. 33.21-22; Num. 23.9; 1 Sam. 24.3; Isa 2.10, 19, 21;
51.1; Jer. 18.14; Ps. 81.17; Job 22.24; 24.8; 28.10; also in the Siloam tunnel
inscription [AHI 4.116.3, 6; cf. 4.402.1]) and this sense also provides a better
basis for the frequent metaphorical use of ‫ צור‬as a divine title (Deut. 32.4
etc.): hence BDB, p. 849, gives ‘rocky wall, cliff’ as the primary meaning
(see further TWAT 6, 973-83 = TDOT 12, pp. 311-21; A. Gray, Psalm 18
in Words and Pictures: A Reading through Metaphor [BibInt 127; Leiden,
2014], pp. 61-63).5
t. Heb. ‫בצור‬. Usually ‫ הכה‬is followed by a direct object (cf. ‫ את־היאר‬in
v. 5) and the other exceptions are, apart from those where ‫ ב‬means ‘among’,
all places where ‫ הכה‬means ‘kill’ or ‘defeat’ persons (e.g. 1 Sam. 18.7; 21.12;
29.5 [the same popular cry in each case]; Num. 22.6; 1 Sam. 14.31; 23.2 [2x],
5: see further DCH 5, p. 689, but in Exod. 12.13 ‫ ב‬probably means ‘in’; cf.
v. 12). Here the usage may be by analogy with that which is regular with ‫;נגע‬
or, as Ges18, p. 816, suggests, ‫ ב‬may have the sense ‘on’. In any case the
anomaly is removed in Num. 20.11 (‫)את־הסלע‬.
Explanatory Notes
1. The beginning of a new episode is marked by a further two-part
itinerary-note (cf. 16.1) bringing the Israelites to Rephidim, which
is also the scene of the subsequent conflict with the Amalekites
according to v. 8. Again there are Priestly features present, as in 16.1
(‘congregation’, ‘according to Yahweh’s command [lit. mouth]’: for
the latter cf. Driver, Introduction9, p. 134, and esp. Num. 9.18, 20,
23), but here too they are likely to be a secondary overlay related
to the incorporation of the Priestly manna-narrative in ch. 16 (see
5
In the similar episode at Kadesh ‫ סלע‬is used instead of ‫( צור‬Num. 20.8,
10-11).
498 EXODUS 1–18
the Explanatory Note on 16.1). The itinerary-note forms part of a

‘chain’ which continues in 19.2, where the Priestly note of arrival
in the wilderness of Sinai in 19.1 is clearly separate. Once again
it is likely that it was extracted from the fuller itinerary preserved
in an edited form in Num. 33.1-49 (cf. vv. 12-14): there the words
‘where there was no water for the people to drink’ also occur, with
only minor differences in the Heb. expressions used. More striking
is the fact that the fuller list includes two other stopping-places,
Dophkah and Alush, between the Wilderness of Sin and Rephidim:
the inclusion of the expression ‘by stages’ here could be the redac-
tor’s way of acknowledging that according to his source further
halts had intervened: the use of the same expression in Num. 10.12
also corresponds to a journey of more than one day (cf. Num. 11.35;
12.16). It was a traditional part of the vocabulary for describing a
longish journey (cf. Gen. 13.3).6
Both in the sequence of the Exodus narrative (cf. 19.2) and in
Num. 33.1-49 Rephidim is the immediately preceding stopping-
place to ‘the wilderness of Sinai’, beside the mountain that is
variously called Sinai, Horeb or ‘the mountain of God’ (18.5). The
name may be related to the root rpd, which in Heb. seems to mean
‘support’ (Song 2.5; 3.10: cf. ESA rfd, Ar. rafada) and ‘spread out’
(Job 17.13; 41.22: cf. Eth. rafada acc. Ges18, p. 1261), but neither
of these meanings provides any clear indication of the origin of the
name. On the basis of its proximity to the traditional site of Mount
Sinai Faran (an episcopal see in the early Christian period and
located by the oasis in Wadi Feiran) or somewhere in its vicinity
was regarded as the location of Rephidim (for the evidence see my
Way of the Wilderness, pp. 32, 40, 43, 45-46): so still by Palmer,
Desert 1, pp. 5-7, 158-62, 168. The name has been thought to be
preserved in Wadi (and Jebel) Refayid, much closer to Jebel Musa
(Abel, Géographie 2, pp. 213, 435); but Noth drew attention to
er-Rafid, a ridge east of the Gulf of Aqaba, in a region where some
have located Mount Sinai (p. 110, ET, p. 138; cf. A. Musil, The
Northern Ḥeǧâz [New York, 1926], p. 269). If nothing else, these
6
On the view of L. Schmidt and Albertz (and here also Dozeman) that the
whole first half of v. 1 was added to an older reference to Rephidim by the Priestly
author of ch. 16 see the introduction to 16.1-36: the objections noted there apply
here too.
17.1-7 499
toponyms give some support to the name Rephidim as also being

an authentic desert place-name, whatever its (and their) meaning
may be.7
2. The story proper begins with two complaints by the people
about their lack of water to drink, to each of which Moses responds
in a different way (vv. 2, 3-4). The first complaint is expressed
by the verb ‘dispute’, which seems a strong word for the request
to ‘give us water’ which follows, and similarly Moses’ response
to the latter is somewhat ‘over the top’. A number of commenta-
tors have intensified the meaning of the word ‘dispute’ (Heb. rîb)
further by seeing it as a technical expression for a legal dispute in
which a person makes a justified claim for a remedy to a griev-
ance which he or she has suffered (Coats, Rebellion, pp. 57-59: so
recently Dozeman and Albertz). Both the verb and the related noun
can bear such a judicial sense (cf. 23.2, 3, 6: TWAT 7, 496-501 =
TDOT 13, pp. 473-79), but there are plenty of cases where it refers
to an everyday quarrel which does not reach a lawcourt (see BDB,
pp. 936-37: cf. 21.18 and Gen. 13.7-8; 26.20-21) and there is no
need to introduce a legal background here. The real reason why rîb
is introduced here is made clearer by the fuller wording of Moses’
response, which is explicitly taken up in v. 7 to explain the two
names which he gives there to the place where the episode is said
to have occurred. It is only here and in the other episode that is
related to the name Meribah (Num. 20.1-13) that rîb appears in the
wilderness narratives. The second part of Moses’ first response adds
a more serious accusation than insubordination to Moses himself as
the people’s leader: they are, he says, putting Yahweh to the test.
The verb (Heb. nsh) is the same one that was used earlier (15.25;
16.4) for Yahweh’s testing of Israel by the commands or instruc-
tions which he gave them. That, it is understood, was his right, but
for Israel to test their God’s power and concern for them is seen
as most inappropriate, and it is similarly regarded in a series of
other passages (Num. 14.22; Deut. 6.16; Isa. 7.12; Pss. 78.18, 41,
56; 95.9; 106.14), most of which appear to refer to this episode.
Only in Judg. 6.36-40 (cf. vv. 16-24) is testing of God allowed to
7
A place byt hrpyd occurs in the Lachish Letters (AHI 1.004.5), but this will
be much further north, probably in the vicinity of Lachish itself (for discussion and
references see Renz, Handbuch 1, p. 419).
500 EXODUS 1–18
be legitimate.8 Generally it is seen as showing a lack of the trust in

God’s care which does not need external confirmation (cf. TWAT 5,
477-86 = TDOT 9, pp. 446-54: Ps. 78.22, 42) and a forgetfulness of
his previous actions on one’s behalf (Ps. 106.13).
3-4. The second sequence of complaint and response begins with
a repetition of the underlying problem, the people’s thirst (v. 1b).
This is sometimes seen as a deepening of the crisis: first the people
perceive that there is no spring or well at their encampment, then
lacking water they become thirsty and complain again (so Nachma-
nides; recently [as one possibility] Propp, pp. 603-605). But this is
an artificial distinction and it is more likely that the repetition is due
to different elements which have been awkwardly combined in the
narrative. The words of complaint (this time the same verb is used
as in 15.24 and 16.2) conform to the description and embody an
implicit accusation of Moses, that he has brought the people into a
place where they will die of thirst. This probably refers to the effect
of his leadership in the Exodus rather than his intention (see Note
i on the translation). ‘To kill…with thirst’ parallels ‘to kill…with
hunger’ in 16.3 and Fritz therefore attributed this verse to a late
redactor who knew the Priestly manna-story (Israel in der Wüste,
p. 11). But in other respects the wording is quite different and
no direct connection need be presumed: both verses belong to a
larger group in which death in the desert has become a recurrent
feature of the people’s complaints (cf. 14.11; Num. 16.13 [‘kill’,
as here]; 20.4; 21.5). Only here and in Num. 20.4 (a related story
about the need for water) do they mention their animals’ needs: for
the animals (Heb. miqneh) which they had with them see also 9.4,
6-7; 10.26; 12.38. As in 15.24 Moses responds by crying out for
help to Yahweh, this time fearing for his life. In what are probably
later reminiscences of the story, the whole people seek Yahweh’s
help and nothing is said of any conflict with Moses (Pss. 105.40-41;
107.5-6; Wis. 11.4).
5-6. Yahweh’s lengthy reply consists of instructions (v. 5) fol-
lowed by reassurance for both Moses and the people (most of v. 6):
there is no rebuke for the people, the emphasis is on how their
8
One might add the unusual case of Isa. 7.10-14, where Ahaz is invited to ask
God for a ‘sign’ and his refusal to do so (because it would involve putting God to
the test) is criticised and overridden.
17.1-7 501
need is to be met (cf. 15.25). In the story as it has been transmit-

ted Yahweh does not act directly but through Moses’ obedience to
his instructions, as in 15.25 again and more explicitly in the con-
test with Pharaoh earlier in the book. In the non-Priestly narrative
Moses’ staff is first a sign to convince the Israelites of his divine
commission (4.1-5, 17, 28-31), then the means to turn the Nile into
blood (7.15, 17-18, 20-21a), then the means by which a wind was
summoned to drive back the waters of the sea (14.16a, 21 [part]),
and now a blow from it opens a spring to provide the Israelites
with water in the desert. The sequence is tied together by the con-
nections made in 7.15 and v. 5 here, which help to distinguish this
strand of the narrative from the Priestly focus on Aaron’s staff in
7.9 etc. In 17.9, as in 4.20, the expression ‘the staff of God’, which
is in Moses’ hand, points to a related tradition (see the Explana-
tory Notes there). The presence of some of ‘the elders of Israel’ (cf.
3.16, 18; 4.29; 12.21), another feature of the non-Priestly narrative,
affirms their authority alongside Moses and also provides additional
witnesses of the miracle (cf. ‘in the sight of’): as the Mekhilta put
it, ‘So that the people should not be able to say, “Perhaps there
were wells [or springs] there” ’ (Lauterbach 2, p. 131). Moses is
told simply to ‘go’, without any specification of a place (cf. Beer,
p. 90). Presumably ‘ahead of the people’ was meant to give him the
direction to take and he would know the place when he saw Yahweh
standing there. It is notable that Yahweh is to stand where the people
are not (yet) present, a sign perhaps that the narrative here was not
composed to answer specifically the uncertainty expressed in v. 7.
Yahweh says that he will be standing on ‘a crag in Horeb’. Apart
from here the name represents the mountain of God in Exodus only
in 3.1 and 33.6; elsewhere the name ‘Sinai’ is preferred (already
in 16.1 [cf. 19.2]; then in 19.1; 24.16; 31.18; 34.29, 32 [all P: cf.
19.20, 23]; and in non-P 19.11, 18; 34.2, 4). ‘Horeb’ occurs pre-
dominantly in Deuteronomy (1.2 etc.) but also in Solomon’s prayer
(1 Kgs 8.9 par. 2 Chr. 5.10), the Elijah story (1 Kgs 19.8), Mal. 3.22
and Ps. 106.19 (with which cf. Deut. 9.8). It is surprising to find
such a specific location suddenly mentioned here when nothing in
the context has yet suggested that the Israelites have reached the
holy mountain. Opinion is divided over whether the expression ‘in
Horeb’ is an original part of the story, but a redactional origin is
perhaps more likely (see e.g. Fritz, Israel in der Wüste, pp. 11-12
502 EXODUS 1–18
[with refs.], for it being a Deuteronomistic addition, which Propp,

pp. 604-606, strongly opposes; also Macdonald, ‘Anticipations of
Horeb’, pp. 12-14), in connection with either the addition of the
Massah-Meribah aetiology (see below on v. 7) or the combination
of the narrative with ch. 18, where ‘the mountain of God’ is referred
to before the formal note(s) of arrival in 19.1-2.
The opening of a water source in an apparently dry place is
not unknown in the south of the Sinai peninsula, as ground water
from the winter precipitation is retained by geological formations
underground (dykes). The Bedouin are aware of such resources and
even phenomena not unlike that portrayed in the Moses story have
been repeated in modern times (in general see Philo, Vita Mosis
1.211; Lipschitz, Sinai, pp. 11-12; A. Perevolotsky and I. Finkel-
stein, ‘The Southern Sinai Route in Ecological Perspective’, BAR
11/4 [1985], pp. 26-41 [more fully in Z. Meshel and I. Finkelstein
(eds.), Qadmoniot Sinay (Tel Aviv, 1980), pp. 385-419 (Heb.)];
I. Beit-Arieh, ‘The Route through Sinai: Why the Israelites Fleeing
Egypt Went South’, BAR 14/3 [1988], pp. 28-37). Specific parallels
are described by Holzinger, pp. 59-60, based on O. Fraas, Aus dem
Orient: Geologische Beobachtungen am Nil, auf der Sinaihalbinsel
und in Syrien (Stuttgart, 1867), pp. 24-25; and by Jarvis, Yesterday
and Today in Sinai, pp. 174-75; Lipschitz, p. 12.
In one of the very few statements in this pericope that recounts
events rather than words, Moses’ obedience to Yahweh’s commands
is briefly reported: ‘Moses did so…’ Such summary notices are
more characteristic of the Priestly narrative in Exodus (cf. 7.10,
20 etc.), but 8.20 and 22 provide non-Priestly parallels and such
expressions are also found in Genesis (e.g. 29.28) and frequently in
Joshua–Kings. ‘In the sight of the elders of Israel’ confirms (as their
earlier mention in v. 5 implies) that the people as a whole are not
thought of as going with Moses to the ‘crag’.
7. It is difficult to identify the subject of ‘he called’ as anyone
other than Moses after his explicit mention just before, and it is
his challenge in v. 2 which provides the basis for the naming here,
rather than anything in the intervening verses. The name Massah
occurs alone only in two brief references to ‘testing’ of Yahweh and
rebellion against him in the wilderness in Deuteronomy (6.16; 9.22)
and in neither case is there any explicit connection with the present
narrative. ‘The waters of Meribah’ are located in ‘the wilderness of
Zin’ in the Priestly version of the same story in Numbers (20.13; cf.
17.1-7 503
Ps. 106.32), where the verb ‘quarrel’ (Heb. rîb) is again prominent
(vv. 3, 13). In Num. 27.14 and Deut. 32.51 the name Meribath-
kadesh makes explicit the association with Kadesh(-barnea) which
is implied by the context in Numbers 20: the same name, without
any allusion to the narratives, appears in Ezekiel’s description of
the boundaries of the land of Canaan in the middle of the southern
border (47.19 [with the pl. Meriboth in MT (cf. LXX)]; 48.28).9
‘The waters of Meribah’ are also included as a place where Yahweh
‘tested’ Israel (Heb. bḥn, not nsh) in a brief summary of the Exodus
story in the early north Israelite Ps. 81.7-8, without any geograph-
ical details and, with Massah, as the place where Yahweh ‘tested’
(Heb. nsh) the priestly tribe of Levi and ‘contended’ with them
(Heb. rîb) in Moses’ blessing of the tribes (Deut. 33.8), an early
poetic collection according to Cross and Freedman (Studies, p. 97;
cf. Canaanite Myth, p. 197) and others. Finally Massah and Meribah
are cited together as a place or places where Israel were rebellious
and disobedient in the wilderness in Ps. 95.8-9, probably the words
of a cultic prophet.
Three features stand out immediately from this varied group of
passages. First, the two names are found separately as well as together
and probably referred originally to different places. Secondly,
Meribah is in later texts associated with the southern boundary of
Canaan and the adjacent part of the desert (‘the wilderness of Zin’),
not with a place near to ‘the mountain of God’. Thirdly, in what are
probably the oldest poetic sources it is Yahweh who does the testing
and contending, not Israel; in one of them (Deut. 33.8) it is the tribe
of Levi, not all Israel, that is involved; and in both Meribah is a
place with ‘waters’, probably an abundant spring or springs. For
further discussion of the complex tradition-history which must lie
behind these texts and its implications for the composition of this
passage see the introduction to this section. What seems most likely
is that vv. 2 and 7 were added to an older story which told simply
of the miraculous provision of water in the desert without naming
a specific location (as in Deut. 8.15; Isa. 35.6b-7; 43.19; 44.3; Pss.
78.15-16; 105.41; 107.35; 114.8: cf. Wis. 11.4). Their motifs were
probably drawn from a quite different context in Numbers 13 and
these were reworked (in a similar spirit to Ps. 78) to make the story
9
It is also sometimes thought to be the original reading for MT’s ‘from ten
thousands of holy ones’ (mēribebōt qōdeš) in Deut. 33.2 (cf. BHS).
504 EXODUS 1–18
into a paradigm case of Israel’s rebelliousness in the desert and

their unworthiness to receive Yahweh’s gracious provision for their
needs.
The concluding question in v. 7 requires fuller comment here:
‘Is Yahweh among us, or not?’ (cf. W. Herrmann, ‘Ex 17bβ und
die Frage nach der Gegenwart Jahwes in Israel’, in J. Hausmann
and H.-J. Zobel [eds.], Alttestamentliche Glaube und Biblische
Theologie [FS H.D. Preuss; Stuttgart, 1992], pp. 46-55 [51-55]).
The words are quite different from those which the people speak
in vv. 2-3, where they are concerned only with their lack of water:
here they (are said by Moses to) raise a fundamental religious issue
which is in fact a recurrent theme of subsequent narratives about the
wilderness journey and one which reflects a central concern of Old
Testament religion more generally: the presence of Yahweh among
his people. The issue surfaces again in the non-Priestly narrative
when the Israelites are about to leave Mount Sinai. When Yahweh
says that he cannot be ‘among’ the people (the same expression as
here, beqereb) because of the idolatry of the golden calf, these are
described as ‘harsh words’ and the people ‘mourned’ (33.3-4, with
a variant version in vv. 5-6). In response to Moses’ impassioned
pleading, Yahweh relents and promises his continued presence with
the people (33.12-17: cf. the variant in 34.9 and later Num. 11.20;
14.14: contrast 14.42-45). In Priestly texts the same idea takes the
form of Yahweh’s ‘dwelling’ in the tabernacle ‘in the midst’ (betôk)
of the people (Exod. 25.8; 29.45-46). This foreshadows, or rather
reflects, the belief in Yahweh’s presence, like other deities, in his
‘house(s)’ or ‘temple(s)’, where he received the people’s worship
and provided for their needs (see in general Clements, God and
Temple; Terrien, The Elusive Presence: Toward a New Biblical
Theology [New York, 1978]). The doubt attributed to the people
here in a time of anxiety and distress is comparable to words
uttered in some lament psalms, when they fear that their enemies
will say ‘Where is your/their God?’ (Pss. 42.4, 11; 79.10; 115.2):
the difference is that here it is regarded as showing lack of faith
and so as reprehensible. Such anxieties, the reassurances offered
to combat them and in turn the challenges to such reassurances by
prophetic voices (Mic. 3.11; Ezek. 8–11) were evidently a recurring
feature of Israel’s religious life (for further evidence see Herrmann,
pp. 52-55).
17.1-7 505
Text and Versions

‫( ויסעו‬17.1) Vulg has a perf. part. to reduce the use of coordinated clauses
in the Heb. Before this word 4QpalExl, 4QpalExm and probably 4QExc had an
open or empty line, so anticipating the division in MT and SP mss.
‫( למסעיהם‬17.1) 4QExc (and apparently 4QpalExl), but not 4QpalExm, had
this word before rather than after ‫ממדבר סין‬, probably as a secondary ‘improve-
ment’ to associate it more closely with the verb and its subject (cf. Num.
10.12). Drew Longacre has pointed out that this agreement of the two mss
is important evidence for a textual affiliation between them, which is likely
on other grounds (‘A Contextualised Approach’, p. 223: see also on ‫ המקום‬in
v. 7). Tgg and Sy render with words related to their equivalents for ‫( ויסעו‬with
a waw in place of the lamed in TgN), but LXX κατὰ παρεμβολὰς αὐτῶν (cf.
its use of σταθμοί in Num. 33.1-2) and Vulg per mansiones suas shift the
focus from the journey-stages to the stopping-places.
‫( על־פי יהוה‬17.1) Most of the Vss naturally render ‫ פי‬as ‘word’ by
metonymy: this will include ‫ מימרא‬in TgO,J and mlth/ʾ in ms. 5b1 of Sy. The
other Sy mss have ʿl mlt pwmh dmryʾ, including the original expression as well
as its explanation. TgN ‫ על פם גזירת מימרה דייי‬goes a stage further (as sometimes
elsewhere) by adding a specific word for ‘decree’ and apparently shifting the
sense of ‫ מימרא‬in a more theological direction (cf. Chester, Divine Revelation,
pp. 304-305, 309-13). On LXX ῥήματος see Fritsch, p. 13: this and similar
renderings are common where the sense is ‘command’, but elsewhere LXX is
ready to speak of God’s στόμα (e.g. Deut. 8.3).
‫( ברפידים‬17.1) TgJ adds ‘a place where their hands ceased from the
commandments of the law and the springs dried up’ (similarly at Num. 33.14).
This explanation of the absence of water (as derivatively in TgJ on 15.22) will
be based on an etymology of the name Rephidim which saw it as a combina-
tion of Heb. ‫ רפה‬and ‫יד‬. In other sources (e.g. MRI [Lauterbach 2, pp. 129,
135]) it is used to explain Amalek’s attack in v. 8: see AramB 2, p. 210 n. 1
for further refs.
‫( לשׁתת העם‬17.1) So also SP, 4QpalExl and TgO,J (Nmg): no other Qumran
evidence survives at this point. LXX τῷ λαῷ πιεῖν may simply be a para-
phrase, but it does agree with MT ‫ לעם לשׁתות‬at Num. 33.14 and might be
due to assimilation. TgN and Sy paraphrase with relative clauses, Vulg with
a gerund. There is no reason to depart from the reading of the three Heb.
witnesses.
‫( וירב‬17.2) Tgg and Sy correctly indicate the meaning ‘quarrel, dispute’,
whereas LXX ἐλοιδορεῖτο (cf. OL), as later in the verse and four times
elsewhere in the Pentateuch for ‫ריב‬, means ‘reproached, abused’ (cf. Wevers,
Notes, p. 263). Vulg iurgatus (again avoiding a sequence of coordinated verbs,
as in v. 1) is ambiguous, as it can mean both ‘dispute’ and ‘blame’.
506 EXODUS 1–18
‫( העם עם‬17.2) TgJ prefixes ‘the wicked of’ as in 14.11 (cf. AramB 2,
p. 200 n. 18). A Geniza ms. reads ‫ על‬for ‫( עם‬BHS), perhaps assimilating to the
construction with ‫ לון‬in v. 3.
‫( ויאמרו‬17.2) Some SP mss have the sing. ‫ויאמר‬, unnecessarily conforming
to the (collective) subject ‫( העם‬cf. Vulg ait). But this is the easier reading
and the other Vss (including LXX λέγοντες) support the pl. Unfortunately
4QExc, the only Qumran ms. extant at this point, does not preserve the end of
the word. Sy as often adds lh to complete the construction.
‫( תנו‬17.2) MT’s reading, which is followed by TgO, inappropriately intro-
duces a pl. addressee, either by confusion with the ending of the next word
or presuming that Aaron is involved (as he had been in ch. 16: cf. Houtman,
p. 361; Albertz, p. 279 n. 1). SP, 4QpalExm, LXX δός and the other Vss
provide strong support for the reading ‫תנה‬, which should be preferred.10
‫( ונשׁתה‬17.2) Most early mss of Sy read just nšth, without waw, as a
separate sentence, ‘Let us drink!’ (the expected d is added in other mss): for
another example of the same tendency see the next note. SP and the other Vss
agree with MT here.
‫( ויאמר‬17.2) Sy (except for 5b1) again omits the conjunction (as in 16.19,
23), but it is present in SP and apparently also in 4QpalExl and is represented
in the other Vss.
‫( מה‬17.2)2o SP (except for Sadaqa), 4QpalExm and 4QExc read ‫( ומה‬cf.
LXX καὶ τί, TgJ and Sy), whereas TgO,N and Vulg agree with MT. The reading
with waw has extensive and early support and might therefore be original,
but in favour of MT is the widespread tendency to add waw and the fact that
‘The abrupt, staccato style fits the narrative better’ (Sanderson, Exodus Scroll,
p. 93).
‫( תנסון‬17.2) SP has the shorter form without nun paragogicum here,
although it agrees with MT’s ‫ תריבון‬earlier in the verse. 4QExc also has the
shorter form, but does not preserve the previous verb. Often (18 cases out
of 36 in Exodus) MT and SP both have the longer form and each text has it
several times where the other does not (in Exodus MT 10x and SP 8x: more
generally see GSH §63b). There is no reason why SP would have omitted
the nun here, whereas MT may have harmonised a text which it saw as
inconsistent. The occurrence of the shorter reading in 4QExc, which is not a
‘proto-Samaritan’ ms. (cf. DJD XII, p. 103; Lange, Handbuch, pp. 154-55),
gives further support to its originality.
‫( את־יהוה‬17.2) The Tgg as often avoid making God the object of human
action by using ‫( קדם‬17.2), ‘before’; TgNmg prefixes ‘the glory of the Shekinah
of’ to the divine name, as TgN itself does in v. 7: it is a characteristic expres-
sion of the Palestinian Targum (Chester, Divine Revelation, pp. 320-22).
‫( ויצמא‬17.3) This and both the other verbs in the verse of which the
people are the subject are rendered in the pl. in TgN and Sy, conforming to the
preceding pl. forms in v. 2.
10
According to BH3 21 Masoretic mss also have this reading.
17.1-7 507
‫( למים‬17.3) Vulg, rather ornately, renders pro penuria aquae (contrast

OL aquam).
‫( העם‬17.3)2o Vulg has no equivalent, deeming the repetition unnecessary.
‫( ויאמר‬17.3) In addition to TgN and Sy (which also add ‘to him’), most
SP mss, LXX (λεγοντες) and TgJ have a pl. form for this verb, perhaps to
match the pl. in v. 2. None of the Qumran mss preserves it.
‫( זה‬17.3) The enclitic particle does not appear here in SP, although it does
in other places where MT has ‫למה זה‬, nor is it represented in Vulg, Sy and
some LXX mss. But it is present in 4QExc (the only Qumran ms. to preserve
this part of the verse) as well as MT and it is presupposed in LXX* (τοῦτο)
and Tgg. Very similar wording, without ‫זה‬, occurs in Num. 20.5; 21.5 and it
is possible that this is the basis for the SP reading: such artificial uniformity
is characteristic of SP.
‫( העליתנו‬17.3) TgO ‫ אסיקתונא‬is a second person pl. verb and presumes a
different vocalisation of the Heb. It does not fit the immediate context, where
Moses alone is addressed, and cannot be right; the translator will have had the
(corrupt) pl. imperative in v. 2 in mind, which he uniquely reproduced.
‫( אתי…בני…מקני‬17.3) 4QpalExl preserves the first of these words with the
unexpected sing. suffix of MT and SP (cf. TgO), on which see Note j on the
translation. 4Q365 and the other Vss have the easier pl. forms (‘us…our’), but
these will be secondary. The ‫ מקנה‬of some SP mss is probably due to a graphic
error, he and yodh being quite similar in the Samaritan script.
‫( ויצעק‬17.4) Tgg and Sy interpret Moses’ cry as a prayer (‫)צלי‬, which is
appropriate when it is addressed to Yahweh (cf. 15.25).
‫( אל־יהוה‬17.4) Tgg have ‘before the Lord’ (cf. v. 2) and here Sy joins
them.
‫( לאמר‬17.4) Sy wʾmr, with a finite verb as e.g. in 14.1 (cf. also the word-
ing of MT in vv. 2-3).
‫( עוד מעט וסקלני‬17.4) 4Q365 reads the verb ‫ויסוקלוני‬, confirming (like
SP) MT’s vocalisation as a pl. by its plene spelling but also exhibiting a
peculiar Qumran orthography in the first stem-syllable (cf. Qimron §311.13g;
Reymond, Qumran Hebrew, pp. 211-21) and the post-biblical representa-
tion of the future by waw plus imperfect. Propp (p. 602) notes a divergent
Masoretic reading ‫( יסקלוני‬with the initial waw replaced by yodh; cited in
C.D. Ginsburg’s edition). The future sense of the Heb. is preserved in most of
the Vss, but TgO (with a perfect verb and ‫ פום‬as in Gen. 26.10; 31.27; Exod.
9.15) and Sy (with a part. and hww) seem to have taken the phrase to refer to
a danger that had already been narrowly escaped: ‘…they would have stoned
me’.11
11
TgJ’s ‫ תוב קליל זעיר‬combines TgN’s ‫( קליל‬cf. Sy) and TgO’s ‫זעיר‬: the effect is
to imply a very real danger for Moses; cf. TgNmg ‫סכינא‬, ‘in danger’, if that is the
correct reading and the annotation does not belong to v. 5 (see below and AramB
2, p. 74 n. e).
508 EXODUS 1–18
SP (but not MT) has a division after v. 4, and 4QExc and 4QpalExm
probably did too; in each case the opening of divine speech in v. 5 will have
been the reason for interrupting the narrative sequence artificially.
‫( יהוה‬17.5) TgNmg again prefixes ‘the Memra of’.
‫( אל־משׁה‬17.5) 4QExc omits the expression, which is not strictly neces-
sary, but it is present in all the other witnesses, which include here 4QpalExm
and 4Q365, and is probably original.
‫( לפני העם‬17.5) TgN curiously (and uniquely) has the Heb. word ‫ לפני‬here
instead of its usual ‫( קדם‬e.g. 16.33-34), another example of its occasional
tendency to reproduce the actual Heb. wording (cf. 16.2). LXX adds τούτου,
‘this (people)’, an unnecessary and inappropriate reproduction of the wording
of v. 4. The uncial mss AFM, as well as the O-text, omit it: the correction may
be pre-hexaplaric (Wevers, Notes, p. 265).
‫( מזקני ישׂראל‬17.5) TgN has ‘the wise’ (‫ )חכימייא‬in place of ‘the elders’ as
elsewhere: see Text and Versions on 3.16. TgNmg has an obscure word attached
to v. 4 (see the fn.) which may be a correction of this rendering to agree with
the Heb. (compare TgN and its mg at Num. 11.26). 4Q365 replaces ‫ישׂראל‬
with ‫העדה‬, introducing the term used in v. 1 and several times in ch. 16; LXX
substitutes τοῦ λαοῦ from earlier in this verse.12
‫( ומטך‬17.5) LXX, Vulg and TgN have no equivalent to the suffix (though
the Three and LXXO add σου), but this could well be due to the translators’
economy with words and all the Heb. witnesses which survive at this point
(4QpalExm, MT, SP) have the suffix.
‫( קח‬17.5) SP has the imperfect ‫( תקח‬cf. TgN ‫)תסיב‬, but the other Vss have
imperatives like MT. None of the Qumran mss preserves this word but in
4QExc it would be in a short lacuna at the beginning of a line and Sanderson
judges that ‘Spacing slightly favours the reading ‫( ’קח‬DJD XII, p. 121).
Neither reading can be called ‘more difficult’ on stylistic or grammatical
grounds, as either a further imperative or an imperfect can follow a sequence
of imperatives (for an imperative even after a preceding object cf. Gen. 43.13
[where SP = MT]; Judg. 6.20; 14.3). But a probably decisive factor is that in
the verse to which reference is made here (7.15) the imperfect ‫ תקח‬is used and
assimilation to that would be so fully in accord with SP scribal practice that
‫ קח‬must be regarded as the more original reading.
‫( והלכת‬17.5) There is some support for a further imperative here (Sy, TgJ,
Vulg) but no need to assume a Vorlage for these Vss different from MT, SP,
4QpalExl and 4QExc: LXX and TgO,N also agree with the common Heb. text.
TgJ adds ‘because of their complaints’, an explanation which MRI (Lauter-
bach, pp. 131-32) attaches specifically to the taking of the staff.
12
R.S. Nam suggested that the change in 4Q365 was ideological, introducing
an expression used for the Qumran community (‘How to Rewrite Torah: The
Case for Proto-Sectarian Ideology in the Reworked Pentateuch (4QRP)’, RdQ 23
[2007], pp. 153-65 [155-56]). If so, it would be an exceptional such intervention
in a (para)biblical ms. at Qumran.
17.1-7 509
‫( הנני‬17.6) Most mss of Sy read whʾ ʾnʾ (but 5b1 has just hʾ), avoiding the
asyndeton of the Heb. LXX ὅδε ἐγώ (cf. Gen. 50.18; Exod. 8.25; Num. 14.40)
aptly uses an old Greek idiom here (cf. LSJ, p. 1197; BDF §289). TgN ‫הא ממרי‬
is a notable instance of the theologically developed use of Memra to represent
God’s presence in a similar way to ‫ יקרא‬and ‫( שׁכינתא‬cf. Chester, Divine Revela-
tion, pp. 300-301, 310-11).
‫( עמד‬17.6) LXX ἕστηκα (again in v. 9), ‘(I) have taken my stand’
(NETS) or ‘(I) stand’ (Wevers, THGE, p. 264, an improvement on his render-
ing in Notes, p. 266), displaces the imminent future sense of the part. intended
by the Heb. and probably expressed by the other Vss, in line with its forced
interpretation of the following words (see below). TgN expands its translation,
as it does below in v. 9, by adding a form of the verb ‫עתד‬, which combines the
senses of ‘standing’ and ‘being ready’ (cf. CAL).
‫( לפניך שׁם‬17.6) 4QpalExm, SP and most of the Vss agree with MT (the
other Qumran mss do not preserve this part of the verse), but LXX* probably
had no equivalent to ‫( שׁם‬like the Sy ms. 5b1) and took ‫ לפניך‬in a temporal
rather than a spatial sense, to avoid what apparently seemed a theologically
inappropriate subordination of God to Moses (Wevers, THGE, p. 264; BAlex,
p. 189).13 Most LXX witnesses, as well as the Three, reflect various kinds of
assimilation to MT (cf. Wevers, Notes, p. 266; THGE, pp. 264-65). TgN has
‫ מתעתד‬where MT has ‫לפניך‬, to deal with the same ‘problem’: it may be just a
double translation as below in v. 9, but ‫ עתד‬is used by TgN in 7.15; 8.16; 9.15
of Moses ‘standing ready’ before Pharaoh (i.e. waiting for him) and this may
be where TgN found its ‘solution’. TgJ added ‘in the place where you shall see
a footprint’, which is based on a derivation of ‫ הצור‬from ‫ צור‬II = ‘draw, mark’
(cf. MRI [Lauterbach, p. 133]; AramB 2, p. 210 n. 4).
‫( על־הצור בחרב‬17.6) Vulg super petram Horeb ignores the ‫ ב‬and presum-
ably intends Horeb as a genitive, ‘upon the rock of Horeb’, a reasonable
paraphrase.
‫( בצור‬17.6) LXX, Vulg and Sy (with l) treat ‫ צור‬as the direct object,
ignoring the unusual ‫( ב‬cf. Note t on the translation), which is present in MT,
SP (no Qumran evidence survives) and the Tgg and is surely original. TgJ
‘(on) it with the rock/stone of your staff’ gives a double interpretation, with
‘it’ referring back to ‫ הצור‬earlier and then an expanded translation alluding to
the tradition that Moses’ staff was made of sapphire (cf. TgJ at 4.20 and Text
and Versions there; MRI [Lauterbach, p. 133]).
13
This is true whether one follows the text of Rahlfs (πρὸ τοῦ σέ with
Vaticanus) or Wevers (πρὸ τοῦ σὲ ἐλθεῖν, with Alexandrinus). πρὸ τοῦ σέ is
not ‘impossible Greek’ (Wevers, THGE, p. 264) – the ellipse of an infinitive of a
preceding verb after πρὸ τοῦ is also attested in Lev. 18.30; Num. 13.22 (BAlex,
p. 189; cf. Fritsch, p. 35) – and the easier reading with ἐλθεῖν may well be
secondary. Houtman (p. 364) favours the old emendation of Z. Frankel, πρότερός
σου, which is ingenious but unnecessary.
510 EXODUS 1–18
‫( ויצאו‬17.6) SP has the sing. ‫ויצא‬, which may be the original reading, even
though none of the other witnesses gives it any support (in LXX and Vulg
the sing. words for ‘water’ naturally take a sing. verb).14 It is the difficilior
lectio, as ‫ מים‬generally has a pl. predicate, but it is not impossible, since in a
few places ‫ מים‬takes a sing. (e.g. Gen. 9.15: see further BDB, p. 565), mainly
where it follows the verb as here (cf. GK §145o). The variation does not seem
to be a regular feature of SP and alternative explanations of it here based on
the wording of 7.15; 8.16; Num. 33.14 are hardly persuasive.
‫( מים‬17.6) TgJ adds ‘to drink’, a clarification that is scarcely necessary
but perhaps based on v. 1.
‫( העם‬17.6) LXXB and other mss read ὁ λαός μου (cf. mss 102 and 104
of OL), which must be a very early reading, as the whole clause was incorpo-
rated in this form in the LXX of Isa. 48.21 (though perhaps not in the original
[second cent. B.C.] translation: cf. J. Ziegler [ed.], Isaias [Göttingen, 1967],
pp. 25-26). Wevers (unlike Rahlfs) doubts if it is from the original translator
here (Notes, pp. 266-67), but it may be: the omission of μου (like that of
τούτου in v. 5) in LXXAFM could be a pre-hexaplaric correction towards MT.
‫( זקני ישׂראל‬17.6) This is the reading of most of the witnesses, including
SP and 4QpalExm: only LXX, Sy and TgN diverge. TgN has its usual ‫חכימייא‬
for ‫זקני‬, as in v. 5 (q.v.). LXX τῶν υἱῶν Ἰσραήλ (BHS mistakenly implies
that LXX also has an equivalent for ‫ זקני‬here) is presumably deduced from the
fact that ‫ העם‬are to drink (for the wording cf. v. 7), but it is the elders whom
Moses was to take with him when he struck the rock on this occasion (v. 5):
it is only in the later episode in Num. 20.7-11 that the people witness this.
Sy combines the readings of LXX and the Heb. in a way that gives priority
to the latter (sbʾ dbny ʾyśrʾl; 5b1 omits bny): it is scarcely original as Propp
thinks (p. 603).
‫( ויקרא‬17.7) Ms. 8b1 of Sy is the only text which specifically identifies
Moses as the subject. The Catena group and some other LXX miniscules have
ἐπωνόμασαν (cf. 16.31), with the intended subject presumably being the
Israelites in v. 6 (LXX).
‫( המקום‬17.7) LXX, Vulg, TgJ and Sy have ‘(of) that place’, as in the Heb.
of Gen. 28.19; 32.3; Num. 11.3, 34 etc. (cf. Gen. 21.31; Num. 13.24). The
addition is unnecessary and not extant in any Heb. source here. In 4QExc
‫ המקום‬probably (to judge from the regular letter-count in col. VIII) stood at
the end of a line and there is a puzzling lacuna of about three letters’ breadth
at the beginning of the following line before the next word. DJD XII, p. 122,
suggests that an indentation due to damage to the leather may be responsible
for the latter; one might alternatively posit that ‫( הזה‬or less likely ‫ )ההוא‬stood
there, added to provide a more specific reference as in the Vss noted above
(LXX occasionally renders ‫ זה‬by ἐκεῖνος). Longacre thinks that in 4QpalExl
14
The word does not survive intact in any Qumran ms., but in 4QpalExl its
final letter may have been waw as in MT rather than aleph (DJD IX, p. 39).
17.1-7 511
too the word following ‫ המקום‬might have begun with he rather than DJD’s
tentative mem (‘A Contextualised Approach’, p. 285: cf. pp. 206-207), which
could then be a further indication of a textual affiliation between this ms. and
4QExc (cf. above on 17.1).
‫( מסה ומריבה‬17.7) Most of the Vss as elsewhere render according to the
sense of these names, using words related to their equivalents for the verbs
in v. 2, rather than just transliterating them: only Sy follows the latter course,
curiously with nsh for ‫מסה‬, perhaps to make the etymological connection
with the explanation clearer. TgN adds a suffix, ‘his/its’ (presumably Israel’s),
taking the final he of the names to represent this; its mg prefixes ‫בית‬, ‘the
place of’. 4QExc writes the first name ‫ ;]מ[שה‬such variation between the
sibilants is found elsewhere, and not only at Qumran (cf. Qimron, pp. 28-30;
Reymond, Qumran Hebrew, pp. 68-70), but here the scribe might carelessly
have repeated ‫ =( משה‬Moses) from the previous verse. Such spelling errors are
rare in 4QExc, but another occurs in the next verse (‫עמלך‬: cf. also DJD XII,
p. 103). Vulg has only Tentatio, ignoring ‫ומריבה‬. This appears to be deliberate:
the same omission occurs in Jerome’s De situ et nom. loc. hebr. (Klostermann,
143.25; as in Eusebius’s original) and Ep. 78.13.2. It may be attributed to
Vulg’s tendency to occasional abbreviation or, perhaps, to a desire to avoid
confusion with the Meribah at Kadesh (Aquae contradictionis: Num. 20.13),
as Wellhausen suggested (Composition, p. 79 n. 2).
‫( ועל נסתם‬17.7) LXX καὶ διὰ τὸ πειράζειν, understanding the subj.
from τῶν υἱῶν Ἰσραήλ: the O-text adds αὐτούς to make it explicit. TgNmg
‘and because the Lord tested them’ must have taken the suffix to represent the
object and assimilates to 15.25, 16.4 and the explanation of the name Massah
in Deut. 33.8.
‫( את־יהוה‬17.7) Tgg (but not Sy) avoid making God the object by using
‫קדם‬.
‫( לאמר‬17.7) Sy wʾmrw: cf. v. 4.
‫( הישׁ‬17.7) TgJ,N amplify with ‫מן קושטא‬, ‘really’, and represent ‫ ישׁ‬with ‫שרי‬,
‘dwell, encamp’, a frequent verb with expressions for God’s presence (see the
next note).
‫( יהוה‬17.7) TgO prefixes ‘the Shekinah of’ (cf. 15.27) and TgJ,N ‘the glory
of the Shekinah of’ (cf. 16.7, 10), as they usually do when God’s presence is
referred to.
‫( בקרבנו‬17.7) LXX ἐν ἡμῖν and Vulg in nobis gloss over the specific term
used: Aq and Theod give the precise rendering.
‫( אם אין‬17.7) 4QExc reads ‫ואם אין‬: ‫ ואם‬is much less common than ‫ אם‬in
alternative questions, especially in the Pentateuch, where it occurs only in
Gen. 17.17 and 42.16 (cf. BDB, p. 50). Although it has no support in other
witnesses, it is the difficilior lectio and as such could be more original; but
there is a little evidence that at Qumran ‫ ואם‬was preferred elsewhere (4Q300
f8.6; 4Q511 f30.4; and especially 4Q365 f32.6-7, where MT and SP have
‫ אם‬in Num. 13.18-19), so that 4QExc may simply have been adapting to its
linguistic context.
512 EXODUS 1–18
All three Qumran mss which preserve this part of ch. 17 had a division
after v. 7 (like MT and SP): in 4QExc and 4QpalExl the verse ends in mid-line,
with an empty line below in the latter case, while in 4QpalExm the verse would
again have ended in mid-line and ‫( יבוא עמלק‬sic) at the beginning of the next
line indicates that the usual enlarged waw of this ms. would have stood in the
(not preserved) empty space.
C h ap t er 1 7 . 8 - 1 6
V i c tory ov er A m alek
The boundaries of the section are not in this case delimited by

itinerary-notes – in fact ‘at Rephidim’ in v. 8 places this episode at
the same place as v. 1 located the previous section (cf. also ‘from
Rephidim’ in 19.2). But the change in topic is clear at both the
beginning and the end and the recurring references to ‘Amalek’ give
the section a strong internal coherence. Both the Masoretic and the
Samaritan traditions have divisions before and after it; at Qumran
the division before it is strongly attested (4QpalExl, 4QpalExm,
4QExc) and there is indirect evidence that a short division occurred
after it in 4QpalExm and 4QExc (the end of ch. 17 and the beginning
of ch. 18 do not survive in the other mss). In addition there is a
division between vv. 13 and 14 in the Masoretic and Samaritan
texts (4QExc marked this too, but 4QpalExm provides no clear
evidence): this is probably due, as elsewhere, to the introduction of
divine speech in v. 14, but modern interpreters have seen a thematic
shift here too (see below).
Verses 8-13 comprise the narrative of a battle with the Amale-
kites, which is (i) introduced in v. 8, perhaps with a proleptic
summary; (ii) Moses explains his plan for action to Joshua, with
Joshua leading a selected group of warriors against the enemy,
while Moses stands on a nearby hill with ‘the staff of God’ in his
hand (v. 9); (iii) the execution of the plan (vv. 10-13) is framed by
the action on the battle-field (vv. 10a, 13); but (iv) the outcome is
influenced decisively by what Moses does on the hill, assisted by
Aaron and Hur (vv. 10b-12). The episode is rounded off in vv. 14-16
by what some have imprecisely called its ‘interpretation’ (Houtman,
p. 376, adding ‘and actualisation’), probably because of the promi-
nence of Yahweh in these verses as speaker and the one who fights
on Israel’s behalf, as he had done earlier against the Egyptians
(cf. 14.14; 15.1, 3, 21). (v) Yahweh instructs Moses to record and
declare his ongoing hostility against Amalek (v. 14); (vi) Moses,
in what could be seen as the original conclusion of the narrative,
514 EXODUS 1–18
builds an altar in honour of Yahweh (v. 15); and (vii) he(?) speaks of
the ancient hostility between Yahweh and Amalek (v. 16). Whereas
the main narrative moves smoothly (with one complication which
is discussed in the Explanatory Note on vv. 10b-12) from start to
finish, these concluding verses comprise at least two, and probably
three, separate sub-units which recognise Yahweh’s role as Israel’s
true leader and protector in quite different ways.
As it stands the section has consequently been described as a
‘didactic theological narrative’ (theologische Lehrerzählung: H.-C.
Schmitt, ‘Die Geschichte vom Sieg’; cf. Albertz’s title ‘lessons…’
[Lehren…] for vv. 14-16 [p. 293]). Yet its narrative section is
striking for the absence from it of the earlier paradigm of Israel’s
crises being laid before Moses and/or Yahweh, so that Yahweh’s
instruction and/or action can play the decisive part in the resolution
of the problem (cf. 14.10-14; 15.24-25a; 16.3-4; 17.2-4). Instead
Moses appears as the leader who already knows what to do and who,
almost like a magician, already has in ‘the staff of God’ the means
to sway the battle in Israel’s favour. Not surprisingly the narrative
has been seen as the most obvious example of a ‘heroic’ portrayal
of Moses’ leadership, in which the involvement of Yahweh (more
precisely ‘God’), such as it is, takes a very distinctive form (so
Coats, ‘Moses versus Amalek: Aetiology and Legend in Exod. xvii
8-16’, in Congress Volume: Edinburgh [VTSup 28; Leiden, 1975],
pp. 29-41; cf. Valentin, Aaron, pp. 168-73). These very different
descriptions of the section highlight some tensions both within it
and between it and other parts of the Exodus-wilderness narrative
which call for an explanation.
The explanation for them has not, however, even in the heyday of
source criticism, been sought in the isolation of two parallel strands
in the section. In particular, it has been agreed that the Priestly
source or redaction is not present here and until the recent upheavals
in Pentateuchal criticism most of the section, at least, was attributed
to one of the older sources.
Even today most scholars would recognise that the narrative has an early
origin, prior to the redactional activity of exilic and later times (see below
on Blum, Levin and Albertz). This is indeed a natural conclusion to draw,
not only from some archaic features in the story but from the fact that the
Amalekites disappear as a threat to Israel after the time of David. As for
its more specific origin, it is easy to forget now that until the mid-twentieth
century the section was mainly attributed to the Elohist source. This was
17.8-16 515
already the view of Knobel (Num.-Jos., p. 532: cf. Exod.-Lev., pp. 158-59)
and Dillmann (pp. 178, 182), who identified ‘the staff of God’ (v. 9: cf. 4.20)
and the mention of Hur (vv. 10, 12: cf. 24.14) as key indicators. Wellhausen
hesitated between J and E, but apparently favoured E, as his observations
about affinities with ch. 32 would suggest (Composition, p. 80: cf. pp. 88,
92). Carpenter/Harford-Battersby, Holzinger, Baentsch (p. 160: ‘unzweideu
tige Beweise’), Driver (Introduction9, p. 30), Gressmann (Mose, p. 155 n. 1;
Anfänge, pp. 107-108) and McNeile represent the general consensus of the
time, which has continued to find minority support in more recent times
(Hyatt, Propp, Baden).
The first challenge came in 1912 from Smend, who allocated both 17.8-16
and the verses which had been associated with it to his new ‘early J’ source
(J1). He cited the representation of Joshua as of like age to Moses rather
than as a younger assistant and the use of ‘raised’ (Heb. hērîm) rather than
‘stretched out’ (Heb. nāṭāh) for Moses’ action with his hand as contrasting
with E and agreeing, at least in the latter case, with J1 (cf. 7.20; Num. 20.11:
Erzählung, pp. 145, 147-48). The mention of Rephidim in v. 8 served to attach
the narrative to the once adjacent itinerary-note in 17.1, which in its original
unexpanded form Smend had, like the rest of the main sequence of itinerary-
notes in Exodus, attributed to J1, and also to 19.2. Although Smend’s ‘third
early source’ always remained a minority view and he was followed closely
here only by Eissfeldt, Beer and Fohrer, his arguments enabled Rudolph to
dispose very quickly of the consensus view (Elohist, p. 37). Noth, as often in
such matters, followed Rudolph’s attribution of most of the section to J (ÜGP,
p. 33; in his commentary, p. 113, ET, p. 141, he was more cautious: ‘It may
derive from J’) and a succession of commentators (e.g. Fritz, Childs, Coats)
were to agree with him, though without producing strong reasons for doing so:
even Fritz’s vocabulary argument (pp. 12-13) is hardly decisive and depends
inevitably on his attributions of other passages.
Closer study of the text led to part or all of some verses being identified
as later additions. ‘In Rephidim’ (v. 8) was apparently first so regarded by
B.W. Bacon in 1894 because of his (plausible) view that the narrative had
once stood in Numbers; for others, such as Noth, dependence on the ‘Priestly’
itinerary-note in v. 1 was sufficient reason in itself.1 Baentsch (pp. 160, 162)
introduced the idea that v. 14 was dependent upon Deut. 25.19 and so was
itself Deuteronomistic: this has been widely accepted. For Noth v. 14 was an
addition to the story, but it is not clear that he regarded it as later than Deuter-
onomy: the same applies to vv. 15-16 (p. 114, ET, p. 143). Fritz apparently
thought that v. 14 was post-Deuteronomic, but that vv. 15-16 were already
1
As a counter-argument it has been suggested that the ‘support’ given to
Moses’ arms in v. 12 was meant to explain the name (e.g. Blum, Studien, p. 152
n. 217), but this is over-subtle: in a real aetiology the connection is made much
more obvious (cf. v. 7).
516 EXODUS 1–18
in J (Israel in der Wüste, pp. 12, 56-57). The mention of ‘the staff of God’ in
v. 9 was deemed secondary by Beer, who also preferred the Samaritan
reading ‘his hands’ in v. 11 (p. 92): the unstated reason in both cases will
have been to remove the (apparent) inconsistency from the original story (cf.
Noth, p. 113, ET, p. 142; Childs, p. 313).
More recent study has seen, on the one hand, increased support for the
substantial unity of the passage, aided no doubt by the priority now given
to later literary layers by many scholars. Blum attributes it all to Kd, who
added v. 14 to an older story; Van Seters sees it as a fabrication by J (dated in
the exile) on the basis of Deut. 25.17-19 (Life, p. 207); Albertz distinguishes
only between an old narrative taken up by KEx in the exile and the mention
of the staff and v. 14, which were added by a later redactor (pp. 283-84).
Houtman (pp. 375-76) and Levin (p. 358) see vv. 14-16 as containing two
separate additions to the main narrative, which Levin (like Kratz, Komposi-
tion, pp. 246-47, 302-304) regards as a late supplement to the wilderness
narrative. Dozeman treats the whole section as an undifferentiated part of the
‘non-Priestly History’ (pp. 392-93).
The heirs of the older literary criticism, who still envisaged (and envisage!)
the use by redactors of parallel Yahwist and Elohist sources, continued not to
impose that pattern on this narrative. To begin with the tensions within the
narrative were resolved, in two different ways, by more far-reaching analyses
than before. Valentin had in 1978 presented an account of its literary growth
which was very similar to Noth’s, with ‘in Rephidim’, the mention of the
staff and the whole of v. 14 seen as later additions to the core. But in addition
he had taken up Keel’s suggestion of an iconographical basis for Moses’
hand-gestures (see the Explanatory Note on vv. 10b-12), which he thought
could explain the origin of this motif. Shortly afterwards Zenger identified
a Grundschrift belonging to the seventh-century ‘JE’ revision of the Yahwist
which spoke into the contemporary situation in the spirit of Isaiah through the
combined motifs of ‘the staff of God’ and the altar of ‘the Lord my banner’
(vv. 8*, 9-10a, 11, 13a, 15, 16a): that is, the whole scenario of Moses’ raised
hands and his need for support was relegated to a later (indeed very late)
redactional addition, along with the mention of Rephidim and the specific
sayings against Amalek (Israel am Sinai, pp. 78-86, 98-100).
Others have more straightforwardly attributed all or most of the section to
J or E. According to W.H. Schmidt it was most likely included in J, except for
v. 14 (Exodus, Sinai und Mose, p. 105 n. 150), and Graupner’s strong if brief
rejection of an attachment to E, since ‘the staff of God’ is to be regarded as
a later addition, suggests that he may have agreed (Elohist, p. 93). The older
view, that the passage is from E, has recently been championed (apparently
without any additions) by Propp (p. 615) and Baden (Composition, p. 125,
with the important new argument that this battle-narrative belongs closely
with 13.18).
17.8-16 517
When virtually all possible alternatives for the literary transmis-

sion of this probably ancient tradition have found some support
over the past century and a half, it might seem wisest to leave the
matter unresolved. But it is noticeable that the earlier challenges to
an attribution to E came either from those, like Smend, who were
looking for material that might be assigned to an additional source
(a hypothesis that has now lost the little support that it had) or
from Rudolph, who in his ‘crusade’ against the existence of E was
ready, where necessary, to grasp at a weak argument or to pass by
indications to the contrary. It cannot be said that a strong positive
case has been made for J and this may be why a number of writers
hesitate to express a confident view. The passage, first of all, has no
obvious connection with the J narrative that precedes it in vv. 3-6
(‘in Rephidim’ connects rather with the redactional itinerary-note
in v. 1; and the terminology for Moses’ staff in vv. 5 and 9 is quite
different), any more than it does with ch. 18 which follows, apart
from (and this may be significant) the recurrence of the divine title
‘God’ (Heb. ʾelōhîm). Internally most of vv. 8-13 hold together well,
but ‘in Rephidim’ (v. 8) and ‘and his people’ (v. 13) are probably
later additions (see the Explanatory Notes). There is undoubtedly
tension between the references to ‘the staff of God’ and Moses’
‘hands’ in the dual, but its origin remains obscure and it could well
go back to a no longer accessible pre-literary stage of the tradition’s
development. The fact that a number of different suggestions have
been made about how the tension can be eased is against regarding
it as the basis for a division of the text at the literary level.2 In
vv. 14-16 there are probably three separate ‘conclusions’ to the main
narrative (see the Explanatory Notes). Verse 14 has since Baentsch
been widely regarded as an addition based on Deut. 25.17-19. There
is a complication in the fact that these verses occur at the very end
of the specific laws in Deuteronomy and are quite unconnected with
the preceding verses: they too look like an addition and the older
view was that they were added on the basis of Exod. 17.14 with some
modifications to fit Deuteronomic sensibilities (cf. Mayes, Deuter-
onomy, pp. 330-31; Weinfeld, Deuteronomy and the Deuteronomic
School, p. 275). However this (possibly complex) issue of literary
dependence is resolved, v. 14 is probably an intrusive element which
2
The more so, since as will appear below, both motifs turn out to have associa-
tions with the same strand of the Exodus narrative.
518 EXODUS 1–18
separates the conclusion of the battle from its immediate celebra-

tion in v. 15. Verse 16 is often associated closely with v. 15, but only
on the basis of an unjustified emendation (see Text and Versions on
v. 16) or, in Houtman’s explanation, through the implausible claim
that the ‘hill’ on which Moses had positioned himself was referred
to as ‘the throne of Yah(weh)’. In fact the concluding words of v. 16,
which should be translated ‘from many generations (ago)’, point
rather to this whole verse having been moved from a quite different
context.
The narrative, therefore, in its oldest accessible form probably
consisted of most of vv. 8-13 and v. 15. The fact that it is a story
of a battle in which Israelites fight links it directly with 13.18.
Within these verses there are also a number of other features, as
the older critical commentators observed, which are characteristic
of passages which were traditionally assigned to the E source. The
choice of Joshua to lead the warriors, while probably related to the
military role which he plays in the conquest narrative, also reflects
the close partnership between Moses and Joshua which is repre-
sented by the designation of the latter as ‘Moses’ (young) assistant’
(Heb. mešārēt: cf. 24.13; 33.11 and elsewhere). Contrary to Smend’s
assertion, there is nothing here to suggest that Joshua was of the
same age or generation as Moses. Aaron and Hur, who accompany
Moses to the top of the hill, are mentioned together elsewhere only
in 24.14 and Hur nowhere else (there is no reason to identify him
with Bezalel’s grandfather in 31.1 or any of the other figures with
the same name). There are few places where Aaron participates
in his own right in the older narratives: he is most prominent, this
time as a leading opponent of Moses, in Exodus 32 (cf. Num. 12).
There Joshua also appears and in Moses’ reply to his statement that
‘there is a noise of war in the camp’ (Exod. 32.17-18) the nouns for
‘victory’ and ‘defeat’ correspond precisely to the verbs used here in
vv. 11 and 13: the words for ‘defeat’ are very rare, especially in this
sense (see Note m on the translation).3
3
This contrasts with Fritz’s list of vocabulary items in the passage which are
(supposedly) characteristic of J (see above), most of which are quite common
expressions in BH generally. Several of the instances in the plague-narratives
which he cited are, according to our analysis, not from J but from E. The use
of Heb. hērîm in v. 11 (rather than nāṭāh), noted by Smend, may be due to the
contextual contrast with ‘let…rest’ (Heb. hēnîaḥ).
17.8-16 519
Whereas Moses uses the divine name in v. 15, in v. 9 he refers to

his staff as ‘the staff of God’, as the narrator does in 4.20 (for the
alternation cf. 3.13-15). That verse, like 4.17, where God speaks
to Moses about ‘this staff’, is taken by us to be from E. A staff,
but described as ‘Moses’ staff’, also appears in a series of con-
nected passages from J (4.1-5; 7.15 [cf. 9.23; 10.13]; 17.5 [and
probably 14.16]). In E (4.17) God has told Moses to use ‘this staff’
to ‘perform the signs’ and this fits well with what he does here.
In the main non-Priestly plague narrative, which we have attrib-
uted to E, Yahweh himself brings the plagues upon Egypt, and the
narrator presumably left it to be assumed that Moses would wield
‘the staff of God’ at appropriate moments: his emphasis there is
entirely on the divine initiative and the narrative passes over what
might have been seen as quasi-magical acts on Moses’ part. Only
when a plague is to cease does Moses intervene, by prayer (8.8, 26;
9.33; 10.18). A similar concern can perhaps be seen in the narra-
tor’s construction of the short but complex story here: he mentions
the miraculous powers of ‘the staff of God’ which Moses holds
(vv. 9-11), but in vv. 12-13 it is a gesture most naturally seen as an
accompaniment of prayer (see the Explanatory Note) that leads to
the eventual victory. In the name he gives to the commemorative
altar Moses also makes clear that it is Yahweh, not he, who has led
the people to victory.
The building of an altar by a leading figure is widespread in
early narratives and appears in passages that have been traditionally
ascribed to both J and E, as well as occasionally in Joshua, Judges
and 1 Samuel (see the Explanatory Note on v. 15). Whether or not
the actual offering of a sacrifice is mentioned does not seem to be
a distinguishing mark of the two Pentateuchal sequences. But they
do differ, so far as the evidence goes, in the way in which they are
associated by the narrators with a deity. In J passages the altar is
said to be built ‘for Yahweh’ (Gen. 8.20; 12.7; 13.18), or the builder
is said to call upon the name of Yahweh (26.25), or both (12.8),
whereas in E passages the altar is named after the deity (Gen. 33.20;
35.7). The present instance clearly follows the latter pattern: the
only difference is that in the two Genesis passages the altar is given
an El-name, whereas here it is given a Yahweh-name, but that is not
surprising when the revelation of the divine name has intervened in
E (Exod. 3.13-15).
520 EXODUS 1–18
There are therefore a number of features in this story which

suggest that it, like 13.18, came from the strand of the older Penta-
teuchal narrative that was derived from the E source. Two further
questions about the story remain to be examined: its original place
in the wilderness narrative and its character. At the beginning of
the twentieth century it was already a common view that the story
really belonged to the closing stages of the wilderness narrative,
where the Amalekites appear as a threat to the Israelites near the
southern extremity of Canaan (Num. 14.25, 43, 45), as they do
again in 1 Samuel 15, 27 and 30 (Carpenter/Harford-Battersby 2,
p. 1107 [with reference to B.W. Bacon, whose Triple Tradition had
appeared in 1894, and to Harford-Battersby himself, presumably
for his articles on Exodus and Numbers in the HDB, which began to
be published in 1898]; Holzinger, p. 55; Baentsch, pp. 160-61). This
naturally led to Gressmann’s association of it with the cycle of Sagen
which centred according to him on Kadesh (although he located
Rephidim somewhat to the north of Kadesh: Anfänge, p. 107) and
to E. Sellin’s view that the narrative records the victory which gave
the Israelites control of Kadesh (Geschichte des israelitisch-jüdis-
chen Volkes [Leipzig, 2nd ed., 1924, 1932] 1, p. 69; cited by Beer,
p. 92, with apparent approval). This, like the Kadesh-theory more
generally, was already viewed with suspicion by Noth and subse-
quent scholars have been even more dubious about it (Blenkinsopp,
in The Pentateuch, pp. 137-38, 162-63 is an isolated exception in
recent times; for an earlier more detailed study see M.L. Newman,
The People of the Covenant [Nashville, 1962], pp. 68-101). But the
weakness of the arguments used to link other stories in Exodus to
Kadesh should not be allowed to obscure the very strong geograph-
ical connection of this narrative with the southern borderlands
of Canaan. Possibly it originated in post-settlement struggles of
Israelite groups with their Bedouin neighbours. Or it may contain
memories of attempts, sometimes successful (cf. Num. 21.1-3),
sometimes not (Num. 14.39-45), of Israel’s ancestors to force their
way into Canaan from the south – a more logical approach for the
bearers of the Exodus tradition than the long detour in the canonical
narrative which brought all the immigrant groups in from the east
across the river Jordan. As for the episode’s position in the extant
text, a possible reason is suggested below (p. 531). More positively,
one might see a clue in its association with other narratives (in ch.
18) about a foreigner’s much friendlier relations with early Israel:
17.8-16 521
was the non-Priestly account (or specifically the E source) using

these episodes to commend a more peaceful relationship with
Israel’s neighbours at a later period (on this see also pp. 553-54)?
There remains the question of the character or genre of the story
in its original or at least its earliest recoverable form. Its military
aspect and the absence of overt divine intervention distinguish it
from the stories of divine provision of food and water which precede
it as well as the sea narrative in 14.1-31. It has more in common
with the conquest-stories of Numbers and Joshua, and even more
with the largely defensive battle-narratives which must underlie
the more highly developed accounts in Judges and 1 Samuel 1–17.
But it has little interest in the actual fighting or in the tactics of the
engagement and the leader of the fighters is celebrated only for his
name. The real focus of the story is on what Moses, not Joshua,
does and his actions are anything but military in the usual sense.
He has understandably been compared to a magician because of the
effects which his actions are implied to have on the course of the
battle (cf. Josh. 8.18-19, 26 perhaps). But the gestures with both his
hands, which recall the accompaniment to prayers and other kinds
of ritual, and the building and naming of an altar at the end give the
story a more recognisably religious character, so that it appears as
a somewhat primitive form of the biblical ‘Yahweh-war’ narratives
in subsequent books.4 Moses’ initiative (and the lack of an explicit
divine one, though note 4.17, 20) and dominance of the narrative
lend some support to Coats’s suggestion that it is a Heldensage (see
above, p. 514), but this can only be maintained if Moses is seen as
a very unusual kind of Held, one who through divine preparation
is equipped, not to engage directly in the fight (like Baal with his
divinely made weapons in his mythical battles) but to invoke the
support of the deity to whom at the end he gives the credit for the
victory – a role which, for all its obvious differences, can be seen as
an extension of the one which he occupies elsewhere as the repre-
sentative of the people before Pharaoh and, in prayer, before God.
The original story, which celebrated the repulse of a particular
attack on the Israelites in their earliest history, has been supple-
mented in ways that made it into a divinely authorised basis for
4
Compare the suggestive references to the passage in Smend, Jahwekrieg, pp.
58-59, 74-75, 92-93 and n. 28 (ET pp. 79-81, 102-103, 128 with n. 28), although
they are not assembled into a complete account of its origin and early history.
522 EXODUS 1–18
continued warfare against the descendants of those attackers. Both

vv. 14 and 16 portray Yahweh as a warrior himself, a divine warrior
(cf. 15.3) who will fight against and defeat (indeed ‘wipe out’)
those who have taken up arms against his people. It is a vision
which was often repeated, both in the Old Testament and among
Israel’s neighbours. But this was a people whose prophets could
also threaten that their own God might turn against them if they
violated his expectations of them (Amos 5.18-20) and could hold
out an ultimate vision of a world in which ‘nation shall not lift up
sword against nation, neither shall they learn war any more’ (Isa.
2.4).
[Here and in ch. 18 the main component of the non-Priestly narrative again
appears to be derived from E and the sections not enclosed in square brackets
are those attributed to it.]
8 Amalek came and fought witha Israel [at Rephidim]. 9 So

Moses said to Joshua, ‘Choose men for us, and go out, fightb
against Amalek tomorrow,c while I station myself on the top of
the hill,d with the staff of God in my hand’.e 10 Joshua did as
Moses had said to him, fightingf against Amalek, and Moses,
Aaron and Hurg went up to the top of the hill.h 11 Whenever
Moses raisedi his hand(s), Israel would prevail,i but when he let
his hand(s) rest,i Amalek would prevail.i 12 Moses’ hands became
tired (lit. heavy),j so they took a rock and placed it under him and
he sat upon it, while Aaron and Hurg supported his hands, one on
one side and one on the other. Then his hands werek firml until
the sun set. 13 Joshua defeatedm Amalek [and his people] with
the edge of the sword.n 14 [Yahweh said to Moses, Write this as
a reminder in a recordo and make Joshua hear itp: (that)q I will
indeed wipe out the memory of Amalek from all the earth (lit.
under heaven).r] 15 Moses built an altar and he called it ‘Yahweh
is my banner’. 16 [He said, (that)s I swear (lit. hand upon throne)t,
Yahweh is at war with Amalek from many generations (ago).u]

a. On the use of ‫ עם‬with ‫ לחם‬Niphal here (but not in vv. 9-10 [cf. 16]) see
Valentin, Aaron, pp. 155-56.
b. The asyndetic construction in the Heb. is frequent with two adjacent
imperatives, especially when the first is a verb of movement (JM §177e: cf.
19.21).
17.8-16 523
c. On early debates about whether ‫ מחר‬belongs with what precedes it or

what follows it see Text and Versions. Since the battle and Moses’ presence
on the hill are simultaneous, it makes little difference, but the time should
be specified for the first action mentioned (cf. Propp, p. 617: also Valentin,
Aaron, p. 142 and n. 2): ‫ אנכי נצב‬can introduce a circumstantial clause (GK
§156c), so there is no need to read ‫( ואנכי‬Holzinger) or ‫( והנה אנכי‬Beer, with
LXX).
d. Heb. ‫הגבעה‬: although no ‘hill’ has been previously referred to (the ‫צור‬
in v. 6 is probably too far away, as well as being different in meaning, to be
meant here), the dialogue assumes that one is adjacent enough to the speakers
to be obviously in mind (cf. Bekins, ‘Non-Prototypical Uses’; JM §137m-n
gives some similar examples). This is distinct from the use of the article to
determine a noun by ‘anticipation’ as in 3.2 and 16.13.
e. A circumstantial nominal clause, attached by waw and with the subject
first (JM §159d).
f. Heb. ‫לחלחם‬: the inf. cons. ‘define(s) more exactly’ how Joshua obeys
Moses’ instructions (GK §114o).
g. The (composite) subject precedes the verb because it differs from the
subject of the preceding clause and there is an element of contrast between
the actions involved; perhaps also because they are simultaneous (JM §118e-f;
155nb).
h. Heb. ‫ראשׁ הגבעה‬, without a preposition: a common idiom (GK §118d, f).
i. All the verbs in this verse are iterative imperfects, indicating the ebb
and flow of the battle. ‫ יניח‬is the standard Hiphil form (contrast Note kk on
the translation of 16.1-36), so the meaning is ‘gave rest to’, not ‘put, placed’.
j. Heb. ‫כבדים‬, ‘heavy’, then ‘burdensome’ (18.18 [leading to ‘exhaus-
tion’]; Num. 11.14; Prov. 27.3). For the sense ‘tired, weary’ cf. Sy lʾy. Heb.
‫ יד‬is always fem. elsewhere: the masc. part. following it is most unusual but
explained in GK §145u by analogy with the ‘dislike’ for the third person pl.
fem. in the imperfect (cf. Jer. 44.19; Prov. 26.23).
k. Heb. ‫ויהי‬: the plural form would be expected (cf. SP and Vss), but when
the predicate precedes the subject, as here, it is sometimes in the ‘simple’ or
uninflected sing. form (GK §145o; esp. with ‫היה‬, JM §150j-m).
l. Heb. ‫אמונה‬, lit. ‘firmness’: for the noun serving in place of an adjective
cf. GK §141c, which suggests that it gives greater emphasis to the attribute.
m. Heb. ‫ויחלשׁ‬. The Vss give the sense as ‘put to flight, defeated’, which is
what the context requires (see Text and Versions): cf. the addition of ‫ ויכם‬in SP
and 4QpalExm. The noun ‫ חלושׁה‬in 32.18 probably has a related sense along-
side ‫גבורה‬. But other occurrences of the root in BH have the sense ‘weakness,
feebleness’ that is found in JAram and Syr. (Joel 4.10; Job 14.10 [of the
dead]). In Isa. 14.12 ‫( חולשׁ‬of Babylon) has been interpreted in both ways,
depending on whether the phrase in which it occurs (‫ )חולשׁ על־גוים‬is related
to Babylon’s former greatness (then ‘conqueror’) or to its present humiliation
(then ‘powerless’): the most recent study of J.B. Burns, ‘ḥôlēš ʿal in Isaiah
524 EXODUS 1–18
14.12: A New Proposal’, ZAH 2 (1989), pp. 199-204, follows the first alterna-
tive but prefers the sense ‘warrior’. Cognates for the sense ‘defeat’ have been
suggested in related languages: Ar. ḫalasa (A. Guillaume, ‘The Use of ‫חלשׁ‬
in Exod. xvii.13, Isa. xiv.12 and Job xiv.10’, JTS N.S. 14 [1963], pp. 91-92;
cf. I. Eitan, ‘Two Unknown Verbs: Etymological Studies’, JBL 42 [1923],
pp. 22-28 [25-28]); Akk. ḫalāšu and Soqotri ḫlš (W. von Soden, ‘Kleine
Beiträge zum Ugaritischen und Hebräischen’, in Hebräische Wortforschung
[FS W. Baumgartner; VTSup 16; Leiden, 1967], pp. 291-300 [296-97]). On
this basis recent lexica have envisaged two homonymous roots (HAL, p. 311;
DCH 3, pp. 246-47; Ges18, p. 362). But the meanings of the Ar. and Akk.
words (‘plunder, seize’ and ‘scrape off’ respectively) are not very close to
‘defeat’ and Soqotri (a modern South Arabian language) is not very likely to
have alone preserved an ancient Hebrew meaning. It is still of course possible
that there were two unconnected homonyms in BH, even if one of them is
difficult (so far) to parallel in another language. But it may be preferable to
view the two senses of ‫ חלשׁ‬as in some way related. Burns (as cited above)
seems to regard the notion of ‘cutting down’ as the unifying element, but it is
difficult to deduce this sense from any of the occurrences. The old view (cf.
BDB, p. 325; Childs, p. 311; Houtman, pp. 384-85) that there was a semantic
development in BH from ‘weak’ to ‘(make) prostrate, inert’ (cf. Job 14.10)
and then to ‘defeat’ may be correct after all: it would match the parallel
development in the case of ‫ גבר‬from ‘be strong’ to ‘prevail, be victorious’
(for which cf. 1 Sam. 2.9; 2 Sam. 11.23; Lam. 1.6; and v. 11 here), and it is
perhaps then not surprising that ‫ גבר‬and ‫ חלשׁ‬occur together both where the
meaning is ‘strong/weak’ (Joel 4.10) and where it is ‘victory/defeat’ (Exod.
32.8: compare also Tgg there). BDB retains the Qal pointing of MT here but is
understandably tempted by the possibility of revocalising the verb as a Hiphil;
a Piel would also be possible. A change to the consonants would be required
to produce a corresponding form in Isa. 14.12 (and perhaps that is why BDB
hesitated), but the meaning (and text) there remains uncertain and ‫ חולשׁ‬in the
sense of (the adj.) ‘prostrate’ could yet be correct.
n. Heb. ‫לפי חרב‬. The idiom is frequent (35x in BH, esp. in Josh. and Judg.)
and ‫ פי‬probably refers metaphorically to the ‘blade’ or ‘edge’ of the sword
(rather than meaning ‘as the sword devours, without quarter’, BDB, p. 805):
the pl. forms ‫( פיות‬Judg. 3.16 [with ‫ ;]שׁני‬Prov. 5.4) and ‫( פיפיות‬Ps. 149.6; Isa.
41.15 is different) are used in this way, and there is a comparable usage of Gk.
στόμα (LSJ, pp. 437, 1649; cf. δίστομος in LXX).
o. Heb. ‫בספר‬. The def. art. is ‘anticipatory’: see Note o on the translation
of 2.11-22. The common rendering ‘in a book’ gives a misleading impression,
as books of the modern form did not exist in biblical times (cf. H.M. Orlinsky,
Notes on the New Translation of the Torah [New York, 1969], p. 172). A ‫ספר‬
might be in the form of a scroll (Jer. 36.2, 4: cf. Isa. 29.11, 12; 34.4) or an
inscribed piece of pottery (‘ostracon’: so especially for letters and lists in
Heb., as archaeological discoveries like the Lachish letters have shown) or
17.8-16 525
even (at least in Phoen. and Aram.: cf. DNWSI, pp. 799, 800) an inscription
on stone. See further in the Explanatory Note. Since the content as well as the
form could be very varied, ‘document’ (NJPS) or ‘record’ is as close as we
can get to the sense.
p. Heb. ‫ושׂים באזני יהושׁע‬. The idiom is unique (more commonly x-‫ באזני‬is
found with ‫[ דבר‬11.2] or ‫[ קרא‬24.7]) and it seems to convey a stronger sense
than simply ‘say to’: compare the use of ‫ שׂים‬with ‫ על־פי‬and ‫ בפי‬for ‘to make
someone remember/speak’. Perhaps ‘impress upon’ comes closest to what is
intended (cf. Houtman, pp. 385-86).
q. Heb. ‫כי‬. A causal rendering (‘For, because’) is excluded because there
is nothing for the following words to explain: most likely this is simply the
‫ כי‬recitativum (‘that’) which is not needed with direct speech but appears
by analogy with the introduction to indirect speech (BDB, pp. 471-72; JM
§157c). In such cases (cf. 3.12; 4.25) it is also possible to see an ‘emphatic’
use of ‫‘( כי‬Surely, indeed’: so already BDB, p. 472 [‘e’]: cf. HAL, p. 448;
Ges18, p. 539) but, especially since support for such an interpretation was
claimed from Ugaritic, there have been counter-arguments seeking to limit
or even exclude its application (Muraoka, Emphatic Words, pp. 158-64;
A. Aejmaleus, ‘The Function and Interpretation of ‫ כי‬in Biblical Hebrew’,
in On the Track, pp. 166-85 [esp. 180-81]). There is certainly no need for
it here.
r. Heb. ‫מתחת השׁמים‬. The implication of this expression seems to be
‘throughout the earth’; only in Gen. 1.9 is there any contrast with what is in
or above the heavens. Similar expressions (sometimes with a different verb)
occur in Deut. 7.24; 9.14; 25.19 (Amalek again); 29.19; 2 Kgs 14.27 (Dtr.);
Lam. 3.66. In BH generally (and including cases with ‫ כל‬before ‫ )השׁמים‬this
is the only occurrence of the phrase that could be pre-Deut. in origin (3x in
Genesis [all P]; 6x in Deuteronomy; 1x in Kings; 3x in Job; 3x in Ecclesiastes;
1x in Lamenations; 1x in Daniel).
s. Here Heb. ‫ כי‬might in theory have a causal sense, since what follows
could be intended to explain the naming of the altar. But the connection is
rather loose (it would be tighter if v. 16 immediately followed v. 14) and as
above (cf. Note q) a ‫ כי‬recitativum is most likely intended.
t. Heb. ּ‫יד על־ ֵכּס יָ ה‬. ‫ ֵכּס‬is a hapax legomenon, with the following ּ‫ יָ ה‬being a
shortened form of the divine name (as in MT at 15.2, the cry ‫ הללו־יה‬and a few
other places). ‫ ֵכּס‬has been explained as an abbreviated form of ‫ ִכּ ֵסּא‬, ‘throne’
(BDB, p. 490, though without any explanation for the shortening), but MT is
probably corrupt and most commentators adopt a reading based on the other
witnesses or a conjectural emendation (see Text and Versions). ‫( יד‬usually
with a pron. suffix) is followed by ‫ על‬in a wide variety of idioms, but there
seems to be no parallel with as abrupt a beginning as this.
u. Heb. ‫מדר דר‬. If the meaning is ‘from generation to generation’ (so LXX,
Vulg) one would expect ‫( מדר לדר‬Isa. 34.10) or ‫מדר אל־דר‬, but a noun without
a preposition can be used adverbially to express movement from place to
526 EXODUS 1–18
place (GK §118f) and ‫ עולם‬is sometimes used without the usual ‫ ל‬to mean ‘for
ever’ (Pss. 21.5; 61.8; 89.2), so ‫ דר‬might mean ‘to generation’. But there seem
to be no parallels to this and when the same noun is immediately repeated it
usually has the effect of expressing totality (GK §123c: cf. 3.15 and elsewhere
‫)יום יום‬, which suggests that the present phrase refers entirely to the past, to
‘war…from of old’ (so Gressmann, Anfänge, p. 107: Deut. 32.7; Isa. 51.9;
58.12; 61.4). There are variant readings of it (see Text and Versions), but they
seem all to be secondary. Valentin has an excellent note on the grammar of
MT (Aaron, p. 144 n. 5), but he emends ‫ דר‬to ‫ לדר‬to fit the context better as
he understands it. See further in the Explanatory Note.
Explanatory Notes
8-10a. The arrival of the Amalekites and their attack on the
Israelites introduces a new kind of threat to the latter, in this part
of their wilderness journey at least. No doubt there were in biblical
times, as there have been subsequently, desert tribes in parts of the
Sinai peninsula who might have resented the need to share its sparse
resources with outsiders. Biblical tradition gives a rare glimpse of
such inhabitants in 1 Kgs 11.18 and the Midianites of Exod. 2.18-22
may have lived in the peninsula. Egyptian texts occasionally refer
to such peoples (see the introduction to 2.11-22 on the story of
Sinuhe); some of the references to Shasu (cf. ABD 5, pp. 1165-67)
could have them in mind. ‘Amalek’, the name used here (and in
vv. 11, 14 and 16) for the people and occasionally elsewhere for its
hypothetical Edomite ancestor (Gen. 36.12, 16; 1 Chr. 1.36), also
occurs frequently as an enemy of the Israelites in later stages of the
wilderness journey (Num. 13.29; 14.25, 43, 45; 24.20; Deut. 1.44;
25.17) and in Judges, 1–2 Samuel and 1 Chronicles (cf. Ps. 83.7).
Their strongest geographical association is with the south of Pales-
tine and the steppe-land and mountains beyond it, in biblical terms
the Negeb (Num. 13.29; 14.43, 45; 1 Sam. 27.8; 30.1) and Mount
Seir (1 Chr. 4.43): their mention after Kadesh in Gen. 14.7 would
also fit this locale (see further ABD 1, pp. 169-71, and the notes on
later verses in this section).
It is possible that ‘and fought with Israel’ is a general introduc-
tory statement, a summary of the fighting which is to be described
in vv. 10-13 (Cassuto, pp. 204-205). Alternatively it may refer to
an initial skirmish, which leads Moses to organise a response for
the main battle on the following day (v. 9). The present text places
17.8-16 527
the battle at Rephidim, where the itinerary-note in v. 1 locates the

previous episode (see the Explanatory Note there for discussion of
its location). But that verse is probably an addition to the narrative
in vv. 2-7 (see the notes there) and here many commentators have
seen the words ‘at Rephidim’ as an even later insertion. In fact the
story about Amalek may, in view of the geographical connections
noted above, have originally belonged to a much later stage of the
wilderness journey (see further the introduction to this section).
On this occasion (contrast 14.10-14) Moses does not simply
trust in God to repel the attack: as in many later narratives the
Israelites will have their own warriors to fight for them, chosen and
led by Joshua, who appears in the narrative here for the first time
and yet without any special introduction (like that, for example,
which he receives in 33.11). The narrator apparently assumes that
Joshua is well known to his hearers/readers, presumably (given the
nature of his task) as the hero of stories in the book which now
bears his name, where he leads the Israelites’ conquering forces in
Canaan. Elsewhere in the Pentateuch, where there is of course very
little fighting to be done, Joshua appears in quite different roles,
chiefly as Moses’ assistant in the non-Priestly narratives (24.13;
32.17-18; 33.11; Num. 11.28; Deut. 1.35) and then as his desig-
nated successor in Deuteronomy and P (Num. 27.18, 22; 32.12, 28;
34.17; Deut. 3.28; 31.3, 7-8, 14, 23; 32.44; 34.9: cf. Num. 13.8,
16; 14.6, 30, 38; 26.65). According to Albrecht Alt’s pioneering
traditio-historical study, Joshua was originally an Ephraimite
(Josh. 24.29-30; cf. 19.49-50), who appeared first as a charismatic
war-leader (Josh. 10.1-15), judge (Josh. 17.14-18) and eventually
the founder of the Yahwistic tribal league centred on the Shechem
sanctuary (Josh. 24.19-28). Only later did he come to be included in
the Benjaminite tribal legends preserved at Gilgal and the national
conquest tradition as a whole, and as a result also as a minor
figure in Pentateuchal tradition (‘Josua’, in P. Volz et al. [eds.],
Werden und Wesen des Alten Testaments [BZAW 66; Berlin, 1936],
pp. 14-58 [13-29]; see further K. Mohlenbrink, ‘Josua im Penta-
teuch’, ZAW 59 [1942/43], pp. 14-58; Noth, ÜGP, pp. 192-94, and
in his commentary, pp. 113-14, ET, pp. 141-42).
The Masoretic division of v. 9 makes ‘tomorrow’ the beginning
of Moses’ declaration of his own intentions (so also most EVV., but
not JB, NEB, REB). It was noticed in antiquity that the word could
528 EXODUS 1–18
just as well be joined to the first half of the verse, and this actually
makes better sense (see Text and Versions and Note c on the trans-
lation). Moses’ taking of a ‘staff’ with him is at first reminiscent
of the previous episode (vv. 5-6), but here it is described as ‘the
staff of God’, an expression which only occurs elsewhere in the
composite account of Moses’ commissioning (4.20), where it takes
up God’s gift of the staff to Moses in v. 17. The distinctive designa-
tion (with ‘God’ [Heb. ʾelōhîm] as a divine title in place of ‘Yahweh’
for the first time since 13.17-19) has often been seen as an indica-
tion that this section was taken from the Elohist source. The staff
is not mentioned in the following verses, which seem to move to a
different mode of securing victory, and the passage may draw on
two once separate traditions in this respect (see the next note and
the introduction to this section).
10b-12. Moses is accompanied on his climb by two other figures,
Aaron and Hur, whose function in this story only subsequently
becomes clear (v. 12). They again appear together in 24.14, where
they remain with the Israelites while Moses (with Joshua) ascends
‘the mountain of God’ (Heb. ʾelōhîm again) to receive the stone
tablets on which the commandments are written: they appear to
deputise for Moses in the judicial role of which more is to be said
in 18.15-16, where once more the title ‘God’ (ʾelōhîm) occurs, as it
does several times elsewhere in that chapter. As in the non-Priestly
narrative generally, there is little or no sign that Aaron as yet has
any priestly functions (see the Excursus in the introduction to
4.10-17). Hur is only mentioned here and in ch. 24 (unless he is to
be identified with Bezalel’s grandfather in 31.2; 35.30; 38.20, who
is presented as a Judahite): his name could be Egyptian, connected
with that of the god Horus (Beer, p. 92; Propp, p. 617), but it is just
as possible (and just as unprovable) that it is related to that of the
Horites in the Edomite genealogy in Genesis 36 or an individual in
that list (v. 22). If there were once more extensive traditions about
him, they are (like much else) no longer recoverable.
As v. 9 has already hinted, it is Moses’ actions on the hill-top –
no more words are attributed to him at this point – which will be
decisive for the outcome of the battle on the plain below. But there
are problems with both the text and the interpretation of vv. 11-12.
In v. 12 the manuscript evidence is uniform and indicates that
both Moses’ hands became tired and needed support from his two
companions to make them steady. On the other hand, v. 11, according
17.8-16 529
to both Qumran manuscripts which survive at this point and the

Masoretic text, says that Moses raised and lowered his ‘hand’ in
the singular. The Samaritan Hebrew text, all the ancient transla-
tions and an early Jewish commentary (the so-called Mekhilta of
Rabbi Ishmael) have ‘his hands’, but this looks suspiciously like a
harmonisation to fit the dual in v. 12 (see further the note in Text and
Versions). It has been suggested that there need be no contradiction
here: ‘his hand’ in v. 11 may refer to the alternating use of each hand
over a long period which led to both hands (or arms) eventually
becoming too tired to use any more (Baentsch; Houtman, p. 375: for
earlier proponents see Propp, p. 618). But the wording in the second
half of v. 12 strongly suggests that both hands were at that point
being raised together. Most probably v. 11 should be understood in
the light of v. 9 and passages like 9.22-23 and 10.12-13 where the
stretching out of hand and staff is synonymous, and v. 12 incorpo-
rates a different narrative motif which did not involve a staff at all
but provided an engaging way to extend and elaborate the story (see
also the introduction to this section on another possible motive for
the combination of the two scenarios).
There remains the question of what interpretation is to be given
to Moses’ gesture(s). Traditionally an expression of prayer to God
has been seen here (as in the Targums: other explanations were
also suggested) and some recent scholars have supported this
view (cf. Houtman, p. 381; H.-C. Schmitt, ‘Die Geschichte vom
Sieg’, pp. 340-41; Van Seters, Life, pp. 205-206). It can claim the
support of some ancient iconographical evidence and of Hebrew
usage (cf. ‘spread out my hands to Yahweh’ in 9.29 and elsewhere).
Many, however, have rejected it (e.g. Childs, Dozeman, Albertz:
see below). Given the presence, it seems, of two different gestures
by which Moses’ influence on the battle is portrayed, it is surely
necessary to consider them separately. The wielding of ‘the staff of
God’ is familiar from earlier in the Exodus narrative (most recently,
though without the specific designation, in 14.16 and 17.5-6),
where it is preceded by a divine command and so does not have
independent magical power. But here there is no such command
and the precise correlation of Moses’ movements with the Israel-
ites’ success or lack of it strongly suggests a magical intention.
A comparison has often been made with Joshua’s wielding of the
sickle-sword (kîdôn) in the battle for Ai (Josh. 8.18-19, 26), but
in the biblical narrative this is preceded by a divine command and
530 EXODUS 1–18
might even be understood as a signal for the army.5 The raising

of both hands that is implied in v. 12 was derived by Keel, whom
Valentin followed closely, from the portrayal of such a gesture in
Egyptian iconography.6 Keel acknowledged that this gesture could
have a wide range of associations – ‘honour, praise, prayer, blessing,
protection and defence’ (p. 95) – and found its origin and apparently
its continuing significance in the practice of incantation.7 But for the
understanding of Exodus 17 he attached particular importance to
memorial reliefs and a scarab in which a smaller figure with raised
hands is shown alongside Pharaoh killing an enemy. The problem
for Keel is that, as he admits, the smaller figure is not seeking
(divine?) support for Pharaoh but asking for Pharaoh’s blessing on
himself, which is a quite different scenario from Exodus 17. As a
result Keel and Valentin have to make the improbable assumption
that the narrative in Exodus 17 was shaped by someone who had
actually seen such a stele or scarab but misunderstood its meaning.
Of course if he did, he may well not have picked up the ‘magical’
implications which Keel apparently sees in the Egyptian examples.
In any case, this again seems a very roundabout way to find an
explanation for the action attributed to Moses, when the raising of
hands in prayer is such a well attested practice (so Keel himself in
Die Welt, pp. 298-301, ET pp. 318-23). It is reflected in Hebrew
and Akkadian idiom (TWAT 3, 424-25, 443-44 = TDOT 5, pp. 396,
416) and provides the most likely interpretation of the outstretched
hands on a small stele from a Late Bronze Age temple at Hazor
(Y. Yadin, Hazor [Schweich Lectures 1970; London, 1972], pl.
XIVa, pp. 71-72); cf. S. Schroer, ‘Zur Deutung der Hand unter der
Grabinschrift von Chirbet el Qôm’, UF 15 [1983], pp. 191-99 [196:
5
Zenger noted the iconographical parallels adduced by O. Keel, in which a
god wields the sickle-sword behind Pharaoh as he fights his enemies (cf. Wirk
mächtige Siegeszeichen im Alten Testament [OBO 5; Freiburg and Göttingen,
1974], pp. 13-82) and suggested that first Joshua and then Moses (with the staff)
were substituted for the god in the Egyptian model (Israel am Sinai, pp. 90-93).
Whatever one may think about the closer parallel in Josh. 8, it is scarcely neces-
sary to invoke such a complex origin for Exod. 17.9-11, which operates with a
common motif of the Exodus tradition.
6
Keel, ibid., pp. 91-109; Valentin, Aaron, pp. 182-88.
7
He found particular significance in a prehistoric rock-drawing that shows
its use by a snake-charmer, which is reproduced in Die Welt, pp. 290-91, fig. 417,
ET pp. 312-13.
17.8-16 531
‘plausibel’]). Against this background the fact that no words spoken

by Moses are mentioned (noted e.g. by Childs, p. 314) need not be
significant.8
13. Another battle between Israel and the Amalekites (joined
on this occasion by Canaanites) is referred to in Num. 14.44-45,
located in or near the southern hills of Canaan. On that occasion the
Israelites are defeated, according to the narrative because they have
attempted to defy Yahweh’s intention to lead them into Canaan by a
different and longer route (Num. 14.25, 39-43). A similar outcome
seems to lie behind the account of an Amalekite attack in the
Exodus period in Deut. 25.17-18 and 1 Sam. 15.2. By contrast the
narrative here speaks of a victory for the Israelites: if, as we have
suggested above (see the Note on vv. 8-10a), the story told origi-
nally of an episode on the southern border of Canaan, it could have
formed part (like Num. 21.1-3 perhaps) of a different account of the
settlement process, according to which (some of?) Israel’s ances-
tors entered Canaan not from the east, across the Jordan, but from
the south (directly from Kadesh?). Then its removal to its present
(rather puzzling) position could be due to a standardisation in the
tradition, which fixed the route of all the tribes’ entry into the land
in a single movement from the east and required the suppression or
displacement of alternative accounts.
An objection to this suggestion might be based on the unusual
verb translated ‘defeated’. Its etymology is debated (see Note m on
the translation) but it may well be related to words meaning ‘weak’.
Some have therefore suggested that its meaning here is really
‘weakened’, so that the victory described was not a decisive one: this
does of course fit well with the fact that the Amalekites continued
to be a troublesome enemy until at least the time of David. Against
such an interpretation may be noted the strong expression ‘with the
edge of the sword’ and the fact that the related noun, which is used
in 32.18, does seem to mean ‘defeat’ in contrast to a victory.9
8
Albertz has recently argued that the Heb. phrase used in v. 11 (rûm Hiph.
with yād) is against an interpretation of the action as an accompaniment of prayer,
since it is never used in this way. This is less of a problem for a ‘mixed’ interpreta-
tion of the passage like that suggested here, but in any case Heb. nāśāʾ and yād are
certainly used in connection with prayer and praise (Pss. 28.2; 63.5).
9
The words ‘and his people’ are unexpected because throughout the passage
so far (and also in vv. 14 and 16) ‘Amalek’ is evidently the name of a people, not
532 EXODUS 1–18
14. So far Yahweh has given no assurances or instructions

about the threat posed by the Amalekite attack, in contrast to the
preceding episodes: the initiative has lain mainly with Moses. Only
now, in what seems to be an afterthought when the battle is over,
does Yahweh declare his intention to eliminate the Amalekite threat
and ensure that it (or is it the narrative of this battle?) is remem-
bered by future generations. The provision for remembrance is not
made for the first time in Exodus: the name of Israel’s God is itself
literally a ‘memorial’ (3.15: see Note g on the translation there),
Passover and other rituals serve as a reminder of the Exodus (12.14,
27; 13.3, 14-16) and a jar of manna is to be preserved to show
Yahweh’s care for his people in the wilderness (16.32-34): compare
also 10.1-2. But the manner and method of the remembrance is
different here. A written record is to be kept and Yahweh’s inten-
tion is to be made known specifically to Joshua, the military leader.
Although the Heb. definite article is used with ‘record’ this need
not mean that the writing was to be in an already existing document
(see Note o on the translation), though some have thought that ‘the
record’ might be ‘the Book of the Wars of the Lord’ (Num. 21.14) or
something similar. In fact very similar words occur in Deut. 25.19,
after the recollection of an Amalekite attack on the Israelites in the
wilderness, and there can be no doubt that there is some connection
between that passage and this. Similar language is found in other
contexts in Deuteronomy and later writings (see Note r on the trans-
lation) and it is understandable that v. 14 is commonly seen as a
Deuteronomistic addition to the story, which looked forward to the
unusually specific ‘reminder’ in Deut. 25.19. Joshua is never said in
the Old Testament to have fought against the Amalekites again, but
the wording here does foreshadow his military exploits against the
Canaanites in the book of Joshua, including the wholesale slaughter
of peoples required in Deut. 7.1-6, 16-24.
15. The initiative returns to Moses, in what is a fitting and
coherent conclusion to the battle-narrative in vv. 8-13. (By contrast,
it has no connection at all with v. 14, which is another sign that
an individual: so generally also elsewhere (e.g. Judg. 6.3), the individual sense
being rare (see the Note above on vv. 8-10a). Perhaps a scribe who had the latter
sense in mind added these words to make clear the full extent of Israel’s victory.
Alternatively it would be possible to translate ‘namely [waw explicativum] his
warriors [for this sense of Heb. ʿām cf. Num. 20.20]’ (so Valentin, Aaron, pp.
161-62; Houtman, pp. 384-85: see also Text and Versions).
17.8-16 533
the latter is an intrusive addition.) An ‘altar’ was etymologically

a place of sacrifice (Heb. mizbēaḥ, cf. zābaḥ, ‘sacrificed’) and it
is no doubt because sacrifice was a central ritual of Old Testament
religion that there are numerous stories of the building of altars by
national heroes in the biblical narratives, and not only in the primary
accounts of the inauguration of Israelite worship. Often these stories
include mention of the sacrifices offered on the altar (Gen. 8.20
[Noah]; 22.9-13 [Abraham]; Exod. 24.4-6 [Moses]; Josh. 8.30-31
[Joshua]; Judg. 6.26-27 [Gideon]). Sometimes, as here, they do
not, but in many of these cases the narratives are brief and serve to
associate a prominent figure of the past with an Israelite sanctuary
of later times (Gen. 12.7 [Shechem]; 12.8 [Bethel]; 13.18 [Hebron];
26.25 [Beersheba]; 33.20 [Shechem]; 35.3, 7 [Bethel]). That may
also originally have been the intention here, though if it was the
sanctuary can no longer be identified. The absence of a specific
mention of sacrifice here is no reason to say, as Houtman does, that
‘the altar was not built for the offering of sacrifices’ (‘ “Yahweh is
my Banner” ’ [for details see below on v. 16], p. 111). Houtman cites
the story in Joshua 22 as a parallel case for an altar not being built for
sacrifices, but that is a ‘theologische Lehrerzählung’ from a much
later period, designed to maintain the monopoly of a central shrine
required by Deuteronomy. Even from that story (cf. vv. 10-20) it is
clear that the normal expectation was that altars were for sacrifice,
and 1 Sam. 14.35 gives another example of such an altar being built
as part of victory celebrations.
Nevertheless that is not what the narrator wished to empha-
sise here: he tells rather of the name that the altar was given (cf.
Gen. 33.20; 35.7). ‘Yahweh (is) my banner’ is a name with clear
military connotations (Isa. 5.26; 13.2-3; 31.8-9; Jer. 4.21; 51.12,
27; Ps. 60.6): other uses of a ‘banner’ (Heb. nēs) are derived from
this (including Num. 21.8, where nēs is used metonymically for
a ‘pole’ on which a banner might be flown). The name is compa-
rable to declarations of confidence like ‘Yahweh is my rock, my
fortress, my deliverer…my shield, and the horn of my salvation,
my stronghold’ (Ps. 18.3: cf. Albertz, p. 294), but it expresses more
clearly Yahweh’s presence in the midst of his people’s army and
the need to follow where he leads.10 Early interpreters did not see,
10
The alternative explanations recorded and added to (also for the next verse)
in TWAT 5, 468-69, 471-72 = TDOT 9, pp. 438, 441 are unnecessarily complex
and far-fetched.
534 EXODUS 1–18
or ignored, these military connotations (see Text and Versions),

preferring more general affirmations of Yahweh’s presence to help,
like the memorial stone Eben-ezer (lit. ‘stone of help’), which is
not called an ‘altar’, in 1 Sam. 7.13. But the specific sense fits the
preceding narrative well and probably reflects the contexts in which
it was originally preserved.
16. The further words attributed to Moses revert, whatever the
significance of the difficult opening words may be (see below), to
the theme of v. 14 by placing the battle-narrative in the context of
a long-lasting hostility to Amalek, which is again given a divine
sanction. As such they too do not belong to the core narrative struc-
ture of the unit, and a precise understanding of the final phrase may
be an additional indication that the verse originated in the account
of a conflict much later in Israel’s history. The common translation
‘from generation to generation’ (found already in LXX) is probably
incorrect (see Note u on the translation and the full discussion by
Valentin, Aaron, p. 144 n. 5) and ‘from many generations (ago)’
should be preferred. This is scarcely appropriate as words spoken
after the first account of a conflict with Amalek and would fit much
better in (at the beginning of?) the narrative of a later battle with
Amalek, such as that in 1 Samuel 15, where it would provide an
alternative way of motivating the people to what now appears in
1 Sam. 15.2. ‘He said’ might then originally have referred to a later
leader like Samuel. It is easy to see how such a statement might then
have been appended, anachronistically, to the story of the ancient
conflict itself, particularly if the exact implications of the final
words were already misunderstood.
The opening phrase of v. 16 has been explained in many different
ways, which have increased since the possibility that the trans-
mitted Heb. readings are all corrupt began to be considered in the
seventeenth century (see Text and Versions). Houtman, who gives
a valuable overview of earlier interpretations, has himself added
a new one (see his commentary, pp. 388-91, and more fully, his
article ‘ “Yahweh is my Banner” – “A ‘Hand’ on the ‘Throne’ of
Yh”: Exodus xvii 15b, 16a and their Interpretation’, in A.S. van der
Woude [ed.], New Avenues in the Study of the Old Testament [OTS
25; Leiden, 1989], pp. 110-20). To begin with explanations that start
from the standard Heb. text (MT), it literally means ‘A hand is/has
been on/against the throne(?) of Yah(weh)’ (see further Note t on
17.8-16 535
the translation). This was taken in the Targums to indicate that war
with Amalek was confirmed by an oath made by God, presumably
on the basis of the Heb. idioms in which an oath was signified by
various actions with the hand (cf. 6.8; Deut. 32.40): this interpreta-
tion was followed by e.g. Rashi, Ibn Ezra, AV and RV. In modern
times it has been suggested that the oath was (to be) sworn by Moses
(Dillmann) or Israel (Beer). Others proposed that the hand was to
be raised towards heaven as a gesture of prayer (e.g. Keil). LXX
and Vulg saw the hand as a symbol of divine power and this meta-
phorical interpretation also lies behind the proposal that the ‘hand’
is Amalek’s, whose challenge to Yahweh becomes the basis for the
latter’s declaration of war (so first Reimarus according to Houtman,
art. cit., p. 115 n. 13; more recently Zenger, Israel am Sinai, p. 98).
Houtman understands ‘hand’ in the sense of ‘memorial’ (cf. 1 Sam.
15.12; 2 Sam. 18.18; Isa. 56.5) and ‘throne’ as a reference to the
‘hill’ earlier in the story, which might be Horeb, the mountain of
God (cf. 17.5; 18.5): the words add a further description of the altar
in v. 15 (and have no original connection with the following clause).
These suggestions all have some degree of plausibility but
equally involve some imaginative elaboration of the text, which
raises doubts about their correctness. A wish for a tighter connec-
tion with v. 15, which may of course be misplaced (see above), has
led to the suggestion that the original text read not ‘throne’ (Heb.
kēs) but ‘banner’ (nēs), a difference of only one letter.11 ‘(A) hand on
the banner of Yah!’ has then generally been understood either as an
oath-formula (Baentsch, Gressmann: cf. modern oaths ‘by the flag’)
or as a rallying-cry when an army set out for battle (Noth, Childs).
Less influential views are that the ‘hand’ is Amalek’s, with the same
implications as above (H.-C. Schmitt, ‘Die Geschichte vom Sieg’,
pp. 336-38) and that the reference is to a ‘votive hand’ portrayed
on the banner as an assurance of divine help (R. Gradwohl, ‘Zum
Verständnis von Ex xvii 15f’, VT 12 [1962], pp. 491-94). It can of
course be objected that the reference to an actual ‘banner’ used in
war which the emendation produces is itself not very close to the
metaphorical use of the term in v. 15 and that the emendation has
11
For an extended argument in favour of this view, based on the structure of
(other) aetiologies for the meanings of names conferred on persons or places (esp.
Exod. 2.10), see Childs, pp. 311-12.
536 EXODUS 1–18
no support in any ancient witness to the text. Similar objections can

be and have been made to other attempts to deal with the exegetical
problem by such conjectures (see Text and Versions and Gradwohl,
pp. 492-93; Zenger, Israel am Sinai, p. 96).
No fully satisfactory explanation for this difficult expression has
been found. As suggested in Text and Versions, even the standard
form of the Heb. text (MT) is probably not the best starting-point
for exegesis. It is arguable that here the Samaritan text, which reads
simply ‘Hand upon throne’, is the original reading. Its brevity makes
it even more obscure but perhaps supports the traditional view that
the phrase was an idiomatic (and otherwise unknown: but note Matt.
23.22) way of introducing an oath. The ‘throne’ is unlikely to be
anything other than the throne of God/a god, presumably in heaven
(cf. Isa. 66.1 etc.): Houtman’s suggestion that ‘the hill’ in the story
might be so described is the weakest part of his explanation. But
in view of the uncertainty about both text and interpretation, the
expression cannot contribute anything to the interpretation of the
verse as a whole. Fortunately, if our earlier observations are correct,
it is peripheral to this narrative.
Text and Versions

The three Qumran mss which preserve the beginning of this section all
have a division before it, marked by an open line in 4QExc and4QpalExm
(the latter has ‫ יבא‬at the beginning of a line and presumably placed its usual
enlarged waw in the empty space) and by an open line and an empty line in
4QpalExl.
‫( ויבא עמלק‬17.8) 4QpalExm marked the ayin with scribal dots, which
puzzled DJD (p. 94): perhaps it was an indication that the exemplar was
faulty here. 4QExc spells the name ‫עמלך‬, which is evidently a simple mistake
based on similarity of sound (cf. ‫ [מ]שה‬for ‫ מסה‬in the previous verse): later
in the section its spelling is correct. TgG(AA),Nmg have ‘…those of the house of
Amalek’, which in view of the ‘early’ date of ms. AA will be an older PalTg
reading than TgN’s ‘chieftains’: the latter provides a subject that better fits its
interpretation of the next clause. Ms. AA introduces the section with ‘Targum
for Purim’ in Judaeo-Arabic, a reference to its use as a reading for that festival
because of the connection seen between the Amalekite king Agag in 1 Samuel
15 (cf. Num. 24.7) and ‘Haman…the Agagite’ in Esth. 3.1 etc. TgJ has a longer
expansion noting Amalek’s origin in ‘the land of the South’ (cf. Num. 13.29),
the distance they covered and the traditional hostility between Esau (to which
Amalek belonged: Gen. 36.16) and Jacob: much of this comes from MRI (ed.
Lauterbach, pp. 136-37).
17.8-16 537
‫( וילחם‬17.8) Although this was perhaps intended originally as a ‘summary

introduction’ (see the Explanatory Note), the Vss (other than TgO,J) found it
problematic before the Israelites had selected their warriors and they adjusted
their translations in various ways. Sy lmʿbd qrbʾ took ‫ וילחם‬to express the
intention to fight (cf. OL ms. 104), TgG(AA),N more plausibly envisaged the
drawing up of battle lines and LXX and Vulg used imperfect verbs to portray
an ongoing conflict of which vv. 9-13 were presumably seen as the climax.
‫( עם־ישׂראל‬17.8) LXX had no explicit equivalent to ‫ עם‬since its verb
(πολεμεῖν) could be used with a direct obj. in Hell. Gk. (LSJ, p. 1432). Symm
reverted to the classical idiom by inserting πρός and so created a precise
match to the Heb. (cf. Vulg contra and Sy).
‫( ברפידם‬17.8) MT’s spelling here is curious and inconsistent with the
fuller forms in 17.1 and 19.2 which the best SP mss (contra von Gall) support.
Of the Qumran mss 4QpalExl probably had a yodh between the daleth and
mem, but the mss preserve no further evidence of the orthography. TgJ has a
long addition indicating that Israelites were killed because of their disobedi-
ence, as in MRI (ed. Lauterbach, pp. 135-37, 139), but it specifically refers to
the Danites, whose idolatry was notorious (see refs. in AramB 2, p. 210 n. 9),
presumably on the basis of Judges 18.
‫( אל־יהושׁע‬17.9). ‫ אל‬is also the reading of 4QpalExl and a number of SP mss,
but the majority of the latter (including Tal, all of Crown’s and Camb. 1846)
have ‫ליהושׁע‬. LXX’s τῷ Ἰησοῦ might seem to support this reading, but the use
of the dative for ‫ אל‬in 8.23 shows that it is inconclusive. SP also has ‫ ל‬where
MT has ‫ אל‬in 18.6 and 35.30, but the reverse is true in 6.6 and 8.5. Here SP
may have harmonised the construction with the following verses, where both
traditions have ‫ל‬.
‫( לנו‬17.9) LXX, Sy and one ms. of TgO read ‘for you’, probably a second-
ary reading influenced by the ‘idiomatic’ reflexive dative construction found
with imperatives (Wevers, Notes, p. 268).
‫( אנשׁים‬17.9) LXX and TgN,G added ‘mighty’, using language similar to
the Heb. of 18.21, 25 (cf. MRI [ed. Lauterbach, p. 141]); TgJ has a much more
expansive addition.
‫( וצא‬17.9) TgJ adds ‘from under the clouds of majesty’, i.e. Yahweh’s
protective presence (cf. MRI ibid.).
‫( הלחם‬17.9) LXX and TgJ,N,F render imprecisely ‘draw up in battle-order’,
the fighting itself being still to come (cf. above on ‫ וילחם‬in v. 8). Sy nʿbd (qrbʾ)
is curious: it presumably means ‘we will make war’, but also serves to delay
the actual fighting.
‫( בעמלק‬17.9) TgJ ‘opposite the camp of Amalek’ and TgN,G ‘with those
of the house of Amalek’ (cf. on v. 8) avoid any suggestion that an individual
opponent was involved.
‫( מחר‬17.9) Whether this word belonged with the following or the
preceding clause was one of five expressions in the Pentateuch which were
said to have been listed by Issi ben Judah as being of uncertain interpretation
538 EXODUS 1–18
in this regard (MRI [ed. Lauterbach, p. 142]; cf. B.Yoma 52b). He seems to
have regarded the connection with what follows as the normal view, and this is
what the MT accents imply. But LXX, by its insertion of καὶ ἰδού after αὔριον
(cf. Sy mḥr wʾnʾ hʾ; but 5b1 has wmḥr ʾnʾ) clearly took the opposite view and
so, to judge from the punctuation marks, did most Samaritan scribes. TgN,G
also add ‘behold’ (hʾ) after mḥr, but without an ‘and’, so they like TgO,J and
Vulg preserve the ambiguity of the original.
‫( אנכי נצב‬17.9) LXX ἐγὼ ἕστηκα, since the perf. of ἵστημι has an intran-
sitive present tense meaning, fits the translator’s more common practice in
rendering the Heb. part., even where as here the context indicates a future
sense (cf. 7.17, 27; 19.9; 23.20: in 3.13; 4.23; 8.25 a Gk. fut. is used): for such
a ‘futuristic use of the present’ cf. BDF §322. Tgg and Sy use Aram. parti�-
ciples, which can have a future sense (Stevenson, pp. 56-57; Brockelmann
§211-12); and Vulg ego stabo clearly expresses it. TgN,G render ‫ נצב‬as TgN
rendered ‫ עמד‬in v. 6 (see Text and Versions there) to give the sense ‘standing
ready’ and TgJ attaches to this a mention of fasting and an allegorical interpre-
tation of ‘the top’ and ‘the hill’ as references to the merits of the patriarchs and
matriarchs respectively, which were both found already in MRI (ed. Lauter-
bach, p. 142-43, 145-46); cf. (without the allegory) TgN on v. 12.
‫( ומטה האלהים‬17.9) Tgg have the same periphrasis as in 4.20 (see Text and
Versions there; also MRI [ed. Lauterbach, p. 142]), but LXX renders straight-
forwardly this time.
‫( בידי‬17.9) Some later mss of Sy (12a1fam) mark this as a pl. form;
compare the note on ‫ ידו‬in v. 11.
‫( משׁה‬17.10)1o TgN adds ‫רביה‬, ‘his master/teacher’, in line with the
frequent designation (e.g. 24.13) of Joshua as Moses’ ‫( משׁרת‬cf. MRI [ed.
Lauterbach, pp. 140-41] on v. 9).
‫( להלחם‬17.10) LXX and Vulg smoothed the connection by rendering with
a finite verb; similarly Sy (though not 5b1), by prefixing wʾzl. LXX and TgN,G
continued even here with ‘drew up battle lines’, but TgJ switched its loyalties
from the PalTg here and followed TgO’s literal rendering.
‫( בעמלק‬17.10) As in v. 9 TgN,G have ‘with those of the house of Amalek’.
‫( אהרן‬17.10) SP prefixes a waw (cf. its variant readings in 1.2-4), and the
Vss other than TgO also add a connective here. No Qumran ms. preserves this
word, but the shorter reading of MT is to be preferred.
‫( ראשׁ הגבעה‬17.10) SP prefixes ‫ על( אל‬according to Sadaqa, probably from
v. 9) to clarify the connection: all the Vss except TgG add a preposition too,
but even more clearly here this was probably dictated by the requirements of
the target languages. MT’s reading is idiomatic (see Note h on the translation)
and superior: again there is no evidence from Qumran.
‫( והיה‬17.11) All the Heb. verbs in the verse represent repeated actions,
including this one (cf. Note i on the translation), and the Tgg convey this well
by the use of Aram. participles. In this case, as often with such introductory
formulae, Sy and Vulg have no equivalent. Rahlfs and Wevers disagree about
the reading of LXX here, the former preferring the imperfect ἐγίνετο of mss
17.8-16 539
ABF and the latter the less well attested aorist ἐγένετο, to which there is a
close parallel in the similar 33.7 (cf. Notes, p. 269 and [better] THGE, p. 226).
LXX’s handling of the tenses later in the verse (see below) is inconsistent,
but this may be due to a growing tendency to favour the Greek aorist over
the imperfect in Koine and in LXX (see Evans, Verbal Syntax, pp. 120-21,
138-40, 198-219).
‫( ירים‬17.11) LXX has the aorist ἐπῆρεν for the Heb. imperfect, Vulg
more precisely the Latin imperfect subjunctive. TgG(AA) ‫ תקף‬is a scribal error
for the ‫ זקיף‬of the other PalTg texts (Klein 1, p. 67).
ֹ‫( יָ דו‬17.11) SP has the dual form ‫ ידיו‬both times in this verse, as the
following context seems to require (cf. v. 12), and the Vss and MRI (ed.
Lauterbach, pp. 143-44) all agree. 4QExc preserves the first occurrence and
4QpalExm the second, both reading ‫ידו‬. At Qumran ‫ו‬- sometimes represents
the suffix of a pl. or dual noun (Qimron §322.14; Reymond, Qumran Hebrew,
pp. 144-47, 159), as indeed it can in BH (like early Heb. inscriptions: cf. JM
§94d n. 2): so in 4QpalExm at 21.6 (cf. Sanderson, Exodus Scroll, pp. 57-58;
DJD IX, p. 103). The sing. vocalisation of MT thus appears isolated and
questionable (cf. Propp, p. 614). One might explain its origin, or its preserva-
tion, from the use of the sing. in v. 9, which recalls passages earlier in Exodus
where Moses holds a staff in his hand (e.g. 4.17, 20; 7.15; 14.16; 17.5). But
it could be ancient, as a prima facie understanding of the consonants in both
MT and the Qumran mss would suggest: and the conflict with v. 12 makes it
the more difficult and so the more likely reading to be original. In any case
the tension between sing. and dual in the passage as a whole is likely to have
been created by the merging of two different traditions about Moses’ action
in this story (see further the introduction to the section). A reflection of this
duality can still be seen in the Jewish interpretations: in TgJ,N,G(AA),F additions
of varying length see it as a gesture of prayer (which ends when Moses lowers
his hands), while in MRI (ibid.) the anxiety to avoid a magical motive leads
to a focus either on the people’s faith when they see it or on their obedience.
Among modern vernacular versions Luther, Tyndale, AV, RV, RSV, JPS, EÜ,
NRSV and ESV have the sing. of MT, and Douai/Rheims, JB, NEB, NIV and
REB the pl.
‫( וגבר‬17.11)1o LXX and Vulg both use the classical imperfect tenses
here to indicate the continuing alternation of supremacy between Israel and
Amalek. TgN,G(AA),F add ‘and were victorious’ to contrast with their portrayal
of the Amalekites later in the verse. TgF(P) also adds ‘in the battle lines’ here,
as TgF(VN) does at the end of the verse, and the puzzling [?‫ בחר]בא‬in TgNmg is
most likely to have been meant as an addition here to parallel (by contrast) the
use of the same expression at the end of the verse.
‫( ישׂראל‬17.11) Here, as with ‫ עמלק‬later in the verse, TgO prefixes ‘those of
the house of’, as in other Tgg in vv. 9-10 and here (‘all those…’ in TgG(AA)).
‫( וכאשׁר יניח ידו‬17.11) On ‫ ידו‬see the note above on its first occurrence
in the verse. LXX καθῆκεν again uses the aorist and Vulg remisisset (with
paululum added on the assumption that Moses would quickly regain his
540 EXODUS 1–18
strength) the pluperfect subjunctive (as in Augustan and later Latin for
repeated action: e.g. Livy 1.31.4; 1.32.13).
‫( וגבר עמלק‬17.11) TgN,G,F add ‫ונפלין בחרבא‬, ‘but they would fall by the
sword’, so that although the Amalekites were on top they suffered casualties;
TgG(AA) makes their situation worse by rendering ‫ וגבר‬not by ‫מתגברין‬, ‘would
prevail’, but (with the probably deliberate omission of one letter) by ‫מתברין‬
‘would be broken, defeated’. On this ‘converse translation’ to remove a
possibly ill-omened statement see Michael Klein on the Targums, pp. 194-95,
and Text and Versions on 12.33.
‫( וידי משׁה כבדים‬17.12) TgO,J, LXX and Vulg render literally ‘heavy’,
presumably with the sense ‘tired, weary’ (Sy) in mind. TgJ has a long addition,
attributing Moses’ problem to his delay in initiating the battle (cf. v. 9 and MRI
[ed. Lauterbach, p. 145]) and the subsequent provision of support to his need
to fast (for which see again below and the note above on v. 9). This contrasts
with TgN,F,G(AA,FF), i.e. the main PalTg witnesses, which replace ‘heavy’ with
‘were lifted up (in prayer)’. TgG(FF) in fact has ‘were spread out in prayer’, like
TgO at the end of the verse, and this whole group seem to have reconstructed
the beginning of the verse on the basis of what they have at its end and so
avoided any reference to Moses’ weakness.
‫( תמכו‬17.12) Both LXX by its apt imperfect (cf. Wevers, Notes, p. 270)
and TgO by its participle indicate that the support continued for some time.
‫( בידיו‬17.12) TgG(FF) and TgNmg specify that both Moses’ hands were sup-
ported. The early SP ms. Camb. 1846 and two later mss in von Gall read ‫בידו‬,
an easy error in view of the Samaritan pronunciation of ‫יו‬- as o (GSH §45aβ).
‫( ויהי‬17.12) The sing. form of MT, although defensible and probably
original (cf. Note k on the translation), was naturally rendered in the pl. in
most of the Vss and the Heb. pl. was also preferred in 4QExc, 4QpalExm and
SP (4QpalExl does not survive for this verse) as the more ‘regular’ form. Only
Vulg presents the sing., as a correction of OL, but in an artificial periphrastic
form: et factum est ut (manus ipsius non lassarentur).
‫( ידיו‬17.12) LXX mirrors, unnecessarily, the fuller expression αἱ χεῖρες
Μωυσῆ at the beginning of the verse.
‫( אמונה‬17.12) LXX (ἐστηριγμέναι), Vulg (see the note on ‫ )ויהי‬and Sy
(bhymnwtʾ) found different ways to render the unfamiliar Heb. idiom (on
which see Note l on the translation). The Tgg saw here a further allusion to
Moses’ prayer, ranging from TgO’s simple ‘(were) spread out in prayer (cf.
TgNmg), through TgJ’s addition to this of Moses’ ‘faithfulness’ and ‘fasting’
and TgG(AA)’s brief mention of the ‘faithfulness’ of the patriarchs, to the very
elaborate development of this latter theme (including the matriarchs as well)
in TgN,F(P) (cf. TgJ on v. 9 and MRI both there and here).
‫( ויחלשׁ‬17.13) The rare word is preserved in 4QExc and SP (but not
4QpalExm, which is fragmentary here) and the Vss all render, as the context
requires, with words for ‘put to flight’ (LXX ἐτρέψατο, Vulg fugavit) or
‘defeat’ (TgO,J, Sy tbr; TgN,G ‫)שיצא‬: cf. their renderings (except for Sy ḥlšʾ =
17.8-16 541
‘weak ones’) of ‫ חלושׁה‬in 32.18 and TgJ ‫ קטיל‬and Vulg qui vulnerabas in Isa.
14.12.12 In Joel 4.10 and Job 14.10 the equivalents are related to the alternative
sense ‘be weak’, while at 32.18 TgO,J ‫ חלשׁין דמתברין‬combines the two senses.
‫( ואת־עמו‬17.13) By what can only be a coincidence these words were
initially omitted by two representatives of the ‘Samaritan’ textual tradition
and then restored, in the SP ms. Camb. 1846 immediately by the original
scribe and in 4QpalExm above the line by a different scribe. In fact the phrase
may be an ignorant addition by an early scribe – it was known to LXX (see
below) and 4QExc – who thought that ‫ עמלק‬was an individual (only elsewhere
in Gen. 36.12 [and 16?]; 1 Chr. 1.36): so Valentin, Aaron, pp. 161-62. After-
wards 4QpalExm has the addition ‫ויכם‬, which subsequently appeared in all SP
mss and forms a more familiar combination with ‫( לפי־חרב‬cf. Num. 21.24;
Deut. 13.16; 20.13; Josh. 10.28). The shorter and unusual text of MT and the
other witnesses (including 4QExc) must, however, be more original. LXX
καὶ πάντα τὸν λαὸν αὐτοῦ makes a typical modification (14.6 is especially
similar); TgJ ‘for he cut off the heads of the warriors of his people’, follows
MRI (ed. Lauterbach, p. 146), which was grappling with the sudden distinction
between Amalek and ‘his people’ (cf. isolated mss of Sy and OL; also NJPS).
‫( לפי־חרב‬17.13) The text is not in doubt (cf. SP, 4QpalExm, 4QExc, Sy,
Vulg), but a variety of interpretations are preserved in the Vss. LXX ἐν φόνῳ
μαχαίρας (only elsewhere in Num. 21.24; Deut. 13.16; 20.13) probably
thought that ‘slaughter’ was appropriate to the arch-enemy Amalek. Tgg took
‫ פי‬as a metonym for ‘word’, as it clearly is in v. 1 and elsewhere: ‘the word
(‫ )פתגם‬of the sword’ (TgO,N,G) might mean ‘the law of war’ (Jastrow, p. 1250).
But TgJ, following MRI again (p. 147), broke up the phrase to introduce a
reference to Yahweh’s word (TgNmg is similar): ‘according to the mouth of the
Memra of the Lord by killing with the sword’.
4QExc divides the text with an open line before the divine speech in v. 14
(cf. MT, SP), but 4QpalExm could only have had a short break here, if that (cf.
DJD IX, p. 94).
‫( יהוה‬17.14) TgG,Nmg prefix ‘the Memra of’.
‫( זאת זכרון‬17.14) LXX (εἰς) and Vulg (ob, here indicating purpose) insert
a preposition to indicate that ‫ זאת‬is a pronoun. TgO ‫ דא דוכרנא‬will intend the
same (Stevenson §5.10) and the other Tgg may do so. Sy (except 5b1 and one
other ms.) dwkrnʾ hnʾ, however, took ‫ זאת‬to be an attributive adjective, ‘this
reminder’. TgN,G add ‫טב‬: see Text and Versions on 12.14 and AramB 2, p. 48
n. 12 and p. 75 n. 13.
12
According to R. Eliezer (as cited in MRI [Lauterbach, p. 147]) the word’s
root letters pointed to the senses ‘made sick’ (‫)ויחל‬. ‘made to tremble’ (‫ויזע‬: perhaps
a contextual interpretation of ‫ )ויחל‬and ‘crushed’ (‫)וישׁבר‬: the latter being the Heb.
equivalent of Aram. ‫ תבר‬and perhaps an indirect testimony to the reading of TgO,J
and Sy.
542 EXODUS 1–18
‫( בספר‬17.14) LXX ἐν βιβλίῳ and TgG ‫ בגו ספר‬clearly identify the
indefinite meaning: the other Vss are ambiguous.
‫( ושׂים באזני יהושׁע‬17.14) On LXX δός for ‫ שׂים‬cf. Wevers, Notes, p. 271: no
difference in Vorlage is involved. For ‫ באזני‬TgO and Sy have ‫קדם‬, which might
seem to mean that the written text was to be placed ‘before’ Joshua. But both
Vss use ‫ קדם‬for ‫ באזני‬with verbs of saying even with human hearers (cf. Gen.
23.16; 50.4; Exod. 11.2), so no difference in meaning from the Heb. need be
intended. TgN,G,J follow the Heb. expression more closely. Sy adds ‘the son of
Nun’ here (but not in vv. 9-10, 13).
‫( כי‬17.14) Vulg and Sy took ‫ כי‬in a causal sense; but the other Vss at least
can be taken to mean ‘that’ (so Wevers, Notes, p. 271 for LXX ὁτι and AramB
ad loc. for Tgg ‫ארום‬/‫)ארי‬.
‫( מחה אמחה‬17.14) TgN,G use the general word ‫שׁיצי‬, ‘destroy’, which they
already used for ‫ חלשׁ‬in v. 13, in line with its wider use in PalTg compared
with TgO,J. In TgNmg a second person pl. form appears: the wording is identical
to TgN at Deut. 25.19 and may be more of a cross-reference than a textual
variant. As often elsewhere Vulg has no separate equivalent to the inf. abs.
(cf. 13.19; 15.26).
‫( מתחת השׁמים‬17.14) LXX ἐκ τῆς ὑπὸ τὸν οὐρανόν understands γῆς and
points to the intended sense of the idiom (Wevers, Notes, p. 272).
‫( מזבח‬17.15) LXX added κυρίῳ, but there is no support elsewhere for this
accommodation to the pattern in Gen. 8.20; 12.7-8; 13.18, and Gen. 26.25
provides a parallel to the shorter expression of the Heb. here.
‫( ויקרא שׁמו‬17.15) LXX, Vulg, Sy and TgJ translate straightforwardly,
and the same understanding is presupposed in both interpretations in MRI
(ed. Lauterbach, pp. 159-60). But TgO,N,G(AA,J) render the text (with minor
variations) as if it read (‫ויקרא בשׁם )יהוה‬, as in Gen. 12.8b, apparently as a way
to avoid too close an association of ‫ יהוה‬with ‫( נסי‬which might also be what
is intended by the paseq in MT): see the next note for a different approach to
achieving the same end.
‫( יהוה נסי‬17.15) The original reading of 4QpalExm was simply ‫יהוה( נסי‬
was added by a later scribe). There is no obvious reason for accidental
omission of ‫יהוה‬, but 4QpalExm probably made a similar slip (also corrected)
in v. 13, which weakens the case for regarding ‫נסי‬, without ‫יהוה‬, as the most
ancient reading. In such namings it is normal for a divine name or title to be
included (cf. Gen. 22.14; 28.19; 33.20). None of the Vss exhibits the deriva-
tion of ‫ נסי‬from ‫נֵ ס‬, ‘banner, ensign’, which is generally accepted today as
being the obvious sense offered by the BH lexicon.13 Perhaps it was ruled out
as being impossible, even idolatrous, as a designation of God: Rashbam seems
13
On the readings of the Vss here and in v. 16 see also Houtman, ‘ “Yahweh
is my Banner’”, pp. 111-14. An exception might be SamTgJ ‫נצועה‬, i.e. ‫= נצוחה‬
‘victory’. But the related verb is used for ‫ חלשׁ‬in v. 13 and the translation may be
based on that.
17.8-16 543
to have been the first to propose it. LXX μου καταφυγή (for the inverted
word-order cf. 15.2) related ‫ נסי‬to ‫נוס‬, ‘flee’, instead (cf. Fritsch, p. 37, and the
use of καταφυγή for ‫ מנוס‬in 2 Sam. 22.3; Jer. 16.19; Ps. 59.16), Vulg exaltatio
mea probably to √‫נשׂא‬, ‘lift up’. The Aram. Vss all saw here the MH/Aram.
meaning ‘sign, miracle’ (cf. MRI) and this was still taken for granted by Rashi
and Ibn Ezra. Tgg other than TgJ, having taken ‫ יהוה‬as explicating the suffix of
‫שׁמו‬, read ‫ נסי‬as an attribute (a part.?) meaning ‘who did miracles for him [sc.
Moses]’. TgJ (cf. Sy) was not so encumbered and continued to be guided by
MRI, most likely to the sense ‘The Memra of the Lord is my miracle’, which
it then explained with ‘because the miracle which the Place (cf. ‫ המקום‬in MRI
here) performed was for my sake’ (AramB 2, p. 211 and nn. 23-24: it is less
likely that TgJ followed the view [attributed to R. Eleazar of Modiʿim] that
God was taken as the subject of ‫ ויקרא‬and said ‘my miracle’ with reference
to himself).
‫( ויאמר‬17.16) LXX has no equivalent, and so links its distinctive rendering
of the verse directly to v. 15 by ὅτι (= ‫)כי‬, ‘because’. The word is present in
4QpalExm and SP as well as MT and is assumed in all the other Vss: LXX
probably omitted it as being unnecessary, since Moses continues to speak.
‫( כי‬17.16) So also SP and 4QpalExm. Vulg quia like LXX took ‫ כי‬to be
causal, and this may also be the intention of TgJ’s ‫ארום‬. TgO,N,G,F have no equiv-
alent, implying a recitative understanding of ‫כי‬. Sy hʾ, ‘behold’, like SamTgJ
‫הלא‬, envisages an ‘emphatic’ use. Finally R. Joshua in MRI (Lauterbach,
p. 160) has )‫לכש(ישב‬, ‘when’ (cf. Segal §513).
‫( יד על כס יה‬17.16) At Qumran 4QpalExm has ‫ יד‬after ‫ כי‬and then a lacuna,
while 4QExc preserves only the ‫ ל‬of ‫על‬, but with too little space between it and
‫ מלחמה‬for MT’s reading: it may, like SP, have read ‫( יד על כסא‬DJD XII, p. 122),
which Sy ʾydʾ ʿl kwrsyʾ also presupposes.14 LXX ἐν χειρὶ κρυφαίᾳ implies
a Vorlage with ‫ יד‬followed by ‫ כסיה‬as a single word, which was probably
read as the pass. part. of ‫( כסה‬cf. GK §75c): ‫ על‬was either absent, ignored or
possibly transposed (Propp, p. 615: cf. ἐν). The Three, Syh and other witnesses
to the O-text add κυρίου to represent the separate word ‫ יה‬of MT. Vulg manus
solii domini took the correction towards MT further, but still ignored ‫על‬.
The phrase was taken, with et added after it, as a parallel subject to ‫ מלחמה‬of
the ‘will be’ which is understood. Early Jewish interpretation (Tgg and MRI
[Lauterbach, pp. 160-61]) mainly saw the phrase as defining the following
words as an oath spoken (in TgJ by his Memra) from the throne of God, with
MT’s reading being presupposed and ‫ יד‬taken as a cipher for an oath because
of its association with swearing in BH (see the Explanatory Note).15 In modern
14
Ephrem’s text read dyh after ʾydʾ, which Weitzman (Syriac Version, p. 289)
thought was the original Sy reading, based on a Vorlage like MT.
15
Two possibly related exceptions are ἡ δύναμις ἐπὶ θρόνον ἄχραντον
(LXXFb), where δύναμις is used for ‫ יד‬in the sense of ‘power’ and ἄχραντον,
‘undefiled’ is (also) a substitute for God’s name (like ‫ יד‬in TgO,J); and R. Joshua’s
544 EXODUS 1–18
times AV and RV adopted the ‘oath’ interpretation, but the emendation of ‫ֵכּס‬
to ‫נֵ ס‬, first suggested by J. Clericus in his Pentateuchus sive Mosis prophetae
libri quinque. 2. Exodus-Deuteronomium (Amsterdam, 1696), p. 77, makes
the words into a battle-cry and has had widespread support (so still e.g. Childs,
pp. 311-12; HAL, p. 465; Propp, p. 620).16 Other suggestions are to replace the
unparalleled ‫ ֵכּס‬with the regular form )‫( ִכּ ֵסּ(א‬Dillmann, Beer) or to read *‫ֶכּ ֶסא‬
= ‘buttock’ (cf. Ar. kusʾun = ‘back part’) as part of an oath formula comparable
to Gen. 24.2, 9 (NEB, REB; Brockington, HTOT, p. 11; D.W. Thomas, cited in
DCH 4, p. 603). But each proposal has its weakness: to read ‫ נֵ ס‬is not only an
unsupported conjecture but one which confuses Moses’ (or Israel’s) ‫נֵ ס‬, which
is Yahweh, with Yahweh’s own ‫ ;נֵ ס‬to ‘restore’ the regular form disregards the
readings of LXX, Sy and SP which know nothing of the ‫ יה‬of MT; while Ar.
kusʾun is not attested in the specific sense of ‘buttock’, which was probably
a product of G.R. Driver’s fertile imagination. The reading which apparently
underlies MT, LXX, Sy and SP (and perhaps 4QExc) is the variant spelling
of ‫ כסא‬as ‫( כסה‬1 Kgs 10.19 [2x]; Job 26.9 [here most likely a misreading of
an original ‫‘ = ֶכּ ֶסא‬full moon’]), which when written plene as ‫ כסיה‬could give
rise to the interpretations of LXX and MT (the latter with division into two
words).17 SP (cf. Sy) simply regularised the spelling by replacing he with
aleph. On the meaning see the Explanatory Note.
‫( מלחמה ליהוה‬17.16) There is no serious doubt about the original text,
on which MT, SP, 4QExc and 4QpalExm agree, with close support from Sy
and Vulg. The rephrasing of LXX (πολεμεῖ κύριος) and the more elaborate
paraphrases and expansions of Tgg can readily be seen to be based on the
same Heb. wording. TgO adds little to MT but characteristically avoids direct
intervention of God in human affairs: ‘(that) it is determined that war shall be
waged before the Lord…’ TgJ simply makes God’s Memra the agent for his
making war. The PalTg, whose versions closely agree, go on, without denying
God’s own involvement, to introduce the exploits of some human intermedi-
aries, first Saul (1 Sam. 15) and then Mordecai and Esther (Esth. 7–9: because
Haman ‘the Agagite’ [3.1, 10; 8.3, 5; 9.24] was regarded as an Amalekite).18
‘When the Holy One, blessed be He, will sit upon the throne of his kingdom
and his reign will prevail…’ (MRI, ibid.): cf. SamTgJ’s non-eschatological variant
‫הלא אתר על כרסי‬, where ‫( אתר‬cf. Heb. ‫ )המקום‬stands for ‫ יד‬and must be another
periphrasis for ‘God’.
16
To harmonise the other way by reading ‫ ִכּ ְס ִאי‬for ‫ נִ ִסּי‬in v. 15 (so Van Seters,
Life, p. 206) is much rarer and was rightly rejected by Noth (p. 115, ET, pp. 143-
44).
17
So it is not the case that ‘‫ יה‬is removed…in the Greek translation’ to avoid an
anthropomorphism (Fritsch, pp. 25-26). According to C.D. Ginsburg (Introduction
to the Hebrew Bible [London, 1897], pp. 382-83) medieval Palestinian scribes
read ‫ כסיה‬as a single word.
18
R. Eliezer (MRI, ibid.) has the war extending from Moses to David (2 Sam.
1.13); see also the note below on ‫מדר דר‬.
17.8-16 545
‫( בעמלק‬17.16) Again there is no doubt about the text: but Tgg amplify to
‘with (those of) the house of Amalek’.
‫( מדר דר‬17.16) There is considerable variation in detail among the
witnesses, but it is probably all due to dissatisfaction with the expression or
the apparent meaning of MT’s reading rather than to an older variant reading.
For this very reason MT is most likely to be original (on its meaning see
Note u on the translation). SP ‫ מדר ודר‬introduces a waw, as was becoming the
preferred idiom already in LBH (cf. GK §123c) and Sy mn dr dryn follows
an Aram. idiom for totality. 4QpalExm just replaces )‫ מ(ן‬with ‫ עד‬to make the
meaning explicitly future (cf. TgN etc. with ‫)לדר דרין‬. LXX ἀπὸ γενεῶν εἰς
γενεάς and, closer to the Heb., Vulg a generatione in generationem are
content to attribute a future sense only to the second ‫דר‬. In TgO,J the problem-
atic ‫ מן‬is understood not temporally but of separation and a form of the verb
‫שׁיצי‬, ‘destroy’, is prefixed to indicate this (cf. the specific interpretations of
the phrase in MRI, ibid., which TgJ combines into a pastiche with a threefold
destruction).
Both 4QpalExm (mid-line) and 4QExc (end of line) seem to have had a
short division before 18.1 (DJD IX, p. 96; XII, p. 122).
C h ap t er 1 8 . 1 - 1 2
Th e C om i n g of J et h r o
an d Fu rth e r C eleb r at i on of t he Exodus
The new section is marked at the beginning by divisions in MT,

SP, 4QpalExm and 4QExc, which correspond to the major change
in subject-matter and the (re-)introduction of Jethro as a leading
participant in the narrative. In MT there is no break after v. 12,
presumably because Jethro continues to play a central role in the
following section. But SP and probably 4QpalExm (see Text and
Versions) did have a division here to mark the fresh topic that is
introduced in v. 13 and the time-lapse before it. Within vv. 1-12 too
SP (after v. 7) and 4QpalExm (after vv. 2 and 11) make sub-divisions
in the narrative, while MT treats it as a single whole. So does 4QExc
apparently, but its preserved text ends within v. 12 and it is impos-
sible to be sure whether it had a break before v. 13 or not.
The section presents a continuous narrative in eight stages,
but it also contains numerous ‘flashbacks’, both to Moses’ family
history (which is also kept in view by the frequent designation of
Jethro as Moses’ father-in-law: see the Explanatory Note on v. 1)
and to the Exodus story itself. (i) Jethro, who is here also called
‘the priest of Midian’, hears the news of Israel’s deliverance from
Egypt (v. 1); (ii) he sets out with Moses’ wife and children, of
whom further details are given, some recapitulated from ch. 2 and
others introduced here for the first time (vv. 2-4); (iii) on arrival at
‘the mountain of God’, where Moses now is, Jethro informs him of
his family’s presence (vv. 5-6); (iv) Moses comes out, greets him
in the traditional way and takes him into his tent (v. 7); (v) Moses
recounts the story of the Exodus and subsequent events to Jethro
(v. 8); (vi) Jethro rejoices at Yahweh’s goodness to Israel in deliv-
ering them from Egyptian rule (v. 9); (vii) still speaking to Moses
(cf. ‘you’) he praises Yahweh and acknowledges his superiority
over all other gods (vv. 10-11); (viii) Jethro brings offerings to
God and is the focal point of the sacrificial meal that follows, in
which Aaron and the elders of Israel also participate (v. 12). It is
remarkable that after Jethro has brought Moses’ family to him, the
18.1-12 547
narrative says no more of them but turns to focus on Jethro himself,

the celebration of the Exodus and the worship of Yahweh. But it is
not a total surprise, for already in v. 1 it is the knowledge (presum-
ably without all the details) of the Exodus that has prompted Jethro
to set out on his journey and the deliverance has been presented as
deriving from the intervention of Yahweh. The fact that Yahweh is
sometimes (vv. 1, 3, 12) referred to here as ‘God’ (Heb. ʾelōhîm), as
he is throughout vv. 13-27, has traditionally been attributed to the
fact that Jethro is a non-Israelite, but in vv. 1-12 at least the occur-
rences are all in the narrator’s words and should be related to his
standpoint rather than Jethro’s.
In modern critical study of Exodus this use of ‘God’ by the narrator was,
as elsewhere, from early on seen as a clue to the older source used by the
compiler at this point. Since the passage contains no distinctive marks of P
(the mention of sacrifices as such not being exclusive to P) and P has in any
case already ceased to use ‘God’ as a title for Yahweh, it was attributed by
Knobel (Num.-Jos., pp. 532-33) to his Rechtsbuch (the E of later scholars),
except for v. 2b, a parenthetical gloss to harmonise this passage with the alter-
native account in 4.20, 24-26, which stated that Moses had taken his family
with him when he returned to Egypt from Midian. The presence of an extract
from E in this passage was soon taken up by Wellhausen (Composition,
pp. 80-81) and Dillmann (pp. 184-85) and was practically unchallenged until
the upheavals in Pentateuchal scholarship in the 1970s: a distinction from the
alternative account (of J) was also seen in the name given to Moses’ father-
in-law (not Reuel, as in 2.18 and Num. 10.29) and in the number of Moses’
children (two, not one as in 2.22 and 4.24-26). But only a few scholars (e.g.
Beer, Hyatt and more recently Propp) attributed the whole section to E.
The occurrences of the name Yahweh alongside ‘God’, often in repetitious
verses (vv. 1, 8, 9, 10), and for some scholars vocabulary arguments as well
were taken to indicate the work of a second author. The more popular of
two explanations for a time was that a J version had been drawn on by the
compiler too (Dillmann, Carpenter/Harford-Battersby, Holzinger, Smend (J2),
Gressmann, McNeile, Eissfeldt, Fohrer [Einleitung11, p. 167]). For example,
Gressmann attributed vv. 1b, 6, 7, 8bβ, 9a, 10a and 11 to J and most of the
rest to E (Anfänge, pp. 86-90). But agreement about the analysis proved very
difficult to reach and according to some scholars several verses were the work
of a redactor, so that even less remained of each of the original versions of
the story. As a result opinion shifted to the view that a single account (attrib-
uted to E) had been modified by the addition of ‘Yahwistic’ sections (so
already Wellhausen, who was followed by Baentsch [pp. 162-65] and then
much later by Noth [pp. 117-18, ET, pp. 146-47; ÜGP, p. 39], Fritz, Childs,
Jenks [pp. 44-45], Schmidt [EdF, pp. 115-16], Houtman [p. 397] and Baden
548 EXODUS 1–18
[Composition, p. 121 with n. 92]). The number and extent of these additions
varied from scholar to scholar: Graupner’s analysis (Elohist, pp. 95-100: cf.
C. Frevel, ‘ “Jetzt habe ich erkannt, dass YHWH grosser ist als alle Götter”.
Exodus 18 und seine kompositionsgeschichtliche Stellung im Pentateuch’,
BZ 47 [2003], pp. 3-22) is perhaps the most thoroughgoing and it leaves only
vv. 1a, 2a, 3abα, 4a, 5-7, 8aαbα, 12 (minus the burnt offering and Aaron) for
the original account.
Rudolph in 1938 had already challenged the fine distinctions which
underlay both these kinds of analysis and argued for the essential unity of the
passage: only vv. 3b-4 were intrusive, having originally appeared in toto in
2.22 (Elohist, pp. 37-39). Since he had already satisfied himself that there were
no divergent traditions in chs. 2–4 and he found (or made) reasons for thinking
that all the occurrences of ‘God’ here had been substituted for ‘Yahweh’ (he
did not explain why), he had no problem in attributing the whole section to J
(compare his similar tour de force on vv. 13-27). Van Seters came to the same
conclusion (Life, pp. 205-12), with the exception of vv. 2b and 3b-4 (p. 210).
The idea that the section is essentially a unity has been taken up by Blum (see
his argument in Studien, p. 158 n. 253), his student V. Haarmann (YHWH-
Verehrer, pp. 72-75) and Albertz (pp. 299-301). Blum initially thought that,
while the passage was inserted where it is at a very late stage of the composi-
tion of the Pentateuch, it came from an old biography of Moses which had also
been used by Kd in chs. 3–4 (pp. 155-56). But he now attributes the composi-
tion too of 18.1-12 to a post-Priestly redactional layer (cf. his ‘Literarische
Verbindung’, pp. 127-30, 136-37), a conclusion which Haarmann and Albertz
have followed. It was also the view of K. Schmid (Erzväter, pp. 235, 252
n. 261) and Kratz (Komposition, pp. 153, 247, 300-301).
On the other hand Levin has found a place for the passage (but only a core
consisting of vv. 5*, 7-8a, 10aαb, 11a, 27: the rest being later additions of
an indeterminate origin) in the narrative of his exilic Yahwist (Der Jahwist,
pp. 359-61), where it offered a positive testimony from an outsider to
Yahweh’s deliverance of Israel, comparable to those of Abimelech, Laban
and others (ibid., pp. 201-202). This implies a similar historical setting to that
envisaged for the story in Van Seters’s Yahwist, Johnstone’s D-work (Chroni-
cles and Exodus, pp. 257-58) and in Dozeman’s Non-P History (pp. 361-62,
400), but they treat it as a unity and do not analyse its origins in detail (see
below on Johnstone’s advocacy of the ‘transposition’ theory; Dozeman allows
for a basis in ‘a independent tradition of worship at the mountain of God’).
Alongside these literary analyses, indirect support for a pre-exilic date for
the passage has come from a number of studies of vv. 13-26 in the context
of Israelite legal history and particularly the history of judicial institutions
(cf. Knierim, ‘Exodus 18’ [see also ZAW 77 (1965), pp. 29-30]; Crüsemann,
Die Tora, pp. 96-121, ET, pp. 76-98; Jackson, Wisdom-Laws, pp. 422-30; see
further his ‘Law in the Ninth Century’). Although the conclusions proposed
apply in the first place to the second half of Exodus 18 (and they will be more
fully dealt with in the next section of the commentary), the close connections
18.1-12 549
between its two parts both at the ‘join’ in v. 13 (and also v. 27) and in
numerous shared features, including the use of the title ‘God’ (Heb. ʾelōhîm),
make it highly probable that at least the nucleus of each part comes from the
same literary and historical setting. In particular vv. 13-26 would ‘hang in the
air’ without an introduction that brought Jethro (who is not even named in vv.
13-26) back into the narrative after his long absence from it, so that vv. 1-12
can be presumed to go back to a time which is at least as early as that which is
presupposed in vv. 13-26. The studies referred to have all argued that the latter
passage reflects changes which took place in the monarchy period, when older
means of settling legal disputes were replaced by a two-level system of courts.
It is evident that an older debate about the unity and composi-

tion of this section of narrative has been largely, but not completely,
superseded by one about its date of origin. The implications of its
connections with passages earlier in Exodus have continued to play
an important part in scholarly discussion, though sometimes with
very different results. We turn now to an evaluation of the main
arguments that have been put forward about these issues. The over-
all coherence of the narrative is a strong argument in favour of its
unity. The disappearance of Moses’ wife and children after v. 6 and
the worship of Yahweh by Jethro at the end are certainly surprising,
but not sufficient to constitute arguments against this. The varia-
tion between ‘God’ and ‘Yahweh’ is unusual, but it is notable that
‘Yahweh’ always occurs when the Exodus is being referred to and
that the statement about his superiority to other gods in v. 11 could not
have been made without using his name. It is only because ‘Yahweh’
sometimes appears in verses which seem repetitive that the possibil-
ity of ‘Yahwistic additions’ gained any plausibility. However, ‘God’
and ‘Yahweh’ appear together only in v. 1 and in v. 8 ‘Yahweh’
appears in both halves of the verse. While some amplification of
the wording might have occurred (especially in v. 10), it is shown in
the Explanatory Notes on vv. 1, 6-8 and 9-11 that grammatical and
stylistic factors may be responsible for much of the character of the
present text (see also Haarmann, YHWH-Verehrer, pp. 72-75).
A different kind of question is raised by vv. 2-4 and their relation-
ship to the earlier references to Moses’ marriage and family in chs.
2–4. Much recent writing on these chapters has been influenced
by newer work on them by Blum and K. Schmid (on which see
the introductions to 3.1-12, 3.13-15, 3.16-22, 4.1-9, 4.10-17 and
4.18-31). Blum initially saw 3.1–4.18 as a largely unified account of
Moses’ commission which was inserted by the author of Kd into an
550 EXODUS 1–18
older narrative which originally proceeded directly from 2.23aα to

4.19 (Studien, pp. 20-28). Subsequently, following the proposal of
Schmid and others that both 3.1–4.18 and 4.27-31 were post-Priestly
in origin, Blum maintained his earlier view for 3.1-22 but accepted
that the whole of 4.1-17 and 4.27-31 was post-Priestly in origin (cf.
‘Literarische Verbindung’, pp. 123-34). Some similarities between
4.27-28 and Jethro’s meeting with Moses in Exod. 18.1-12 have
led Blum (ibid., pp. 136-37) and more recently Haarmann (YHWH-
Verehrer, pp. 75-77) to speak of them as Parallelszenen and to
deduce a similarly late date for 18.1-12. These theses about Exodus
3–4 have been discussed earlier in the commentary and found to
be unpersuasive. The older view that the main narrative portions in
those chapters (as well as in ch. 2) derive from two parallel accounts
of the episodes concerned remains compelling and will form the
basis of our treatment of 18.1-12.
The naming of Moses’ father-in-law as Jethro in 18.1ff. attaches
this narrative to the strand of Exodus 3–4 to which 3.1 (where ‘the
mountain of God’ is also mentioned) and 4.18 belong, while distin-
guishing it from 2.15b-22 (and also from Num. 10.29-32), where
Zipporah’s father is called Reuel (v. 18). These are respectively the
E and J strands of the narrative. An ascription of 18.1-12 to E fits
well with its use (even if alongside the divine name) of ‘God’ (Heb.
ʾelōhîm), while its attribution of two sons to Moses rather than one
is a further difference from J (2.22; 4.25-26). The Parallelszene in
4.27-28 has already been assigned to E: it also shares with 18.1-12
the presumption that geographically ‘the mountain of God’ is
somewhere between Midian and Egypt (whatever exact location
for the mountain, if any, is presupposed). It is also noteworthy
that Aaron is prominent in both episodes (cf. 18.12), whatever the
significance of that may be. In its present form 18.2 implies that
Moses had initially taken Zipporah and the boys back with him
to Egypt but had then sent her back to Midian (unless [though it
is unlikely] the final words of the verse mean ‘after she had been
left behind’ in Midian).1 It is in fact J which speaks of Moses
taking Zipporah and his (one) son back with him (4.20a, 24-26)
and nothing is ever said afterwards about their being sent back
1
This was Eissfeldt’s view (pp. 144*, 272*: Verlassung) and it is mentioned
as a possibility by Propp, p. 629, although he interprets the words in yet another
way (see Note c on the translation).
18.1-12 551
to Midian until here: in E Moses seems to return to Egypt alone

(4.18, 20b). The words ‘after she had been sent back’ are placed
awkwardly in the middle of the list of those whom Jethro brings
with him and look like a secondary addition made to reconcile the
different scenarios in the two sources when they were combined
(RJE).2 If so, then here too there would be nothing in the original
account to suggest that Moses took his family back with him to
Egypt: they remained in Midian.3
As for the date of the narrative, a final decision will have to await
consideration of vv. 13-26. But it has no Deuteronomic or Priestly
features, nor any other characteristics that would require a late date.
Haarmann cites three features of it as confirming the post-Priestly
setting that he has deduced from his (and others’) assessment of the
parallels with 4.27-31 – an argument which we have already rejected
– but none of them is at all conclusive. The fact that Moses is credited
with two sons rather than one may perhaps indicate a later develop-
ment in the tradition than 2.22 and 4.25-26, but nothing more: the
inclusion of both names in 1 Chr. 23.15 (also in v. 17 and 26.25) at
most shows that the Chronicler knew the Pentateuch in something
like its present form. Similarly the occurrence of the verb ‘acted
arrogantly’ (zîd: v. 11) in Neh. 9.10, which is not its earliest attesta-
tion, is probably due to the author’s familiarity with an older text,
as often elsewhere in this prayer (cf. M.J. Boda, Praying the Tradi-
tion: The Origin and Use of Tradition in Nehemiah 9 [BZAW 277;
Berlin and New York, 1999], p. 119). Finally, the ‘awkwardness’ of
the present position of Exodus 18 proves nothing about its origin,
whether it has been moved from an original location elsewhere in
the Pentateuch (as many have thought) or always formed a prelude
to the main Sinai narrative (for further discussion of this issue see
below). One other parallel, mentioned by Childs (p. 323: though
not as an argument for a late date), is between Jethro’s confession
and the post-exilic Ps. 135.5: but again there is no need to envisage
dependence on this late text, as the idea is found earlier elsewhere
(see the Explanatory Note).
2
So also Blum (Studien, p. 158 n. 253) and Haarmann (p. 71), even though
they do not accept the explanation in terms of J and E.
3
Proposals for other secondary elements in the section, such as the elimina-
tion of some designations of Jethro by Noth (pp. 116-18, ET, p. 144-46) and the
excision of ‘a burnt offering’ and ‘Aaron’ in v. 12 (Valentin, Aaron, pp. 385, 389-
91; Graupner, p. 99) are unsupported by convincing arguments.
552 EXODUS 1–18
The placement of this section (and even more so vv. 13-27) before
the main Sinai narrative begins in 19.1 has long been a subject of
discussion. Already in MRI some comments on the passage place
Jethro’s arrival after the revelation of the law (Lauterbach, pp. 162,
164) and this view was taken by several of the medieval rabbis: Ibn
Ezra was the most insistent (cf. Rottzoll 2, pp. 483-87). He cited (1)
the offering of sacrifices without any prior mention of the building
of an altar (v. 12); (2) the proclamation of statutes and laws as
having already begun (v. 16: this like [4] can more conveniently be
discussed in the introduction to the next section); (3) Jethro’s arrival
when the people were already encamped at ‘the mountain of God’
(v. 5: ‫העד הנאמן‬, ‘the surest proof’); and (4) the clear implication
of Num. 10.29 (Jethro being identified with Hobab by Ibn Ezra)
and Deut. 1.9-18 that Jethro was (still?) with the Israelites shortly
before the departure from Sinai/Horeb and that this was when the
changes made according to vv. 17-26 here were put into practice.
Much the same issues led most of the critics of the late nineteenth
and early twentieth centuries to a similar conclusion, even though
the development of source criticism caused it to be expressed in
a different way (and occasionally to be rejected altogether: cf.
Smend, Erzählung, pp. 152, 155; Eissfeldt, p. 60): it was thought
to be in the E source, with its mention of ‘the mountain of God’
in v. 5, that the episode(s) had once appeared later in the narrative.
Growing interest in the preliterary stages of the tradition even led
to its original location being found not at ‘the mountain of God’
at all, but at Kadesh (Gressmann, Anfänge, pp. 89-94: so already
Wellhausen, Prolegomena4, pp. 348-49, ET, pp. 342-43). But later,
with Noth (ÜGP, pp. 150-55) and those who followed him, it was
the sharp differences in character and content between the Jethro-
traditions and the main Sinai-tradition that attracted most attention,
rather than speculation about where they might once have stood
in a written account (so e.g. in Schmidt, Exodus, Sinai und Mose,
pp. 115-18: and see the next section of this introduction).4 Only in
the studies of Van Seters (Life, pp. 208-12) and Johnstone (Chron-
icles and Exodus, pp. 257-59), with their renewed focus on the
written stages of composition and the Priestly author as giving fresh
4
For Blum, who already in Studien, pp. 152-63, saw the chapter as an isolated
and late post-Priestly addition to the Pentateuch, this is even more naturally the
case.
18.1-12 553
shape to the earlier narratives, and the commentary of N. Sarna

(1991, pp. 97-98) did the hypothesis of an original placement later
in the Sinai-narrative reappear. Other recent commentators such
as Houtman (pp. 400-401) and Propp (pp. 627-28) have strongly
opposed such an idea.
Looked at in the light of the scholarship of the past century, the
old arguments may seem no longer quite as compelling as they once
did. The itinerary notes in 19.1-2a are attributed by many to a Priestly
or Deuteronomistic hand (most likely v. 1 is from P and v. 2a from
Dtr), and the older story seems to have arranged narratives about the
wilderness in a sequence that did not associate them with a series
of places along a route. True, 19.2b probably belongs to neither
of the later compositions, but if the older narrative was composed
from two or more earlier sources, the ‘premature’ mention of Israel
at ‘the mountain of God’ in 18.5 could be due to the difficulties of
reconciling their different presentations of events. This might also
explain the appearance of Hobab (most likely the son of Moses’
father-in-law in view of 2.18) at the end of the Sinai-narrative in
Num. 10.29: the J and E versions may have had different settings
as well as different names and different roles for Moses’ Midianite
in-laws in their narratives. The same might apply, to anticipate a
later discussion, to Deuteronomy’s placement of the appointment of
judges when the Israelites were about to leave Mount Horeb/Sinai,
especially as it avoids all mention of Jethro’s role (as of Moses’
Midianite connections in general).
Moreover, the present placing of Jethro’s reappearance is from
various points of view explicable and even appropriate. As Ibn Ezra
already saw, it provides a welcome antidote to the xenophobic atti-
tude to Israel’s neighbours in 17.8-16. It is natural too that Jethro
should restore Moses’ wife and family to him as early as possible.
And where it is, as Sarna and especially Dozeman (pp. 361-63,
400-401) and Albertz (p. 298) have emphasised, the story plays
an important transitional role by, on the one hand, presenting a
further celebration of the Exodus deliverance and, on the other (in
vv. 13-26), introducing more strongly than hitherto (cf. 15.25; 16.4,
28) the themes of worship and law that are to be so central in the
non-Priestly Sinai narrative.5
5
The fullest explanation of this aspect of the whole chapter, from a mainly
literary perspective, is provided by E. Carpenter, ‘Exodus 18: Its Structure, Style,
554 EXODUS 1–18
The contribution of the story to the history of religion has been

assessed in two very different ways. Traditionally it was seen as an
exemplary account of the conversion of a foreigner to Judaism (as in
the Targumim and MRI: see Text and Versions passim), even though
Jethro does not adopt any distinctively Jewish practices, such as
circumcision (contrast 12.44, 48). Recently this approach has been
taken up, in the context of a post-exilic origin which he envisages
for the passage, by Haarmann, but in a modified form which sees
Jethro as analogous to the ‘righteous Gentile’ of later rabbinic law:
he does not become a member of the people of God but he worships
the God of Israel (YHWH-Verehrer, pp. 33-42, 91-94). In this his
portrayal is comparable to those of Rahab, Naaman and the sailors in
Jonah 1 and to the foreign worshippers who are envisaged in 1 Kgs
8.41-43 and Isa. 56.1-8. Comparable perhaps, but also distinct, both
in the likelihood that Jethro’s story (like Naaman’s?) comes from
an earlier period and in the special relationship to Moses in which
he is allowed to stand. Both these characteristics have encouraged
another understanding of the Jethro traditions which has been much
more widespread, though by no means uncontroversial, in modern
scholarship.
The narratives of chs. 2–4 and the setting of Moses’ call by
Yahweh to lead his people out of Egypt in the midst of them have
already drawn attention to the traditional, and perhaps historical
and religious, connection between Moses, Israel and the Midianites
(see briefly in the introduction to 2.11-22). Since the mid-nineteenth
century this has, with other converging evidence, been taken to point
to Midianite (or Kenite: see below) religious beliefs and practices
as the seed out of which Israel’s very different worship of Yahweh
grew (on the origins of the hypothesis see J. Blenkinsopp, ‘The
Midianite–Kenite Hypothesis Revisited and the Origins of Judah’,
JSOT 33 [2008], pp. 131-53 [132-33]). Classic expositions of this
view were given by K. Budde, Religion of Israel to the Exile (New
York, 1899), pp. 17-25, and H.H. Rowley, From Joseph to Joshua
(London, 1948), pp. 148-60, and more recently it has received the
support of scholars such as R. Smend (Jahwekrieg, pp. 95-96, ET,
pp. 131-34), Schmidt, (Exodus, Sinai und Mose, pp. 110-30) and
J. Blenkinsopp (art. cit.: for criticism of the theory see below).
Motifs and Function in the Book of Exodus’, in id. (ed.), A Biblical Itinerary:
In Search of Method, Form and Content. Essays in Honor of George W. Coats
(JSOTSup 240; Sheffield, 1997), pp. 91-108.
18.1-12 555
The status of Jethro (or Reuel as the name of Moses’ father-in-

law is given in the alternative tradition in Exod. 2.18 and Num.
10.29) as a Midianite priest and his apparently leading role in the
sacrifice to Yahweh in Exod. 18.12 have been the central pillars of
the theory. But support for it has also been claimed from Yahweh’s
close association with Sinai, Horeb and ‘the mountain of God’
(whether these are different names for the same mountain or, as
some have held, different places) in the region to the south of
Palestine, and especially his designation as ‘the one of Sinai’ (Heb.
zeh sînāy) in Judg. 5.5 and Ps. 68.9; from Egyptian texts refer-
ring to the šśw yhw3 in the same region (on whom see Albertz,
Religionsgeschichte, p. 83, ET, p. 51, and the references in n. 51);
and (in relation to the Kenite version of the theory) from the ethno-
graphic element in the story of Cain as a worshipper of Yahweh
in Gen. 4.1-16 and the associations of descendants of Hobab the
Kenite, who is also described as the father-in-law of Moses (a
change from Num. 10.29), with tribal traditions of early Israel
(Judg. 4.11; cf. 1.16). In addition Schmidt has argued that Moses’
Midianite connections provide the necessary traditio-historical
bridge between Yahweh the God of Sinai/Horeb, who was previ-
ously unknown to Israel, and the understanding of Yahweh as
the God of the Exodus and the God of Israel (Exodus, Sinai und
Mose, pp. 124-30). Such connections become the more plausible
in the context of the broader arguments of M. Leuenberger, ‘Jhwhs
Herkunft aus dem Süden. Archäologische Befunde – biblische
Überlieferungen – historische Korrelationen’, ZAW 122 (2010),
pp. 1-19.
The theory has, on the other hand, often been questioned.
Already Rowley had to answer criticisms from scholars such as T.J.
Meek and M. Buber and more recently it has been challenged by
C.H.W. Brekelmans (‘Exodus XVIII and the Origins of Yahwism
in Israel’, OTS 10 [1954], pp. 215-24), R. de Vaux (Histoire anci-
enne 1, pp. 313-21, ET, pp. 330-38), Childs (pp. 322-24) and
Haarmann (pp. 77-81, though only with reference to Exod. 18.1-
12). Others (e.g. Van Seters, Life, pp. 208, 211) pass it by in favour
of a much more speculative political interpretation. Nevertheless,
while it cannot be more than a hypothesis, it does provide a plau-
sible response to the otherwise unanswered question about the
pre-Mosaic worship of Yahweh and a satisfying understanding of
Exodus 18.1-12 as reflecting, in the person of Jethro and the words
and actions attributed to him, Israel’s take-over of an existing cult
556 EXODUS 1–18
and its transformation into the historical and covenantal religion of

the Pentateuchal tradition.6
In one sense, and it is an important one, 18.1-12 provides a
second celebration of the Exodus as Yahweh’s deliverance of Israel
(vv. 8-12 with v. 1 as preparatory). Indeed it can be seen as the first
comprehensive celebration of the Exodus story as a whole, since the
hymns of Exodus 15 confine their theme to the specific, and final,
overthrow of Egyptian power at the sea, which is only one aspect
(if our reading of v. 11 is correct) of Yahweh’s mighty acts here.
But there is much more to the story than even this, for it is not by
chance that the celebrant is a Midianite, indeed a Midianite priest.
He cannot, like Rahab or Naaman, be merely an example of the
‘righteous Gentile’ of rabbinic thought. In him, somewhat veiled it
is true, and in his words ‘Now I know that Yahweh is greater than
all the gods’, now (that is) that he has heard the Exodus story, there
is also preserved the memory that Israel itself, who continued to tell
that story, was indebted to a foreign people for the name of its God
and the knowledge of his original dwelling-place on earth.
1 Jethro the priest of Midian, Moses’ father-in-law, heard ofa all

that God had done for Moses and Israel his people, (namely)
thatb Yahweh had brought Israel out from Egypt. 2 So Jethro,
Moses’ father-in-law, took Zipporah the wife of Moses [after
she had been sent awayc] 3 and her two sons, of whom the name
of one was Gershom, for he (sc. Moses) had said, ‘I have been
a displaced persond in a foreign land’. 4 The name of the othere
was Eliezer, for (he had said), ‘The God of my father is my helpf
and he delivered me from Pharaoh’s sword’. 5 Jethro, Moses’
father-in-law, and his sons and his wife cameg to Moses, to the
desert where he was camping at the mountain of Godh. 6 He said
to Moses, ‘I, your father-in-law Jethro, am comingi to you, and
your wifej and her two sons with her’. 7 So Moses came out to
6
By contrast, the short-lived explanation of the sacrificial meal in v. 12 as
a ‘covenant meal’ cementing an alliance between Midian and Israel (cf. Brekel-
mans, art. cit.; F.C. Fensham, ‘Did a Treaty between the Israelites and the Kenites
Exist?’, BASOR 175 [1964], pp. 51-54; A. Cody, ‘Exodus 18.12: Jethro Accepts
a Covenant with the Israelites’, Bib 49 [1968], pp. 153-66) and the specific
proposal to connect the episode with Israel’s sojourn at Kadesh, recently revived
by Blenkinsopp (art. cit., pp. 144-48), are quite unfounded.
18.1-12 557
meet his father-in-law. He bowed downk and kissed him, they

asked each other about their welfarel and they went into the tent.
8 Moses recounted to his father-in-law all that Yahweh had done
to Pharaoh and Egypt for the sake ofm Israel, all the hardshipn
which had come upon them on their journey, from which Yahweh
had delivered them. 9 Jethro rejoicedo at all the good which
Yahweh had done for Israel, in thatp he had delivered them from
the powerq of the Egyptians. 10 Jethro said, ‘Blessed be Yahweh,
who has delivered you [pl.] from the power of the Egyptians
and from the power of Pharaoh, who delivered the people
from under the power of the Egyptians!r 11 Now I know that
Yahweh is greater than all the gods, sfor through what they had
used arrogantly <against Israel he has prevailed> over thems.’
12 Then Jethro, Moses’ father-in-law, broughtt a burnt-offering
and sacrifices for God, and Aaron and all the elders of Israel
came to eat a mealu with Moses’ father-in-law before God.

a. Heb. ‫וישׁמע‬. Here ‫ שׁמע‬has the rarer, extended sense ‘hear of’, as e.g. in 1
Sam. 3.11; 1 Kgs 19.11.
b. The ‫כי‬-clause is here explicative of what had been said in general terms
before; compare the similar amplification with ‫ אשׁר‬in v. 9.
c. Heb. ‫אחר שׁלוחיה‬. The noun ‫ שׁלוחים‬occurs only three times in BH, always
in the pl.: in 1 Kgs 9.16 and Mic. 1.14 the association with ‫נתן‬, ‘give’, indicates
the sense ‘parting gift’, specifically ‘dowry’: the context of judgment in Mic.
1.14 has suggested the rendering ‘divorce’ to some (Ges18, p. 1358: cf. below
on the use of the verb ‫)שׁלח‬, but the expression is in any case metaphorical and
an ironic ‘dowry’ is at least as plausible. The word occurs as both sing. and pl.
in MH for ‘sending away’ and also ‘allowing to go, release’ (Jastrow, p. 1563),
but never ‘divorce’. In Ug. the unique ṯlḫh in KTU 1.24.47 (a broken context)
has been taken to mean ‘(her) dowry’ in view of the setting of a wedding
(e.g. Gibson, CML2, pp. 129, 160), but the verb for ‘send etc.’ is always spelt
šlḥ in Ug. and ṯlḫh is probably not connected with it: it may be the name
of a deity (DULAT, p. 906). Here the use with ‫ אחר‬implies that an event or
action is meant (see GK §124f for other such plurals, even if the explanation
given is not always appropriate), the nature of which could correspond to
any of the many senses of ‫ שׁלח‬Piel (note ‫ וישׁלח‬in v. 27): there is no particular
reason to suggest that a formal divorce is intended (as with the verb in Deut.
22.19, 29; 24.1, 3; Jer. 3.1, 8; Mal. 2.16), and the references to Zipporah
as still Moses’ wife (vv. 2, 5 and 6) are against it: cf. Houtman, p. 404.
On the precise meaning and circumstances see further in the Explanatory
Note. In view of the difficulties with the normal understanding of the phrase,
558 EXODUS 1–18
G. Del Olmo Lete suggested that ‫ שׁלוחים‬might have the sense ‘dowry’ here
too (perhaps derived from the sense ‘give, bestow’ for ‫שׁלח‬, as sometimes in
Ug.: so A.S. van der Woude, ‘I Reg 20 34’, ZAW 76 [1964], pp. 188-91 [190];
cf. DULAT, p. 816) and that ‫ אחר‬here means ‘with’ as occasionally in Ug. and
elsewhere in BH (‘ “ʾaḥar šillûḥèhā” (Ex 18,2)’, Bib 51 [1970], pp. 414-16).
The latter suggestion was made for a number of Heb. instances without refer-
ence to Ug. by R.B.Y. Scott in ‘Secondary Meanings of ‫אחר‬, after, behind’,
JTS 50 (1949), pp. 178-79 (cf. HAL, p. 34), and was subsequently taken up in
the light of Ug. usage by M. Dahood (e.g. Psalms III [AB 17A; Garden City,
1970], p. 390: cf. DULAT, pp. 39-40) and more fully by Del Olmo Lete in ‘La
preposición ʾaḥar/ʾaḥarê (cum) en ugaritíco y hebreo’, Claretianum 10 (1970),
pp. 339-60. But the Ug. passages cited (KTU 1.24.32 and 1.14.4.46 par.) do
not in fact require this sense and serious doubts were raised by E. Jenni about
its presence in Heb. (THAT 1, 112 = TLOT 1, p. 84): it has not been included in
Ges18 or DCH. Attractive therefore as Del Olmo Lete’s translation ‘with her
dowry’ might be, it rests on uncertain linguistic foundations. Propp (p. 629)
adopts the ‘dowry’ interpretation (like some older Jewish interpreters) but
renders ‫ אחר‬in the usual way: this will hardly do, as it leaves the text stating
the obvious.
d. Heb. ‫גר‬. On the meaning see Note dd on the translation of 2.11-22.
e. Heb. ‫האחד‬. When the expression is repeated, as also in 2 Sam. 14.6 and
1 Kgs 12.29, the meaning is ‘the one…the other’; in Exod. 26.19, 21, 25 the
idiom is related but different.
f. Heb. ‫ בעזרי‬is a further case of beth essentiae, as in 6.3 (see Note d on the
translation of 6.1-9) but with what BDB, p. 88, calls the ‘primary predicate’,
which expresses the nature of the subject rather than an attribute: this is found
especially often with expressions of help and support (cf. Pss. 54.6; 118.7;
146.5; prob. Hos. 13.9).
g. Heb. ‫ויבא‬, with the sing. verb agreeing with the first element of a
composite subject that follows (GK §146f).
h. Heb. ‫הר האלהים‬, with the location expressed by an ‘accusative of local
determination’ (JM §126h), perhaps uniquely with ‫( הר‬contrast 4.27). The
preceding ‫ שׁם‬may have made the omission of ‫ ב‬easier here.
i. Heb. ‫בא‬: the part., as most often, has a present durative sense, implying
that Jethro here sends a message ahead of him.
j. Heb. ‫ואשׁתך‬. For the waw concomitantiae cf. GK §154a note (b) and more
fully Driver, Samuel2, p. 55 (also BDB, p. 253). This is a stylistic feature and
need not mark a secondary addition.
k. Heb. ‫וישׁתחו‬. On the discussion over whether the root is ‫ שׁחה‬or ‫ חוה‬see
Note cc on the translation of 4.18-31. The form here is the apocopated third
person sing. masc. of the waw consecutive imperfect. As is pointed out in
GK §75kk, the vocalisation at the end developed in the same way as in the
noun ‫( שׁחו‬and possibly also the short form ‫[יהו‬-] of the divine name Yahweh).
18.1-12 559
Although the verb is widely used of the worship of a god (as in 4.31 and
probably 12.27; cf. 20.5 etc.), it also represented a gesture of respect towards
a fellow human being (11.8).
l. For ‫ שׁלום‬in the sense of ‘well-being, welfare’ cf. 4.18: the sense ‘peace’
may be a secondary development (see TWAT 8, 12-46 [18] = TDOT 15, pp.
13-49 [19]). To ask after someone’s ‫ שׁלום‬is a common idiom in BH (e.g.
Gen. 43.27; Judg. 18.15; 2 Sam. 11.7) and it is also attested in epigraphic
Hebrew (AHI 2.18.3), where ‫ שׁלום‬appears in a variety of epistolary formulae
of greeting (see AHI, p. 495; 2, p. 218).
m. Heb. ‫על אודת‬, a compound preposition incorporating a noun
‫ א(ו)דה‬which does not occur separately in BH. The plene spelling in the first
syllable is found only here and in the textually corrupt 2 Sam. 13.16. Ar.
ʾaddā, ‘cause’, provides the most likely clue to its etymology. The meaning
is occasionally ‘about’ (with ‫ הגיד‬and ‫ )דבר‬but most often ‘because of’ with
reference to a prior cause (for complaint, punishment or the giving of a name)
or to personal concern, as here and in Gen. 21.11, where ‘for the sake of’ is the
closest Eng. equivalent. It does not express purpose as ‫ בעבור‬can and is closest
in meaning as well as form and syntax to the compounds of ‫ על‬with ‫ ָדּ ָבר‬and
‫( ִדּ ְב ָרה‬BDB, p. 184). Half the ten secure occurrences in BH are in passages
which have traditionally been ascribed to E (Gen. 21.11, 25; Exod. 18.8; Num.
12.1; 13.24); most of the rest are in Deuteronomistic contexts.
n. Heb. ‫את כל־התלאה‬. ‫תלאה‬, from √‫לאה‬, ‘(be) weary’, is used again of
Israel’s trials in the wilderness in Num. 20.14 and in a similar phrase of
Israel’s sufferings at the hands of powerful enemies in Neh. 9.32 (cf. Lam.
3.5). The absence of a preceding waw (which some witnesses supply: see
Text and Versions) implies that the expression is in apposition to the previous
object of ‫ויספר‬, but ‫ בדרך‬might reasonably be taken to include the Israelites’
journey after leaving Egypt as well, thus extending the scope of Moses’ report.
o. Heb. ‫ויחד‬, from ‫ חדה‬II (the even rarer ‫ חדה‬I is a by-form of ‫חדד‬, ‘be
sharp’): on the spelling of the apocopated form cf. Job 3.6 and GK §21f, 28e,
75r. The verb is rare in BH, being elsewhere confined to poetry (Ps. 21.7; Job
3.6) and the related noun ‫ חדוה‬occurs only in LBH prose (Neh. 8.10; 1 Chr.
16.27): both are more characteristic of Aram. The narrator’s choice of this
verb (rather than ‫ שׂמח‬or ‫ )גיל‬might have been intended to stress the emotion
of the occasion or to reflect Jethro’s foreign origin. See further Note t on the
p. Heb. ‫ אשׁר‬here introduces a clause (like the second half of v. 1) which is
not so much causal as explicative of what ‫ הטובה‬refers to (‘in that…’, ‘namely
that…’): cf. Gen. 42.21; 2 Kgs 17.4.
q. Heb. ‫מיד‬, lit. ‘from the hand of…’; so again three times in v. 10 and often
with this metaphorical sense.
r. The two relative clauses in this verse are unusually repetitive, despite
small variations, and the Heb. texts may preserve two alternative versions
560 EXODUS 1–18
of the wording. Another possibility is that the narrator has deliberately intro-
duced into Jethro’s hymn of praise what might be called ‘prose parallelism’
(cf. vv. 1 and 9: also Houtman, p. 408), in imitation of the widespread poetic
variety: for further discussion of this phenomenon see J.L Kugel, The Idea of
Biblical Poetry: Parallelism and its History (New Haven, 1981), pp. 59-87.
For a yet further (but less likely) explanation of the duplication see Text and
Versions on the end of v. 11.
s. Heb. ‫כי בדבר אשׁר זדו עליהם‬, lit. ‘…for by the word/thing (by) which
they acted proudly against them’. The incompleteness of the sentence is
intolerable and a well-known crux. The problem cannot be solved by treating
‫ כי‬as emphatic (RV: presumably ‘precisely [because]’, cf. Keil) rather than
subordinating, since there are no parallels to such a placement of the particle:
for a careful study of the circumstances in which the emphatic use may be
present see Muraoka, Emphatic Words, pp. 158-64. Nor can Durham’s trans-
lation ‘for in this thing they have acted rebelliously against them’ be derived
from the text as we have it, as he partly acknowledges (pp. 239-40). Some
text has evidently been lost or displaced from the end of the verse, as Driver
saw: see Text and Versions for a discussion of various proposals to fill the
gap. Our translation is based on a new suggestion (which follows broadly the
treatment in TgO) that between ‫ זדו‬and ‫ עליהם‬there originally stood the words
<‫ >על־ישׂראל הוא גבר‬or something similar, which was omitted when an early
scribe’s eye jumped from the first ‫ על‬to the second one.
t. Heb. ‫ויקח‬. The standard translation ‘took’ of older EVV., even with ‘for
God’ following, leaves an awkward gap in the proceedings before the eating
of the sacrificial meal: when ‫ לקח‬is used of a sacrifice it normally represents
a preliminary to the act of offering itself (e.g. Gen. 8.20). Some of the Vss
have ‘offered’ instead (see Text and Versions: cf. RSV), but no emendation is
needed: in a ‘pregnant’ construction before a preposition (GK §119ee, gg) ‫לקח‬
sometimes includes the action which follows the taking and can be translated
‘bring, brought’ (NIV, NRSV: cf. 1 Kgs 17.10-11 and other exx. in BDB,
p. 543). It is even used of a sacrifice in parallel to ‫ הביא‬in Lev. 12.6-8, but there
the actual offering is mentioned separately.7
u. Heb. ‫לאכל לחם‬. ‫ לחם‬alone can mean not only ‘bread’ but ‘food’ more
generally (see Note x on the translation of 16.1-36) and the combination ‫אכל‬
‫ לחם‬could be used of a meal that included meat as well (Gen. 31.54). Here
the ‫ זבחים‬most likely were (or included) animal sacrifices, which were mainly
eaten by the worshippers (see the Explanatory Note on 10.25-26).
7
A. Cody, ‘Exodus 18.12’ (see n. 6), argued that ‫ ויקח‬referred to Jethro’s
‘acceptance’ of sacrifices presented by the Israelites, but this would only make
sense in a narrative which mentioned their presentation and, as Valentin pointed
out (Aaron, pp. 387-88), neither Gen. 21.27-31 nor Josh. 9.14 provides any
support for Cody’s view.
18.1-12 561
Explanatory Notes
1. The new chapter begins unexpectedly with a change of scene
and the reintroduction of a group of people who have not featured
in the narrative since the early chapters of Exodus. The group is led
by Jethro, who is first called ‘the priest of Midian’, as he had been
in 3.1 (in 2.18 ‘the priest of Midian’, at first anonymous in v. 16,
is given the name Reuel, which appears again, with that of Hobab
his son, in Num. 10.29), and is then designated as ‘Moses’ father-
in-law’, as in 3.1 and 4.18: it is this designation which continues
to be used in the rest of the section (vv. 2, 5, 6, 7, 8, 12) and also
in the second half of the chapter (here without his name: vv. 14,
15, 17, 24, 27). The reason given for his reappearance in the narra-
tive is that he has ‘heard’, presumably in his homeland of Midian,
about what has happened to his son-in-law and to the Israelites
since he bade Moses farewell in 4.18. How he has heard the narrator
does not say, but from the fact that he knows where to find Moses
(v. 5) and that he brings Moses’ wife and children with him (vv. 2-4)
one might well infer that Moses had sent him a message. What he
has heard is at first stated in general terms, as befits the resumption
of an association that has been broken off, and then specifically
as Israel’s departure from Egypt through Yahweh’s mighty inter-
vention on their behalf, which has been the main theme of the
intervening narrative (on the rendering ‘namely that’ see Note b on
the translation). It is this ‘deliverance’ from Egypt which continues
to be the central topic later in the section (vv. 8-10): v. 4 also hints at
it, even though the reference there must be to something else). Only
in v. 8 is there any allusion to events subsequent to the deliverance:
‘all the hardship…on the way’. Although the divine name continues
to be used throughout the section in connection with the Exodus
events, whether Moses or Jethro is the speaker, the title ‘God’ (Heb.
ʾelōhîm) is also used both here and later in the actual narrative (vv. 5,
12): similarly throughout the second half of the chapter, even in the
dialogue. This avoidance of the divine name where there is no doubt
that Yahweh is meant, which has also occurred earlier in Exodus
(e.g. in 1.15-22; 3.1-6, 13-15; 13.17-19), has not found a convincing
overall contextual explanation and can only be described as a choice
to use the divine title ‘God’ instead by the author of some parts of
the Exodus narrative.
562 EXODUS 1–18
2. The brief account of Jethro’s journey to see Moses (v. 5) is

delayed because of the need to enumerate those whom he ‘took’
with him, which was clearly felt to be an important opening to the
episode: after all, it provides an additional reason for Jethro to make
the journey.8 According to the narrative here Moses’ wife Zipporah
(first mentioned in 2.21) and their children were still in Midian, and
taken alone 4.18, which mentions only Moses as returning to Egypt,
is compatible with their never having left there. But the final phrase
of the verse, ‘after she had been sent away’ (its most probable inter-
pretation: see Note c on the translation), assumes that Zipporah had
initially accompanied Moses on his journey back to Egypt (cf. 4.20,
24-26) and then been sent home by him, presumably to protect her
(and the children) from the dangerous situation in Egypt. Nothing is
said of this in the Exodus narrative; Rabbinic tradition associated it
with Moses’ meeting with Aaron in 4.27 (cf. MRI, pp. 167-68: see
further Text and Versions). Because of other inconsistencies with
(and in) the early chapters of Exodus these words (which also seem
like an afterthought) are commonly seen as a harmonising addition
to reconcile the differing accounts in the sources used in the compo-
sition of the book (see further the introduction to this section).
3-4. One of these inconsistencies concerns the number of
children that Moses and Zipporah have: here there are two, but only
Gershom is mentioned earlier – with his name and its explanation in
2.22 (see the note there) and again, referred to simply as ‘her son’,
in 4.25. Eliezer and the explanation of his name appear for the first
time here, at least in the extant Heb. texts: in the Peshiṭta and many
manuscripts of the LXX and Vulg these details are added after 2.22
to reconcile the two versions of the story (see Text and Versions
there). The plural ‘his sons’ in 4.20 may be due to the same concern
at an earlier stage of the textual tradition.
8
There is nothing in the Heb. to suggest or even allow the translation ‘Jethro
had taken…’, as if some earlier action were referred to (so RSV, NEB, REB, ESV;
NIV ‘received’ implies the same interpretation). Childs attributes this to ‘an old
exegetical tradition’ which sought ‘to avoid the difficulty in the chronology of
the narrative’ (p. 320) and the rabbinic commentary of Bekor Shor (twelfth cent.)
seems to be meant: it is associated there with an unusual understanding of the
final words of the verse, and the EVV. in question may also have been seeking
a smoother connection with these words. Childs rightly rejects the idea, for the
reason given above, and it is not adopted in NAB (according to him), NJPS, JB or
NRSV, nor in any modern scholarly commentary known to me.
18.1-12 563
Eliezer’s name means ‘(My) God is a help(er)’ or possibly, if ‘El’

is taken as the name of a god, ‘El is a help(er)’. Abraham’s servant
bears the same name in Gen. 15.2 and it occurs several more times in
genealogies and lists in Chronicles and Ezra. The name ‘Eleazar’ has
essentially the same meaning and is attested in earlier as well as later
texts, most notably as the name of Aaron’s third son and successor
as chief priest (Num. 20.25-26, 28): it also appears occasionally in
Heb. inscriptions (AHI, p. 283; AHI 2, p. 133). In 1 Chr. 26.25-28
(cf. 23.15-17) Moses’ son Eliezer is given a list of descendants
which ends with a Levite who is represented as the keeper of the
war booty of King David and others: it is probably a late invention.
The explanation of the name that is provided in v. 4 goes beyond
its etymological basis in two ways, both with some specific corre-
spondence to the traditions about Moses. The expression ‘the God
of my father’ in words that are presumably attributed to Moses takes
up or, given the implied chronology, anticipates the parallel to ‘my
God’ (ʾēlî) in 15.2, and is also related to the divine address to Moses
in 3.6, ‘I am the God of your father’: on its associations with family-
based religion in early Israel see the Explanatory Note on 3.6. The
reference to deliverance ‘from the sword of Pharaoh’ may allude to
the circumstances which brought Moses to Midian (cf. 2.15; 4.19) or
to his escape as a baby from Pharaoh’s attempt to have male Israelite
children drowned (cf. 1.15-21; 2.1-10): ‘the sword’ could easily be a
metaphor for any kind of killing. The time of Moses’ confrontation
with Pharaoh immediately prior to the Exodus can hardly be meant,
since Eliezer would have been born long before that; but ‘delivered’
(Heb. nāṣal Hiphil) is the same verb that is used of the Exodus in
vv. 8-10 and the echo may be deliberate (for other occurrences in
Exodus see 2.19; 3.8; 5.23; 6.6; 12.27).
5. The main narrative of the section is resumed with the arrival of
Jethro and Moses’ family at their destination. This time the children
are mentioned first, probably just to make a smoother connection
with the previous verses, and here they are called Moses’ children
rather than Zipporah’s as in vv. 3 and 6: this was unavoidable when
mention of her was delayed. They arrive ‘in the desert’ – Midian
itself seems to be thought of as outside the desert (cf. 3.1) – at ‘the
mountain of God’, where Moses had come with Jethro’s flock in
3.1 before his commission by Yahweh to lead the Exodus. There
it was identified with Horeb (as again in 1 Kgs 19.8), and Horeb
is named again in 17.6 as the place where Israel’s need for water
564 EXODUS 1–18
has recently been met: nothing has been said of any movement by
the people since then (on the terminology used and its significance
see the Explanatory Note on 3.1). But the words ‘in Horeb’ in 17.6
are probably a secondary addition (see the Explanatory Note on
17.5-6), so that originally no earlier indication will have been given
of the Israelites’ arrival at the mountain. In fact their encampment
there is only recounted (at some length) in 19.1-2, where the main
Sinai narrative begins. So 18.1-12 appears to be either premature
in its present position (and it has long been suspected that it and,
with even greater reason, vv. 13-27 really belong to a later stage
of the story) or an extract from a parallel account of the wilderness
journey where it was preceded by a note of Israel’s arrival at the
holy mountain. The issues involved are complex: for a fuller discus-
sion see the introduction to this section.
6-8. Moses has been increasingly in view in the preceding verses
(in v. 5 he is named twice and referred to by pronouns three times)
and now he becomes directly involved in the narrative, occupying
the leading role in vv. 7-8 before it reverts to Jethro in v. 9. First
Jethro announces his approach, apparently through a messenger
since he and Moses are not yet face-to-face in v. 6 (see Text and
Versions). The presence of other members of the family may be
assumed here without anyone being named (cf. Hobab in Num.
10.29). The wording of messages, both oral and written, naturally
retained the first-person reference to the sender of the message
(cf. Gen. 32.3-5; 2 Kgs 10.6: similarly in the letters that have been
preserved in epigraphic Hebrew and in documents from the neigh-
bouring cultures [see, e.g., D. Pardee et al., Handbook of Ancient
Hebrew Letters (Chico CA, 1982), passim]).9 The description of the
meeting between Moses and Jethro – Zipporah and the children are
no longer of any concern to the narrator – is unusually fulsome (cf.
L. Köhler, Der hebräische Mensch: eine Skizze [Tübingen, 1953],
pp. 64-66, ET pp. 77-78): for the kiss compare Gen. 33.4; Exod.
4.26; for the enquiry after the other’s welfare 1 Sam. 25.5; 2 Sam.
11.7; also 2 Sam. 20.9; 2 Kgs 4.26, where the question ‘Is it well
(with you)?’ is cited (see further I. Lande, Formelhafte Wendungen
der Umgangssprache im Alten Testament [Leiden, 1949], pp. 5-8);
9
In some later versions of the biblical text this formulaic pattern was found
difficult and ‘I…(am)’ was replaced by ‘Behold…(is)’ (Heb. hinnēh: see Text and
Versions).
18.1-12 565
for the invitation inside Gen. 24.31. The bowing down is an addi-
tional act of respect or homage (cf. Gen. 27.29; Exod. 11.8). Within
Moses’ tent food might be prepared and served (Gen. 18.6; Judg.
4.19), shade found (Gen. 18.1) and a comfortable place to rest
provided (Judg. 4.18). But all the narrator is interested in here
(v. 8) is what Moses has to tell Jethro, presumably amplifying what
he knows already (v. 1). The topics are packed closely together in
this short summary of what must have been a long and complex
story. The emphasis falls initially on Yahweh’s treatment of the
Egyptians, but this overlaps with Israel’s own sufferings on their
journey (see Note n on the translation for the sentence structure)
and so probably embraces both the plague-narrative and the Isra-
elites’ narrow escape at the sea. The final summary, ‘and Yahweh
delivered them’, introduces the keynote which, thrice repeated,
provides the new focus for Jethro’s celebration of the Exodus itself.
9-11. Jethro’s response is first sheer exhilaration at Yahweh’s
beneficent treatment of the Israelites in Egypt, a ‘deliverance’ which
was necessary because of the powerful grasp (lit. ‘hand’) of the
Egyptians from which Israel has now been released. This theme
remains central and, with only slight variation, is twice repeated in
Jethro’s thanksgiving to Yahweh (v. 10). In the Psalms rejoicing and
praise regularly go together (e.g. Pss. 40.17; 66.6-8; 70.5; 97.12),
and the opening formula ‘Blessed be Yahweh’ is often used in
hymns and thanksgivings (e.g. Pss. 28.6; 68.20). It became increas-
ingly popular in late and post-biblical texts (cf. Gunkel, Einleitung,
p. 40; W.S. Towner, ‘ “Blessed be YHWH” and “Blessed art thou,
YHWH”: The Modulation of a Biblical Formula’, CBQ 30 [1968],
pp. 386-99). ‘Blessed’ (Heb. bārûk) in such phrases is the equiva-
lent of ‘to be praised’, just as the active verb ‘bless’ (Heb. bērēk) is
frequently a word for praise and thanks addressed to God (e.g. Pss.
26.12; 68.27; 103.1: cf. THAT 1, 373-76 = TLOT 1, pp. 281-82).
Such formulae also, it is true, appear in narratives of everyday life
(e.g. Gen. 24.27; 1 Sam. 25.32, 39; Ruth 4.14: Albertz, p. 306), but
in this case the sequel (esp. v. 12) and the national perspective evoke
a cultic setting. As Houtman has pointed out (p. 404), the repetitive
parallelism of this verse recalls the style of much Hebrew poetry:10
10
Cf. Knierim, ‘Exodus 18’, pp. 148-49 (cited approvingly by Blum, Studien,
p. 158 n. 253), who attributes the ‘redundant’ formulation to an ‘elevated’ literary
style.
566 EXODUS 1–18
the variation between ‘you’ (though it is plural in the Heb.) and ‘the
people’ could pick up ‘Moses and Israel his people’ in v. 1.
The overcoming of Egypt’s power is given a fresh twist in
Jethro’s justification of his new conviction that Yahweh is the
most powerful of all the gods (v. 11). Here, if our restoration of the
damaged text is even approximately correct (see further in Text and
Versions), Yahweh shows not only his power but his justice and his
superior manipulation of nature by using the very means by which
the Egyptians had tried to destroy Israel to bring about their own
downfall. Most likely the waters in and around Egypt are meant,
first those of the Nile in which Israelite babies were to be drowned
and then those of Yam Suf, in which Pharaoh’s elite troops had
perished.11 The formula ‘(Now) I know’ again, naturally enough,
echoes similar expressions in responses to evidence of God’s power
to act (Dozeman, pp. 403-405), whether in a royal psalm (Ps. 20.6),
a prophetic narrative (1 Kgs 17.18) or (cf. Childs, p. 323) a psalm
of praise (Ps. 135.5). In Exodus it recalls 5.2 (by contrast) and 8.6.
The content of Jethro’s exclamation also resembles some psalmic
language (e.g. Pss. 97.9; 135.5) – as well as the hymns of praise in
Exodus 15 – so that, while it falls short of the purest monotheistic
language in the Old Testament (like Isa. 44.8), it attributes to him a
religious outlook common to many Israelites.
It is entirely appropriate that in vv. 9-11, and also in v. 8, the name
of Yahweh should be so prominent, even in the work of a narrator
who continues to use the title ‘God’ in his own words (see the note
on v. 1). At first sight it suggests that ‘the priest of Midian’ had
previously been a worshipper of ‘other gods’ and now had come to
recognise (and worship: v. 12) Yahweh. But this is no longer the only
possible explanation. Traditionally Jethro has been regarded as the
archetypal Gentile convert or proselyte (see e.g. Text and Versions
on vv. 6-7) and, taken alone, the narrative is easily understood in
this way. But since the later nineteenth century a variety of evidence
has led to the widespread view (though it is contested by some) that
Yahweh was originally the god of the Midianites (or Kenites) and
that it was through Moses’ acquaintance with ‘the priest of Midian’
that he, and subsequently Israel as a whole, came to worship Yahweh
11
Or is it the latter in both cases (Baentsch, p. 165)? ‘Arrogance’ would fit the
headlong pursuit into the sea very well.
18.1-12 567
(see Blenkinsopp, ‘The Midianite–Kenite Hypothesis Revisited’,

and our fuller discussion [with further bibliography] in the intro-
duction to this section). On this view the present narrative would be
showing how Jethro, already a worshipper of Yahweh according to
his own (probably very different) tribal tradition, came to a much
higher understanding of Yahweh’s nature and power.
12. Jethro’s rejoicing and praise, as was customary in Israel and
other peoples of the ancient world, leads into sacrificial worship (cf.
H. Ringgren, Israelitische Religion (Stuttgart, 1963), pp. 151-83, ET
pp. 166-200). This connection may first have been made in relation
to the celebration of harvest, but here it is given a specifically histor-
ical foundation as a response to the Exodus deliverance. ‘Brought’
is a possible translation of the Heb. wayyiqqaḥ (from lāqaḥ, more
commonly used for ‘took, received’ [see Note t on the translation])
and will mean ‘offered’ (which is the meaning given by TgO, Sy
and Vulg). The less specific word may have been used (perhaps
by a late scribe rather than the original writer) to avoid implying
the participation of Aaron and the elders in a ritual presided over
by a foreign priest. The offerings comprise, as often elsewhere (cf.
10.25; 20.24; 24.5; 32.6), both a burnt-offering (Heb. ʿōlāh) and
the ‘sacrifices’ (zebāḥîm; sometimes called šelāmîm, ‘offerings of
well-being’: see the Explanatory Note on 10.25) from which only
portions were burnt as offerings to God, the remainder being, as
here, cooked and eaten by the worshippers. It is a little curious that
Moses is not mentioned among the participants in the meal and this
has occasionally been taken as a sign that Aaron, not Moses, was
once Israel’s leading representative in this tradition. But he did not
need to ‘come’ and his presence is probably simply assumed from
the earlier part of the narrative. It has been suggested (e.g., though
he was not the first, by Brekelmans, ‘Exodus xviii’, pp. 218-20; cf.
Cody, ‘Exodus 18.12’) that the meal was a covenant-meal, in this
case to seal a treaty between Israel and Midian: but there is nothing
specific to covenant-making here (e.g. the swearing of oaths) nor any
sign elsewhere of a treaty relationship between Israel and Midian,
and the idea has rightly been abandoned (cf. Houtman, p. 411).
‘Before God’ may be intended to imply the existence of some kind
of (Midianite?) cultic site or enclosure at ‘the mountain of God’,
at least an altar comparable to those which are built in 24.4 and
32.5. But at this holy place (cf. 3.5) the presence of God would be
568 EXODUS 1–18
sufficiently recognised even without a built structure. This ‘worship

at the mountain’ could be meant to be understood as that which was
promised in 3.12 (so Childs, p. 327; Propp, pp. 633-34), but if so the
prominence of Jethro in it would be a highly unexpected element
and it is probably better seen as at most a prelude to the main act
of worship in (part of) 24.1-11 (similarly Graupner, p. 102 n. 358).
Text and Versions

18.1 was evidently preceded by a small mid-line division in 4QpalExm and by
an open line in 4QExc.
‫( כהן מדין‬18.1) On the renderings of Tgg, which avoid words for ‘priest’,
see Text and Versions on 2.16 and one opinion in MRI (2, p. 166).
‫( חתן משׁה‬18.1) For ‫ חתן‬LXX, Theod and Vulg have, as in 3.1 and
elsewhere, the more general equivalents γαμβρός and cognatus. Many LXX
mss adopted Symm’s more precise πενθερός (cf. Tgg, Sy). Aq’s νυμφευτής
seems to be an innovative attempt to preserve in Gk. the etymological
relationship between ‫ ח ֵֹתן‬and ‫( ָח ָתן‬for which Aq, like Symm and Theod,
had νυμφίος in 4.25-26; cf. his use of νυμφεύω, which can mean ‘give in
marriage’, for ‫ חתן‬Hithpael in 1 Sam. 18.21), since it seems never to be used
for ‘father-in-law’ elsewhere.
‫( אלהים‬18.1) The reading of MT is supported by SP, 4QpalExm (4QExc
does not survive for the middle of the verse), Vulg and Sy. Tgg substitute the
divine name, as they do from Gen. 1.1 onwards, while LXX’s κύριος and ‫יהוה‬
in a Geniza ms. (BHS) are probably assimilations to the close context.
‫( למשׁה‬18.1) LXX* had no equivalent, but the word is present in
4QpalExm, SP and the other Vss (as well as the O-text of LXX). The omission
in LXX could well be due to the exclusive concern with Israel as a whole
later in the verse.
‫( כי‬18.1) LXX γάρ and probably Vulg eo quod (though it can mean ‘in
that’: OLD, p. 1565) took ‫ כי‬to be causal here.12 Tgg are ambiguous, because
‫ארום‬/‫ ארי‬can like ‫ כי‬mean both ‘because’ and ‘that’ (but AramB’s renderings
prefer variants on the latter); Sy d points more clearly to ‘that’ (cf. ὡς in LXX
mss).
‫( את־ישׂראל‬18.1) TgN adds ‫ פריקין‬as it tends to do where the verb is ‫הוציא‬
(cf. Text and Versions on 3.10). Sy as often has the fuller expression bny ʾysryl
here.
12
Wevers’ suggestion (Notes, p. 274) that the γάρ-clause and also the ‫כי‬-clause
in the Heb. were intended to be the introduction to v. 2 cannot be right as it stands:
Yahweh’s deliverance of Israel from Egypt could hardly be the reason why Jethro
‘had taken Zipporah etc.’ from there.
18.1-12 569
Before 18.2 4QpalExm has a small mid-line division which is not present
in MT, SP or 4QExc (compare the similarly ‘odd’ division before 16.33 and
others listed in DJD XII, pp. 60-61).
‫( יתרו חתן משׁה‬18.2) Vulg, having attached v. 1 closely as a subordinate
clause to v. 2, does not repeat the subject with the verb.
‫( אחר שׁלוחיה‬18.2) 4QpalExm apparently read ‫ אחרי‬for the ‫ אחר‬of the other
Heb. texts (but only the final letter [partly] survives), perhaps to make its
prepositional character clearer. ‫ שׁלוחיה‬was generally recognised by the Vss
to have a different sense from the other two occurrences of ‫ שׁלוחים‬in BH:
only the corrector of LXXF with προῖκα(ς) preserves the rendering ‘dowry’
here. There was a rabbinic debate over whether a divorce was meant (MRI,
pp. 167-68): LXX τὴν ἄφεσιν αὐτῆς and Vulg’s paraphrase quam remis-
erat (sc. Moses) seem not to imply this but simple ‘sending away’, likewise
TgO,J. Support for ‘divorce’ is stronger in TgN and Sy, since their ‫ שׁבק‬was
used specifically in this sense. One might deduce the same for the (τὰς)
ἐξαποστολὰς (αὐτῆς) of Aq and Theod, since the corresponding verb is used
in Deut. 24.4; but their motive for altering ἄφεσιν could have been purely
linguistic (cf. 1 Kgs 9.16 in LXXA [KR?] and Mic. 1.14LXX). The occasion
for Zipporah’s departure is specified by TgJ and MRI (ibid. R. Eleazar) as
being on Moses’ journey back to Egypt; LXX may have thought the same,
if Wevers was correct to suppose that Zipporah was the intended subject of
ἀπῆλθεν in 4.26 (Notes, pp. 55-56). Josephus (AJ 3.63) seems not to reckon
with any separation at all: probably δεχόμενος is his rendering of ‫ויקח‬, but he
provides no equivalent for ‫אחר שׁלוחיה‬.
‫( בניה‬18.3) LXX and TgN have ‘his (sc. Moses’) sons’ and Vulg’s eius is
ambiguous, but the Heb. texts and the other Vss all agree on ‘her sons’. The
change to ‘his’ will be for contextual reasons (Wevers, Notes, p. 275).
‫( האחד‬18.3) LXX and TgN again agree in adding ‘of them’ to bring out the
connection to ‫ בניה‬made by the relative ‫( אשׁר‬cf. Vulg quorum).
‫( גרשׁם‬18.3) LXX and Vulg represent the name as Gersam/n (cf. 2.22),
which the consonantal Heb. text would allow, but the -ō- of MT is supported
by Syh, TgO,J and Sy.
‫( כי אמר‬18.3) LXX freely (and without any concern for the syntax)
renders with λέγων, a common equivalent for ‫ ;לאמר‬Vulg dicente patre
improves both the grammar and the reference.
‫( גר‬18.3) The Vss mostly render as in 2.22, with TgN again adding ‫ותושׁב‬
(see Text and Versions there); but here it uses ‫ גיור‬for ‫גר‬. This can just mean
‘stranger’, but the other Tgg generally keep it for non-Israelites.
‫( נכריה‬18.3) TgJ again adds ‘which was not my own’, as in 2.22 (read ‫דידי‬
with the CAL text).
‫( ושׁם האחד‬18.4) The Heb. idiom, to repeat ‫ אחד‬with the second of a pair,
is naturally rendered by words for ‘the other’ in TgN, Sy and Vulg. LXX
curiously has the more distant τοῦ δευτέρου, which is perhaps modelled on
1.15 (where Heb. has ‫)השׁנית‬, as its usual practice in such cases is to repeat
‘one’ like the Heb. (cf. TgO,J).
570 EXODUS 1–18
‫( כי‬18.4) TgN and Vulg both add ‘he said’ to introduce the explanatory
direct speech.
‫( אלהי אבי‬18.4) TgN prefixes ‘the Memra of’. For ‫ אבי‬Tgg have the regular
Aram. equivalent ‫אבא‬/‫ אבה‬but Sy, which reserves this form for the vocative
(Gen. 22.7) and generally uses a suffixed form (cf. 15.2), here has ʾbhy, ‘my
fathers’.
‫( בעזרי‬18.4) LXX βοηθός μου and Vulg adiutor meus appropriately use
a personal equivalent for the abstract noun, while the Aram. versions retain
the idiom with ‫ב‬.
‫( מחרב פרעה‬18.4) The unusual expression of MT is supported by all the
available witnesses (including 4QpalExm) except for LXX, which has ἐκ
χειρὸς Φαραώ as in v. 10 (and also vv. 8-9 in LXX). The translator evidently
preferred the milder expression (BAlex, p. 193).
‫( ויבא‬18.5) Wevers prefers the majority reading καὶ ἦλθεν to LXXB’s καὶ
ἐξῆλθεν (see Notes, p. 276). Vulg (venit) ergo… indicates the resumption of
the main narrative after the digression about the boys’ names, following (or at
least agreeing with) Symm’s stylish ἦλθεν οὖν (Wevers, Notes, p. 276 n. 8).
‫( ובניו ואשׁתו‬18.5) LXX does not represent the suffixes, which are no
longer necessary and might even give the mistaken impression that Jethro was
bringing his own wife and sons (Wevers, Notes, p. 277). The Three and the
O-text restored them, Vulg only the first in accordance with good Latin style.
TgJ ‘Moses’ sons and his wife’ was perhaps reinforcing the precedence given
to the children by their prior position, as well as the fact that Moses was their
father (MRI, p. 172).
‫( הר האלהים‬18.5) Most of the Vss insert prepositions meaning ‘at’ or
‘near’, but TgO and Sy ‫( לטורא‬as in 3.1) perhaps see the mountain as Jethro’s
destination rather than Moses’ location. As in 3.1 Tgg replace the designation
of the mountain (and in TgN even the mention of it) with references to the
appearance of God’s glory there (which TgJ emphasises was ‘to Moses at the
beginning’) or (in TgN) to its ‘dwelling’ there: the latter, as Chester points out
(Divine Revelation, pp. 159-65), indicates (like TgN on 4.27 and Num. 10.33)
a desire to foreground the similarity of the desert mountain to the temple on
Mount Zion.
‫( ויאמר‬18.6) MT’s vowels indicate an active verb, of which the subject
can only be Jethro (as in v. 5): the words that he says, ‘I (am) your father-
in-law Jethro’, leave no alternative. Since he only meets Moses in v. 7, this
appears premature. The other witnesses present two different solutions to the
problem. Vulg’s unusual rendering mandavit cannot have its regular sense
‘commanded’; it must mean ‘sent a message’ (cf. LS, p. 1107; OLD, p. 1071),
which corresponds to, and is probably derived from, the explanation given in
MRI (Lauterbach, p. 172) that Rashi, Ibn Ezra and Rashbam were to follow.
LXX, Sy and SamTg read the verb as a passive (‫ )וַ יֵּ ָא ֶמר‬with an associated
variation in the words spoken (see below on ‫)אני‬. LXXO ἀνήγγειλαν δέ, ‘and
they reported’, is a partial correction towards MT’s reading.
18.1-12 571
‫( אל־משׁה‬18.6) Most SP mss read ‫( למשׁה‬compare its variant reading in

17.9), but both 4QpalExm and 4QExc agree with MT. LXX’s dative Μωυσῇ
could be based on either reading (see Text and Versions on 17.9). The
presence of ‫ אל‬in the ‘proto-Samaritan’ ms. 4QpalExm suggests that the ‫ ל‬of
the medieval SP mss is secondary, perhaps due to the influence of Aram., in
which ‫ אל‬was not used. LXX λέγοντες (cf. 12.43; 15.21) and Vulg dicens (cf.
ait in v. 4) add an introduction to the direct speech which follows.
‫( אני‬18.6) LXX ἰδού, Sy hʾ and SamTg ʾh have words meaning ‘behold’
instead of the self-referential pron., and evidence of the Heb. Vorlage of this
variant (‫ )הנה‬exists not only in SP but in 4QpalExm. Together with the reading
of ‫ ויאמר‬as a passive this produces a straightforward text, and some earlier
commentators (e.g. Beer, Hyatt) and translations (RSV, JB, NEB, REB)
adopted it. But as the more difficult reading MT’s ‫אני‬, which appears already
in 4QExc (as well as Tgg and Vulg), must be original (with the active pronun-
ciation of ‫)ויאמר‬.
‫( חתנך יתרו‬18.6) 4QExc and LXX (also ms. 5b1 of Sy) agree in inverting
the order of these words, to conform to the priority of the PN in vv. 1, 2 and
5. The variation here in MT and the other witnesses is an argument for its
originality: the prefixing of the relational word is also natural as a way of
self-introduction.
‫( בא אליך‬18.6) TgJ adds ‫לאתגיירא‬, which can mean just ‘to stay as a guest’
(CAL: cf. TgO on 12.48), but in view of the use of the corresponding active
form by TgJ in v. 27 (and Jethro’s attitude throughout the chapter) must here
refer to conversion to Judaism. Both this and Jethro’s self-effacing depend-
ence on the claims of Moses’ wife and children which follows once again
take up themes from the commentary in MRI (Lauterbach, pp. 172-73), where
they are reinforced by a divine word of encouragement for Jews to welcome
converts.
‫( ואשׁתך‬18.6) LXX* ignored the suffix as in v. 5, but it was represented in
most of the mss and in the other Vss.
‫( ושׁני־בניה‬18.6) LXX οἱ δύο υἱοί σου, with which Sy and Vulg agree,
follows the lead of v. 5 in emphasising that the boys are Moses’ children (too).
‫( עמה‬18.6) LXX and Sy (according to the vocalised text) render ‘with
him’, which in their version of the words spoken (see above on ‫ )אני‬will refer
to Jethro. This would not have required a different Vorlage, only reading the
form with the old third person m.s. suffix, as in pre-exilic inscriptions and
occasionally in MT, e.g. Exod. 22.26 (cf. GK §91e, JM §94h).
‫( ויצא משׁה‬18.7) TgJ adds ‘from under the clouds of majesty’, i.e. outside
the camp, as in 17.9 where MRI had prompted the addition.
‫( וישׁתחו‬18.7) SP as elsewhere (e.g. 34.8) has a fuller spelling ‫ וישׁתחוי‬for
this form (cf. GSH §64c, 98fβ: the pl. form in the apparatus of BHS is an
error). It also adds ‫למשׁה‬, making Jethro the subject of this verb (and presum-
ably the next one) and removing the idea that Moses paid homage to a foreign
priest. The addition appears nowhere else, but it could have been present in
572 EXODUS 1–18
a lacuna in 4QExc (DJD XII, p. 122): this seems less likely in 4QpalExm.
MRI (Lauterbach, pp. 173-74) was aware of the potential ambiguity of MT,
but concluded that Moses continued to be the subject (which would in any
case be the natural supposition) because ‫ אישׁ‬was used of him in Num. 12.3.
TgN, as earlier in 11.8, found the practice of bowing down before another
human inappropriate and again substituted ‘asked about (his welfare)’, despite
the repetition that this produced. LXX and Sy added ‘to him’, presumably
meaning Jethro and acknowledging that such respect was not reprehensible
(cf. MRI, p. 174).
‫( וישׁק לו‬18.7) After the initial gestures of welcome, TgJ added ‫וגייריה‬, ‘and
he made him a proselyte’, fulfilling the desire that TgJ had attributed to Jethro
in v. 6 (see the note on ‫)בא אליך‬.
‫( וישׁאלו…לשׁלום‬18.7) LXX ἠσπάσαντο ἀλλήλους here, as in the A-text
of Judg. 18.15, chooses to emphasise the function of the words spoken;
likewise Vulg, though as elsewhere it lets the warmth of the greeting show
by adding verbis pacificis. Aq and the O-text added εἰς εἰρήνην, but perhaps
more for textual accuracy.
‫( ויבאו‬18.7) There is a well attested variant ‫ויביאהו‬, ‘and he (sc. Moses)
brought him (sc. Jethro) in’ (4QExc, SP, LXX: in 4QpalExm only the first three
letters of the word survive, leaving its reading uncertain). This makes good
sense too and reinforces the portrayal of Moses’ welcome of his father-in-law.
But it is perhaps more likely to be an ingenious ‘improvement’ of the MT
reading than vice versa: in an unvocalised text ‫ ויבאו‬could have been misread
as a Hiphil form and the final ‫ו‬- as a form of the object suffix (cf. Jer. 23.6).
Vulg cumque intrasset need not be evidence for a reading ‫וַ יָּ בֹא‬: the sing. may
simply be due to the combination of the clause with v. 8.
‫( האהלה‬18.7) TgJ adds that it was ‘a place for teaching’, as is found
already in MRI (Lauterbach, pp. 174): the gloss was probably inspired by what
Moses does in v. 8.
After v. 7 SP has a division which is not present in MT and probably did
not appear in 4QpalExm or 4QExc (cf. DJD XII, p. 122).
‫( לחתנו‬18.8) LXX ignored the suffix, as Greek grammar allows, but the
Three and the O-text added αὐτοῦ.
‫יהוה‬1o (18.8) TgNmg prefixes ‘the Memra of’, as again at the end of the
verse.
‫( ישׂראל‬18.8) Sy as often prefixes bny (cf. 17.6; 18.1).
‫( את כל־התלאה‬18.8) This is also the reading of SP, 4QExc (the phrase is
not extant in 4QpalExm), Tgg and Vulg, but LXX and Sy prefix ‘and’, as do a
Geniza ms. and some other Heb. mss Tolerable sense can be made of the verse
without this (see Note n on the translation) and it is probably secondary. TgN
renders ‫ התלאה‬in the pl., but it is always sing. in BH.
‫( אשׁר מצאתם‬18.8) Sy dlʾyw offers a free rendering which makes a neat
etymological link with ‫התלאה‬.
18.1-12 573
‫( בדרך‬18.8) TgJ spells out the episodes involved, from the Red Sea to
the battle with Amalek, developing the interpretation already given in MRI
‫( ויצלם‬18.8) LXX inserts ὁτι for stylistic reasons.
‫( יהוה‬18.8) LXX adds ‘from the hand of Pharaoh and from the hand of the
Egyptians’, anticipating Jethro’s response in v. 10.
‫( ויחד‬18.9) LXX ἐξέστη is a general word for mental disorientation,
perhaps chosen because the translator was puzzled by the rare Heb. verb and
its form here: there is no need to posit a different Vorlage. MT’s reading is
confirmed by SP, 4QpalExm and 4QExc and the other Vss render it correctly.
‫( על־כל הטובה‬18.9) LXX, Sy and Vulg render in the pl., a reasonable
interpretation of the sing. form (which is attested in 4QExc as well as MT
and SP).
‫( יהוה לישׂראל‬18.9) LXX αὐτοῖς κύριος, with a change of word-order in
accordance with Greek style and the substitution of a pronoun for the proper
name of the original. TgNmg prefixes ‘the Memra of’ as usual.
‫( אשׁר הצילו‬18.9) LXX (ὅτι) and Vulg (eo quod) both render ‫ אשׁר‬as a
conjunction, ‘because, (in) that’. The Leiden text of Sy reads wpṣy with most
of the mss, but this probably originated in a scribal error due to the similar
phrase in the previous verse: dpṣy, the reading of 7a1 and 8b1, is more likely
to be original. (on this pair see Koster, Peshiṭta, pp. 148-49, 169-70).
‫( יתרו‬18.10) Vulg does not repeat the subject, which is the same as in v. 9.
‫( יהוה‬18.10) TgJ prefixes ‘the name of’, reflecting Jewish liturgical
formulae which have their roots in the Bible (Ps. 113.2; Job 1.21: cf. Chester,
Divine Revelation, pp. 343-47).
‫אשׁר‬1o (18.10) LXX ὁτι, ‘because’, as in v. 9, might reflect the frequency
of ‫ כי‬near the beginning of hymns of praise in the Psalter. The rarer ‫אשׁר‬, which
all the other witnesses have or presuppose, is undoubtedly original.
‫( אתכם‬18.10) LXX τὸν λαὸν αὐτοῦ is taken from the second half of the
verse (which LXX omits: see below), with the addition of the possessive ‘his’,
as in Vulg and Sy there (from v. 1?), for greater specificity. A few mss replace
it with αὐτούς or ὑμᾶς.
‫( מיד‬18.10) For both occurrences TgN has the dual form of ‫יד‬, which (like
Tg ) it uses after ‫ מן‬much more frequently than BH (cf. OL).
J
‫אשׁר…מצרים‬2o (18.10) LXX has no equivalent to the second relative

clause, but it shows that the translator knew it by the introduction of τὸν λαὸν
αὐτοῦ earlier in the verse. So this is not a case of parablepsis but of delib-
erate omission to avoid what was judged an unnecessary repetition (likewise
Wevers, Notes, p. 279). The O-text added a rendering of the omitted words to
agree with MT and the other Heb. witnesses (most of the words are present in
4QpalExm and there is plenty of room for them in a lacuna in 4QExc).
574 EXODUS 1–18
‫( יד‬18.10) TgO,J recognise the figurative sense here by using ‫מרותא‬,

‘dominion’, while TgN is even more specific with ‘the yoke of servitude’ (cf.
MRI [Lauterbach, p. 176]), which it also has in 6.6-7, where ‫ מתחת‬again occurs
in the Heb.
‫( ידעתי‬18.11) TgO ‫ידענא‬, like the other Vss, renders the Heb. precisely:
it is the first person sing. of the ‘participial tense’ (cf. Stevenson §21.7). TgN
added ‫וגלי הוא לי‬, which could be no more than a rephrasing of the Heb. (cf.
Chester, Divine Revelation, p. 20); but in the context of TgN’s wider use of the
verb (ibid., pp. 246-49) the addition of the expression to the straightforward
rendering that precedes may suggest that Jethro is made to acknowledge a
revelatory basis for his knowledge.
‫( גדול יהוה מכל־האלהים‬18.11) TgO ‘is great and there is no god except him’
makes Jethro into a monotheist; on the other hand MRI (Lauterbach, p. 176)
takes the text at its face value and observes his inferiority to Naaman in 2 Kgs
5.15. An even greater disdain for Jethro’s conversion is shown by Philo (De
Ebr. 36-45). It is not so clear what the close renderings of the Heb. in the other
Vss imply, but Chester is probably right to say that the strong repudiation
of polytheism elsewhere in TgN and TgJ makes it most improbable that their
wording here intends anything different (Divine Revelation, pp. 332-34).13
‫( כי בדבר אשׁר זדו עליהם‬18.11) Apart from Tgg the Vss attempt a literal
rendering of the difficult Heb. (which SP and 4QpalExm [with a lacuna in
the middle] also attest – 4QExc preserves only [most of] the final word) and
show that it must be the starting-point for any solution to the text-critical
problem. LXX ἕνεκεν τούτου ὅτι and Vulg eo quod make the first three
words a way of saying ‘because’ (cf. Symm, Theod, if the citations from them
are complete), while Sy mṭl trʿytʾ d and especially Aq ὅτι ἐν τῷ ῥήματι ᾧ
preserve more of the original wording. For ‫ זדו‬LXX has ἐπέθεντο, ‘attacked’,
which is clearly based on the use of ‫ זיד‬in 21.14, where LXX uses the same
Gk. equivalent. But its other occurrences show that ‘arrogance’ is an essential
component of its meaning and this was recognised by the Three and Vulg
(superbe egerint). Sy’s ʾtrʿyw (which gave rise to its equivalent for ‫)דבר‬,
here ‘think, desire, plan’, normally lacks a specifically negative overtone but
seems to be related to TgO,N ‫חשׁיבו‬, which introduces a fuller description of the
Egyptians’ evil plan: perhaps that was thought to be sufficiently conveyed in
Sy by the hostility implied in ʿlyhwn, ‘against them’ (sc. Israel). The Tgg are
all much more explicit about what they understood this to mean, even TgO:
‫ארי בפתגמא דחשׁיבו מצראי למדן ית ישׂראל ביה דנינון‬, ‘for by the means which the
Egyptians planned to use to punish Israel, by that he punished them’ (cf.
MRI [Lauterbach, pp. 176-77], where ‫ חשׁב‬is also used). TgJ specifies water
as the means (and makes a characteristic reference to the ‘wickedness’ of the
13
Josephus’s treatment of the chapter (AJ 3.63-74) practically ignores Jethro’s
religious belief.
18.1-12 575
Egyptians’ action) and TgN identifies the events in question as the drowning of
the Israelites’ children and (through a decree of God’s Memra) the catastrophe
for the Egyptian chariots at the Red Sea. All these renderings assume that
there is an unspoken declaration of God’s judgement at the end of the verse.14
To summarise, while LXX, Vulg and Sy create the appearance of a complete
sentence by effectively ignoring either ‫ כי‬or ‫בדבר אשׁר‬, the Tgg (and OL) do
so (more successfully, it must be said) by adding a whole clause about God’s
intervention (see also the next note on 4QpalExm).
In modern times C.F. Keil sought to justify the former approach by taking
‫ כי‬as an emphatic repetition (see Note s on the translation), but he had in effect
to supply a verb to complete the sense and this is a solution born of despera-
tion. An alternative might be to regard ‫ כי‬as a gloss on the unique expression
‫בדבר אשׁר‬, which could have been intended (like ‫ על דבר אשׁר‬in Deut. 22.24;
23.5; 2 Sam. 13.22) as an expression for ‘because’. But both of these inter-
pretations leave Jethro with a statement that makes no real sense: he believes
that Yahweh is the most powerful of the gods because the Egyptians acted
arrogantly towards his people. Something more is surely needed to justify his
exaltation of Yahweh.15 Some scholars have found the ‘missing link’ nearby in
the biblical text, in v. 10b, which is not needed there and appears at first sight
not to have been in LXX’s Vorlage. Dillmann (p. 186) suggested that v. 10b
was originally (presumably without ‫ )אשׁר‬the conclusion of v. 11, was acciden-
tally omitted and was then reinserted in the wrong place (i.e. where it is in the
standard Heb. text: cf. BH3, Beer, Childs, RSV, NRSV). This is an ingenious
theory, but it is not easy to see any reason why the original omission at the end
of v. 11 would have taken place and the fact that (as we have shown above) the
LXX translator very probably had the whole of v. 10 in his Vorlage removes a
vital piece of supporting ‘evidence’. As a result other scholars have concluded
that what is missing can no longer be restored (so Baentsch, Holzinger, Noth:
cf. BHS). But perhaps some progress towards a solution can still be made.
An omission of text can often be traced to the recurrence of the same
word or group of letters in the original, and the readings of the Targumim,
with their theory of a correspondence between wrongdoing and its punish-
ment, which is grounded in the most obvious reading of ‫בדבר אשׁר‬, imply
just such a recurrence. It is unlikely that any of them is a translation of the
original Heb. text, because none of them provides a plausible rendering of
14
For completeness we may note that OL uses a different expedient to
complete the sense, by substituting qui liberαvit famulos suos de manu eorum
for ‫כי בדבר‬. On the verse as a whole see J.R. Baskin, Pharaoh’s Counsellors:
Job, Jethro and Balaam in Rabbinic and Patristic Tradition (Chico CA, 1983),
pp. 54-55, 61-66.
15
For some less likely solutions than those discussed here see Houtman,
p. 409, to which should be added Weimar, Berufung, p. 28 n. 25, and Blum,
Studien, p. 158 n. 253.
576 EXODUS 1–18
‫זדו‬. It is notable that proper names are absent in the sentence as it stands, so
the omitted text probably stood before ‫ עליהם‬and not after it (this is a further
objection to Dillmann’s view), and included at least one proper name. It is
also likely that it began with ‫על‬, because that would explain how it came to
be omitted (by homoeoarkton). <‫זדו >על ישׂראל‬, ‘they (sc. the Egyptians in
v. 10) acted arrogantly against Israel’, has a good parallel in Jer. 50.29: ‫אל‬
‫( יהוה זדה‬sc. Babylon).16 ‫עליהם‬, in the corresponding act of punishment, must
now refer to the Egyptians and before it there must have been an expression
for Yahweh’s intervention against them which could be followed by ‫על‬. It
is unlikely that this included the verb ‫זיד‬, as it could scarcely be applied to
Yahweh. But there are at least two other possibilities. One is an expression
incorporating the verb ‫פקד‬, which is commonly followed by an accusative of
the sin punished and ‫ על‬before the perpetrator of it (e.g. Hos. 1.4; 2.15; Amos
3.2, 14). There are many possible alternatives for the word for the Egyptians’
sin, but of course ‫ זדון‬would have particular point: so perhaps the text read:
‫אשׁר זדו >על ישׂראל הוא פקד את־זדונם< עליהם‬. Alternatively, and more briefly, one
might supply the verb ‫גבר‬, ‘prevailed’: then ‫אשׁר זדו >על ישׂראל הוא גבר< עליהם‬.
For ‫ גבר‬with ‫ על‬of enemies cf. 2 Sam. 11.23 or, in the Hithpael with a divine
subject, Isa. 42.13. An advantage of this, perhaps, is that Neh. 9.10b could be
seen as a paraphrase of it (cf. n. 16).
Between vv. 11 and 12 4QpalExm had a short mid-line division. There is
no obvious reason for this, but it is where Tgg detected a gap in the Heb. text
which they filled in various ways (see the previous note) and possibly the
scribe of 4QpalExm was showing his awareness of it in this way.
‫( ויקח‬18.12) While SP, 4QExc, LXX and TgJ,N agree with MT, in TgO, Sy17
and Vulg words meaning ‘offered’ appear and some modern commentators
have emended to ‫ וַ יַּ ְק ֵרב‬on this basis (cf. BH3; RSV). The change is unneces-
sary, as ‫ ויקח‬in effect has the sense ‘brought’ here (see Note t on the translation)
and the Vss cited in support of it may simply have been recognising this.
‫( חתן משׁה‬18.12) The first word of the verse that is preserved in 4QpalExm
is ‫ [עלה‬and there is probably insufficient room in the lacuna for all of the
standard Heb. text. The scribe probably omitted ‫ חתן משׁה‬on this occasion, as
all the witnesses had done in vv. 9-10. On this occasion the shorter reading is
probably secondary.
16
‫ כי הזידו עליהם‬in Neh. 9.10 (with the Egyptians as the subject) is even closer
and could well be dependent upon Exod. 18.11 (Boda, Praying the Tradition,
p. 119). But the relationship of the two texts would be just as easy to envisage if
Exod. 18.11 were still in its original complete form as that is reconstructed here.
17
This is the reading of the mss. But Weitzman has pointed out that Ephrem’s
commentary cites the reading wnsb (as in TgJ,N), which ‘agrees with MT and
seems original’, while Ephrem’s comment suggests the interpretation qrb (Syriac
Version, pp. 289-90).
18.1-12 577
‫( עלה‬18.12) All the Vss render in the pl. (Sy yqdʾ šlmʾ is evidently a
calque from Gk. ὁλοκαυτώματα), giving ‫ עלה‬a collective sense (which it
probably has in 2 Chr. 7.1). This is possible, and narrative texts generally
have the pl. form (cf. 10.25); but a single animal could well have sufficed on
this occasion.
‫( וזבחים‬18.12) On the rendering ‫ נכסת קודשׁין‬which all the Tgg (inc. TgN)
have here see Text and Versions on 10.25.
‫( לאלהים‬18.12) TgO,J,Nmg use ‫ קדם‬in place of ‫ ל‬and TgN inserts ‫ שמיה ד‬to
avoid the idea that God receives offerings directly. Sy substitutes the divine
name (i.e. mryʾ) for ‫ אלהים‬here (likewise Tgg, as usual, as they do again at the
end of the verse).
‫( ויבא‬18.12) Vulg veneruntque was required by the composite subject by
the strictness of Latin grammar and need not reflect a different Vorlage.
‫( וכל זקני ישׂראל‬18.12) SP ‫ומזקני ישׂראל‬, ‘and some of the elders of Israel’, is
a puzzling variant. All the witnesses at 17.5 refer to only ‘some’ of the elders
accompanying Moses when he struck the rock at Rephidim (cf. 24.1, 9; Num.
11.16, 24), and SP seems to have presumed that the same would be the case
for this celebration with Jethro. Whether this has also something to do with the
section on ‘The Status of the Elders’ in Memar Marqah 3.3, where the twelve
tribes are divided into two groups, blessed and cursed, must remain uncertain,
as there appears to be no reference to the story of Jethro in that work. The
phrase is not preserved in 4QpalExm or 4QExc and the variation is too short for
inferences based on the space available to be valid (contra DJD XII, p. 122),
but all the Vss presuppose ‫וכל‬. TgN as usual renders ‫ זקני‬by ‫( חכימיא‬cf. 17.5-6);
Sy as often inserts bny afterwards.
‫( עם־חתן משׁה‬18.12) Vulg cum eo, evidently regarding a repetition of
Jethro’s status as unnecessary.
‫( לפני האלהים‬18.12) TgJ adds ‘and Moses was standing and ministering
(‫ )משמש‬before them’ in wording modelled on Gen. 18.8 and MRI (Lauterbach,
pp. 177-78), to account for the fact that Moses is not named among those
who ate.
After v. 12 SP has a division before the second episode involving Jethro
and it is likely that 4QpalExm did too, although DJD IX does not comment: in
the lacuna at the end of col. XVIII 19 there is room for about ten characters
and only the word ‫ ויהי‬is needed there before ‫ ממחרת‬at the beginning of the
next line. In 4QExc there was at most space for a ‘very short’ interval at this
point (DJD XII, p. 121), but the fragmentary state of the ms. here (after which
continuous text is no longer preserved) makes even this uncertain.
C h ap t er 1 8 . 1 3 - 2 7
J et hr o ’ s A d vi ce a bou t a J u d i c i a l S yst e m
for I s r a el an d H i s D epart u r e
The boundaries of this section are unusually clear. The second

episode of Moses’ renewed encounter with Jethro is marked off by
a change of time, which nevertheless by its very character (‘On the
next day’) makes it the direct sequel of the first. A new topic, the
judging of the people’s disputes, is also introduced. The section is
rounded off with the adoption of a new judicial system (vv. 24-26)
and Jethro’s departure for his homeland (v. 27). In 19.1-2 the repet-
itive itinerary-note marks a fresh beginning on the literary level,
even though geographically arrival at ‘the mountain’ (19.2) seems
to overlook the fact that Israel is already encamped ‘at the mountain
of God’ (18.5). In the Masoretic text the whole of ch. 18 is treated
as a single unit and there is no further division marked until after
v. 27. The Samaritan Pentateuch does have a division before v. 13
as well as after v. 27 and it also subdivides the section internally
after v. 16, where Jethro’s advice to Moses begins, and after v. 24,
which states Moses’ acceptance of it. 4QpalExm may well also have
had a division before v. 13 and there is strong indirect evidence that
it did after v. 27 (DJD IX, p. 99: see Text and Versions). The other
Qumran mss offer no clear evidence about their division of the text.
The overall structure of the narrative in most of the textual wit-
nesses is straightforward and clear. On the day following Jethro’s
arrival (i) Moses is occupied with judging the people’s disputes
(v. 13), which sets the scene for what follows; then (ii) Jethro ques-
tions Moses about his current practice, in a manner that already
expresses an element of criticism (cf. ‘Why?’: v. 14), and (iii)
Moses explains his role as a mediator of God’s responses to the peo-
ple’s disputes (vv. 15-16: on the nature of these responses see the
Explanatory Note on vv. 13-16); Jethro then (iv) highlights the prac-
tical disadvantages of the present arrangements (vv. 17-18), and (v)
urges Moses to take his advice and introduce a new two-tier system
of adjudication, in which judges chosen by Moses and of suitable
character and ability will deal with the more straightforward cases
18.13-27 579
and only the most serious will be referred to Moses for resolution
(vv. 19-23); Moses (vi) accepts this advice (v. 24) and (vii) proceeds
to put the new system into operation (vv. 25-26); this done (viii) he
sends Jethro on his way to his homeland (v. 27).
In 4QpalExm, however, the first part of (vii), v. 25, is replaced
by a much longer account in which Moses first presents the new
system to the people for their approval and then speaks directly to
the judges about their responsibilities. This version of the appoint-
ment of the judges is in essence identical to that which appears in
Deut. 1.9-18 (without any mention of Jethro at all) and the same
expansion of the Exodus story was taken up in the Samaritan tradi-
tion, where it is attested not only in the Hebrew Pentateuch text
but in Greek and Aramaic translations (for the details see Text and
Versions on v. 25). There is no doubt that this longer version of
the text (which is comparable in character to the similar expan-
sions of the plague-story in chs. 7–11 in these textual witnesses) is
a secondary development designed to reconcile (in some measure
at least) the Exodus story with what happens in Deuteronomy. The
further question of whether the Exodus version (with Jethro as the
initiator of the new system) or that in Deuteronomy (where Moses
presents it as his own idea) is the older will be discussed later in this
introduction.
When it has been examined in its own right, Exod. 18.13-27
has been found to be a generally coherent narrative which forms
a plausible sequel to vv. 1-12. Here too the narrator (and in this
case the human speakers too) uses ʾelōhîm, ‘God’, consistently in
place of the divine name Yahweh. Jethro is referred to throughout as
Moses’ father-in-law (cf. v. 14 etc. with v. 1), but not by name as he
had been several times in vv. 1-12. Critical study of the section has,
however, raised some problems about the unity of vv. 13-27 and
their relationship to vv. 1-12, which of course deal with a different
topic.
For Knobel (Num.-Jos., pp. 532-33) and very many scholars after him the
passage’s repeated and exclusive use of ʾelōhîm, in the context of the Penta-
teuch’s wider practice, was in itself a clear indicator that, as in at least the
majority of vv. 1-12, they again had before them an extract from the E source.
Many maintained this without qualification (‘ein im Ganzen heiles Stück aus
E’: Wellhausen, Composition, p. 80; cf. Noth, p. 117, ET, p. 146), but gradu-
ally some commentators identified certain verses which they considered to be
secondary additions: v. 20 (Holzinger), vv. 21b, 25-26 (Baentsch), vv. 21b,
580 EXODUS 1–18
25b (Knierim, ‘Exodus 18’, pp. 167-71: similarly Hyatt, Fritz [Israel in der
Wüste, p. 14 – ‘26b’ is a misprint] and see further below). This kind of analysis
was taken further by M. Rose (Deuteronomist und Jahwist: Untersuchungen
zu den Berührungspunkten beider Literarwerke [ATANT 67; Zurich, 1981),
pp. 229-31: vv. 16b, 20, 21b, 25b) and C. Schäfer-Lichtenberger (‘Exodus
18 – Zur Begründung der königlichen Gerichtsbarkeit in Israel-Juda’, DBAT
21 [1985], pp. 61-85: also v. 15b), and especially by Graupner, who included
vv. 19b, 25a and 26 in the secondary layer(s) (Elohist, pp. 107-109: cf. his
‘Exodus 18,13-27 – Ätiologie einer Justizreform in Israel?’, in S. Beyerle et
al. [eds.], Recht und Ethos im Alten Testament. Gestalt und Wirkung [FS H.
Seebass; Neukirchen, 1999], pp. 11-26). A different approach to such supposed
unevenness in the text was taken by Smend (Erzählung, pp. 154-55, followed
by Eissfeldt and Fohrer, but not in this case by Beer) and Gressmann (Mose,
p. 168 n. 2; Anfänge, pp. 87-88), who in somewhat different ways unravelled
both a J and an E version of the episode: their arguments were convincingly
rebutted by Rudolph (Elohist, p. 39). Rudolph himself remarkably attributed
the passage (except for v. 20, which was too ‘sittlich’ for this!) to J, claiming
that the recurrent use of ʾelōhîm was simply due to the fact that Jethro was a
foreigner (p. 40), despite acknowledging that vv. 10-11 had shown Jethro as
(now) having no difficulty with the use of the divine name. A similar view was
later taken by Van Seters (Life, p. 209).
Since the late 1980s scholarly assessment of this section has become more
fragmented, due to the influence of wider currents in recent Pentateuchal
scholarship, which do not always push in the same direction. An example
is the way in which comparisons with Deuteronomy have been handled.
A forerunner of this recent preoccupation was the monograph of E. Junge,
Die Wiederaufbau des Heerwesens des Reiches Juda unter Josia (BWANT
75; Stuttgart, 1937), which observed the correspondence between the appar-
ently military titles given to the new judges in vv. 21b and 25b and what he
took to be a revival of the old tribal conscript army in the time of Josiah.
He saw this as indicating that the passage could at least not have reached its
completed form until then (pp. 56-59). Dozeman, who assigns it to his older
‘Non-P History’ but emphasises the close associations of that work with
Deuteronomy (pp. 15-16, 362), seems to incline to a similar view. But others
have not hesitated to place it much later, reversing the older view that Deut.
1.9-18 was dependent upon Exodus 18: so Van Seters (Life, pp. 215-17: post-
Deuteronomistic Yahwist) and Albertz (‘Hexateuch redactor’, contemporary
with Nehemiah: p. 314). Johnstone also holds that Exodus 18 was based on
Deut. 1.9-18 (or perhaps its narrative reflex which once stood in Numbers?),
but that it owes its present form and position to the P-writer/redactor whose
hand Johnstone believes he can trace in some features of it (Chronicles and
Exodus, pp. 258-59). Blum’s present position is ambivalent, in more than
one way, but he certainly concluded earlier and apparently still holds that the
passage (like vv. 1-12) formed no part of either Kd or Kp: its isolation from
18.13-27 581
the surrounding narratives and its contradiction of their geographical sequence

show that it was inserted into the Pentateuch at a very late stage (Studien,
pp. 153-63). In his original study he did not presume that it had only been
created then: he saw it, with parts of Exodus 3–4 (and vv. 1-12) to which it is
closely linked, as originating in ‘a separate connected Moses-tradition’, which
had been drawn upon by Kd and so was presumably a survival from pre-exilic
times. In a more recent study and in the work of his pupil V. Haarmann this
proposal seems to have been abandoned for vv. 1-12 (see the introduction to
that section), but it is not clear whether he still holds to it for vv. 13-27.
In stark contrast to the contemporary scholars who envisage an exilic
or later date for the creation of 18.13-271 stand not only those who have
continued to uphold the existence of E as a source-document in the tradi-
tional sense (such as Propp, Graupner and Baden) but another group who
have concluded that the passage or most of it has its origin in the monarchy
period for quite different reasons, not literary but legal-historical. They take
their lead from Knierim’s 1961 study (and those of some followers such as
H. Reviv [‘The Tradition concerning the Inception of the Legal System in
Israel: Significance and Dating’, ZAW 94 (1982), pp. 566-75] and C. Schäfer-
Lichtenberger), which placed the episode in the context of what can be known
about the development of courts and legal practice in early Israel and under
the monarchy. It is important to emphasise that Knierim’s argument was based
on much more than a simplistic trust in and use of the Chronicler’s account
of Jehoshaphat’s legal reforms in 2 Chronicles 19 (contrary to the brief
summary by Van Seters in Life, p. 218 n. 31).2 In seeking to identify the Bild
which Exodus 18 presents, he noted the strong cultic context which vv. 1-12
give to Moses’ judging and that the process of decentralisation is initiated
‘from above’ (by Moses advised by Jethro). Since there is no (other) explicit
evidence of local courts in the Judges period (p. 157), whereas by the time
of the eighth-century prophets and Deuteronomy they are well established
(pp. 158-60), the central authority which represents (one might better say, is
represented by) Moses’ role in establishing them on an official basis must be
that of the king. The evidence of Isaiah and Micah shows that the ‘reform’
occurred before their time and it is only in trying to specify its occasion more
precisely that Knierim introduced into the discussion 2 Chronicles’ account
of Jehoshaphat’s action (esp. in 19.5-11), which he held (like Rudolph, but
against the view of many other scholars) to contain historical information from
an older source. Finally there is the question of the description of the judges as
‘officers for thousands etc.’, expressions which have generally been taken to
1
To those mentioned already may be added Levin (Jahwist, pp. 359-61) and
Kratz (Komposition, pp. 246-47).
2
Blum in his earlier work referred very positively to the work of such scholars
as confirming that the tradition in Exod. 18.13ff. underlies that of Deut. 1.9-18 and
is an independent, older narrative (Studien, p. 157 n. 247).
582 EXODUS 1–18
indicate a military background (cf. 1 Sam. 8.12; 22.7 etc.). Knierim discusses
this several times (pp. 149-51, 154-55, 167-72) and eventually concludes, for
both literary and historical reasons, that vv. 21b and 25b are later additions to
the text, which reflect the combination of military and judicial roles by royal
officials (śārîm) from at least the eighth century onwards.3
More recently three scholars working more broadly on Israelite legal
history have reached similar conclusions to Knierim. One, Crüsemann, devel-
ops the argument that prior to the monarchy there were, contrary to almost
all previous opinion, no local courts at all: the evidence is that disputes were
resolved by other means (Die Tora, pp. 80-95, ET, pp. 63-76). It was only
under the monarchy that they were established, by royal authority: before
the fortification of cities there could be no ‘justice in the gate’, because
there were no gates. 1 Kings 21 shows that this had already taken place by
the mid-ninth century in the northern kingdom (pp. 99-100, ET pp. 78-79):
Exodus 18 provides the legitimation for this change and comes from the
monarchic period, unlike the related but different passages in Deut. 1.9-18
and Num. 11.11-12, 14-17, 24-25, which are later (pp. 104-13, ET pp. 83-90).
Crüsemann, like Knierim, also finds historical material in 2 Chronicles 19
and regards Exod. 18.21b and 25b as later additions. Bernard Levinson, by
contrast, in a paper given in 2001 and first published in 2005, finds no old
source behind 2 Chronicles 19 and apparently regards vv. 21b and 25b as
integral parts of Exodus 18 (see “The Right Chorale”, pp. 62-68, esp. p. 62
n. 29 and pp. 65-66). But he emphasises the debt of Israel’s bipartite judicial
system to ancient Near Eastern culture that emerges from Exodus 18 and that
it was heavily revised and ‘corrected’ in the ‘much later’ account in Deut.
1.9-18. Thirdly B.S. Jackson agrees, with his own reasons, that there were
no local courts until the monarchy period: the dual system was established
then, probably by Jehoshaphat (Wisdom-Laws, pp. 411-30; for the basis of
2 Chr. 19 in older tradition see p. 412 n. 128).4 Deuteronomy provides further
evidence of its (intended) operation. Jackson notes that the Exodus 18 account
preserves the memory of an earlier, purely oracular, mode of arbitration and
that its recommendations for the future conduct of both ‘central’ and ‘local’
practice differ somewhat from those in Deuteronomy and 2 Chronicles. He
does not discuss the date of Exodus 18 in detail, but he appears to hold that it
reflects an earlier stage of development than Deuteronomy, which ‘suppresses
the role of Jethro’ (p. 423), as would fit these differences. He also sees that
Moses’ acceptance of the advice of a foreigner was based on practical grounds
and would have served to support such openness in later times (pp. 422, 423
n. 181).
3
By contrast Knierim regarded it as possible that vv. 16b and 20, which added
complications to the older tradition, might have been introduced by the author of
E himself (pp. 154-55).
4
Jackson amplifies his argument in ‘Law in the Ninth Century’.
18.13-27 583
Cook (‘The Tradition of Mosaic Judges’) wrote before Levinson’s and

Jackson’s studies were published and he appears not to have known Crüse-
mann’s book. But he too follows the general line of Knierim’s approach,
although with two important modifications. First, he places Exodus 18 within
an ‘E-stream’ of texts and traditions, in which Num. 11.11-12, 14-17, 24-25
occupies an even earlier position; and secondly he makes the important point
that Exodus 18 is not simply providing validation for the monarchic changes
in judicial arrangements (in which vv. 21b and 25b reflect a later stage of
development than the rest of the passage): as with other features of the E
source it also seeks to ‘temper’ and ‘control’ the decentralisation of power
by making ethical demands of the judges (cf. v. 21). To him this suggests, as
H.W. Wolff proposed for the prophet Micah (cf. 3.1-4, 9-12), an origin for the
author among conservative village elders.
Our assessment of this review of earlier scholarship can appro-

priately begin with a consideration of the arguments for a late dating
of the passage as a whole in the exilic or post-exilic period, since
this has recently become a widely held view (though as we have
just shown it is by no means universally held at the present time).
The first scholar to isolate Exodus 18 from the main components
of the Pentateuch was Erhard Blum.5 In 1990 he presented some
arguments for its late incorporation into the Pentateuchal narrative.
At that time they were not intended to be arguments for the late
origin of the material concerned, which Blum then believed to be
part of an older Moses-tradition also used by Kd in Exodus 3–4.
As noted above, he has subsequently abandoned that earlier dating
for vv. 1-12 but he has not made clear what he now thinks about
vv. 13-27. However, it is hard to see how he could maintain an
early origin for these verses when they depend upon the preceding
section which he holds to be post-exilic, especially as he has already
conceded that they are an isolated late addition to the Exodus story.
His reasons for this latter conclusion therefore need some examina-
tion here. He makes two points: (i) 18.13-27 (like vv. 1-12), at least
in their present position, stand outside both of the main composi-
tional layers which he has detected in Exodus, Kd and Kp (Studien,
pp. 153-55); (ii) these verses include some expressions which are
‘late’ (p. 158).
5
Rose had questioned the arguments for an E origin (pp. 224-26) in 1981, but
he saw the passage as mainly ‘ancient’ (p. 225).
584 EXODUS 1–18
In the former case, the lack of any connection with Kp may read-
ily be granted6: this is of course just as true of much of the older
material which Blum assigns to Kd, both earlier in Exodus and
in the Sinai-pericope. The situation with Kd is more complex, as
Blum himself has to recognise. There are in fact narrative connec-
tions with Kd in Exodus 3(–4), in the figure (and name) of Moses’
non-Israelite father-in-law; and the similar episode that is recapitu-
lated in Deut. 1.9-18 (similar even in some details of wording) must
point to some kind of literary connection between that passage and
this, even if the nature of the connection has for the moment to
remain uncertain (but see below). The older medieval and source-
critical concern about the location of this episode before the Sinai
legislation, to which Blum also refers, was based on what is prob-
ably a misunderstanding of vv. 16 and 20, as if they presupposed
an already existing body of law: if these verses refer either to a
simple delivery of verdicts or to a ‘cumulative’ judicial tradition
based on precedents (see Note l on the translation and the Explana-
tory Notes), one might actually expect them to precede rather than
follow any ‘code’ of law, especially as they deal with legal matters
from a very different point of view from what follows in chs. 19–24.
The supposed linguistic signs of lateness do not count for much.
The words for ‘commandments’ in vv. 16 and 20 (if that is what
they mean: see above) do not recur as a pair in either of the verses
that Blum cites (15.26; 16.28) and both of them appear elsewhere in
pre-exilic texts. The verb ‘teach’ in v. 20 does indeed also occur in
2 Chr. 19.10 in a similar context (though not certainly in the same
sense), which Blum finds ‘noteworthy’ (p. 158); but there is noth-
ing to suggest that the Exodus author took it from there rather than
vice versa.
An argument for a late date that Blum did not use (because he
believed that in essentials the reverse was the case: likewise Rose,
pp. 224-26, 263) was that the Exodus version is dependent upon
Deut. 1.9-18. But this argument was soon to be deployed by Van
Seters and Johnstone, and it has been revived by Albertz. Since
Deuteronomy 1–3 is generally attributed to a late supplement to the
Deuteronomic law-code, this argument would if accepted indeed
imply a post-exilic origin for Exod. 18.13-27. But, as Levinson has
6
The ‘problems’ with the gap in the itinerary between 17.1 and 19.1-2 (p. 154)
probably have more to do with the itinerary-material itself than with the narrative
texts which it frames.
18.13-27 585
well remarked at the end of a detailed discussion (“Right Chorale”,

p. 66 with n. 36): ‘The highly selective, point for point adjustment
both of the chronology and the aetiology of the judicial adminis-
tration can only be explained in terms of the authors of Deut. 1
consciously seeking to revise and correct the narrative of Exod. 18’.
In a footnote he adds: ‘It is all but inconceivable that this argument
could be reversed, making Exod. 18 the later text that revises and
corrects Deut. 1. The move from a problem-free to a problematic
text is most unlikely.’ This is surely correct: when Deuteronomy
has presented an aetiology of the two-tier judicial system, which it
commends in 16.18-20; 17.2-12, in terms of Moses’ initiative and
in the ‘logical’ place after the covenant at Horeb, who is going to
manufacture a narrative which makes this the idea of his foreign
father-in-law and then insert it at such a puzzling point in the
narrative? It is not surprising that a succession of commentators
on Deuteronomy have taken the same view as Levinson (e.g. S.R.
Driver, Deuteronomy [ICC; Edinburgh, 3rd ed., 1902], pp. xiv-xix,
9-10; Mayes, Deuteronomy, pp. 118-25; Weinfeld, Deuteronomy
1–11, pp. 139-40; E. Nielsen, Deuteronomium [HAT; Tübingen,
1995], p. 26).7 A further factor which they mention adds weight
to this conclusion: it is not difficult to detect the use of another
non-Priestly section alongside Exod. 18.13-27 in Deut. 1.9-18,
namely Num. 11.11-12, 14-17, 24-25. It is the use of these two
passages in combination which causes the Deuteronomist to
compose his narrative in two parts, vv. 9-14 (based on Num. 11) and
vv. 15-18 (based on Exod. 18). The ordering of the two sub-sections
is deliberate, for it ensures that the judges will be tribal leaders,
chosen by the tribes, and not (royal) officials imposed from above.
So determined were the authors to achieve this, that the reference to
‘judges’ in v. 16 comes without any preparation, a clear sign that the
two traditions have been secondarily tacked together.8
7
The puzzlement of von Rad on this point (Deuteronomium [ATD; Göttingen,
1964], pp. 28-29, ET, pp. 39-40) has prompted hesitation by some others, but it
arises only from what he recognises to be the freedom with which the Deuter-
onomist has handled the older tradition and would in no way lend support to the
opposite view.
8
In any case Deut. 1.9-18 is the ‘meeting-point’ of Exod. 18.13-26 and Num.
11 in most scholars’ understanding of the three passages. Even Cook’s unusual
view that Exod. 18 is an elaboration of the tradition in Num. 11 has to reckon with
renewed contact between the two passages in Deuteronomy (pp. 298-99).
586 EXODUS 1–18
It would (or should) take some very powerful arguments to

overturn such a strongly founded consensus, but those offered are
weak and inconclusive. Van Seters suggests that in Exodus 18 ‘the
presentation of Moses’ office, some tasks and functions seems to
be confused’ and needs to be seen as the result of a combination
of Exod. 33.7-11 and Deut. 1.9-18 by his late Yahwist. His under-
standing of some details in Exodus 18 is probably mistaken, which
undermines the charge of ‘confusion’ (see the Explanatory Notes).
Significantly he does not explain why the ‘late Yahwist’ introduced
Jethro as the key player in the story and he does not provide an
adequate answer to Knierim’s (and now others’) proposal for a
background to the passage in the monarchy period. His linguistic
argument (taken from Rose) that Gen. 26.5 and Ps. 105.45 require
a late date for vv. 16 and 20 carries no weight if the legal terms are
understood here as we have suggested.9 According to Johnstone the
Exodus passage is a Priestly recasting of Deut. 1.9-18 and passages
in Numbers 10–11. His argument is briefer than that of Van Seters
but vulnerable to the same criticisms. He notes differences from
Deuteronomy and Numbers but only considers the P-edition as an
alternative (later) source for them. The vocabulary shared by Exodus
and Deuteronomy which he observes could as easily be borrowed
by Deuteronomy as by Exodus. It is true that the verb ‘teach’ is used
a number of times by Ezekiel (e.g. in 3.17-21), whose language is
often similar to P, but the meaning is different there and the verb
also occurs in non-Priestly passages (see Note t on the translation):
it is never used in P itself and cannot be regarded as an indication
of Priestly authorship. Finally, Albertz also cites the occurrence of
vocabulary shared with Deuteronomy 1, but it is not distinctively
Deuteronomic language and the transmission could just as easily
have gone the other way. Albertz does provide an explanation for the
introduction of Jethro into the Exodus version – it was to challenge
the xenophobic attitudes of Nehemiah and those who thought like
him in the fifth century B.C. – but it is an odd point of view to
attribute to his ‘Hexateuch redactor’, whose main contribution was
to add the highly xenophobic book of Joshua to Israel’s developing
9
Even if they are understood in the usual way, their occurrence in two prob-
ably late texts is not a sufficient proof that two other texts which use them are
also late.
18.13-27 587
corpus of origin traditions!10 In any case Albertz does not explain

adequately why Jethro’s involvement should be associated with the
reform of the judicial system: in fact he pays very little attention
to this central topic of the passage and says nothing about those
scholars who have based their understanding of its origin on a more
serious engagement with this.
It remains most probable that the passage is earlier, not later,
than Deut. 1.9-18. Before a more precise conclusion can be reached
about its date and literary affinities, it is necessary to consider the
question of its literary unity. There is no need to say more about
theories that found extracts from two sources interwoven here, as
Rudolph’s critique of this approach in 1938 justifiably dissuaded
almost all subsequent scholars from adopting it.11 But a number of
scholars (though not as many as for vv. 1-12) have concluded that
some verses or parts of verses are secondary additions to the original
narrative. The most recent commentaries (Propp, Dozeman and
Albertz) do not take this view, but it has continued to be supported,
in some cases affecting more of the text than before (cf. Schäfer-
Lichtenberger and Graupner).
The first unit to be suspected of a later origin than the rest was
v. 20, which Holzinger (pp. 61-62) found to be a confusing and
moralising addition (cf. Rudolph, p. 39): despite taking the very
similar expressions in v. 16b to refer to verdicts, here he thought
they meant the ‘Mitteilung eines Rechtscodex’. It would surely be
more consistent to understand the expressions in the same way both
times. Schäfer-Lichtenberger and Graupner read both of them as
references to legal instruction (like Rose, Deuteronomist, pp. 229-
30) and so see v. 16b as also a later addition. In the context this is
surely not the primary meaning (see the Explanatory Note), and if
10
Levin (p. 361) asserts that it is obvious (‘liegt auf der Hand’) that vv. 13-26
validate a ruling from the Second Temple period: he thinks it concerns the delega-
tion of priestly(!) arbitration to local officials. This is of course pure speculation
and surely most improbable in Second Temple times: one might have thought that
he would have considered alternatives. His only other argument is that v. 18 ‘cites’
Num. 11.14. This is plainly not the case, and if there is any literary connection
between the verses, it is not easy to see what it is: even Levin seems to be unsure
(cf. p. 374).
11
Fohrer is the only exception known to me (Einleitung in das Alte Testament
[Heidelberg, 11th ed., 1969], pp. 161, 167).
588 EXODUS 1–18
it is present at all it can probably seen as an anticipation of Moses’

later role by the original author, as most commentators do. Baentsch
introduced what has been the most widely followed proposal for
a secondary addition: the list of specific titles for the judges in
vv. 21b and 25b (p. 168: likewise Junge, Knierim, Hyatt, Fritz,
Rose [pp. 230-31], Schäfer-Lichtenberger, Crüsemann [Die Tora,
pp. 105-106, ET, p. 84], Cook, Graupner). The main reason given is
that the titles belong to a military and hierarchical organisation that
does not fit the simple two-tier judicial system proposed by Jethro.
It is also unlikely that they correspond (apart from the word śar)
to any actual feature of the judicial system that was in place under
the monarchy: neither Deuteronomy 16–17 nor 2 Chronicles 19
mentions them. They might best be seen as adding some ‘colouring’
to fit the setting in Israel’s early history in which the author has
placed the passage (which Deut. 1.15 retained), and this might as
easily have been done by the original author as by a later redac-
tor.12 Baentsch also saw vv. 25a and 26 as redactional additions, on
the grounds that they were superfluous, only reflecting what had
already been said; Graupner follows suit, but because of the differ-
ences from vv. 21-22 which correspond to Deut. 1.15 and 17 (the
addition of ‘heads’ and the replacement of ‘great’ by ‘difficult’).
Neither argument carries much weight. Schäfer-Lichtenberger
saw the prophetic character of v. 15b as incompatible with Moses’
judicial role, apparently overlooking the same implication in v. 19b;
Graupner excised v. 19b too. This is unduly ‘purist’: an origin tradi-
tion which elsewhere portrayed Moses as enjoying close intimacy
with God (see especially 33.7-11) could perfectly well have attrib-
uted his ability as a judge to this.
The section may therefore be treated as substantially a unity,
though it is possible that vv. 21b and 25b are later additions. Its
literary affiliation is clearly to the non-Priestly version of the Exodus
story and our argument thus far has pointed to it being earlier, not
later, than Deuteronomy 1. It is attached to vv. 1-12 by the time
reference in v. 13 and by the fact that the author does not consider
it necessary to give the name of Moses’ father-in-law again. It also
12
Baentsch and Crüsemann both found the attachment of the list in v. 21
awkward, because ‘over them’ might seem to refer to the officials. But the context
really leaves no doubt that the officials are the object of ‘appoint’ (see Note x on
the translation).
18.13-27 589
shares with vv. 1-12 the repeated use of ‘God’ (ʾelōhîm) in place
of the divine name, here without exception, in contrast to most of
the Exodus story. This encourages an association of it with other
passages in Exodus which do likewise (and further passages which
are closely associated with them) and have been attributed to the
E source, an association which is strengthened, in part through
vv. 1-12, by narrative continuity and similar terminology to parts
of chs. 3–4. In addition the need for the judges to ‘fear God’ (v. 21)
recalls the prominence of this quality in 1.15-21 (see the Explana-
tory Note on 1.17) as well as in some sections of Genesis which have
commonly been attributed to E (20.11; 22.10; 42.18); and, while it
is not restricted to E, the idea of God being ‘with’ a person to help
and protect him (v. 19) is certainly present there (Gen. 21.20, 22;
Exod. 3.12). A broadly pre-exilic date is also supported by the now
wide acceptance among specialists of the view (see above) that in
the background of the reform attributed to Jethro and Moses should
be seen the change to a two-tier judicial system at some point in the
monarchy period. It is unlikely that the actual pressures that led to
such a system only had their effect after centuries of monarchic rule,
so that a date for the reform as late as the time of Josiah (Junge) is
improbable. The local courts seem to have been in existence by the
eighth and probably the ninth century. Some scholars have placed
the reforms very early, in the time of David, citing 2 Samuel 14 as
evidence of tensions then (Reviv, Schäfer-Lichtenberger); others,
treating 2 Chronicles 19 as having at least some basis in a histor-
ical memory, have proposed the reign of Jehoshaphat (Knierim,
Crüsemann [Die Tora, pp. 113-16, ET, pp. 91-93], Jackson: against
Junge’s association of Chr.’s source here with changes in Josiah’s
time see Williamson, 1 and 2 Chronicles [NCB; London, 1982],
pp. 287-89), but the ongoing debate about the authenticity of the
Chronicler’s account makes it impossible to be certain about this.
Nor do we know whether such procedures were perhaps first intro-
duced in the larger northern kingdom before they reached Judah.
In any case a general dating of the passage in the monarchy
period is supported by the portrayal of Jethro as a kind of court
adviser (see the Explanatory Notes on vv. 17-23) and it may well
be that the attribution of the proposal to a foreigner was designed to
deal with possible criticism of what was known to be a change that
was modelled on a foreign precedent. Comparisons have sometimes
been made with the fourteenth-century Edict of Horemheb (for the
590 EXODUS 1–18
text see ARE 3, §§45-67, esp. §63), in which the king of Egypt
appoints two suitably qualified officials (viziers) in ‘the two great
cities of the South and the North’ (the former certainly Thebes, the
latter either Heliopolis or Memphis), to whom citizens could bring
their disputes for resolution. This is of course long before any likely
borrowing of the system by the Israelites. But the Edict is only the
best-known example of a much wider body of evidence of local
courts and (necessarily in such a large country) two central courts
in Egypt (see for a summary and further references OEAE 2, pp.
277-82). It is in every way possible that the inspiration for the court
system in Israel and Judah came, like much else in their government
and administration, from Egypt.
Exodus 18.13-26 has often been described as an ‘aetiology’ of
the Israelite judicial system (e.g. by Knierim), but this designation
has also been criticised, both by Childs and by Graupner (the latter
especially in ‘Exodus 18,13-27 – Ätiologie einer Justizreform in
Israel?’, but also in Elohist, pp. 109-10), though perhaps for the
wrong reason. The chief reason why the designation is inadequate,
though it has some truth in it, is that it fails to do justice to the
intention of the narrative to regulate the institution, not simply to
validate it. This is particularly evident in the qualities required of
the judges in v. 21. Stephen Cook has quite rightly seen the passage
as belonging to a ‘stream’ of tradition which has connections with
eighth-century prophecy and Deuteronomy, which he refers to as the
‘E-stream’. This is a body of literature from the monarchy period
which, among other things, shares a strongly ethical character. To
modify the alternative designation proposed by Graupner, the story
may best be described, not as ‘eine theologische Lehrerzählung’
(Elohist, p. 109), but as ‘eine ethische Lehrerzählung’.
As such it fits very well into a central and distinctive feature of
the material traditionally ascribed to E. It is visible, for example,
in the shaping of the ‘wife-sister motif’ in Genesis 20 and in the
midwives’ refusal to kill Israelite babies in Exod. 1.15-21 (see
further e.g. Fohrer, Einleitung11, p. 171 [ET p. 157]; Graupner,
Elohist, p. 393). Graupner has with good reason associated this
emphasis with wisdom teaching (Elohist, p. 395), and we have noted
already that Jethro appears here in the role of the wise counsellor.
But just such an intense ethical concern, including its application to
the law-courts, is also a prominent characteristic of eighth-century
prophecy and in the case of Amos and Micah it has been specifically
18.13-27 591
linked to their closeness to wisdom teaching and forms of speech

(cf. H.W. Wolff, Amos’ geistige Heimat [WMANT 18; Neukirchen,
1964]; ‘Michas geistige Heimat’, in Mit Micha reden: Prophetie
einst und jetzt [Munich, 1978], pp. 30-40). This may justify us in
seeing a contrast between the two Jethro episodes in Exodus 18 in
relation to their originality. There seem to be good reasons to trace
in vv. 1-12 (and 27) a strong element of ancient (and surprising)
tradition about close contacts in early times between Israel (and
specifically Moses) and the Midianites and their religion. In vv.
13-26 the impression that emerges is rather one of an engagement
with an institution and perhaps its corruption which belonged to the
author’s own time, just as is more explicitly the case in the parenesis
in 23.1-3, 6-9 and later in Deut. 1.16-18 and 16.19-20. The benefit
to the people promised in Exod. 18.23 depends on the observance of
the whole of ‘Jethro’s’ advice – not just the institutional changes but
also the behaviour of the judges themselves as it is presented in the
qualities listed in v. 21 – and that is what Moses himself is said to
have done (v. 24). As a later writer summed it up: ‘Justice, and only
justice, you shall pursue, so that you may live and occupy the land
that the Lord your God is giving you’ (Deut. 16.20). Graupner may
well be right, therefore, to affirm that ‘no originally independent
tradition can be isolated from 18.13ff.’ (Elohist, p. 110), at least
as far as the role of Jethro is concerned.13 Instead we may see the
author taking the opportunity provided by the tradition of a visit by
a foreign notable to portray him as both a wise adviser about admin-
istrative arrangements and, under God (vv. 19, 23), a teacher of the
ethical values which are essential if a society is to flourish.
13 On the next day Moses sat to give judgement fora the people
and the people stood byb Moses from the morningc tilld the
evening. 14 Moses’ father-in-law saw all that he was doinge for
the people and said, ‘What is thisf that you are doing for the
people? Why do you sit alone, whileg all the people stand by
you from morning to evening?’h 15 Moses said to his father-
in-law, ‘Because the people come to me to seek God.i 16 When
theyj have a dispute which comesj to me, I judge between the
13
The qualification is important, for the picture of Moses’ roles as a judge and
a kind of ‘seer’ in vv. 13, 15-16 and 19 coincides with what is implied by passages
such as 24.14 and 33.7-11 and may well have a basis in older tradition.
592 EXODUS 1–18
two partiesk and I declare the decisions and instructions (or ‘the
statutes and laws’) of God.l’ 17 The father-in-law of Moses said
to him, ‘What you are doing is not good. 18 You are tiring out
both yourself and this people who are with you. The taskm is too
heavy for youn – you cannot do ito on your own. 19 Now listen
to my words of advicep and may God be with you!q You should
continue to be the people’s representativer with Gods and bring
their disputes to God. 20 You should teacht them the decisions
and instructions (or ‘the statutes and laws’)l and make known
to them the way in whichu they should proceed and the action(s)
which they should perform. 21 But you should selectv from the
whole people capable menw who fear God, honest men who hate
unjust gain, and appoint them over themx (sc. the people) as
officers for thousands, hundreds, fifties and tens. 22 They shall
give judgment for the people at any time. Any serious (lit. great)
disputey they shall refer (lit. bring) to you, but minor disputes
they can resolve themselvesz. Lighten the load on yourselfaa and
let them bear it with you. 23 If you do this and God so commands
youbb, you will be able to survivecc and all this people will come
todd their place in well-beingee. 24 Moses listened to his father-
in-law and did all that he said. 25 Moses chose capable men
from all Israel and made them leaders over the people as officers
for thousands, hundreds, fifties and tens, 26 and they would
give judgementff for the people at any time. The more difficult
disputes they would refer (lit. bring) to Moses, but every minor
dispute they would resolve themselvesgg. 27 Then Moses let his
father-in-law go and he made his journeyhh to his own land.

a. Heb. ‫לשׁפט את־העם‬. For the phrase (used of Solomon) see 1 Kgs 3.9 (cf.
2 Kgs 15.5): both here and there the context shows that ‫ ָשׁ ַפט‬can be used gener-
ally of acting as a judge and not only of the specific verdicts, favourable or
unfavourable, that might be given. This wider sense is even more evident in
other constructions which do not include a direct object, as in v. 16 below (cf.
TWAT 8, 417-18 = TDOT 15, pp. 420-21).
b. For ‫ על‬in the sense ‘by, beside, with’ (as again in v. 14) see BDB,
pp. 755-56.
c. Heb. ‫מן־הבקר‬, with the full spelling of ‫ מן‬retained as generally before the
art. (GK §102b; JM §103d). As it happens the three other cases with ‫ הבקר‬all
depart from that rule (2 Sam. 2.27; 24.15; 1 Kgs 18.26).
18.13-27 593
d. Heb. ‫( עד‬so MT, but see Text and Versions), as opposed to the ‘idiomatic’
‫ ועד‬which appears in 9.25; 11.7; 12.12; 13.15; 23.31a; 28.42. But ‫ עד‬without
waw occurs again in the next verse, as well as 22.3; 23.31b; 27.21. Overall the
variation does not seem to be due to the date of a text’s composition.
e. Heb. ‫עשׂה‬, with the participle here indicating continuing action in the past
(from the narrator’s point of view): JM §121f.
f. Heb. ‫הדבר הזה‬. The ten(!) occurrences of ‫ ָדּ ָבר‬in vv. 13-27 illustrate well
some of its varied uses (on which see more generally BDB, pp. 182-84; TWAT
2, 111-33 = TDOT 3, pp. 103-25). Here, as in vv. 17 and 23, it does little more
than reinforce the demonstrative (‘this thing’); in vv. 16, 19, 22 (2x) and 26
(2x) it has the technical legal sense of ‘dispute’; while in v. 18 it refers to a
‘task’ (cf. 5.13, 19). All these uses relate to the second major sense of ‫ ָדּ ָבר‬,
‘thing, matter’, except perhaps in v. 23, where it could be paraphrased ‘what I
have said’ and so be derived from the primary sense ‘word’. There is a parallel
use of Akk. awātu for ‘legal dispute’; compare also ‫ בעל דברים‬in 24.14 with
Akk. bēl awātim (AHw, pp. 89a, 119b). This need not of course imply any
linguistic borrowing.
g. The clause is a circumstantial clause, highlighting the consequences
which make Moses’ present practice questionable.
h. This time, in contrast to v. 13, ‫ בקר‬and ‫ ערב‬are used without the article,
as in 16.6-7, 19-20. The explanation here may be that a colloquial form of
speech is being imitated or that the indeterminate forms imply a regular occur-
rence (cf. ‫ לילה‬when used adverbially): JM §137p suggests that it is due to the
interrogation.
i. Heb. ‫לדרשׁ אלהים‬. The expression is widespread (usually with ‫ יהוה‬rather
than ‫אלהים‬, but cf. 1 Sam. 9.9; Pss. 14.2; 53.3; 69.33; 1 Chr. 21.30; 2 Chr.
19.3; 26.5; 30.19 for the latter; also Job 5.8 with ‫)אל‬ ֵ but it is used in different
senses (TWAT 2, 318-27 = TDOT 3, pp. 298-304; cf. TWAT 1, 763-67 = TDOT
2, pp. 236-39 on the parallel idioms with ‫)בקשׁ‬. ‫ דרשׁ‬can be used of worship at a
sanctuary (Ps. 24.6 and Isa. 58.2 are clear cases: cf. Amos 5.4-6), but ‫( בקשׁ‬esp.
with ‫ )פנים‬is more common in this sense. It is also an expression for repentance
(e.g. Isa. 55.6; Hos. 10.12) and piety in general (Ps. 9.11 and frequently in
Chronicles). The sense which comes closest to the present context, however,
is the request for divine guidance, most often through the consultation of
a prophet (1 Sam. 9.9 etc.), which can also be expressed by ‫שׁאל‬. But this
seems to be the only place where ‫ דרשׁ‬is used of a regular legal procedure: for
comparable (but perhaps different) practices described in other ways see the
Explanatory Note.
j. Heb. ‫להם…בא‬. The difficulty of these words is reflected in the varied
readings to which they gave rise (see Text and Versions). The pl. suff. of ‫להם‬
refers back to the collective ‫העם‬, as often, but the reversion to a sing. form in
‫ בא‬is surprising; unless, perhaps, the subject is indefinite (GK §144d) or ‫ בא‬is
a participle attached to ‫( דבר‬Childs, p. 321, comparing 22.8, Houtman, p. 414,
594 EXODUS 1–18
Propp, p. 626; similarly Joosten, Verbal System, p. 302 n. 101: cf. ‫ בא‬Hiphil
with ‫ דבר‬as object in vv. 22, 26). Our translation follows the latter view, but
possibly the words ‫ בא אלי‬are a misplaced alternative reading for ‫ יבא אלי‬in
v. 15: they are not essential to the sense here. On the use of ‫ דבר‬here see Note
f.
k. Heb. ‫בין אישׁ ובין רעהו‬, lit. ‘between a man and his companion’.
l. Heb. ‫את־חקי האלהים ואת־תורתיו‬: cf. v. 20. This wording, as usually under-
stood, seems to confuse legislation with the resolution of individual cases,
unless the ‘making known’ refers to the citation of already existing legislation
in support of Moses’ verdicts. ‫ תורה‬does occasionally mean a specific verdict
or ruling (Deut. 17.11; Hag. 2.11; perhaps Isa. 2.3 par. Mic. 4.2; Job 22.22: cf.
the use of ‫ ירה‬in Mic. 3.11) and might do so here (TWAT 8, 606-607 = TDOT
15, pp. 617-18). The application of ‫ חק‬to a one-off ‘decree’ or ‘prescription’
(cf. Ps. 2.7; Job 23.14; Gen. 47.22; Exod. 5.14) might justify a similar sense
being given to it here in a specifically legal context: cf. the further examples of
‘concrete meanings’ in TWAT 3, 150-52 = TDOT 5, pp. 141-42. In particular
the related form )‫(־אוֶ ן‬
ָ ‫ ִח ְק ֵקי‬in Isa. 10.1 is best understood, in the context of v. 2
and other passages in Isaiah, to refer to unjust verdicts rather than unjust laws
(see the discussion in H.G.M. Williamson, Isaiah 6–12 [ICC; London, 2018],
pp. 469-73, which speaks of ‘decisions and judgments’ by the central court
in Jerusalem). This is certainly how one authority cited in MRI (Lauterbach,
pp. 180, 182) thought that ‫ חק‬here could be understood. For further discussion
see the Explanatory Note and Text and Versions.
m. Heb. ‫הדבר‬. On the sense here see Note f.
n. Heb. ‫ממך‬. For the sense ‘too…for’ (an extension of the comparative use
of ‫ )מן‬cf. 1 Kgs 19.7; 2 Kgs 6.1 and BDB, p. 582 (6d).
o. Heb. ‫עשׂהוּ‬.ֲ For the rarer form of the inf. constr. of a Lamed He verb,
without the usual ending ‫ת‬-, see GK §75n, where it is observed that in the
Pentateuch such forms are found only in passages traditionally attributed to
the E source (Gen. 31.28; 48.11; 50.20: cf. also GK §69m n. 2).14 The other
occurrences are in Ps. 101.3 and Prov. 16.16a; 31.4. The ‘very remarkable’
(GK) form with the suffix arises logically from the fact that this type of inf.
ends in a vowel (cf. GK §58a).
p. Heb. ‫שׁמע בקולי איעצך‬. It is difficult to be sure whether ‫ איעצך‬is an asyn-
detic relative clause subordinate to ‫בקולי‬, as in the translation given here with
some support from the MT accents and Vulg atque consilia (cf. JB ‘Take
my advice’ and Propp, p. 632; for the syntax see GK §155f; JM §158a); or
an independent clause with a cohortative verb, ‘Let me advise you’, as it
is commonly understood. Constructions of the former kind are much more
14
The only one of Graupner’s ‘exceptions’ to this observation that concerns a
Lamed He verb is Gen. 48.11 (Elohist, p. 108 n. 386), and his reasons for denying
that verse to E are not strong (ibid., pp. 358-59).
18.13-27 595
common, at least with a verbal clause, in poetry, though v. 20 below and

1 Sam. 6.9 provide clear examples in prose.
q. Heb. ‫ויהי אלהים עמך‬. Again there is uncertainty over the syntax (cf.
Propp, p. 632): is this a wish, as the jussive verb would most naturally suggest
(cf. IBHS §34.3 with parallels), or an assurance, like ‫ יהי‬in 7.9 after two
imperatives (cf. JM §116d, i and the more common construction with two
imperatives linked by waw)?15
r. Heb. ‫היה אתה לעם‬. The independent pronoun as usual gives some emphasis
to the subject (GK §135a; JM §146a-b), but in the circumstances there may
be more emphasis on the verb which precedes it: Moses is to ‘continue’ with
the mediatory role that he has been undertaking (cf. 1 Sam. 17.56; 23.22 for a
similar weaker role for the pronoun; elsewhere, even when postponed, it does
express a contrast: 20.19; Deut. 5.27; 1 Sam. 20.8). ‫ ל‬has the sense ‘on behalf
of’ here, hence ‘representative’ is an appropriate paraphrase (cf. 4.16a; 2 Kgs
4.13; Isa. 6.8).
s. Heb. ‫מול האלהים‬. Elsewhere, whether alone or preceded by a preposition,
‫ מול‬always refers to physical location, ‘in front of, opposite, towards’, perhaps
with an element of proximity implied (cf. Num. 22.5). In the absence of other
theological occurrences it is difficult to determine its relationship to ‫לפני‬,
which can itself indicate varying degrees of closeness to God. What follows
(‫ )אל־האלהים‬certainly implies the ability to enter God’s presence in some sense.
Ehrlich’s view (pp. 332-33) that ‫ האלהים‬here means ‘the judges’ is unfounded:
even in the passages where this interpretation has had more support (e.g. 21.6;
22.7, 8, 27 [cf. BDB, p. 43]) ‘God’ must be meant.
t. Heb. ‫והזהרתה‬. The final he, while not necessary in this form, is ‘frequently
added’ in MT according to GK §9d (and is found even in some pre-exilic
inscriptions: cf. AHI 1.003.8; 2.007.5, 040.9). ‫ זהר‬Hiphil must, from the
context, have the sense ‘teach, inform’ here, but the lexica continue to give
‘warn’ for most or all of its other occurrences (similarly for the Niphal). In
some of these, at least, the meaning ‘teach, inform’ is possible too: 2 Kgs
6.10; Ps. 19.12; Qoh. 4.13; 2 Chr. 19.10; Sir. 35.22 [32.21]; 11QT 51.5. It
has even been argued that the 15 occurrences in Ezekiel (all in 3.17-21 and
33.1-9), where the ‘watchman’ motif has seemed to require the sense ‘warn’,
may be a specialised use based on ‘teach, inform’ (TWAT 2, 544-50 = TDOT
4, pp. 41-46; J. Tropper, ‘Hebräisch zhr2 “kundtun, warnen” ’, ZAH 8 [1995],
pp. 144-48, with fuller examination of the Semitic cognates). ‫זהר‬, ‘shine’
(Dan. 12.3: cf. Ezek. 8.2), is a distinct homonym and not relevant (Tropper,
pp. 147-48). M. Mishor, ‘On the Language and Text of Exodus 18’, in S.E.
Fassberg and A. Hurvitz (eds.), Biblical Hebrew in its Northwest Semitic
Setting (Jerusalem and Winona Lake, 2006), pp. 225-29 (228), is in no doubt
that ‫ זהר‬here is an Aramaism and, following a hint of B. Jacob (p. 521), sees
15
Graupner (Elohist, pp. 108-109) adds a further possibility, that ‫ ויהי‬intro-
duces a purpose clause.
596 EXODUS 1–18
it as one of a number of expressions (cf. Note v below) which show that the
author of Jethro’s speech ‘tried to create the illusion that another language was
being spoken by using rare terms or constructions. It seems that he did not
care much about the difference between Aramaic and Midianite.’
u. Here (but not in the next clause) the relative clause is attached without
‫( אשׁר‬as in Gen. 39.4 and Exod. 9.4 perhaps), as more commonly in poetry
(see Note p above).
v. Heb. ‫ואתה תחזה‬. The anteposed independent pronoun as elsewhere gives
emphasis, but to the new element in Jethro’s instructions as a whole. The
sense ‘select, look out’ is not found elsewhere with ‫ חזה‬but it is well attested
with its synonym ‫( ראה‬e.g. 2 Kgs 10.3, also with ‫)מן‬. In v. 25 the more specific
word ‫ בחר‬is used. Mishor (see Note t) supposes that the author ‘was deter-
mined to insert the Aramaic [sic] verb into Jethro’s speech’.
w. An ‫ אישׁ־חיל‬is sometimes a warrior (e.g. Judg. 3.29), but ‫ גבור־חיל‬is more
frequent in this sense (and more explicit) and ‫ חיל‬can denote other kinds of
ability or worth, as it clearly does here (cf. Gen. 47.6; Prov. 31.10; Ruth 3.11).
x. Heb. ‫ושׂמת עליהם‬. The object of ‫ ושׂמת‬must (and can) be understood from
the preceding clause and ‫ עליהם‬refers back to ‫העם‬, with a pl. ending as in v. 20.
y. Heb. ‫כל הדבר‬. ‫ הדבר‬is presumably collective, like ‫ בן‬and ‫ בת‬in 1.22,
which also have the article: cf. GK §127b and several more examples in BDB,
p. 481.
z. Heb. ‫ישׁפטו־הם‬. The imperfect here is permissive (JM §113l) and the
independent subject pronoun underlines the contrast with the previous clause.
aa. GK §110i takes ‫ והקל‬as a case of the imperative expressing a conse-
quence when following waw.
bb. Heb. ‫וצוך אלהים‬. The words seem surprisingly like an afterthought and
they gave trouble to some of the Vss (see Text and Versions). But their sense
and relevance is clear enough. Graupner’s view of them as a relative clause
without ‫( אשׁר‬Elohist, p. 109) is baseless and of no help.
cc. Heb. ‫עמד‬. For the sense see especially Ezek. 22.14, and also Deut. 25.8;
Ruth 2.7.
dd. One might expect ‫ אל‬rather than ‫( על‬cf. Text and Versions), but the
latter is quite frequently used loosely of motion to a person or a place even
where none of the special senses of ‫ על‬apply (BDB, p. 757).
ee. For this common sense of ‫ שׁלום‬cf. 4.18 and 18.7.
ff. Heb. ‫ושׁפטו‬. The perfect consecutive (like the imperfects later in the
verse) indicates a repeated occurrence, even though the previous verb had
expressed a single action (GK §112g).
gg. Heb. ‫ישׁפוטו הם‬. On ‫ הם‬see Note z above. In ‫ ישׁפוטו‬the internal u-vowel
is unusual but found again in MT at Prov. 14.3 and Ruth 2.8 (cf. GK §47g and
Judg. 9.8, 12K). These may be survivals of the most ancient pronunciation
of Hebrew. It is surprising for there to be any vowel at all (except in pause:
GK §29m), but such orthography is more frequent at Qumran (cf. Meyer, II,
pp. 101-102; Qimron, pp. 50-53; Reymond, Qumran Hebrew, pp. 209-11), at
18.13-27 597
least in mss exhibiting what E. Tov has called ‘the Qumran scribal practice’.
The reason for this and its relationship to the Masoretic forms remains
uncertain.
hh. Heb. ‫וילך לו‬. The ‘centripetal lamed’ puts an emphasis on the subject’s
involvement in the action denoted by the verb (JM §133d note; cf. BDB,
pp. 515-16; GK §119s).
Explanatory Notes
13-16. The transition to the second narrative about Jethro
and Moses is made by the common Heb. formula wayehî, ‘and it
happened’, followed by a new specification of time, ‘on the next
day’, which is close enough to make a connection with vv. 11-12 as
well as to mark a fresh beginning (cf. Gen. 19.34; Exod. 32.30: less
close are Exod. 9.6; 32.6). What follows indicates the new activity
which is to be the focus of the ensuing conversation between Jethro
and Moses: Moses’ role as the judge of disputes for the people.
This role has not been mentioned before, but it is presupposed in
24.14, where Aaron and Hur (cf. 17.10, 12) are appointed to act as
his deputies while he is up on the mountain with Joshua.16 Moses
himself acts as a judge in several other passages (Lev. 24.10-23;
Num. 9.6-14; 15.32-36; 27.1-11; 36.1-12), but these ‘narratives of
desert adjudication’ belong to a different and much later stage of
Pentateuchal legal history (cf. Crüsemann, Die Tora, pp. 121-31,
ET, pp. 98-107; Jackson, Wisdom-Laws, pp. 425-30). The contrast
with what Jethro proposes in v. 22 (cf. v. 26), a procedure that is
available ‘at any time’ (this is the sense of bekol-ʿēt, not ‘for ever’,
as some have proposed [e.g. Knierim, ‘Exodus 18’, p. 151: correctly
Graupner, Elohist, p. 109]), confirms what would in any case have
been probable, that Moses only acted in this way from time to time
and not every day. For judges ‘sitting’ cf. Judg. 4.4-5; Isa. 28.6;
Ps. 122.5: of God as a judge in Joel 4.12; Ps. 9.5. The (hyperbolic)
mention of ‘the (whole) people’ standing and waiting all day (cf.
v. 14) for their cases to be heard serves to accentuate the disadvan-
tages of having only one judge.
16
Some (including Cook, ‘The Tradition of Mosaic Judges’, p. 294) presume
that it is also in view in Num. 11.11-15, 24-30, but judging is not mentioned there
and the issue is the burden of leadership more generally.
598 EXODUS 1–18
Jethro’s questions in v. 14 lead to a clarification of the process

of ‘Mosaic judgement’ (vv. 15-16). First of all, Moses is the person
through whom the people ‘seek God’, that is, in this context, they
seek a divinely approved resolution to their disputes.17 The expres-
sion ‘to seek (dāraš) God/Yahweh’ has a much wider use elsewhere
in the Old Testament (see Note i on the translation and C. Wester-
mann, ‘Die Begriffe für Fragen und Suchen in AT’, KuD 6 [1960],
pp. 2-30; repr. in his Forschung am Alten Testament: Gesam-
melte Studien II [Munich, 1974], pp. 162-90): it is used generally
of worship and then of repentance and prayer. But it most fre-
quently denotes the seeking of divine guidance in a great variety
of circumstances, both through a prophet (e.g. 1 Sam. 9.9; 1 Kgs
22.7-8; 2 Kgs 22.13-14; Jer. 21.2; Ezek. 14.7) and at a shrine (Gen.
25.22; Ps. 34.5): sometimes the two coincide. Exodus 33.7 uses
the close synonym biqqēš for people going to inquire of Yahweh
through Moses at ‘the tent of meeting’. Neither verb seems to be
used elsewhere specifically for the solution of legal disputes, but
prophet(esse)s are said to have acted as judges in early Israel (Judg.
4.4-5 [Deborah]; 1 Sam. 7.15-17 [Samuel, at places known to be
cultic centres]) and some laws in the Book of the Covenant provide
for a divine decision to be obtained at a shrine (Exod. 22.7, 8, 9-10;
cf. 1 Kgs 8.31-32).18
Here the process concludes (as also in v. 20) with a phrase that is
generally understood to mean that Moses reveals ‘the statutes and
instructions of God’ (NRSV and other EVV.: so also the ancient
Vss): this is what the words ḥōq and tôrāh usually mean. They might
then refer to the existing laws according to which Moses gives his
(i.e. God’s) verdict – leaving on one side the problem that the main
body of laws is yet to be delivered (but perhaps see 15.25 and 16.4,
28) – or they might be new laws (which seems to be the general
view now), presumably generalisations from the particular cases
brought to Moses: Houtman (pp. 414-15: cf. Albertz, p. 309) gives
the clearest explanation in these terms. But either way, the process
of law-giving seems to be more drawn out than it is in the standard
Exodus accounts of ready-made legal collections being revealed by
17
Throughout vv. 13-27 Heb. ʾelōhîm, ‘God’ is used in place of the divine
name Yahweh (vv. 15, 16, 19 [3x], 21, 23: cf. vv. 1a, 5, 12 [2x]).
18
For a much later relic of this practice see 28.30 and the related passages,
including Ezra 2.63/Neh. 7.65.
18.13-27 599
God to Moses and/or the people as part of the covenant-making

process. This is not the only difficulty with the usual interpretation,
for the context, which is very much about legal process (jurisdic-
tion) and not about legislation, leads us to expect a reference to
verdicts, not new laws. A case can be made that the words might
mean this here (see Note l on the translation).
Although this is not the usual view today, there is some support
for it in the Mekhilta of Rabbi Ishmael, a third-century A.D. Jewish
commentary (Lauterbach, pp. 180, 182), where R. Joshua is said to
have taken the words to mean ‘interpretations’ and ‘decisions’, both
words befitting what judges do; while R. Eleazar of Modin is credited
with the same interpretation of ‘instructions’, but with a reference
to the sexual laws in Leviticus 18 for ‘statutes’ (because the same
Heb. word is used for them in Lev. 18.30; cf. Eleazar’s interpretation
of its occurrence in 15.26, ibid., pp. 94, 96). Wellhausen took this
view for both words (Rechtsentscheidungen: Composition, p. 81 n.
1); also Holzinger, Baentsch and (for ‘instructions’ [tôrōt]) McNeile
and Hyatt, likewise RSV, JB. Jackson has recently supported the
rendering of tôrōt (but not that of ḥuqqîm) here as ‘decisions’
(Wisdom-Laws, p. 422, with the interesting shift to ‘precedents’
on p. 424). The case for this sense with tôrōt is certainly stronger,
particularly in view of its appearance in Deut. 17.11 and Hag. 2.11,
but there are also uses of ḥōq for a ‘decree’ or a ‘decision’ which
might justify its application to a judge’s verdict in a judicial context
like this (cf. Isa. 10.1). If the objection is raised that there were other
more obvious words to use for a verdict (Heb. mišpāṭ, dābār), the
answer might be that the very ambiguity of ḥōq and tôrāh allowed
some continuity to be expressed between Moses’ (original?) judicial
role and his better known part in the origin of Israel’s laws.
17-19a. Jethro’s criticism of what Moses is doing is based on its
practical disadvantages: it is ‘not good’ (lōʾ ṭôb) because it is weari-
some for Moses and the people, who must wait for their disputes
to be resolved: there is no issue of principle involved. ‘The most
common meaning of ṭôb in the OT is utilitarian’ (TWAT 3, 324 =
TDOT 5, p. 304). Not surprisingly ṭôb is particularly frequent in
the pragmatic teaching of Proverbs and Ecclesiastes, especially in
comparative sentences (e.g. Prov. 15.16; Eccl. 4.6, 9: see further
TWAT 3, 333-35 = TDOT 5, pp. 313-14). It is notable that Jethro
shows concern not only for his son-in-law’s well-being but for
that of the people as a whole (the latter’s importance in vv. 13-26
600 EXODUS 1–18
is emphasised by Knierim, ‘Exodus 18’, p. 149 and n. 10). This

too corresponds to the ethos of the wisdom literature, which is not
purely individualistic but values highly the benefits (or otherwise)
of an individual’s actions for the whole community (cf. W. McKane,
Proverbs [OTL; London, 1970], pp. 11-12, 21-22). Jethro presents
himself to Moses as an ‘advisor’: when he says, according to the
literal meaning of the Heb. word, ‘(with which) I will advise you’,
he is using the verb which generated (as its active participle) the
word for a ‘counsellor’ (yōʿēṣ): such figures played an important
part in the government of the Israelite kingdoms (e.g. 2 Sam. 17;
Isa. 3.3; and numerous passages in Chronicles), as well as in private
life (Prov. 15.22; 24.6). Counsellors were meant to be wise, and in
this section Jethro perhaps appears as one of those ‘wise men of the
East’ with whom Solomon was compared (1 Kgs 4.30).
Jethro concludes his introductory remarks with what is probably
a wish (rather than an assurance: see Note q on the translation) that
God will be ‘with’ Moses, presumably to help him to carry out the
action which he is about to recommend. This is not an expression
that is typical of the wisdom literature – its theology is of a different
kind – but it is widespread, especially in narrative texts, and such
wishes seem to have been common as words of encouragement to
those in danger or undertaking a new responsibility (1 Sam. 20.13 [cf.
17.37?]; 1 Kgs 1.37), as well as in everyday life (Ruth 2.4).19 State-
ments that Yahweh/God is, has been or will be ‘with’ an individual
or group are even more widespread: there are numerous examples in
Genesis (e.g. 21.20, 22; 26.28; 31.3; 39.3, 21, 23) and Moses himself
has received such an assurance in Exod. 3.12 and 4.15. In 2 Chr. 19.6
Jehoshaphat assures his officials that Yahweh will be ‘with’ them
in their work as judges – the passage is probably dependent at that
point on this one. According to H.-D. Preuss such formulae belong
to ‘a fundamental nomadic structure of Israelite faith and thought’;
‘Yahweh’s presence is not primarily a spiritual gift…it is a concrete
promise that Yahweh will be with his own’ (TWAT 1, 492-93 = TDOT
1, p. 456). Certainly it is well rooted in the older narrative tradition
and is not to be limited to any particular theological ‘stream’ (see
further Preuss, ‘…ich will mit dir sein!’, ZAW 80 [1968], pp. 137-73;
TWAT 1, 485-500 = TDOT 1, pp. 449-63).
19
In Exod. 10.10 Pharaoh expresses such a wish ironically.
18.13-27 601
19b-20. Jethro’s specific recommendations begin with Moses’

own role, which looks at first as if it is to continue unchanged:
Moses is to go on acting as the people’s intermediary by taking
disputed cases for resolution by God. The same uncertainty exists
in v. 20 about the meaning of ‘statutes and laws’ as in v. 16, but the
context again seems to point to a reference to verdicts (‘them’ will
refer to the litigants). The omission of ‘of God’ scarcely makes any
difference: if the cases have been ‘brought to God’ (v. 19), it will be
his decisions that are reported to the people. When compared with
Moses’ description of his task in v. 16, Jethro’s restatement of it is
fuller: he adds ‘and make known to them the way in which they
should proceed and the action(s) which they should perform’. These
words might point to an additional task which Moses is to under-
take, the instruction of the people about their general behaviour:
‘way’ (Heb. derek) and ‘proceed’ (Heb. hālak) are often used in this
way. But they could equally refer to the amplification of a simple
verdict, in the form of detailed instructions about the requirements
imposed on the disputing parties. In one respect the language fits
this better, as ‘action(s)’ is a singular noun in the Heb. (maʿaśeh): it
is occasionally used collectively for ‘actions’ more generally (e.g.
Mic. 6.16), but the singular meaning is the more obvious (cf. Gen.
44.15). In this case Jethro does not (yet) modify what Moses is to
do to any significant extent: he keeps the limitation of it (like any
wise advisor) until the end, where he underlines its practical value
again (vv. 22-23).
21-23. The new element in Jethro’s proposal is to establish a
much larger body of judges who will each take responsibility for
the primary adjudication of disputes in a section of the people. It is
highly likely that behind this story the outlines of a judicial reform
can be detected, which at a later stage of Israel’s history created
the system of local courts that is visible in texts from the monarchy
period as the usual setting for the ordinary populace to seek arbitra-
tion in their legal disputes (for fuller discussion of this view see
the introduction to this section). The responsibility for the selection
and appointment of these judges is given here to Moses himself
(v. 21: contrast Deut. 1.13), not the people or their representatives.
They are to possess certain qualities to fit them for their task and
prevent injustice from being done. Apart from the last one, these
qualities are not special to good judges, they are characteristics that
are admired in anyone, according to the Old Testament’s scheme
602 EXODUS 1–18
of values. The first two also appear in the description of the ideal
wife in Prov. 31.10-31: she is ‘capable’ (v. 10: cf. 12.4) and she
‘fears the Lord’ (v. 30). ‘Capable men’ (literally ‘men of ability’,
Heb. ʾanšê ḥayil) were often warriors, but this is unlikely to be the
meaning here and, as the examples from Proverbs show, ḥayil could
refer to other types of ability, as it also does in those who were to
be appointed to official positions (śārîm: see further below) in Gen.
47.6. Sometimes it appears to mean ‘worth’ or ‘respect’ (1 Kgs 1.42,
52; 2 Kgs 2.16). ‘Fear of God/Yahweh’, which is also often praised
in Proverbs (1.7; 8.13 etc.), introduces a religious characteristic, but
it is one with a strongly ethical dimension, as can be seen from its
application to the Hebrews’ midwives earlier in Exodus (1.17, 21:
see the notes there and more generally TWAT 3, 876-93 = TDOT
6, pp. 297-315). The two final qualities are specifically ethical and
probably belong closely together. ‘Honest men’ are men of ʾemet,
a word which covers a range of meaning from ‘faithfulness, relia-
bility’ to ‘truth’, the latter especially where speaking is involved
(Prov. 12.19, where the opposite is ‘falsehood’, šeqer). The quality
is naturally highly valued in judges elsewhere: ‘judgments that are
in accordance with ʾemeth correspond to the actual facts so that they
prove to be right and just’ (TWAT 1, 336 = TDOT 1, p. 312, with
reference to Zech. 7.9; 8.16: cf. also Prov. 29.14). In this context
the contrast is with partiality and unfairness (Exod. 23.6-7). Such
judges will also ‘hate unjust gain’ (beṣaʿ), a word that is sometimes
paired or paralleled with words for ‘bribe’ (1 Sam. 8.3; Isa. 33.15;
Prov. 15.27) or is associated with judicial practice in other ways
(Jer. 22.16-17): more direct criticism of bribery appears in Exod.
23.8; Isa. 1.23; Mic. 3.11 etc.
The judges are to be ‘officers’ or ‘officials’, Heb. śārîm: in the
original the word is repeated before each of the groups. The same
expressions are used below in v. 25 and in the similar (probably
later: see the introduction to this section) passage in Deut. 1 (v. 15).
Elsewhere they refer to the commanders of military units, mainly
in Chronicles but occasionally in some older texts (of a ‘fifty’ in
1 Sam. 8.12; 2 Kgs 1.9-15 [6x]; Isa. 3.3; of a ‘hundred’ in 1 Sam.
22.7 [perhaps also in 8.12]; 2 Kgs 11.4-15 [4x]; of a ‘thousand’ in
1 Sam. 8.12; 17.18; 22.7: no clear case exists with ‘ten’, but some
have taken Ishmael’s ten supporters to be such a ‘unit’ of the army
[Jer. 41.1-2; cf. de Vaux, Institutions 2, p. 27, ET, p. 226]). Such
military units are sometimes referred to without explicit mention
18.13-27 603
of their commanders (1 Sam. 29.2; 2 Sam. 18.4 – thousands and

hundreds only: likewise in Chr.) and the same expressions also
appear without a clear military reference as divisions of a tribe,
whether in the form of a ‘clan’ (Judg. 6.15; 20.10; 1 Sam. 10.19;
23.23) or of a ‘city’ (Mic. 5.1; cf. Amos 5.3?). There must, it seems,
at some point have been a correspondence between the terminology
for tribal divisions as such and those of the tribal contingents of the
early Israelite (‘conscript’) army and this terminology, like other
features of the tribal system that were gradually displaced by the
organisation of the monarchic state, evidently remained in use for
some time afterwards (cf. Wellhausen, Abriss der Geschichte Israels
und Juda’s, in Skizzen und Vorarbeiten I [Berlin, 1884], pp. 3-112
[12]; ET in Prolegomena, p. 436: ‘In time of war they [sc. the elders
and heads of houses] each commanded his own household force,
and in peace they dispensed justice each within his own circle’).
The śārîm (sing. śar), whether mentioned as a group or with
definitions of their different responsibilities, seem to have been
essentially the officials of the royal bureaucracy and its military and
judicial branches (see the studies of U. Rüterswörden, Die Beamten
der israelitischen Königszeit. Eine Studie zu śr und vergleichbaren
Begriffen [BWANT 117; Stuttgart, 1985], and R. Kessler, Staat und
Gesellschaft in vorexilischen Juda. Vom 8. Jahrhundert bis zum
Exil [VTSup 47; Leiden, 1992], for wide-ranging discussions of
the evidence; the article on śar in TWAT 7, 855-79 = TDOT 14,
pp. 190-215, provides much valuable information). A few earlier
occurrences use śar in the more general sense of a ‘leader’ (Num.
21.18; Judg. 5.15; 8.6, 14; 9.30; 10.18), but most of the occurrences
in the Pentateuch are likely to be anachronistic retrojections of a
later situation into the early period or the application of the later
term to corresponding types of officials in neighbouring countries
such as Egypt. It is clear that the śārîm of the monarchy period
had, among many others, judicial roles (Isa. 1.23; 10.1-2; Jer.
26.10-16; cf. Exod. 2.14); a seventh-century Heb. inscription from
Meṣad Ḥašavyahu supplements the biblical evidence in a vivid
way, providing a record of a judicial petition made to a śar (AHI
7.001; cf. AHI 2, 99.008, though its genuineness has been disputed).
For discussion of the original setting and purpose of this part of
Jethro’s advice see the introduction to this section. The new judges
are to be available ‘at any time’ (v. 22), which would avoid the
long delays mentioned earlier (for evidence of similar complaints
604 EXODUS 1–18
in the monarchy period see 2 Sam. 15.1-6). There is also to be a

distinction between ‘serious’ and ‘minor’ cases (lit. ‘great’– in v. 26
the word ‘difficult’ [Heb. qāšeh] is substituted – and ‘small’): the
former, of which there would presumably be fewer, will be referred
by the judges to Moses, but the latter they can deal with themselves.
Exactly the same distinction operated in ancient Egypt, where there
was a similar two-level structure of law-courts (ḳnbt: cf. OEAE 2,
pp. 277-82).
In his summing-up (v. 23), which is again dominated by practical
concerns, Jethro recognises the need for divine approval of his
proposals (for other, less satisfactory, interpretations see Text and
Versions). This fits in with his (and no doubt the narrator’s) overall
concept of wisdom, which has already led him to mention God
several times (vv. 19, 21). Presumably he assumes that Moses will
consult God before taking action – or maybe he envisages that the
success or otherwise of the change will show whether God approves
of it. A similar submission to God’s sovereignty is to be found in
even the older sections of Proverbs (16.1-3; 19.21: 3.5-8 gives a
stronger warning against a wisdom that claims independence from
God). The precise way in which the people are said to benefit has
been understood in different ways, because of some ambiguity in
the word ‘place’ (Heb. māqôm). Some have taken ‘their place’ to
mean the land of Canaan, citing 23.30 and Num. 10.29 in support.
But it is hard to see this reform as a condition for God’s overall
plan for his people. It is better to see ‘place’ being used here as a
general expression for ‘home’ (cf. NRSV and Judg. 7.7; 1 Sam.
2.20), which had the advantage that it neither tied the assurance
too tightly to the desert situation (as ‘tent’ would have done) nor
anticipated too crudely (like ‘house’) the future to which it also, and
perhaps primarily, was meant to apply.
24-26. Moses accepts Jethro’s advice, like a king being guided
by his wise courtier(s). All that is peculiar, despite the close family
relationship between them, is that Jethro is a foreigner. Perhaps there
is a hint here that the reform to which this episode alludes (and which
it validates) was based on a foreign model, just as it has often been
thought that the administrative structure of the royal bureaucracy in
later times was patterned on the much more complex Egyptian royal
administration (cf. 2 Sam. 15.18-22; 23.24-39: Albertz, Religions-
geschichte, pp. 167-69, ET pp. 111-13; H. Donner, Geschichte des
Volkes Israel und seiner Nachbarn in Grundzügen 1 [Göttingen,
18.13-27 605
2nd ed., 1995], pp. 229-32). In the context of the Exodus story
an Egyptian ‘adviser’ could hardly be introduced, but Jethro as a
trusted relative of Moses could readily serve as ‘a wise man of the
East’, to bring about the favourable reception of a constitutional
reform with a foreign origin.
27. It is not made clear how long it is supposed to have taken
Moses to put Jethro’s proposals fully into operation: in any case the
summary in v. 26 (where the Heb. imperfect tenses indicate repeated
action) is, no doubt deliberately, open-ended at least through Moses’
lifetime. But the implication of the narrative’s location where it is
must be that Jethro departs for Midian before the theophany and
covenant-making begin. This is different from the truncated ending
of a narrative about Moses’ Midianite relatives which has been
placed at the end of the Sinai-narrative in Num. 10.29-32. This
placing is not accidental, since the surviving part of that narrative is
entirely concerned with whether a Midianite relative of Moses will
accompany the Israelites on their onward journey, namely Hobab
the son of Reuel, Moses’ father-in-law. There is clearly a connec-
tion there with the alternative account of Moses’ stay in Midian
in 2.15-22 (and probably 4.24-26), where Moses’ father-in-law is
called Reuel and not Jethro. The Hobab story, which has appar-
ently lost its ending as well as its beginning, is complicated further,
but also perhaps explained, by the fact that Hobab reappears in the
book of Judges, now as the ancestor of a group of Kenites who
have settled in northern Canaan (Judg. 4.11; cf. 5.24). One can then
imagine, or even infer, that he was identified with the Kenite father-
in-law of Moses whose descendants entered Canaan, according to a
separate tradition (Judg. 1.16: cf. LXX and B. Lindars, Judges 1–5:
A New Translation and Commentary [Edinburgh, 1995], pp. 35-36),
with the tribe of Judah and settled in the far south of the land near
Arad.
The narratives associated with the name of Jethro are evidently
quite distinct from these fragments of tradition about Reuel and
Hobab, with which they share only the name of Moses’ wife and
his eldest son and the role of Moses’ father-in-law as a Midianite
priest. There can be no real doubt that pieces of at least two different
‘jig-saws’ have been selected and combined in the present text of
the Hebrew Bible. If this results in problems that can no longer be
fully solved, it is also as good an illustration as any that behind
the Pentateuch there once lay a much richer store of traditional
606 EXODUS 1–18
narrative, from which the epic narrators (the ‘sources’ of the present
text) have selected their material before they adapted it to their own
purposes. If, or where, they created and shaped their material ex
nihilo, they would surely have made a much tidier job of it.
Text and Versions

‫( ויהי ממחרת‬18.13) In 4QpalExm ‫ ממחרת‬is at the beginning of a line, so
‫ ויהי‬would have been in the lacuna at the end of the previous line: whether
it was separated from v. 12 by a division is not certain but probable. After
‫ ממחרת‬there is some empty space as if, oddly, there was a division here. LXX
μετὰ τὴν ἐπαύριον (similarly at 32.20) is probably a literalistic rendering
(cf. Wevers, Notes, p. 282), not to be translated ‘After the next day’. TgJ
specifies ‘the day after the Day of Atonement’ (cf. MRI [Lauterbach, p. 179]):
the rationale for this chronology is worked out by Rashi in his comments here
and on 33.11.
‫( ויעמד‬18.13) LXX, Sy and Vulg understandably render this as a past
continuous verb (cf. v. 14b), even though it would be an unusual sense for the
Heb. form here. TgN and Sy have the verb in the pl.
‫( העם‬18.13)2o LXX πᾶς ὁ λαός makes a typical heightening of the
account (cf. 17.13 and Text and Versions there), perhaps in this case derived
from v. 14b.
‫( על־משׁה‬18.13) 4Q365 seems to have read ‫ אל‬for ‫על‬, which is probably
just an extension of a fluidity in the use of these prepositions in some biblical
books (cf. BDB, p. 41, and esp. 2 Kgs 11.14). TgJ,N ‫ קדם‬adapts the picture a
little.
‫( מן־הבקר‬18.13) 4QpalExm has ‫מן בקר‬, without the art., as in MT at v. 14
(see Note h on the translation: SP also harmonises, but in the opposite direc-
tion).
‫( עד־הערב‬18.13) SP (cf. Sy) prefixes waw, as also in v. 14: the ‘idiomatic’
construction is sometimes found in MT (e.g. 9.18; 11.7), but it is more frequent
(though not universal) in SP. 4QpalExm has only traces of ‫ ערב‬preserved after
a lacuna: presumably it again omitted the article.
‫( חתן משׁה‬18.14) LXX has Ιοθόρ (its regular spelling of Jethro’s name)
instead, while 4Q365 adds ‫( יתר‬a form which appears in extant Heb. texts
only otherwise at 4.18) to ‫חתן משׁה‬. The name had appeared in v. 12: its (surely
secondary) introduction here may be due to the beginning of a new episode,
if it is not just a case of assimilation to the near context.
‫( לעם‬18.14)1o Vulg in populo (cf. in plebe later in the verse) and TgJ ‫לעמיה‬
made minor changes to the Heb.
‫( ויאמר‬18.14) Sy, as often, adds lh to identify the addressee (cf. 17.2-3).
‫( הדבר‬18.14) LXX* (and Vulg) found an equivalent to be unnecessary,
but the O-text inserted τὸ ῥῆμα (from Aq: Symm, Theod have ὁ λόγος) to
represent it.
18.13-27 607
‫( יושׁב‬18.14) In 4Q365 the word is secondarily inserted above the line: the
space available in the lacuna that precedes suggest that the omission by the
original scribe extended back to the second ‫( עשׂה‬DJD XIII, p. 273).
‫( לבדך‬18.14) TgN added ‘to give judgement’ (cf. v. 13) for clarity.
‫( נצב‬18.14) 4Q365 has the pl. ‫( נצבים‬cf. Tgg, Sy), which would be
possible in BH (GK §132g): but the agreement of MT and SP (cf. LXX, Vulg:
4QpalExm is not extant here) suggests that it is a secondary development. Vulg
praestolatur, without a separate equivalent for ‫עליך‬, can mean ‘stand ready’
but in Vulg it generally has the sense ‘(a)wait’, which would highlight the
effect of Moses’ practice on the people well.
‫( מן־בקר עד־ערב‬18.14) SP adds the article to each noun, assimilating to
v. 13, as well as reading ‫( ועד‬for the latter cf. 4Q365, Sy). LXX has δείλης
(instead of ἑσπέρας) for ‫ערב‬: the precise meaning of δείλη was ‘afternoon’
but it was used in later class. Gk. in formulae like this (LSJ, pp. 373-74) and
stands for ‫ ערב‬several times elsewhere in LXX (cf. δειλινόν in 29.39, 41).
‫( ויאמר‬18.15) The word begins a new line in 4Q365 and the end of v. 14
according to MT/SP would come well before the end of the previous line.
So either 4Q365 had additional text here, which seems rather unlikely, or a
division which does not appear in MT or SP was marked. For some unparal-
leled breaks elsewhere in 4Q365 see DJD XIII, pp. 259-60.
‫( לחתנו‬18.15) LXX* had just τῷ γαμβρῷ, as the possessive pronoun
could be understood, but the Three add αὐτοῦ, as does the O-text, to agree
precisely with Heb.
‫( לדרשׁ אלהים‬18.15) The Vss all add a word to specify what was sought
from God, taking their lead from v. 16: LXX κρίσιν and Vulg sententiam both
specify a legal decision (see respectively Wevers, Notes, p. 283, and OLD,
p. 1736);20 Tgg ʾūlpān, ‘instruction’, which TgO also has at Gen. 25.22 (TgN,J
are different there) and Sy mltʾ, ‘a word’, use more general expressions but
probably have the same object in mind. By adding ‫מן קדם‬, ‘from before’, Tgg
as often avoid the impression that God interacts directly with humans: AramB
2, p. 78 n. 9, suggests that in them ‘Moses is depicted as a Torah scholar or
rabbi’ (cf. v. 19 and 19.3-4).
‫( כי‬18.16) LXX (ὅταν γάρ) and Vulg (cumque) make an explicit connec-
tion with v. 15, as do numerous mss of SP (‫ )וכי‬and Sy, though not the best
ones. Such connections are more likely to be added than removed and the
asyndetic construction of MT and Tgg is both effective and most likely
original.
‫( דבר‬18.16) Only Sy and SamTg use a regular equivalent: the other Vss
have renderings that reflect the legal character of the context (see Note f on
the translation).
20
This is not the introduction of an ‘intermediary’ to avoid anthropomorphism
(Fritsch, pp. 56-57), in LXX at least, as the translator has no difficulty with κύριον
as a direct object of ‘seek’ in 33.7.
608 EXODUS 1–18
‫( בא‬18.16) The Vss all have the verb in the pl., which is certainly expected
after ‫להם‬, and SP reads ‫באו‬. 4QpalExm has a waw before and not after the verb:
there is a gap between two fragments of the ms. at this point and Sanderson
originally restored ‫ו[ב]א אלי‬
ׄ (Exodus Scroll, p. 333 etc.). In DJD IX, p. 96, she
reads ‫ו[בו]א אלי‬
ׄ and comments that ‘the plene waw is based on other usage in
this scroll’. It is hard to be sure what she means: the only other occurrence of
‫ בוא‬is at 10.1, where it is an imperative and that would not fit here. Perhaps she
understands the form as an inf. abs., to which the same spelling conventions
would apply: one might then see it as being used in place of a finite verb (GK
§113y). But she gives no reason for not reading ‫ו[ב]א‬ׄ as she had done earlier,
which would at least get round one difficulty with MT’s perfect tense or parti-
ciple (see Note j on the translation). At any rate it is clear that 4QpalExm did
not have the SP reading: it is actually closer to LXX’s καὶ ἔλθωσιν and might
(according to either of Sanderson’s restorations) represent LXX’s Vorlage:
given LXX’s freedom later in the verse (see below), even a change from sing.
to pl. by the translator would be possible. But whether the introduction of the
conjunction in these two witnesses is connected or independent, it is most
likely secondary, as often elsewhere (see the note on ‫ כי‬above). MT is the only
reading that cannot be explained from the other evidence, as well as being
unquestionably difficult, and as such probably represents the oldest reading
we have. On ways in which it might be explained see Note j on the translation.
‫( ושׁפטתי‬18.16) Vulg ut iudicem draws out the element of purpose that
is implicit in the people’s coming (cf. v. 15b) but as a translation it is free:
the perfect consecutive here expresses recurrent action (so recently Joosten,
Verbal System, pp. 300-301).
‫( בין אישׁ ובין רעהו‬18.16) Vulg inter eos is a neat and precise equivalent to
the Heb., which can scarcely be said for LXX’s free ἕκαστον: the Aram. Vss
follow the idiom of the original.
‫( והודעתי‬18.16) The Vss (except Vulg) add a pl. pronoun referring to the
people, and this is presumably the intention behind SP’s ‫ והודעתיו‬with its sing.
suffix.
‫( ואת־תורתיו‬18.16) As for ‫את־חקי האלהים‬,21 the Vss use their standard
equivalents rather than words indicating ‘decisions’ in particular cases.22 More
striking is LXX’s sing. τὸν νόμον, which matches its renderings in 16.28;
18.20; Lev. 26.46 and must reflect the growing use of the sing. for ‘the Law’
as a whole; AramB renders TgJ’s ‫ אורייתיה‬in the sing. too, although the form
is ambiguous (Stevenson §14) and TgO’s similar form is probably meant as a
21
The printing of the phrase in DJD IX, p. 96, makes it seem that in 4QpalExm
some additional text might intervene between ‫ חוקי‬and ‫האלהים‬, but on the photo-
graph it is clear that the space is no larger than between other words.
22
This contrasts with the Tannaitic interpretations of ‫ תורתיו‬that are given in
MRI (Lauterbach, p. 180): ‫ההוריות‬, ‘the decisions’ (see above, p. 599).
18.13-27 609
pl. (cf. v. 20). TgN’s expansion ‫גזרת אורייתא‬, ‘the decree(s) of the Law’ (for the
expression see 12.43, 49 as well as v. 20 here), is ambiguous but could follow
the pl. form of MT and SP (the Qumran mss do not preserve this phrase) while
making room for the dominant usage.
‫( ויאמר‬18.17) In 4QpalExl the verse opens a new line, which could mean
that the scribe marked a division here (as in SP) by leaving the end of the
previous line blank. There is a stronger reason, based on the available space,
to presume this in 4QpalExm: a line begins with ‫אליו‬, and reconstructing
the previous line according to the text of MT/SP would leave it about five
character-spaces short of the regular line-length for this column. Vulg, which
is very free in this verse, defers its inquit idiomatically until the middle of
Jethro’s answer.
‫( חתן משׁה‬18.17) Vulg omits these words, the subject being self-evident.
Sy ḥmwhy lmwšʾ breaks up the construct phrase and turns mwšʾ into an
explicit addressee, having advanced its pronominal expression lh to anticipate
it earlier in the verse, according to Syriac idiom (Brockelmann §216).
‫( לא טוב‬18.17) LXX οὐκ ὀρθῶς introduces with an adverbial expression
a reshaped rendering in which, as in Vulg and Sy, the relative construction is
abandoned. Tgg ‫ תקין‬and possibly Sy špyr give the same sense ‘right(ly)’ to
‫טוב‬: in LXX ὀρθῶς several times renders ‫ היטיב‬or ‫ טוב‬in this way (e.g. Gen.
40.16; Deut. 5.28). SP agrees with the wording of MT throughout, as do
4QpalExl and 4QpalExm for the first half of the verse at least.
‫( נבל תבל‬18.18) Tgg represent the idea of weariness well, but Sy mistakes
the root for ‫ נָ ַבל‬I with its mṣtʿrw mṣtʿr (from ṣʿr Ethpaal, ‘be dishonoured’).
LXX φθορᾷ καταφθαρήσῃ ἀνυπομονήτῳ uses a related noun for the
emphatic inf. abs. (its more common equivalent in the Pentateuch: cf. H.StJ.
Thackeray, ‘Renderings of the Infinitive Absolute in the LXX’, JTS 9 [1908],
pp. 597-601 [598], with other exx. of the simplex noun in 21.20; 21.28),
but unusually supplements it with an adjective to intensify the meaning still
further – comparable perhaps to the more common addition of an adjective to
a cognate object of a verb in both Heb. and Gk. The choice of καταφθείρω
as an equivalent to ‫ נָ ֵבל‬is in itself imprecise and an exaggeration. Vulg stulto
labore consumeris retains the sense and structure of LXX/OL, but by intro-
ducing labore takes more account of the true meaning of the Heb.
…‫( גם אתה‬18.18) TgJ has ‘both you and Aaron and his sons and the
elders’, based on the ‘inclusive’ understanding of ‫ גם‬in the interpretation given
in MRI (Lauterbach, p. 181), where Aaron’s sons are specified as Nadab and
Abihu: this points to Exod. 24.9-14 as the source of the expansion. A larger
body of judges (a prototype Sanhedrin?) is assumed to exist already and to be
spared from over-work: the wider benefit is set aside. By contrast LXX and Sy
prefix ‘all’ to ‘this people’ (cf. vv. 13-14): the later Jewish revisers remove it.
‫( כבד ממך‬18.18) LXX βαρύ σοι (cf. μέγα σοι in Isa. 49.6) seems at first
sight to miss the significance of the comparative ‫מן‬, to which Gk. has a closely
corresponding idiom (cf. LSJ, p. 2040, s.v. ὥστε). But the Greek idiom could
610 EXODUS 1–18
be abbreviated and elsewhere LXX shows awareness of the sense of the

Heb. in various ways (Gen. 18.14; Num. 11.14; Deut. 1.17; 17.8; 2 Kgs 6.1):
perhaps, therefore, the simple adjective conveys the meaning ‘too heavy’ here.
Vulg’s freer ultra vires tuos brings it out more clearly.
‫( הדבר‬18.18) LXX’s addition of τοῦτο, like its omission of the pronoun
in the following clause, shows its readiness to modify the wording when it
thought fit.
‫( עשׂהו‬18.18) SP ‫ עשׂותו‬replaces the unusual form with its regular counter-
part.
‫( עתה‬18.19) The Heb. texts (including 4QpalExl) all make the transition
to Jethro’s instructions with the simple adverb (cf. Tgg, Sy), as in Gen. 31.13
and Num. 22.11. LXX νυν οὐν (as often for ‫ )ועתה‬and Vulg sed introduced the
stronger link with what precedes, whether causal or contrastive, that Greek
and Latin canons of style normally prefer.
‫( איעצך‬18.19) All SP mss and recent edd. spell the word ‫( אעיצך‬so also
at Num. 24.14), making the root Ayin Yodh rather than Pe Yodh (cf. the Ayin
Waw by-form in MT at Judg. 19.13; Isa. 8.10, in JAram and probably in
Deir Alla 2.9 [cf. DNWSI, p. 834]). For the occurrence of such variations in
Samaritan Heb. see GSH §61a. Vulg has paraphrased with atque consilia mea
(a second object to audi), which lends some support to the suggestion (in Note
p on the translation) that ‫ איעצך‬is a relative clause without ‫אשׁר‬: atque, like
LXX’s καί, is a further example of the quest for explicit linkages.
‫( ויהי‬18.19) Both LXX (καὶ ἔσται) and Vulg (et erit) take the jussive in a
future indicative sense (cf. Note q on the translation): the forms in the Aram.
Vss are ambiguous.
‫( היה‬18.19) As elsewhere for this form, SP has ‫( הוי‬cf. Gen. 12.2; 17.1;
27.29; Exod. 24.12; 34.2: and for the fem. in Gen. 24.60), with variants ‫הוה‬
and ‫הוא‬. Both waw as the middle root-letter and the orthography reflect the
influence of Aram. on Samaritan Heb. (GSH §32aε), but for other forms of the
verb the traditional spelling was retained (cf. ‫ ויהי‬earlier in the verse).
‫( מול האלהים‬18.19) The physical proximity implied by ‫( מול‬see Note s on
the translation) is avoided by LXX, as in 17.6, but in a different way that fits
the present context well: τὰ πρὸς τὸν θεόν/κύριον is a common device of
the translators (Fritsch, pp. 55-56), which is used here (as in 4.16: cf. Deut.
1.36; 9.7, 24; 31.27) in an adverbial sense (cf. Rom. 15.17; Heb. 2.17; 5.1:
BDF §160).
‫( והבאת אתה‬18.19) LXX καὶ ἀνοίσεις and Vulg ut referas may omit any
equivalent to ‫ אתה‬because it was not strictly necessary, but for LXX at least the
reason might here lie in its Vorlage, as 4QpalExl seems to have had a shorter
text at this point (DJD IX, p. 40) and omission of ‫ אתה‬by homoeoteleuton is
the most likely explanation (cf. Wevers, Notes, p. 286 n. 18).
‫( את־הדברים‬18.19) LXX, even though it is even less necessary in Gk. than
Heb., added the possessive pronoun αὐτῶν (but Origen’s obelus is preserved
in Syh), as did TgJ and Sy, presumably to reinforce Moses’ representative role.
18.13-27 611
‫( והזהרתה‬18.20) SP omits the ‫ה‬-, which it retains only with forms of ‫נתן‬
(GSH §62b). No Qumran ms. preserves the word. The Aram. Vss use the
cognate verb, for which a sense ‘teach, explain’ is recognised in Tgg here
(Jastrow, p. 382; CAL: cf. Vulg ostendasque); in Sy dnṭrwn, ‘that they should
keep’, is added and Payne Smith (p. 111: cf. CAL) recognises only the sense
‘warn, admonish’. LXX καὶ διαμαρτυρῇ (second person sing., fut. mid.) is
commonly taken (e.g. by Wevers, Notes, p. 286; Muraoka, Lexicon, p. 117)
to mean ‘testify, bear witness’ here, but both classical and Septuagintal usage
show that it can also refer to other kinds of solemn affirmation or declaration
(LSJ, p. 403; Exod. 19.10, 21; Deut. 32.46). Aq substituted διαστέλ(λ)ου,
‘command’, probably not to change the meaning but to clarify it.
‫( אתהם‬18.20) SP substitutes ‫אתם‬, as it does in Gen. 32.1; Num. 21.3, to
create complete consistency in the use of this form throughout the Pentateuch.
4QpalExl is very difficult to decipher at this point, but its mem is more likely
to be an object suffix than (as DJD IX, p. 40, proposes) the final letter of an
independent word, because of the limited space available.
‫( את־החקים‬18.20) LXX adds τοῦ θεοῦ to agree with v. 16.
‫( ואת־התורת‬18.20) 4QpalExl, TgO,J and Sy correspond to the MT vocalisa-
tion, but LXX (with the addition of αὐτοῦ as in v. 16) and SP have the sing.
and TgN repeats its longer equivalent: see Text and Versions on v. 16. For this
and the previous phrase Vulg has caerimonias et ritum colendi, probably in
line with a Jewish exegetical tradition that is reflected in TgJ’s treatment of the
second half of the verse (see below): due recognition is then given to the place
of ritual as well as moral teaching in the Mosaic law (which is assumed to be
meant). Both caerimonia and ritus occur frequently, with various equivalents,
in the Vulg Pentateuch. Two examples of a quite different branch of Jewish
exegesis are attested in MRI (Lauterbach, p. 182): both agree that the second
phrase refers to ‘decisions’, i.e. particular rulings (in line with the narrative
context), but the first is applied to ‘interpretations’ or to the sexual laws in Lev.
18 (by gezerah shawah on ‫ חקות‬in v. 30).
‫( את־הדרך ילכו בה‬18.20) As one would expect, the Vss insert a rel. pron.
or particle before the verb (as in the following expression), but there is also
strong attestation for ‫ אשׁר‬at this point in SP, 4QpalExl and 4QpalExm (for
other places where SP makes such an addition see GSH §58b: its motive is
probably consistency in each case, since it tolerates a similar omission in 9.4
and in Num. 23.13). The more difficult reading of MT (on which see Note u
on the translation) is as usual to be preferred. TgJ follows an interpretation
of these words in MRI (ibid.), which applies them to the care of the sick and
the deceased, but places before this ‘the prayer which they shall pray in their
synagogues’, which may claim a Mosaic origin for the Shemoneh Esreh.
‫( ואת־המעשׂה אשׁר יעשׂון‬18.20) TgJ again follows an interpretation in
MRI (ibid.) which opts for a specific application, in this case to a distinc-
tion between strictness and leniency in judicial practice. For other relevant
rabbinic texts see AramB 2, p. 213 n. 24.
612 EXODUS 1–18
‫( ואתה‬18.21) Vulg has only autem, omitting the independent pronoun.

‫( תחזה‬18.21) TgJ ‫ברור‬, ‘select’, anticipates the sense of v. 25. SP, LXX
and possibly 4QpalExl (DJD IX, p. 40) add ‘for yourself’, to indicate the role
of those chosen as assistants to Moses; TgN adds, equally unnecessarily, ‫גברין‬,
‘men’, before the list of attributes in which it will twice appear again.
‫( יראי אלהים‬18.21) 4QpalExl has ‫ יהוה‬in place of ‫ אלהים‬to agree with
the more widespread form of the expression. LXX θεοσεβεῖς was found
insufficiently precise by the Three, who replaced it with φοβουμένους τὸν
θεόν (cf. Vulg).
‫( אנשׁי אמת‬18.21) LXX δικαίους employs, freely, a general word of
ethical approval: the other Vss are more precise. Vulg in quibus sit veritas,
with its two additions of et, turns the list into a more stylish description.
‫( שׂנאי בצע‬18.21) LXX renders ‫ בצע‬with ὑπερηφανίαν, ‘arrogance’,
which is far from the sense required and never renders ‫ בצע‬elsewhere.
Sanderson, Exodus Scroll, pp. 63-64, has argued strongly that LXX’s (fairly)
precise rendering of the other attributes suggests that it had a different (if
inferior) Vorlage from MT, SP and 4QpalExm (4QpalExl does not preserve
the word), but she does not say what it was. Wevers (Notes, p. 287) prefers
to think that the translator was exposing the underlying motivation of ‫בצע‬.
Whether the variant originated in Heb. or Gk., it reflects the same antipa-
thy to arrogance that appears in the Psalms and Ben Sira. The other Vss
follow the reading ‫ בצע‬more closely but, whereas TgJ, Sy and the original
reading of TgN do justice to its connotation of dishonesty, TgO ‫( קבלא ממון‬to
which a ‘censor’ modified TgN) and Vulg avaritiam (its regular equivalent
for ‫בצע‬: cf. πλεονεξίαν in the Three) exclude the ‘greedy’ in general from
consideration.23
‫( ושׂמת‬18.21) The object is supplied in 4QpalExl (‫ )אתם‬and LXX (αὐτούς):
cf. Vulg’s ex eis, which actually displaces any equivalent to ‫עליהם‬.
‫( שׂרי אלפים‬18.21) Vulg’s tribunos (from OL, like centuriones) is a precise
equivalent, since there were six in a legion of 6000 men. LXX could find older
military terms for all four categories of officers.
‫( שׂרי מאות‬18.21) SP prefixes a waw, dividing Moses’ assistants into two
pairs: similarly 4QpalExm, LXX, Vulg, TgN and Sy, which (like MT in Deut.
1.15) also add waw before the next phrase. MT’s reading there and probably
here (DJD IX, p. 40) is supported by 4QpalExl (and cf. v. 25). 2QExb has the
spelling ‫מאיות‬, with consonantal yodh (GK §97g; Reymond, Qumran Hebrew,
p. 122), which was formerly described as ‘Aramaising’: it also occurs as the
Ketib in 2 Kgs 11.4, 9-10, 15 and in 11Q 19 42.15, all as it happens in the
same phrase as here.
23
MRI (Lauterbach, p. 183) is closer to the former interpretation with its
limitation to behaviour ‫בדין‬, ‘in court’.
18.13-27 613
‫( שׂרי חמשׁים‬18.21) Latin military terms failed Jerome here, as the Roman
army did not have units of fifty men: Vulg’s quinquageniarios was an old
adjective that was now put to new use.
‫( ושׂרי עשׂרת‬18.21) OL had rendered LXX’s δεκαδάρχους with decuri-
ones, a military term that was used by Julius Caesar in his Bellum Civile.
Jerome preferred decanos, a more recent coinage employed by the fourth-
cent. writer Vegetius (LS, p. 516) which had already found its way into
monastic use (Jer., Ep. 22.35) and was later adapted to designate an ecclesi-
astical or academic ‘dean’.
‫( ושׁפטו‬18.22) Vulg qui iudicent neatly creates both a tighter connection
and an expression of purpose, which is presumably continued by the subjunc-
tives in the rest of the verse.
‫( בכל־עת‬18.22) LXX πᾶσαν ὥραν follows the use of the accusative of
ὥρα for an instant of time, which was already common in class. Gk. (LSJ,
p. 2036; cf. BDF §161).
‫( כל־הדבר הגדל‬18.22) LXX intensified ‫ גדל‬with ὑπέρογκον, ‘excessive,
very large’, no doubt implying that very few cases would now come to Moses
(see also the note below on ‫)הקטן‬: the Three replaced it with the more precise
μέγα. Sy rendered ‫( כל‬which is the only word in the verse to survive in
2QExb) with kd nhyʾ here and with kd later in the verse, turning the anteposed
objects into temporal clauses, perhaps for easier comprehension.
‫( כל־הדבר הקטן‬18.22) LXX, Vulg and TgN put their equivalents in the pl.,
again as a way of indicating that most cases would no longer come before
Moses.
‫( והקל‬18.22) The same Heb. text appears in SP and (though no trace of
the waw survives) in 4QpalExl, but among the Vss only Sy ʾql and possibly
TgNmg render it as it is generally understood now, i.e. as an imperative sing.
(on the idiomatic use see Note aa on the translation). LXX and TgO,J have third
person pl. future verbs and TgN and Vulg have an impersonal third person
sing., perhaps because they understood the Heb. as an inf. abs., which could
stand for any finite form of the verb (cf. GK §113y-z and Rashi ad loc.): in the
context a future sense would fit very well.
‫( תעשׂה‬18.23) TgJ adds a clause to make explicit that Moses’ ‘freedom’
from judging is meant. MRI (Lauterbach, p. 184) improbably connects the
condition with what precedes it.
‫( וצוך אלהים‬18.23) TgO,N is content with a literal rendering which implies
(as MRI [Lauterbach, p. 185] comments) that Moses needs to consult God
first. By omitting the conjunction and using a participle for ‫צוך‬, Sy is probably
not making the apodosis begin here (Wevers, Notes, p. 288: cf. below) but
treating the words as a parenthesis in which Jethro claims divine support for
his recommendation (cf. the comment in MRI on ‫ אמר‬in v. 24). The other
Vss do see these words as the first consequence of compliance with it. LXX,
perhaps puzzled by them, has κατισχύσει σε, ‘will strengthen you’, which
614 EXODUS 1–18
is clearly a deduction from the next clause (and not based on a different
Vorlage). Vulg implebis imperium Dei appears to be a paraphrase of the
interpretation presupposed in Sy (which Childs and Houtman also favoured
in modern times), while TgJ’s addition of ‘the commandments’ introduces its
application of the next clause to Moses’ imminent task of receiving the Law
on behalf of the people.
‫( ויכלת עמד‬18.23) The sense of ‫ עמד‬is evidently ‘endure’ and even LXX’s
curious παραστῆναι (which looks like a lazy repeat of the more appropriate
use of the verb for ‘stand by’ in v. 13) could perhaps have meant this (see
Diodorus Siculus 17.43 and other exx. in LSJ, p. 1341). TgJ adds ‘to hear them
(sc. God’s commandments)’ and this midrash evidently lies behind Vulg’s
otherwise remarkable et praecepta eius poteris sustentare.
‫( וגם כל־העם הזה‬18.23) The introductory ‫ גם‬was presumably again the
basis (as in v. 18) for TgJ’s application of these words to ‘Aaron and his sons
and all the elders of the people’ (cf. MRI, p. 185) and of the following phrase
to their ‘court-house’.
‫( על־מקמו‬18.23) SP and 4QpalExm read the more usual ‫ אל‬and LXX’s εἰς
could well be based on such a Vorlage. It might be the original reading (so
Sanderson, Exodus Scroll, pp. 58-59), but MT’s ‫ על‬also appears (with little
doubt) in 4QpalExl (cf. TgO,J) and the use of ‫ על‬to indicate motion towards is
frequent enough in BH for it not to be a problem (cf. Note dd on the transla-
tion). In fact it is more likely that ‫ אל‬would be secondarily substituted for it
than vice versa. There is some indication in the Vss of an understanding of
‫ מקמו‬to refer not to the land of Canaan but to individuals’ ‘homes’: Vulg and
TgN have ‘places’ in the pl. and Sy ʾnš lbyth is even more explicit (cf. OL’s
added [i]n domus suas and Wevers, Notes, p. 289, on LXX’s τὸν ἑαυτοῦ
τόπον). The non-standard renderings of ‫ יבא‬in Vulg (revertetur) and TgO (‫יהך‬,
rather than ‫ייתי‬: on the variation see AramB 7, p. 51 n. 11) may point in the
same direction (on TgJ see the previous note).
Little survives of 18.24 in 4QpalExm but DJD IX, p. 98, calculates that
both it and v. 25 opened a new line, with empty space and an enlarged waw at
the end of the previous line (cf. the division before v. 25 in SP). It also appears
that there was a space between v. 24a and v. 24b. In 4QpalExl the text seems
to have continued without such spaces at least until the end of v. 24: after that
no evidence survives, as the rest of the chapter is lost.
‫( לקול־חתנו‬18.24) LXX* did not represent the suffix, but αὐτοῦ appears
in the O-text (with an asterisk) and some other witnesses. Vulg quibus auditis
abbreviates the first half of the verse even more. TgO,J interpret ‫ קול‬with ‫מימר‬
in a non-technical sense (as e.g. in Gen. 3.17: cf. TgNmg here).
‫( כל אשׁר‬18.24) LXX ὅσα conveys the sense of the MT reading (cf. 6.29;
9.19; 20.17 etc.).
‫( אמר‬18.24) Sy as often adds lh (cf. v. 14). MRI (Lauterbach, p. 185)
records the view that the subject of this verb is God.
18.13-27 615
In place of 18.25 SP and 4QpalExm have an adaptation of the parallel story

in Deut. 1.9-18, where Moses recalls the installation of additional judges
without any reference to the role of Jethro. This expansion is also preserved
in translation in SamTg and in a Syriac version which appears in the margins
of Syh (for the text of the latter see Lagarde’s 1892 edition; a Latin translation
is provided in the apparatus critici of the Septuagint editions of Brooke-
McLean and Wevers). It has the same general characteristics as the expansions
of the plague narratives in these texts, but in this case the ‘borrowed’ text
comes from a different book of the Pentateuch (as happens again in Exod.
20). The wording of Deuteronomy has undergone minor revisions to fit the
new context.24 Some pronominal morphemes are changed (first person sing.
to third person; second person pl. to third person) because the section is no
longer part of an address by Moses to the people but a report of what he and
they say and do. The words ‫( בעת ההיא‬Deut. 1.9, 16, 18) are omitted because
the account is no longer retrospective. The introduction is expanded to
‫ויאמר משׁה אל העם‬. Otherwise the match to the SP text of Deuteronomy is very
close, with only two significant variations.25 There is no equivalent to ‫ אתי‬in
Deut. 1.14 (probably a more drastic instance of the removal of first person
forms) and in Moses’ opening words ‫ אנכי‬is added before ‫( לבדי‬as in the same
expression in Num. 11.14). Possibly both of these variations belong to a
secondary stage of the expansionist text: in 4QpalExm there is a lacuna where
‫ אתו‬would appear and the ms. clearly did not include ‫( אנכי‬DJD IX, pp. 98-99).
They are both reflected in the Syriac version, so they cannot be too late. This
text at several points follows variations in the LXX translation, as elsewhere,
but not invariably (see the general Introduction and J. Joosten in Lange and
Tov, Textual History, IA, pp. 237-38 [1.3.2.2]). The older recension of SamTg
(J) also has both the variations from Deuteronomy, which is significant in
view of its otherwise generally careful adherence to the Heb. The substituted
text does include a close equivalent to the second half of v. 25 (but not the
first half) in MT and the non-Samaritan Vss. This should not, however, be
understood as a sign that the scribe responsible for it saw the MT of this verse
as his basis for an ‘exegetical’ development. The author of Deuteronomy 1
had already incorporated the sentence in question (cf. 1.15) as well as some
other features of Exod. 18.13-26 into his drastic revision of the narrative, and
the addition of the latter now provided an important corrective to the older
text. The result was to make the judicial reform a matter negotiated between
24
For a helpful presentation of the changes see J.H. Tigay, ‘An Empirical
Basis for the Documentary Hypothesis’, JBL 94 (1975), pp. 329-42 (333-35).
25
There are three further variations from the MT of Deuteronomy at places
where it differs from the SP text: the expansion has no waw before ‫ משׁאכם‬in 1.12
or ‫ שׂרי חמשׁים‬in 1.15 (both in lists) and the pl. imperative ‫ שׁמעו‬appears in place of
MT’s inf. abs., which is probably the original reading in this case.
616 EXODUS 1–18
Moses and the people (and the people now ‘choose’ their judges, not Moses),
with a strongly Yahwistic foundation (from Deut. 1.10-11) and additional
qualifications and instructions for Israel’s future judges (Deut. 1.13, 15-17).
In the first or second century B.C. Jethro’s advice could no longer be elimi-
nated, but it could be effectively marginalised as a mere preliminary to Moses’
consultation of the people (see further the observations of Mayes, Deuter-
onomy, pp. 118-19, 121-25, and Weinfeld, Deuteronomy 1–11, pp. 137-41, on
the redactional aims of the Deuteronomist here). The substituted text also has
a harmonising motive but, given the fact that it appears in Deuteronomy at a
later point in the story when Israel was departing from, not arriving at, Mount
Sinai/Horeb, it achieved its purpose at the cost of introducing a new element
of anachronism into Exodus 18.
‫( ויבחר משׁה אנשׁי־חיל‬18.25) Vulg et electis viris strenuis uses the ablative
absolute construction again to tighten the sentence structure (cf. v. 24) and so
avoids the repetition of Moses’ name (as it does again in v. 26), while at the
same time removing one of the contradictions between Exodus and Deuter-
onomy here. strenuis is a typical stylistic variation from the rendering of
‫ אנשׁי־חיל‬in v. 21. TgN again has an unnecessary ‫גברין‬.
‫( ויתן‬18.25) Symm κατέστησεν is the more usual word for appointments
(cf. LXX in v. 21 for ‫ )ושׂמת‬and the O-text adopted it in preference to LXX’s
unspecific ἐποίησεν here (cf. Vulg constituit).
‫( ראשׁים על־העם‬18.25) LXX’s abbreviation to αὐτοὺς ἐπ’ αὐτῶν is
conforming to the wording of v. 21; Vulg’s principes populi is a good ad
sensum equivalent.
‫( שׂרי אלפים וג׳‬18.25) The treatment of the list in the Vss on the whole
corresponds to v. 21 (see the notes there), but TgO,J group the officials into two
pairs (as in SP) by adding a waw before the second group and TgJ inserts the
numbers for each category on the basis of the total population of 600,000 in
12.37 (as in MRI [Lauterbach, p. 183]).
‫( ושׁפטו‬18.26) MT’s perfect consecutive clearly implies repeated action,
like the imperfects later in the verse, and the Vss mostly reflect this. SP has
‫וישׁפטו‬, the imperfect consecutive (cf. SamTg ‫)ואדינו‬: it does not generally differ
from MT in such cases (cf. 16.21; 17.11; 33.7-11) and its reading is no less
possible here, whether it be seen as treating the action of judging as a single
whole or as an ‘unmarked iterative’ (Joosten, Verbal System, pp. 174-75,
with a few examples). It is thus impossible to be sure which reading is more
original, but MT might perhaps be seen as a secondary ‘marking’ to fit the
following verbs.
‫( הקשׁה‬18.26) Tgg, Sy and Vulg agree with MT, but SP has ‫הגדול‬, an
assimilation to v. 22. LXX τὸ ὑπέρογκον, also used for ‫ הגדול‬in v. 22, could
be a rendering of either reading: Aq and Symm, with σκληρόν, and Theod
δυσχερές clearly represent MT.
‫( וכל‬18.26) TgN, Sy and Vulg omit ‘all’, probably to match the parallel
phrase.
18.13-27 617
‫( הקטן‬18.26) LXX ἐλαφρόν, ‘easy’, and Vulg faciliora depart from

their renderings in v. 22: this is perhaps an indirect indication that LXX also
presupposes the reading ‫ הקשׁה‬earlier in this verse.
‫( ישׁפוטו הם‬18.26) SP spells the verb in the normal way, without the first
waw. Vulg iudicantes neatly removes the need for et before faciliora and,
unlike v. 22, an equivalent for ‫ הם‬is deemed unnecessary.
4QpalExm has a long mid-line break after 18.26, to which there is no
equivalent in MT or SP. It does serve to mark v. 27 as the conclusion to the
whole chapter.
‫( את־חתנו‬18.27) LXX τὸν ἑαυτοῦ γαμβρόν, with a possessive pronoun
for the first time since v. 7 and this time in its more emphatic reflexive form.
Vulg, which had substituted pronouns in vv. 15 and 24, has cognatum here,
but without a possessive.
‫( וילך לו‬18.27) 4QpalExm reads ]‫ [וילך אל‬here, i.e. without ‫לו‬, which both
MT and SP have (on its function see Note hh on the translation). LXX and Sy
do not represent ‫ לו‬here: in the same idiom at Gen. 12.1 and 22.2 LXX does
likewise, but Sy follows MT. It is not surprising that Vss and even a Heb. ms.
would sometimes ignore an idiom that might well have become unintelligible
(cf. Wevers, Notes, p. 291) and there is no reason to depart from the MT/SP
reading.
‫( אל־ארצו‬18.27) TgJ ‘to make proselytes of all the people of the (ed. pr.
“his”) land’ extends a theme which it introduced in vv. 6-7 and picks up the
rationale which MRI (Lauterbach, pp. 185-86, cf. 191) saw for Jethro/Hobab’s
decision to leave in Num. 10.30.
After these words there was a vacat in 4QpalExm: while there is no firm
proof for DJD’s inference that in it 19.1 began a new line (IX, p. 99), it is
inherently likely and in any case the location of ‫ וילך‬directly above ]‫השׁל[ילשׁי‬
in the next line makes it virtually certain that there was an empty space (about
half a line) between 18.27 and 19.1.
618

G. I. Davies - A Critical and Exegetical Commentary On Exodus 1 - 18 Volume 2 - in Two Volumes - Commentary On Exodus 11 - 18-Bloomsbury Publishing PLC (2020)

Uploaded by

Copyright:

Available Formats

G. I. Davies - A Critical and Exegetical Commentary On Exodus 1 - 18 Volume 2 - in Two Volumes - Commentary On Exodus 11 - 18-Bloomsbury Publishing PLC (2020)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

G. I. Davies - A Critical and Exegetical Commentary On Exodus 1 - 18 Volume 2 - in Two Volumes - Commentary On Exodus 11 - 18-Bloomsbury Publishing PLC (2020)

Uploaded by

Copyright:

Available Formats

The

FORMERLY UNDER THE EDITORSHIP OF

J. A. EMERTON, F.B.A., C. E. B. CRANFIELD, F.B.A.

Commentary on Exodus 11–18

First published in Great Britain 2020

Copyright © G. I. Davies, 2020

All rights reserved. No part of this publication may be reproduced or transmitted

ISBN: HB: 978-0-5676-8871-2

Series: International Critical Commentary

Typeset by Forthcoming Publications (www.forthpub.com)

Introductory Note vii

PASSOVER AND DEPARTURE

Chapter 12.28-42, 50-51

THE DELIVERANCE AT THE SEA

THE JOURNEY TO THE MOUNTAIN OF GOD

This volume represents the second part of the commentary on

Moses’ confrontations with Pharaoh have reached a decisive point

of a division before v. 4 in 4QpalExl and 2QExa has been inferred

Knobel pioneered the long-standing division of the passage between

Most analyses of this passage have taken their starting-point at

grammatical references (cf. ‘you’ in vv. 1, 7 and 8) in the two

Blum and Van Seters hold) or an originally independent source.

Verses 9-10 were originally the introduction to a second

1 [Yahweh said to Moses, ‘One more plaguea I will bring on

Notes on the Translation

The statement has often been seen as premature, since it appears

than Israel (Nachmanides), to whom it refers in vv. 2-3a: mention

Jeremias, Theophanie, pp. 7-12, 148; TWAT 3, 803-804 = TDOT 6,

this more detached and perhaps more basic use of cry-language is

be a reference to his repeated not listening in the earlier narratives of

Text and Versions

Emendations to a sufﬁxed form of ‫ כל‬or a form of the verb ‫ כלה‬have been

‫( וכלי זהב‬11.2) In LXX and Vulg there is no equivalent to ‫ כלי‬this time,

translator’s inference from a passage such as 10.16-17, the addition of πάντων

‫( אני יוצא‬11.4) LXX’s εἰσπορεύομαι (followed, in a future form, by OL)

‫( המפתים‬11.10) LXX preﬁxes τὰ σημεῖα καί, as in v. 9, and TgN has

I nstr u cti o n s fr om Y ah weh a b ou t t he

The boundaries of this section are clearly deﬁned by the introduc-

between these sections follows v. 13 or v. 14: see the Explanatory

There is good reason to attribute the whole passage to the Priestly

be as late as the Holiness Code, but the mention of ‘holy occasions’

deliverance for the Israelites is going to be provided need by no

as an assurance that he can do the same again. The Exodus tradi-

1 Yahweh said to Moses and Aaron in the land of Egypt, saying:

Notes on the Translation

w. The three short phrases, coordinated like a series of nouns by waw

this among its examples of conditional clauses, but since no uncertainty is

identify the different months of the year. There is no verb in the

frequently in the Priestly sections throughout the Tetrateuch (so

‘Unblemished’ (v. 5: Heb. tāmîm) is a requirement for sacriﬁcial

blood of circumcision in 4.25-26 (but not in the Priestly instruc-

roasting as opposed to boiling was that the animal could be cooked

explanation corresponds closely but not exactly to the non-Priestly

Egypt is based on Yahweh’s remembering of his covenant with

14. There is some disagreement over whether this verse belongs

possible. The Priestly writers were probably aware of recent moves

the evening of Passover may be an addition to the seven full days

Text and Versions

‫( תותירו‬12.10) The original LXX reading ἀπολείψετε is preserved only

Introductory Note vii