A Revised Chronology of Ghazālī's Writings

Author(s): George F. Hourani

Source: Journal of the American Oriental Society , Apr. - Jun., 1984, Vol. 104, No. 2
(Apr. - Jun., 1984), pp. 289-302
Published by: American Oriental Society

Stable URL: https://www.jstor.org/stable/602173

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

American Oriental Society is collaborating with JSTOR to digitize, preserve and extend access
to Journal of the American Oriental Society

This content downloaded from

194.27.127.99 on Thu, 26 May 2022 13:21:49 UTC
All use subject to https://about.jstor.org/terms
 A REVISED CHRONOLOGY OF GHAZALI'S WRITINGS

GEORGE F. HOURANI

STATE UNIVERSITY OF NEW YORK AT BUFFALO

The author's first "Chronology," published in the Journal, volume 79 (1959), is here completely
rewritten. The method remains the same, using only the evidence of historical sources and cross-
references from GhazalI's books, which are plentiful, and avoiding inferences from any observed
or supposed development in GhazalT's thought. The conclusions are more accurate and definite
than before, owing to more careful observations and with the help of studies by Bouyges,
Badawi, and Lazarus-Yafeh not available to the author at the time of his original article.

THE ORDERING OF GHAZALT's WORKS is a task which Dissatisfied with all preceding attempts known to
naturally attracts attention for several reasons: the me, I worked out the chronology afresh in an article
existence of a biographical framework in his own published in 1959,2 following the method to be
Munqidh, the multitude of cross references in his described below. The result was a further advance in
books, and above all the necessity of establishing an accuracy, presented in a handy format. But my effort
order as a basis for understanding the development of was surpassed in the same year by the publication of
his thought, which so clearly did not remain constant an entire book on the subject whose existence had
throughout his life. been unknown to me: that of Bouyges, which he had
Pioneer attempts at a chronology were published in completed in 1924 but which had lain among his
the first half of the twentieth century by Goldziher, papers until after his death, when it was published by
Massignon, Asin Palacios and Watt.' Watt's list was Allard.3 Bouyges' book is very detailed and remains
an advance on anything previously published in the an indispensable aid to the study of GhazalT. Yet
general correctness of its order and the presence of many of his conclusions need to be reconsidered half
many references. But it was incidental to the main a century after he wrote, and even twenty years after
purpose of his article, and fell short of desirable Allard's edition with its supplementary notes.
fullness in omitting the works of fiqh, not connecting
Since 1959 two more books have been published
the works listed with biographical data such as known which have made important contributions to the ques-
dates in GhazalT's career, and not containing dis- tion of chronology. One is Badawi's bibliography of
cussion of doubtful points. Moreover, Watt's groups the manuscripts and printed works of GhazalT.4 The
are not purely chronological but are defined by the other is a collection of Ghazalian studies by Lazarus-
topics and doctrines of the works. While the four Yafeh.5 There have also been many new editions of
groups correspond roughly with four periods in GhazalT's works and a few monograph studies, all of
GhazalT's life, there is some overlap in time between which will be mentioned below in the text or notes.
particular works in different groups. The sum of these contributions and my own further
observations since 1959 have made it rewarding and

I I. Goldziher, Die Streitschrift des GazalT gegen die

Bdtinijja-Sekte (Leiden, 1916), pp. 25-29; L. Massignon, 2 G. F. Hourani, "The chronology of GhazalT's writings,"
Recueil de textes inedits concernant ihistoire'de la mystique Journal of the American Oriental Society, 79 (1959),
au pays d'Islam (Paris, 1929), p. 93; M. Asin Palacios, La pp. 225-33.
Espiritualidad de Algazel, I (Madrid, 1935), pp. 35-36; 3 M. Bouyges, Essai de chronologie des oeuvres de al-
W. M. Watt, "The authenticity of the works attributed to al- GhazdlT (Algazel), ed. M. Allard (Beirut, 1959).
GhazdlI, " Journal of the Royal Asiatic Society (1952), 4 A. Bddawf, Mu'allafdt al-Ghazali, 2 vols. (Cairo, 1961).
pp. 24-45, including brief "Notes on chronology" (pp. 43- 5 H. Lazarus-Yafeh, Studies in Al-GhazzalT (Jerusalem,
44). 1975).

289

This content downloaded from

194.27.127.99 on Thu, 26 May 2022 13:21:49 UTC
All use subject to https://about.jstor.org/terms
290 Journal of the American Oriental Society 104.2 (1984)

even imperative to offer a new chronology, which have led to the order of the works in the lists. Serial
makes use of all the preceding work and should give numbers are provided only for a few small groups
the most accurate account possible until new evidence within which an order can be determined. It will be
is discovered. This article is based on my previous one evident that many uncertainties remain, but the
but modifies it extensively. chronology as a whole should furnish a sufficient
The aim is to determine the order and dates of basis for placing Ghazali's works and thought in a
GhazalT's works in so far as they can be learned from meaningful biographical context.
the more conclusive kinds of evidence: his own refer-
ences to previous or projected works, and biographi- For the purpose in view GhazAli's life can be
cal data gathered from his Munqidh and other early divided into four periods, distinguished as phases of
sources.6 Only by limiting ourselves to such evidence teaching activity and retirement.
in the first place can we provide a solid basis for (1) An early period of teaching and writing extends
tracing the development of GhazalT's thought. To from an unknown date preceding the death of
reverse this procedure by making inferences from the Juwayni, the Imam al-Haramayn (d. 478=1085/86),
intellectual contents of works to their dates is bound to GhazalT's departure from Baghdad at the end of
to lead to arbitrary and erroneous results, and has 488 (1095). Born at Tis in 450=1058, he had gone to
already done so. After a chronology has been estab- Nishapuir as a youth to study under the Imam, and it
lished by sound methods the intellectual biography of is known that he started teaching and writing there
Ghazal! can be studied more accurately. Such studies while the Imam was still alive.7 Only one work is
in turn may at length allow scholars in later times to definitely assignable to this time, Mankhil (see below
refine the chronology by making use of a more or less and note 10). After Juwayni's death GhazalT was
known order of development in his thought. attached to the camp-court (ma'askar) of the Seljuq
Problems of authenticity will not be treated to any sultan's great wazlr, Nizam al-Mulk, and he remained
great extent. I have listed those writings which are there in high favor for some six years. In 484=1091/92
generally accepted as genuine, and whose positions in he was appointed to the chair of Shafi'ite law at the
the series can be determined to some extent by the Nizamiyya College in Baghdad, and he taught there
kinds of evidence being used. Five works which I for four years, 484-88=1091/92-1095.
reject as spurious are listed because they are well GhazalT informs us in two places that he wrote on
known and call for brief accounts. They are bracketed law and jurisprudence in this first period of teaching.
and inserted at places where they would have been (i) Mustasfd, I, 38 states: "In the prime of my youth
most likely to occur if they had been genuine. Beyond ... I composed many books on law and jurisprudence
these works there are many others listed by Bouyges (fT fura'i l-fiqhi wa usalihi); then I turned to the
and BadawT of which little is known; an attempt to science of the way of the afterlife and acquaintance
decide their authenticity and dates would have led us with the inner secrets of religion." As we know from
too far afield and would probably have been almost Munqidh, the second sentence fixes a latter limit
fruitless. Nor have I recorded the many variant titles around 488-1095 for most of Ghazali's legal writings,
of works, as Bouyges and BadawT have done; I have with the exception of Mustayfd itself which is much
used the title given in the edition referred to. later. (ii) Munqidh, 85 states that he worked on
The system of numbering adopted in my previous philosophy "in my spare time between writing and
article has been abandoned. It gave a misleading lecturing on the scriptural sciences (al- 'ulumi sh-
impression of accuracy which could not be sustained. shar'iyya)," i.e., law and jurisprudence. Since this
The present list is in fact more accurate, and the statement is in the context of Baghdad, we know that
expanded discussions give all the considerations at least some of his treatises on law and jurisprudence
which

6Sources have been given by D. B. Macdonald, "The life 7 Ibn cAsakir, TabyTn kadhb al-muftard, ed. A. F. Mehren,
of al-GhazzalT," JAOS, 20 (1899), pp. 7 1-132, and "al- "Expose de la reforme de l'Islamisme," 3rd International
Ghazzali," Shorter Encyclopaedia of Islam (Leiden, 1953), Congress of Orientalists, Transactions, II (Leiden, 1879),
pp. 111-14; and by F. Jabre, "La biographic et l'oeuvre de p. 322.

GhazalT reconsiderees a la lumiere des Tabaqdt de SobkT," 8 All page references are to the edition mentioned for each
.41langes de l'Institut Dominicain db'tudes Orientales, Ibook in the list, except in a few cases where a reference could
(Cairo, 1954), pp. 73-102. be obtained only indirectly through another modern author.

This content downloaded from

194.27.127.99 on Thu, 26 May 2022 13:21:49 UTC
All use subject to https://about.jstor.org/terms
 HOURANI: A Revised Chronology of GhazdlT's Writings 291

were written there. Apart from this fact and the one (3) He came out of retirement in Dhui al-Qa'da,
early work of NTshapur, we cannot distribute the 499=July, 1106, to resume teaching in Nishapur. The
works on these subjects precisely among the three end of this period is not known, but 503=1109/10 is a
sub-periods at Nishapur, the camp-court and Baghdad. fair guess, because he completed Mustasfd, a major
The years in Baghdad were also a time of extensive work on jurisprudence, on Muharram 6, 503=August
writing on new subjects, as will be seen from the list 5, 1109, and writing on this subject does not seem like
between Maqdsid al-faldsifa and Mlzdn al- 'amal. an occupation of retirement.
Munqidh, 79ff. specifies the order of his studies of (4) A final retirement at TUs lasted until his death
four subjects in his wide-ranging search for religious on Jumada II, 14, 505=December 18, 1111. Only one
truth: theology ('ilm al-kaldm), philosophy, Isma'ilism work, 1ldm, can be placed with certainty in this
and Sufism. He does not say that he wrote anything period, a fact which confirms the opinion that the
on theology before writing on the other three; in fact period was quite short.
theology was a part of his previous education, and his
one book on it, the Iqtisdd, probably belongs to the AL-MANKHOL FT USOL AL-FIQH, in MS.
latter end of the Baghdad years. There is some over-
lap between the books on philosophy and on Isma'il- Mentioned in Mustawsf I, 3 as a concise work on
ism, but the order seems to be correct in a rough law. SubkT says GhazalT wrote it during the lifetime of
fashion. As for Sufism, GhazqlT studied it but had not his teacher Imam al-Haramayn, i.e., before 478=
yet started writing on it at this time, although M-zdn 1085/86.'1 We place it first because this early time of
al- 'amal as a work on ethics is in accord with composition is not specified for any other work. But it
GhazalT's concern for practice while he approached cannot be proved that no other work belongs to this
Sufism, as shown in Munqidh, 122. period.
(2) A period of retirement extends for eleven lunar Mankhal" means "sifted" and points to a summary.
years, from his departure from Baghdad in Dh5 al-
Qa'da, 488=November, 1095, to his return to teach- 1. MA'AKHIDH AL-KHILAF
ing at NTshapur in Dhi! al-Qa'da, 499=July, 1106
2. LUBAB AN-NAZAR
(Munqidh, 153). The order of his travels and resi-
dences during these years can be traced fairly well 3. TAHSTN AL-MA~'AKHIDH
from statements in Munqidh and by Subki, but the
4. AL-MABADTV WA AL-GHAYAT
lengths of the sub-periods cannot be established
accurately.
All are lost.
He started with nearly two years in Syria, between
Mi'yar al-'ilm, 23, which belongs to the end of
Damascus and Jerusalem, and probably wrote all or
the first period, 488=1095, lists in order these four
most of his greatest work, Ihyd' u/am ad-dTn, in
works on methods of legal debate. As Badawl sug-
these two cities.9 From there he proceeded on a
gests (p. 33), the order of the list is likely to be the
pilgrimage to Mecca and Madina, probably that of
chronological order of composition; at least 3. must
the end of 490=November, 1097. Then he went back
follow 1. But the group as a whole cannot be related
to Baghdad, then to Iran, ending with residence in a
to most of the other early works on law.
monastery in his home town, Ths. (Thus he had left
the Syrian area well before the arrival of the First
SHIFA' AL-GHALIL FT AL-QIYAS WA AT-TA cLIL, in MSS.
Crusade in 1099, which he never mentions in his
works.) If he spent two years in Syria out of the
(Various titles, including callI for ghalTl).
eleven years of this whole period, that leaves nine
years to be divided between the pilgrimage, Baghdad,
cities of Iran and TUs, but where he spent most of this 10 Tj ad-DTn as-Subki (d. 722= 1370), Tabaqdt ash-
time remains a mystery. A reasonable answer would Shdfi'iyya al-kubra ed. A. F. M. al-Halii and M. M. at-
be TUs, where he would have had a permanent resi- TanahT (Cairo, 1388=1968), VI, 225. Confirmed by Yafi'T,
dence, but there is no firm evidence known to me. Mir'dt al-jandn, fol. 257b, quoted by M. Smith, Al-GhazdlT's
life and personality (London, 1944), p. 16.
As in Mustaasf and SubkT; better than manhal, "emaci-
9 See below under ar-Risdla al-Qudsiyya and IhydO' forated," as in Macdonald, "Life," pp. 105-6, and Asin,
more details on these years. Espiritualidad, I, p. 29, note 1.

This content downloaded from

194.27.127.99 on Thu, 26 May 2022 13:21:49 UTC
All use subject to https://about.jstor.org/terms
 292 Journal of the American Oriental Society 104.2 (1984)

After MabadT' and close to it, see Mihakk, 91. Bouyges, pp. 13-14, gives indecisive reasons for
Before MustazhirT: fol. 79a=Goldziher, Streitschrift, placing it after al-BasTt and al-Wasft. But in fact it
p. 29 and Arabic, p. 52. cannot be related definitely to any of the other early
fiqh books.
Three other works can be related to each other but
may be before, after or interspersed with the preceding TAH-DHIB AL-USOL, lost.
group. None of them is mentioned before Ihy '.
In Mustasfd I, 3 and 111, Ghazali refers to this
(I) AL-BASTT, in MSS. book as his own, and as a work onfiqh more detailed
than Mustasfd. Bouyges knew the first reference but
Ihyd', I, 108, ed. Cairo, 1326=1908/09, an indirect rejected it, p. 70, because he could not believe that
reference taken from Bouyges. (Most references to Ghazal! wrote a book on fiqh longer than Mustasfd.
Ihyd' below will be to the 'Irdql edition of Cairo, Badawli pours scorn on Bouyges' reasoning, pp. 210-
1356/ 57.) 11. Anyhow, the matter is settled by the second
Also mentioned in Jawdhir al-Qur'an, 22, as a reference (I, 111), where GhazalT writes "We have
work of fiqh of his earlier life. discussed at length in TahdhTb al-uszl ..."
Regarded as a summary of Juwayni's Nihayat al- There is no way to date this last work with any
matlab (Bouyges, p. 12), but itself of considerable accuracy, but it is more likely to be a product of the
length. first period, with most of the other legal works.

(II) AL-WASTT, in MSS. There follow five books on philosophy and Isma'il-
ism which can be related to each other serially with
Ihyd', I, 108 (ed. Cairo, 1326). Also mentioned in some accuracy and dated to the later part of the four
Jawdhir al-Qur'dn, 22, as a work offiqh of his earlier years in Baghdad. There are no references which
life. might relate this group to the previous one, but books
A summary of al-Basqt, as GhazalT mentions in the on philosophy and Isma'ilism are generally to be
preface to al-Waslt (see Bouyges, p. 13). placed after those on law because Ghazali gives this as
the order of his studies in Munqidh, 79ff. Some
(III) AL-WAJITZ FT FIQH AL-IMAM ASH-SHAFIcT. (Cairo, overlap is not precluded.
1317=1899/ 1900).

I. MAQASID AL-FALASIFA, ed. M. S. Kurdl (Cairo,

Ihyd', I, 108 (ed. Cairo, 1326) and Jawahir, 22; in
1355= 1936), 3 parts.
both cases third after al-Baslt and al-Waslt. A sum-
mary of Shafi'ite law, perhaps made as a text-book
As it was written in Baghdad (Munqidh, 85), it
for students.
could not have been started earlier than 484=1091/92.
Bouyges, pp. 12 and 49, dated it much later,
Maqdsid, i, 2-3 and iii, 77, as well as Munqidh, 84-
between Ihyd' and Jawahir, on the ground of a note
on a Cairo MS, later than 656=1258, which gives the
85, make it plain that the book was written as a

date as 495=1101. I do not think this is sufficient

background to Tahdfut, which means it was com-
reason to ignore the above references by GhazAlT to
pleted hardly later than 486 (ended Jan. 20, 1094), in
view of the evidence on the dates of Tahdfut and
al- WajTz in close association with al-Basft and al-
MustazhirT. It must have been written during the "less
Waslt. Bouyges' dating is also rejected implicitly by
than two years" when GhazalT was studying philos-
Badawi, p. 25, and by Lazarus, pp. 210, 378-79, 407,
on grounds of style and content.
ophy in his spare time with the primary aim of
understanding it (Munqidh, 85).

KHULASAT AL-MUKHTASAR WA NAQAWAT AL-MUcTASAR,

in MS. II. TAHAFUT AL-FALASIFA, ed. M. Bouyges, Biblio-
theca Arabica Scholasticorum, II (Beirut, 1927).
Mentioned in Ihya', I, 30 (cIraqi ed., see 10). The
Mukhtasar was by Ismac'l al-MuzanT, an early After Maqdsid (Maqdsid, i, 2-3 and iii, 77; Munqidh,
Shaficite lawyer (d. 877). This is a resume of it, 84-85). MS Fatih, 2921 (Istanbul) records that Tahifut
perhaps made as a textbook. was completed on Muharram 11, 488=January 21,

This content downloaded from

194.27.127.99 on Thu, 26 May 2022 13:21:49 UTC
All use subject to https://about.jstor.org/terms
 HOURANI: A Revised Chronology of GhazdlT's Writings 293

1095;12 so most of it was probably written in 487= worked on Isma'ilism (al-Btiniyya) after philosophy.
1094. In Munqidh, 85 GhazAl! says he spent "nearly a And Goldziher saw an allusion to Tahafut in Mus-
year" in critical reflection on philosophy, after the less taghirr, fol. 196, where GhazalT mentions a philo-
than two years spent in understanding it. sophic doctrine which he had refuted ft al-kalam.'3
There are numerous later references to Tahafut. (In Jawdhir al-Qur'dn, 21 he refers to his Tahdfut as a
The most pertinent to chronology are Mi'ydr, 22, 161, work of kalam.)
174, all describing Micydr as a supplement on logic to The temporal relation of the MustazhirT to Micydr
Tahdfut. (See below, Micydr, for more on their and Mihakk is less clear. We have shown that all
relation.) three overlap with Tahdfut. They may also overlap
with each other. But, on the basis of Munqidh's
III. MIC YAR AL-cILM FT FANN AL-MANTIQ, ed. M. S. account of Ghazal!'s work, Mustazhirl can be placed
Kurdi (Cairo, 1329=1911). after the other two, although not conclusively. This is
the order given by Bouyges, Badawt and Lazarus.
Tahifut, 17 and 20, anticipates it as an appendix, Munqidh, 119, lists MustazhirT first among five
under the title Micydr albcaql. Tahdfut, 213, refers to books against the Isma'T1Ts, and since he numbers the
it as Maddrik al-cuqal, and implies that it has already five in order and gives the places of the other four, the
been written (sannafndhu). Micydr, 22, 161, 174, justi- order of his list must be accepted as chronological.
fies itself partly on the ground that it explains the
technical terms in Tahdfut. Thus the relation of the HUJJAT AL-HAQQ, lost.
two works is close. There is nothing unusual about an
author writing a new book before the last one has Mentioned in Jawahir al-Qur'dn, 21. Listed second
been revised or published. in Munqidh, 119, as a reply to criticisms by the
Mentioned in several later books, e.g., Iqtisdd, 9. Ism'TlITs made against him in Baghdad. This does not
prove that he wrote it in Baghdad, but it suggests that
IV. MIHAKK AN-NAZAR FT AL-MANTIQ, ed. M. HalabT he was there, or had been there recently. Macdonald
and M. QabbanT (Cairo, n.d., Adabiyya Press). thought "perhaps during his second residence there,"'4
but there is no way to decide.
P. 131 mentions Micydr as still unpublished, await-
ing corrections; it is made clear that Micydr was AL-IQTISAD FT AL-IcTIQAD, ed. I. A. 1ubukqt and
H. Atay (Ankara, 1962).
substantially written first but published later. The two
books are mentioned together in Iqtisdd, 9 and else-
where. Mentions Tahafit (105, 215), Micyar and Mihakk
(15) and MustazhirT (239); so cannot be earlier than
V. AL-MUSTAZHIRT, or FADA'IH AL-BATINIYYA WA 488=1095. There is probably a forward reference to it
FADAVIL AL-MUSTAZHIRIYYA, selections, ed. I. Gold- in Tahifut, 78, though the title given there is Qawacid
ziher, Streitschrift des GazdlT gegen die Bdtinijja- al-caqd'id. He says there that after finishing Tahdfut
Sekte (Leiden, 1916). (Complete text, ed. A. BadawT, he hopes to write a constructive work on doctrine, as
Fadd'ih al-Batiniyya, Cairo, 1964.) the present one is critical. Such an intention seems
fulfilled more specifically in Iqtisdd than in the actual
The MustazhirT can be closely dated by its refer- Qawdcid al-caqd'id, which is later and is but a part of
ences to two caliphs. It refers to the cAbbasid Mustaz- Iha' Is
hir as holding his office (fols. 3b-4a), and his
accession was on Muharram 15, 487 (February 4,
1094); and to the Fatimid Mustansir as still alive 13 Streitschrift, p. 28.
(fol. 18a), and he died on Dhu al-Hijja 17, 487= 14 "Life," p. 88.
December 29, 1094. Thus the book must have been at 15 The text of Tahafut, 78, should not, however, be
least begun before Tahdfut was completed (January, emended, as is done by S. Van den Bergh in his translation
1095, see above). But it was probably completed after of Ibn Rushd's Tahdfut at-tahdfut (London, 1954), I, 68=
Tahdfut. For Munqidh, 79 and 109, shows that he Bouyges, TT, 116. There is no textual authority for any
reading but Qawd'id al-'aqd'id. Ghazdli may well have
changed his mind about the title of a book (cf. Mi'ydr,
12Bouyges, Introduction to Tahdfut al-faldsifa, pp. ix, xiii. above).

This content downloaded from

194.27.127.99 on Thu, 26 May 2022 13:21:49 UTC
All use subject to https://about.jstor.org/terms
294 Journal of the American Oriental Society 104.2 (1984)

Mentioned in Ihyd', I, 68 and 169; therefore earlier in Baghdad, 488= 1095. This dating is confirmed by
than the earliest part of Ihyd '. most recent scholars (Bouyges, Badawl, Sherif,
These data fix the place of Iqtisad in the sequence Lazarus) and by Goldziher among the older ones, on
of writings almost precisely (except for its relation to grounds of content: that it reflects an attitude "still
MTzan al-'amal, to be discussed below). But we still before his complete conversion to Sufism, yet already
need to know whether it was one of the latest works standing at its gate," as Goldziher put it.'6 Such a date
of Baghdad or one of the earliest of the period of also accords perfectly with the order of studies out-
travel. We know from the passage in Tahafut that lined in Munqidh, 78ff., because a book on ethics
GhazalT intended to write such a book "after finish- would be a natural product of the time when GhazdIT
ing this book," as a completion of a trilogy whose first was involved in the study of Sufism, a practical
two parts were Maqdsid and Tahdfut. But did he discipline.
actually write it or complete it in Baghdad? That A date after Ghazalt's departure from Baghdad was
seems more likely than the alternative, for it is hard to favored by me in my former article, but I am now
believe that this prosaic piece of kallm was one of the convinced of the earlier date, because of the closeness
first products of his new life as a SufT. Bouyges, p. 34, to Micydr as shown and the opinion of recent scholars
placed it in Baghdad because he found in it the same as explained above.
concerns as those of Tahafut, still very much alive. Since Iqtisdd and MTzdn are both to be placed in
(Badawt is silent on this question, pp. 87-88. Lazarus, Ghazalt's final year at the Nizamiyya College in Bagh-
p. 440, places it in Baghdad without explanation.) dad (488=1095), we have to ask which is earlier.
The best available time would be the first half of There are no indications from references. But some
488=1095, before his nervous crisis became acute in simple considerations seem to indicate the priority of
Rajab (July). But even this time seems crowded, Iqtisdd: that it is a sequel to Tahdfut, while MTzdn is a
between the completion of Micydr, Mihakk and sequel to Micydr, which is itself an appendix to
MustazhirT and his new attention to Sufism (Munqidh, Tahdfat. In other words, Ghazali would more natu-
122ff.); and Iqtisid itself is neither short nor easy. It is rally have completed his writing on metaphysics and
possible, then, that it was completed during the second theology before turning to ethics, just as he says he
half of 488=1095, in spite of the crisis, for GhazalT did in Munqidh, 122: "When I had finished with these
does not say he was inhibited from writing, only from sciences [theology, philosophy and Isma'ilism], I
lecturing (Munqidh, 128). As we shall see, MTzdn al- turned my attention to the way of the SiifTs, and I
camal was probably written in these six months. We learned that their way is completed only by both
may wonder whether the prolific and intense writing knowledge and practice."
of 477-88= 1094-95 was a contributing cause of the Now that both Iqtisdd and MTzmn have been placed
strain which led to a breakdown of health. with some confidence in the period when GhazaIt was
approaching or actually immersed in the intense spiri-
MTZAN AL-cAMAL, ed. M. S. KurdT and M. S. NucaymT
tual crisis of his life, the importance of these two
(Cairo, 1328=1909/10). works for understanding the evolution of his thought
will readily be understood. Both of them therefore
M-zdn is anticipated at the end of Micydr (195) as a deserve more serious studies than they have hitherto
companion work. Both knowledge (cilm) and practice received, and they should be read in the context of the
(camal) are required for happiness in this world and author's revealing account of his state of mind at the
the next, and as Micydr gives the criteria for sound time, narrated in Munqidh, 122-30.
knowledge, so another book is to be written which I have postponed until the end consideration of the
will give the criteria for sound action. MTizdn refers to authenticity of M-zdn, brought into question in a
Micydr six times, (3, 28, 56, 64, 153, 156). Therefore it notable article by Montgomery Watt in 1952, because
is certainly later than Micydr.
But how much later? Unfortunately there is not a
single reference to MTzdn in any subsequent work of 16 I. Goldziher, Die Richtungen der Islamischen Koranaus-
Ghazalt, so we cannot fix a late limit by the usual legung (Leiden, 1920, 1952), p. 205. Bouyges, pp. 28-30;
kind of evidence. But the fact that MTzdn refers to Badawi, pp. 79-81; M. A. Sherif, GhazdlM's theory of virtue
MTcydr so frequently and to no other of his books (Albany, 1975), pp. 6, 171; Lazarus, pp. 211, 259, 300;
proves that MTcydr was very much in his mind when M. Abul Quasem, The Ethics of al-GhazdlT (Delmar, N.Y.,
he wrote Mrzdn, and points to a date in his last year 1978), p. 37, n. 25.

This content downloaded from

194.27.127.99 on Thu, 26 May 2022 13:21:49 UTC
All use subject to https://about.jstor.org/terms
 HOURANI: A Revised Chronology of GhazdlT's Writings 295

Watt's arguments were to a large extent bound up Again, Bouyges mentions a reference to MTzdn by
with the question of date. Watt accepted the view of Abil Bakr at-Turtiishl (1059-1126), a contemporary
several of the older scholars (Gairdner, Asin Palacios, of GhazalT.
Massignon, Hachem) that, if MTzan is genuine, it Finally, Watt's conclusion of 1952 has been rejected
must belong to a late period-after Munqidh accord- by four recent scholars aware of it: Allard in an
ing to Hachem. He then proposed that it could not editorial note on Bouyges, Badawi by implicit accep-
belong to that period, because of certain assertions in tance of MTzdn, Lazarus by acceptance and a generally
MTzan about the primacy of reason and the non- critical view of Watt's interpretations, and Sherif in a
resurrection of the body which are incompatible with detailed appendix refuting Watt directly. Sherif points
GhazalT's later views as known from other late works. out that Watt himself in a later article assumes MTzdn
However, the reasons given by those scholars for a as genuine except for certain parts.'
late date are slim. Gairdner and Massignon gave no
reasons. Asin thought Mlzdn must be later than Ihy AR-RISALA AL-QUDSIYYA, ed. and Eng. tr. A. L.
because Ihyd' has no references to it-a dubious Tibawi, Al-GhaziilT's tract on dogmatic theology
argument from silence. Hachem placed it after Mun- (London, 1965).
qidh because it denies the resurrection of the body:
precisely one of the reasons for which Watt declared it Mentioned in Ihyd', I, 169 and 180, as a short
could not be after Munqidh, and for which he dis- treatise (risdla) written for the people of Jerusalem and
missed it as mostly spurious!' But Watt's arguments, subsequently incorporated into Qawd'id al- 'aqj 'id,
based on the assumption of a late date as the only the second of the forty "books" of Ihyd . The
possible one, are undercut by the solid reasons given colophons of two manuscripts, Cairo, Majami' 66
above for dating MTzin much earlier, close to MTcydr. and London, SOAS 45818, state that the work was
Watt's other reason for rejecting MTzan is that its completed in the Aqsa Mosque of Jerusalem (Tibawi,
organization is very confused and self-contradictory, pp. 10-11 and Bouyges, p. 35). Thus it is easy to fix
unlike most of GhazalT's books. Watt concluded that, the locale of the treatise, in Jerusalem, and its place in
while some pages may have been taken from GhazaIT, the serial order, after Iqtisdd and MTzin and before
as a whole "the MTzin is an unintelligent compilation Ihyd'. It can also be stated with confidence that it was
from very varied sources." But may not the confusion the first surviving work completed during the period
be seen rather as confirming the dating of the book to of retirement, which began in DhO al-Qa'da, 488=
the period of GhazaIT's nervous disorder? The dis- November, 1095 (Munqidh, 153).
organization will in that case be of the deepest interest, The date of GhazalT's stay in Jerusalem is uncertain,
when it is re-examined in light of the biographical since there is a discrepancy between his own account
context of 488=1095. of his travels in Munqidh, 130-31, and that of
Further, if MTzan is a forgery, the cross references Subkl.'9 According to GhazalT, he spent "nearly two
cited above will be difficult to explain, although not years" in Syria (ash-Shdm), starting in Damascus and
quite impossible. The forward reference in Micydr proceeding to Jerusalem, and staying a considerable
would state an intention of GhazaIi which he never time in each city. Thus he would be in Jerusalem
carried out, or which he completed in a book that was during 490=1097. Subki gives him only a few days at
soon lost. The six Mlzdn references back to Micydr, a the beginning in Damascus, then a residence in Jeru-
work of lesser renown, would be the only ones salem followed by a second residence in Damascus.
inserted by the forger, for no apparent purpose,
whereas other certainly spurious psuedo-Ghazalian
books usually try to impress readers by referring to 18 Bouyges, p. 30, on TurtUIshI, referring to a text quoted by
IhyW', the most famous of GhazalT's works. M. Schreiner, Zeitschrift der Deutschen Morgenlandischen
GeselIschaft, 52 (1898), p. 503. M. Allard in Bouyges, p. 30;
BadawI, pp. 79-81; Lazarus, pp. 211, 259, 281, 300, etc.;
1 Watt, JRAS (1952), W. H. T. Gairdner, "Al-GhazAll's
Sherif, GhazdliTs theory of virtue, pp. 170-75. Watt, "al-
Mishkdt al-anwar and the Ghaza1I-problem," Der Islam, 5 GhazalT," Encyclopaedia of Islam, 2nd ed.
(1914), p. 131; M. Asin Palacios, Los precedentes musul- 19 Tabaqdt, VI, 197. F. Jabre, "La biographic et l'oeuvre de
manos del Pari de Pascal (Santander, 1920), p. 15; Massignon, GhazAll,'' MIDEO, I, 94-97, tabulates the statements of the
Recueil, p. 93; H. Hachem, French tr. of Mzlzan, Critre de sources and offers a tentative reconstruction of the chronol-
laction (Paris, 1945), Introduction. ogy.

This content downloaded from

194.27.127.99 on Thu, 26 May 2022 13:21:49 UTC
All use subject to https://about.jstor.org/terms
296 Journal of the American Oriental Society 104.2 (1984)

By this timetable he would be in Jerusalem in 489= in Jerusalem, according to his own chronology; or not
1096. The difference is of small significance, as it does before 489=1096, according to SubkT (see above on
not affect the order or time of other books, except the the Risdla).
Ihyd' in a limited way, as will be explained. Mentioned in most later works; among the earlier
of these, it is referred to in Bidayd, 14, 28, 33, 34, 39;
AR-RADD AL-JAMTL cALA ILAHIYYAT cTSA BI-SARTH 34 refers to sins of the heart which are dealt with in
AL-INJTL, ed. R. Chidiac in Bibliotheque de l'cole the
des third quarter of Ihyd'. Maqsad, 63 refers to Ihyd',
hautes etudes: Sciences religieuses, 54 (Paris, 1939). Book 37, i.e., near the end of the fourth quarter.
The date of completion cannot be determined accu-
I formerly accepted this work as genuine on the rately from GhazalT's references alone. All we know
basis of arguments by Chidiac in the introduction to by this method is that at least five works were written
his edition. Since then, however, Lazarus has proved between the completion of Ihya' and the return to
decisively that it is inauthentic (pp. 458-77). Her Nishapir in Dhii al-Qa'da, 499=July, 1106. (See below
reasons may be summed up in a few sentences. 1. The on Biddya, Madnin, Maqsad, Jawdhir al-Qur'dn and
book is not mentioned in any list of Ghazalt's works K-mIyd). Thus the completion of Ihyd' could hardly
before the beginning of the twentieth century. It is have been later than the end of 498=August/ Septem-
only mentioned, with two quotations, in a book by a ber, 1 105.21
Coptic theologian in the thirteenth century. 2. There But external evidence takes the probable date of
are no cross-references by GhazaIt to or from the completion back to the two years in Syria. Subki
22
Radd, contrary to his usual habit. 3. The style and reports that Ghazali read Ihyd ' publicly on his
vocabulary of the book are very different from those return to Baghdad, which followed his pilgrimage to
of his authentic books. 4. The author shows detailed Mecca and Madina in DhM al-Hijia, the last month of
knowledge of Christian doctrines and sects, quotes 490=November/ December, 1097. SubkT's assertion is
from the Bible giving chapter and verse, and even confirmed by a younger contemporary of GhazaIT,
quotes two sentences in Hebrew and one in Coptic Abil Bakr Ibn al-'Arabi (468-543=1076-1148), an
(transcribed into Arabic)! All this goes far beyond the Andalusian scholar who writes that he heard GhazalT
knowledge displayed by GhazaIt in all his authentic read the Ihyd' in Baghdad.23 Now the reading in
writings. 5. The author shows little knowledge of Baghdad still does not provide a firm date for the
Islamic Traditions or theology. Lazarus concludes completion of Ihyd', for we do not know how long
that Ghazali could not have composed the Radd, and
that the author "may have been a converted Muslim,
a former Christian, probably a Copt" (p. 475).
If Ghazali did not compose the Radd, there is little 2 A statement in Ihya', VII, 157 that "nearly 500 years
solid evidence that he visited Egypt, and I accept the have now passed" since the hijra gives a date before 500=
20
opinion of Jabre that he never did so. 1106/07 for the completion of the second quarter of Ihyd'.
But this is of little help, because the date is vague and in any
IHYA' cULOM AD-DIN, c'Iraq ed., 16 vols. (Cairo, case gives no clue to the completion of the third and fourth
1356-57= 1937-39). quarters. Moreover, external evidence, to be mentioned
below, indicates a date of 490=1097 or soon after.
I, 169 and 180, mentions ar-Risala al-Qudsiyya; I, 22 Tabaqdt, VI, 200.
68 and 169, mentions Iqtisad. 23 Al-'Awdsim min al-qawdsim, Cairo: Dar al-kutub, MS
Since ar-Risala al-Qudsiyya was inserted into Book 2 22031 B, fol. 7; MS dated 536=1141/42. See Jabre, pp. 87-88.
of Ihyd', it is most probable that the whole of Ihyd' Ibn al-'ArabT asserts that GhazAlT entered the Suff path in
was composed after the Risala, or at the outside that 486=1093/94, which is two years too early; and he gives the
Book 1 alone was prior. This would place the begin- month of his meeting with Ghazall in Baghdad as JumadA
ning of Ihy ' not before 490=1097, when GhazaIT was II, 490=June, 1097, which does not allow GhazalT nearly two
years in Syria plus the pilgrimage at the end of 490=1097.
But these inaccuracies in dating, no doubt dependent on
20 Jabre, MIDEO, I, 97. Subki, Tabaqdt, VI, 199, reports
memory, do not invalidate Ibn al-'ArabTis claim that he
an Egyptian journey, but as a matter of doubtful hearsay (wa attended GhazalI's Baghdad reading.
yuhka canhu hikdydt), see Tibawi, p. 4. No other source Ibn al-'ArabT had performed the pilgrimage in 489=1096,
mentions Egypt. a year before GhazAlT; Ibn Khallikdn, I, 619.

This content downloaded from

194.27.127.99 on Thu, 26 May 2022 13:21:49 UTC
All use subject to https://about.jstor.org/terms
 HOURANI: A Revised Chronology of GhazdlT's Writings 297

GhazalT spent in Hijaz, whether he returned directly learned about chronology from such evidence, and we
from there to Baghdad,24 how long he stayed in can only say with confidence that the work is later
Baghdad before the reading, and even whether the than Ihyd'.
entire book had been completed at the time of the
reading. But Ibn al-Athir (555-630=1160-1233) re- AL-IMLA' FT ISHKALAT AL-IHYA', printed at the end
ports that GhazaIT read it in Damascus, i.e., at some of Mhy ' in the 'IrdqT edition, XVI.
time before the fall of 1097 when he would have to
start on his pilgrimage. It also seems more likely that 2, mentions criticisms of fIyd' by ignorant readers
he wrote and completed this book in 489-90=1096-97 and a ban on it somewhere. The criticisms imply a
in the tranquillity of the Umayyad Mosque in Damas- certain interval after the publication of MhyA'. If the
cus and the Dome of the Rock in Jerusalem (Mun- ban refers to the burning of GhazaIT's books in the
qidh, 130-31), rather than after the pilgrimage when Maghrib, which took place on the orders of the
he began to be distracted by concern for his family Almoravid amfr al-muslimrn 'AlT b. Yiisuf b. Tashfin,25
and other matters not specified (Munqidh, 131). that would date Imld' later than 500=1106/07, when
'AlT's reign began. But we cannot be sure that GhazalT
I place next a group of works which cannot be is referring to this event.
related to any other book but only to Ihyd' as preced-
ing them. Thus the order in which these works are Three works that are probably spurious are placed
listed is not significant of temporal relations between here because they all refer to Ihyd'. If any of them is
them or with other late works. genuine, this is the only indication of date.

AR-RISALA AL-WAcZIYYA, in the collection Al-Jawdhir [AL-HIKMA FT MAKHLOQAT ALLAH, ed. M. QabbanT
al-ghawalr min rasd'il al-imam hujjat al-isldm al- (Cairo, 1321 = 1903/ 04).
GhazdlT (Cairo, 1353=1934), pp. 151-55. Also called
RISALAT AL-WAcZ WA AL-IcTIQAD. See BadawT, Nos. According to Asin, Ifyd' promises this work,26 but
49, 145. I have been unable to trace his reference and it is not
mentioned by Bouyges or Badawl. There is no other
155 refers to Qawd'cd al-caqd'id, i.e., Ifyd', Book indication of its date.
2. Bouyges accepts it as probably genuine, merely
Its authenticity is not questioned by Bouyges, p. 64, because "le Hikma serait digne d'Algazel" (p. 89).
or BadawT, pp. 190, 313. Lazarus says it "abounds Badawl places it among works of doubtful authenti-
in expressions and images typical of Al-Ghazallt" city (pp. 257-59). Lazarus barely refers to it (pp. 31,
(p. 138). 452) and raises no questions. On the basis only of its
title and table of contents I am inclined to doubt its
AYYUHUA AL-WALAD, ed. T. Sabbagh (Beirut, 1959), authenticity. "The wisdom of God's creation" sounds
with introduction by G. H. Scherer. more like the philosophic view of God's providence.
The Qur'an emphasizes rather the benevolence of
29 and 59, mentions Jhyd'. God in making the world fit for the use and enjoy-
A prologue usually attached to this opuscule, but ment of man (see especially siira xvi). The contents as
not written by Ghazall, makes out (5) that it is given by BadawY show a treatment of the wisdom in
addressed to a former student, who had spent most of the parts of nature in turn, from the heavens to the
his life studying various sciences but would now like
to receive some advice to prepare him for life beyond
the tomb. This suggestion of an older man does not 25 cAbd al-Wdhid al-MarrdkushT, al-Mucjib fTi talkhTs
ride well with the title referring to a child, which is akhbar al-Maghrib, ed. M. S. al-cAraydn and M. A. al-
repeated in the text (9, 11, etc.) Nothing can be cAlami (Cairo, 1368=1949), p. 173; ed. M. Dozy, 2nd
(Leiden, 1881), p. 123. Bouyges, pp. 75-78, gives more
references to the book burning and discusses its date further,
24 SubkT inserts a possible visit to Egypt, but this is doubt-
but it is unnecessary to pursue this question since we cannot
ful; see note 20 above. Ibn al-AthTr, al-Kamil ft- at-ta'rTkh definitely connect GhazalI's Imla' with it.
(Cairo, 1883/84), Part 10, p. 87, sub anno 488, takes him 26 Espiritualidad, IV, 80, referring to IhkvJ, iv, 90, perhaps
directly from the pilgrimage to Baghdad. in the Cairo edition of 1316=1898/99 in 4 vols.

This content downloaded from

194.27.127.99 on Thu, 26 May 2022 13:21:49 UTC
All use subject to https://about.jstor.org/terms
298 Journal of the American Oriental Society 104.2 (1984)

species of animals and plants. This would require an fourth quarter, on "Things leading to salvation."
interest in and knowledge of natural science which Mentioned in Arba'Tn, 29.
GhazaIT does not generally display. The organization
is too systematic.] QAWASIM AL-BATINIYYA, ed. A. Ates in Ilahiyat
Fakultesi Dergisi, III (1954), pp. 23-54.
[MARAQT AZ-ZULFX, lost.
A passage quoted by Ates places it after Mus-
Considered doubtful by Bouyges, p. 159, and tazhirT.28 Mentioned in Jawahir al-Qur'n, 21; and
Badawl, No. 302. But they give references to HajjT Qistas, 66, as concerning the Ismac'lTs. Not mentioned
Khallfa and other Arabic bibliographers. A possible in Munqidh, 119, among the books against the
reference to it in Imla', printed in the margin of Ismac'lTs which are listed there.
Murtada az-Zab-IdT, Ithaf as-sada (Cairo, 1311= Badawi, p. 86, places it immediately after Mus-
1893/94), IV, 397.] tazhirT and fHujjat al-Haqq, i.e., before GhazalT left
Baghdad. Thus the three polemical works against the
[AL-MADNON BIHI 'ALA GHAYR AHLIHI (Cairo, 1309= Isma'TlTs would belong to the same period. This can-
1891/92). not be disproved, because Qawdsim contains no
reference to Ihyd'. But it makes even heavier the load
30, mentions MIAyd', as the only other book of his to of writing done in the final year at the Nizamiyya
date containing these truths. If Madnan is authentic, College.
this would give a time after Ihyd'. But if it is not
authentic, it only shows that forgers liked to refer to JAWAB MUFASSIL AL-KHILAF, lost.

IWy ', a fact which is confirmed by several other

cases. Mentioned with Qawdsim in Jawdhir al-Qur'dn, 21
Arba'ln, 25, mentions "some of our books which and Qistas, 66 as refuting the Isma'TlTs; also referred
are restricted for those who are unqualified to use to in Qistds, 84. Munqidh, 119 describes it as the third
them" (ba'di kutubind 1-madnani bihd 'ald ghavri book against the Isma'clfs, after MustazhirT and
ahlihd). But this is not a title, for biha refers to kutub, Hujjat al-haqq and before Kitab ad-Dadj; since Dadj
plural. is not mentioned in Jawahir al-Qur'dn, among these
Without going into the long history of doubts polemical works, Bouyges, p. 56, rightly infers that it
about the authenticity of this Madnan, I accept the had not yet been written. Thus Jawdb is to be placed
conclusion of Lazarus, pp. 251-53, that it is spurious, between Hujjat al-haqq and Jawahir al-Qur'dn.
as shown by a philosophical vocabulary not found in Munqidh, 119, calls it "a reply to criticisms made
the authentic works, except where Ghazali is ex- against me in Hamadan."
pounding the theories of philosophers, as in Maqdsid A residence in Hamadan is likely to have been after
and Tahdfut]. his second residence in Baghdad. This would place it
later than IWyd', as Bouyges points out, p. 45. But it is
There follows a group of works which contain uncertain whether he passed by Hamadan on his way
references bracketing them between Ihyd 'and Jawdhir home to Khurasan, as Bouyges supposes, p. 4, or
al-Qur'an or its sequel Arba'Tn. They cannot be returned to Hamadan some time later with a Seljuq
related to each other. army which captured it from Barkiyaruq around
493=1099/1100, as Jabre conjectures.29
AL-MAQSAD AL-ASNA FT SHARH MAcANT ASMAD ALLAH
AL-HUSNA, ed. F. A. Shehadi (Beirut, 1971).

115, 127, 161, mention of MyJ D; 127, refers to

Book 36. Mentioned in ArbacTn, 13, 25. Bouyges,
p. 46, note 4, also gives a reference from Jawahir al-
Qur'dn (Cairo, 1329=1911), 57. Mentioned in Mish- 27 Misprinted as Mawdhim in an older edition (Cairo,
kat, 122. 1324=1906/07), IV, 116. See now Tabaqit, VI, 226, note 2.
28 See G. Makdisi, Ibn 'AqTl et la resurgence de l'Islam
BIDAYAT AL-HIDAYA (Cairo, 1353=1934). traditionaliste au XIe sicle (Damascus, 1963), pp. xiv, 212,
n. 2, 290, n. 2.
Mentions Ih V d' in several places; 34 mentions the 29 "Biographie," pp. 98-99.

This content downloaded from

194.27.127.99 on Thu, 26 May 2022 13:21:49 UTC
All use subject to https://about.jstor.org/terms
 HOURANI: A Revised Chronology of GhazdlT's Writings 299

JAWAHIR AL-QUR'AN (Cairo, 1352=1933, Rahma- order. The whole context is an account of a historical
niyya Press). development, from the Kharijites to Ghazal!, which is
given in a true temporal sequence, allowing for the
24, mentions the 40 books of Ihy '. overlap of the SufTs with the other schools, especially
After Biddya and Maqsad, because they are men- the Ash'arites. Then, when he comes to Ghazali, he
tioned in Arba'Tn, the sequel to Jawdhir (Jawdhir, 6). mentions the first three works in their correct chrono-
(Maqsad is also mentioned in Jawdhir (Cairo, 1329= logical order, just as we have shown it. Moreover the
1911), 57, according to Bouyges, p. 46, note 4). chronological order is made explicit by Ibn Rushd,
References to Jawdhir in Qistds, 70, 92, 93. who begins the sentences referring to Tahafut, Jawdhir
Described in Musta.sfd, I, 3 as before the return to and Mishkat with the word thumma, "then." It might
teaching at NTshapUr, DhM al-Qa'da, 499=July, 1106. be suggested that Ibn Rushd is using thumma as
Farabi does in his Philosophy of Plato, to connect a
AL-ARBA'-TN FT USCL AD-DrN (Cairo, 1344=1925, series of philosophical topics in a logical order which
Istiqama Press). is unrelated to the chronology of Plato's dialogues.
But, whereas the logical order is obvious in FardbT, in
29, mentions Biddya. 13, 25, mentions Maqsad. Ibn Rushd there is no clue to one and the only order
305, describes itself as a sequel to Jawdhir al- which is obvious is a chronological one. As for Ibn
Qur'dn. Rushd's knowledge of this, it seems to have been
Before Qistas, because that work mentions the com- rather accurate, since little fault can be found in it.
panion volume of Arba'cn, Jawahir al-Qur'dn (see BadawT's only textual argument that the order is
above). not chronological is the fact that after this sequence of
Ghazali's books from Maqdsid to K-mlyd Ibn Rushd
MISHKAT AL-ANWAR, ed. A. A. `AfTfT (Cairo, 1964). mentions Faysal at-tafriqa, whereas Faysal is known
to be earlier than Munqidh from a reference in
46, mentions Ihyd', Book 21. 47, mentions Micyar Munqidh, 99. But Ibn Rushd mentions Faysal more
and Mihakk. 56, mentions al-Maqsad al-asnd. Thus it than half a page later than the other books, in a new
is certainly after Maqsad. context in which he is no longer explaining what
But we can probably date it later, if we can give a Ghazal! did but the harmful consequences that fol-
chronological interpretation to a passage in Ibn Rushd's lowed from it, and here chronology is of no concern.
Kitdb al-Kashf 'an mandhij al-adilla where he refers Thus the reference to Faysal in no way invalidates the
to several of Ghazali's works in close succession.30 Ibn chronological character of the preceding series of
Rushd is describing how philosophy was gradually books listed.
made known to the Muslim public, first by the The evidence of Ibn Rushd, then, is that Mishkit is
Kharijites, then by the Mu'tazilites, Ash'arites and later than Jawahir al-Qur'dn. Such evidence cannot
Sf-is. Then GhazalT let the whole cat out of the bag weigh as much as that of the direct cross-references in
in his Maqdsid, Tahafut, Jawdhir al-Qur'dn, Madnan, GhazalT's own books, which we have been using for
Mishkdt, Munqidh and KTm-ya, mentioned in this the most part. But we have seen that Ibn Rushd's
order. If it was intended as a chronological order and knowledge was quite good, as should be expected
if Ibn Rushd was well informed on this point, Mishkat from such a careful scholar writing less than 70 years
can be placed after Jawdhir al-Qur'dn. (Madnan can after GhazalT's death.
be ignored if it is spurious, even though Ibn Tufayl There is no way to attain further accuracy on the
and Ibn Rushd thought it genuine.) This conclusion date of Mishkat between Jawdhir al-Qur'dn and
was accepted by Bouyges, p. 66. GhazalT's death, using the kinds of evidence to which
But Badawl, pp. 147-48, in commenting on Jawdhir this article is limited. Therefore several attempts to
al-Qur'dn, denied it, saying that "the text shows no base a late date on grounds of content, such as an
indication whatever that Ibn Rushd in this passage is advanced mystical doctrine in Mishkdt, need not be
enumerating GhazAlT's books in their chronological discussed in this place.3' We might just as well have
order." I believe that BadawT is mistaken about this,
and that there are clear indications of a chronological
3 See W. H. T. Gairdner, "Al-GhazalT's Mishkat al-anwdr
and the GhazAlT-problem," Der Islam, 5 (1914), p. 121;
Macdonald, "al-Ghazzd1l," Shorter Encyclopaedia of Islam;
30 Ed. M. J. Muller (Munich, 1859), pp. 70-71. Bouyges, p. 65; Watt, "Authenticity," JRAS (1952), p. 44;

This content downloaded from

194.27.127.99 on Thu, 26 May 2022 13:21:49 UTC
All use subject to https://about.jstor.org/terms
300 Journal of the American Oriental Society 104.2 (1984)

inserted Mishkat at any later point, but have pre- Mentioned in Munqidh, 99. Lazarus, p. 37 perceives a
ferred to do so at the earliest possible point, after close linguistic affinity between Faysal and Munqidh.
Arba 'n which is the close sequel to Jawdhir al- Another late reference in Mustasfd (Cairo, 1322=
Qur'dn as has been seen. 1904/05), I, 185.

KITAB AD-DARJ (or AD-DURJ) AL-MARQOM BI AL- KTMTYA-YI SAcADET.

JADAWIL, lost.
Not having access to this work in the Persian
Mentioned in Munqidh, 119, as a retort to feeble original, I rely on the accounts of Bouyges, pp. 59-60,
criticisms by the Ismacills in TUs against GhazalI. and BadawT, pp. 172-78. The latter mentions some
Thus it must have been composed while he was living editions published in India, of which the latest is
in TUs before his return to NishapUr in 499=1106, or Lucknow, 1291=1874
very soon after that. But it can be fixed more accu- This large book is a Persian version of IAyd', thus
rately than that. Munqidh lists it as the fourth in likely to have been composed after GhazalT's return
temporal order of the anti-Ismac'l! works, before al- from the Arab countries. The preface refers to Jawahir
Qistds al-mustaqlm. On the other hand, in the earlier al-Qur'Dan, making it later than Jawdhir-Arba'hn.
book Jawdhir al-Qur'dn, 21, where Ghazal! lists three From the other direction, Mustasfd, I, 3 mentions it
books as directed against the IsmaFlT1s, (Mustazhiri, as before the return to teaching in NTshapUr (Dhi! al-
Hujjat al-haqq and Qawdsim al-Bdiiniyya), Dard is Qacda, 499=July, 1106). So it can be confidently
not mentioned; from this silence Bouyges, p. 56, assigned to the first period of retirement at TUs.
rightly concluded that it had not yet been written, for The reference to this title in Munqidh, 159, is
Ghazali is not reticent about his own books. probably meant for another work, as will be explained
In brief, Darj can be firmly placed after Jawdhir- under the next title.
Arba'Tn and before Qistds, and belongs to the resi-
dence in Tiis before 499=1106. "REFUTATION OF THE PERMISSIVISTS"

AL-QISTAS AL-MUSTAQTM, ed. V. Chelhot (Beirut, A Persian "Refutation of the Permissivists," ed. and
1959). German tr. 0. Pretzl, "Die Streitschrift des GazalT
gegen die Ibahija," Sitzungsberichte der Bayerischen
The place of this work in the order can be deter- Akademie der Wissenschaften, Philosophisch-histor-
mined precisely by references. 70, etc. mentions ische Abteilung, 1933, Heft 7 (Munich, 1933).
Jawdhir al-Qur'dn, thus posterior to it and its sequel
ArbacTn. Munqidh, 119, names it after Darj as the As Ibahiyya is written in Persian it is likely to be
fifth of the polemical works against Ismacilism. later than GhazalT's return from the Arab countries.
On the other side it is referred to in Faysal, 88 and Munqidh, 154 declares: "As for the delusions of the
96. No other known work could have intervened Permissivists (ahl al-ibaha), we have grouped their
between these two, except Mishkat whose exact posi- errors into seven kinds and exposed them in K-mWya
tion remains undetermined as explained above. as-sacdda." Now K-mTyd contains passages which can
be considered answers to the Ibahls, but not systema-
FAYSAL AT-TAFRIQA BAYNA AL-ISLAM WA AZ-ZANDAQA, tically under seven heads. Ibahiyya on the other hand
in Al-Jawlhir al-ghawdlT (Cairo, 1353=1934). does accuse them of eight errors, which it answers in
turn. As GhazalT habitually relied on his memory for
This book too can be placed easily, although less facts, it is probable that in Munqidh he is referring to
exactly at the latter end. 88 and 96, mentions Qistds. this book and slipped in citing its title and the number
of errors.
Because of these uncertainties, and the lack of other
references by GhazalT, Badawi suspends judgement
Lazarus, pp. 299-300. Lazarus points out that the descrip- about authenticity (pp. 467-68). But the fact remains
tions in Mishkdt of intuitive mystical knowledge (dhawq), that the book that exists corresponds in its subject
which Watt took as evidence of a late date, are already and organization to GhazAll's description. A single
anticipated in different language in earlier works such as reference in Munqidh should be enough. I prefer to
Ihyd' and even Tahdfut. regard it provisionally as genuine, although subject to

This content downloaded from

194.27.127.99 on Thu, 26 May 2022 13:21:49 UTC
All use subject to https://about.jstor.org/terms
 HOURANI: A Revised Chronology of GhazdlT's Writings 301

further study. Bouyges-Allard and Lazarus are silent teaching again at the Nizamiyya College of Nishapu-r
about the work. (499=1106), and Nishapuir was in Sanjar's province.
Pretzl judges that Munqidh shows fairly certainly This new view, endorsed and elaborated by Laoust,
that the book was written at the Nizamiyya College in does not directly bring about any progress in the
Nishapflr after 499 (1106).32 He does not give reasons, dating of the work. But it does create a presumption
but it is likely that he relies on Munqidh, 151, where that NasThat al-mu/ak is connected with the program
Ghazalr says he was commanded by the Sultan to of religious revival for which Ghaz5lT was brought
return urgently to NishapUr to combat the spread of back from his retirement to an active life of teaching
religious indifference. The preceding pages of Mun- and writing in NTshapUr. We can find a pattern of
qidh, however, show that he had already been con- missionary writing in the three Persian works, K-mTyI,
cerned with this trend in his previous retirement, so he Ibahiyya and NasTha, and the two Arabic works of
may also have been writing about it then. Thus the NTshapilr, Munqidh and Mustasfd. While the NasTha
book cannot be assigned with certainty to before or could have been written in TUs, either immediately
after the return. before 1106 or soon after 1109, the book is more
suited to the intervening period in Nishapu-r, when
NAsTHAT AL-MULCK, Persian, ed. J. K. S. HumVT Ghazal! was working in a more urban and political
(Tehran, 1937-39). environment.
There are no references to indicate its temporal
In considering this work I have relied on the English relations to Munqidh and Mustasfd.
translation by Bagley, as well as on the discussion of
it in his introduction and in recent works by Lambton AL-MUNQIDH MIN AD-DALAL, ed. J. SalTM and
and Laoust.33 K. 'AyyAd (Damascus, 1939).
It has been noticed that the presentation of the faith
of Islam in this work follows closely the organization 99, mentions Faysal, and 154, mentions one or the
and contents of KTmTya-yi sa cdet.34 Beyond that, the other of the two Persian works as explained above.
dating of the work depends mainly on the interpreta- 153, mentions the month of his return to NTshapur,
tion of the dedication and the related historical cir- DhU al-Qa'da, 499=July, 1106. 67, GhazalT gives his
cumstances. age at the time of writing, "over 50," i.e., after the
The introductory remarks of the Arabic translation, beginning of 500=Sept. 2, 1106.
which is early, state that the work is addressed to 153, GhazRlT refers to himself as teaching, so he was
"Muhammad b. Malikshah, sultan of the East and the still at NishapUr before his final retirement to TUs.
West." This Great SeljUq sultan did not accede to the This fact and his age allow us to place Munqidh
full title until 499=1105. On this basis it has generally somewhere between late 1106 and 1109, as proposed
been concluded that the book is posterior to that year. by Bouyges, pp. 70-71, and confirmed by Poggi in the
But Lambton has now argued convincingly that the most thorough examination of the question to the
work is addressed to Sanjar, the brother of Muham- present time.35
mad who was provincial governor of KhurasAn from
1103 on and who was also entitled "sultan." For the AL-MUSTASFA MIN 'ILM AL-USCL, 2 vols. (Cairo,
Persian original addresses itself to "the sultan of the 1356=1937, TijAriyya Press).
East" only. Moreover, it was Sanjar's wazir Fakhr al-
Mulk who invited or commanded Ghazqli to take up The date of completion is given in the colophons of
two manuscripts of 1182/83 and 1193, as well as by
Ibn Khallikan: Muharram 6, 503=August 5, 1109.
This was noted by Bouyges, pp. 73-74, with refer-
32 Sitzungsberichte (1933), p. 16. ences.
3 F. R. C. Bagley, English tr., Ghazdlt's Book of counsel I, 3, mentions the return to NishApUr in 1106, also
for kings (London, 1964), pp. xvi-xxvi. A. K. S. Lambton, Jawdhir al-Qur'dn and Krmryd. Bouyges, p. 58, n. 3,
"Justice in the medieval Persian theory of kingship," Studia also gives a reference to Faysal (in an older edition).
Islamica, 17 (1962), pp. 91-119. H. Laoust, La politique de
GazdfI (Paris, 1970), pp. 144-47.
34 Bagley, pp. xxiv-xxv, referring to HumA'T and to a 35 V. J. Poggi, Un classico delta spiritualiti musulmana
doctoral thesis by H. Spencer (Edinburgh University). (Rome, 1967), pp. 3-15.

This content downloaded from

194.27.127.99 on Thu, 26 May 2022 13:21:49 UTC
All use subject to https://about.jstor.org/terms
302 Journal of the American Oriental Society 104.2 (1984)

There are no cross references between Munqidh 14th of that month (December 18, 11 1 1). The colophon
and Mustasft. As a conjecture I would put the com- is in a very early manuscript, Istanbul: Sheh-d Ali
pletion of Mustawsf later, because the phrase "over 1712:1, which gives its own date of completion as the
50" for Munqidh (67) puts it closer to 1106 than 1109, middle of Sha'bdn, 507=1113.36
and more time would be needed to complete the
substantial MustawsJ. But the writing of these two
books may well overlap.
36 Bouyges was the first to notice this colophon (pp. 80-
ILJAM AL-cAWAMM CAN cILM AL-KALAM, (Cairo, 82), but he misread the year of the manuscript as 509, and
1309=1891/92, Maymuniyya Press). this error has been repeated by later scholars. Sab' ("seven")
is quite clear in my enlargement taken from the Arab League
This work is dated precisely by a colophon as microfilm collection, listed as TawhId 34 in F. Sayyid,
having been completed in "the first days of Jumada Fihrist al-makhtutit al-musawwara (Cairo: Arab League
II, 505," i.e., a few days before Ghazall's death on the Cultural Commission, 1954).

This content downloaded from

194.27.127.99 on Thu, 26 May 2022 13:21:49 UTC
All use subject to https://about.jstor.org/terms