Outline of Stratificational Grammar-07-Lamb-Sydney Lamb
Outline of Stratificational Grammar-07-Lamb-Sydney Lamb
Outline of Stratificational Grammar-07-Lamb-Sydney Lamb
by
Syd n ey M . La mb
Ill
With a n Appe n d ix by
Leo n ard E. Newell
Copyright
by Georgetown University
ii
Preface
iii
C O NTENTS
Page
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
I. Introduction . . . . . . . . . .
Linguistic Analysis . .
Tactic Analysis . . . .
Realizational Analysis
.
. .. . . . . .. . . . . . . . . . . . . . .. . .. . . . . .
. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . .
1
3
4
5
III.
Linguistic Description . . . . . . . .
Effective Information
Surface Information . . . . . . .
Superficial Information . . . .
Some Practical Considerations
.
Bibliography
Exercises
. . . . . . . . . . . . . . . . . . . . . . . . 21
. . . . . . . . . . . . . . . . . . . . . . . . 28
. . . ..... .
. .
. . 29
.
. . . . . . . . . . . . . . . . . . . . . . . . 31
Structure
. .
. .
. . . 32
.
. . . . . . . . . . . . . . . . . . . . . . . . . . 41
.
. . . . . .
..
.
41
.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
.. . .
.
. . . . . . . . . . . . . . . . . . . . . . . . . . 46
. . . . . . . . . . . . .. ... . .. . .. . . . 54
Concerning Description . . . . . . . 5 6
. ,. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
.
. . . . . . . . . . . . . . . . . . . . . . . . 11
..
. . . . . . . . . . 18
iii
71
107
I. INTRODUCTION
Introduction
rOne of the reasons for the complexity of linguistic structure, i.e. for the
fact that it comprises several stratal systems rather than just one or two,
is that sounds and meanings- are, by their natures, patterned separately from
each other; they each have their own structural relationships. Phonological
systems must be adapted to the articulatory and auditory organs, while
semological systems must be adapted to thought patterns and to the
phenomena, events, and relationships about which people think and talk.
Moreover, phonological systems must conform to the fact that speech takes
place in time, which is linear; but no such restriction is imposed upon
semological systems, and on the contrary, the things to which semological
systems must relate - events, phenomena, experiences, relationships, and
the like - are often multidimensional. Thus a close correspondence be
tween semological and phonological systems would be impossible. The
same situation exists for written languages because writing systems are
generally based upon spoken languages, so that they tend to have close
correspondence to phonological systems but not to semologies. This situa
tion may be contrasted with that of notational systems which have been
developed independently of spoken languages and which have taken advan
tage of the two-dimensional character of paper and blackboards to allow
for closer correspondence to meaning than would be possible for one
dimensional expression systems. Examples are the notational systems used
for diagrams of organic compounds in chemistry, electronic circuit dia
grams, and flow diagrams for computer programs. {It is this same conideration which underlies the desirability of a two-dimensional diagram
ing notation, such as that used in this Outline, for linguistic structure.)
Another factor is that a language has the property of allow_ing its
speakers to speak of anything in the range of their experience and imagina
tion while employing only a very small number of different fundamental
phonological elements. Thus the number of elements at the lowest stratum
of a language is only around fifteen, while the elements at the highest
stratum number well into the thousands.
The concept of stratification in language is by no means original with me,
nor are many other features of this system, which is no more than the result
of adding a few extensions and refinements to a synthesis of Hjelmslev's
glossematics ( 1 943 ) and one of the standard versions of American struc
tural linguistics.1 The present formulation differs from earlier ones mainly
in recognizing more strata and in that it provides a more precise specifica
tion of the relationships on which the concept of stratification is based. It
derives partly from the work of Hockett (e.g. 1947c, 1 954, 1 9 5 8 ) , but
differs from it by recognizing more strata. That earlier, "proto-stratifica
tional," framework had just two strata, phonemic and morphemic, which
were approximately at the levels of the hypophonemes and the lexons,
respectively, of the system presented here.
linguistic Analysis
l i n guistic A nalysis
I ntroduction
The old Hindu saying quoted above actually expresses what is perhaps
the most important and powerful single principle of linguistic analysis.
Another important principle is that linguistic analysis should be based on
such criteria that any (competent) independent investigator applying them
to the same material would arrive at an equivalent solution. A third is that
a language should be described in its own terms, and not by the application
of an external mold. This last principle also applies to areas within linguistic
structure. Thus, for example, the pattern of arrangement of morphemes
should be described in its own terms, not by the application of phonological
criteria. Similarly, it is both reprehensible and unnecessary to attempt to
define units of the tactics of phonemes, such as syllables, by the use of
phonetic entities, such as chest pulses. In other words, a good procedure
of tactic analysis is one which allows the patterns of arrangement to reveal
themselves, while that analysis is bad which applies to a system, from the
outside, a predetermined structure.
It is convenient to . distinguish two general types of linguistic analysis,
concerned with two kinds of linguistic patterning. They may be called
TACTIC ANALYSIS and REALIZATIONAL ANALYSIS.
Tactic Analysis
Let us consider further the algebraic example given above. The second
expression, i.e.
ab (c+d +e +f +g)
is simpler than the first. The simpler version shows the result of finding
a recurrent partial similarity and isolating it. The same type of process lies
at the basis of tactic analysis. A simple illustration is provided by the
following forms:
big boy
sad boy
wet boy
big man
sad man
wet man
big dog
sad dog
wet dog
big cat
sad cat
wet cat
One way to describe the fact of the occurrence of these forms in English
is that just used: one may list the forms. For this description it was
necessary to make 24 item-identifications. Another way to account for the
same data is as follows:
boy
big
man
sad
dog
wet
cat
Realizational Analysis
or simply (big, sad, wet) (boy, man, dog, cat) . The latter description
requires only seven item-identifications. The excess surface information
of the first description is seventeen item-identifications, i.e. 24 minus 7; and
if one cares to count, one will see that those 1 7 excess identifications are
entirely accounted for as repetitions in the first description. When one is
dealing with classes of hundreds of items, the difference is of course more
striking. Moreover, the former approach to description is not even
theoretically possible in general, since the number of combinations gen
erated by a tactic system as a whole is infinite.
Simple as it appears, this process of isolating recurrent partial similarities
is the basis of tactic analysis. If the traditional equipment of tactic analysis
is reduced to its essentials, it turns out to involve little more than applying
a simple factoring operation like that performed by the ninth-grade algebra
student. This process leads the analyst to distribution classes and construc
tions which describe arrangements in the simplest possible terms. Questions
concerning immediate constituency are automatically answered; and the
criteria which have been put forth for determination of immediate con
stituents (cf. Wells 1 94 7 , Nida 1 949: Chap. 4), insofar as they are valid
at all, are derivable from a more basic simplicity principle.
Realizational Analysis
Analysis concerned with the parts of a stratal system other than its
tactics may be called REALIZATIONAL ANALYSis. Such analysis involves
four basic types of operations: HORIZONTAL GROUPING, HORIZONTAL SPLIT
T I NG, VERTICAL SPLITTING, and V ERTICAL GROUPING.
An example of
HORIZONTAL GROUPING is the recognition that elements c, a, t combine, in
that order, to form a structural unit cat to be identified in such forms as
cats, catlike, catgut, etc., but not in cattle or caterpillar. HoRIZONTAL
SPLITTING is illustrated by the analysis of French au as a realization of the
underlying combination a [e. VERTICA L SPLITTING involves distinguishing
two or more higher level units (e.g. the past participle lexon and the past
tense lexon of English) realized by the same lower level unit (e.g. -ed) .
Note that au is a realization of the combination ale, i.e. of a AND le, whereas
any occurrence of -ed is a realization of either the past participle lexon O R
the past tense lexon. V ERTICAL GROUPING involves the recognition that two
or more units (e.g. -ed, -en) are alternate realizations of a single higher
level unit (the past participle lexon). In practice, none of these four
operations can be performed in isolation, nor is there any sequence in which
they must be performed; rather, they are generally carried out together,
with reference to one another.
These operations are motivated by the simplicity principle and by the
need to account for the linguistic data. Failure to carry one of them out,
where the possibility is available, would mean either that the description
Introduction
Realizational Analysis
Criteria of this type are to be used for testing proposed descriptions rather
than for prescribing procedures of analysis. Such criteria specify the prop
erties of an acceptable solution. With this approach, the analyst has
much
greater freedom. He can jump around from one level of analysis to another
at will. He can use meaning whenever he wants to (notwithstanding earlier
teachings to the contrary).
mind all along as he conducts his analysis, performing it so that the require
ments will be satisfied. This approach does not in any way deny the value
of various valuable analytical tools and techniques which a linguist should
know; but these belong to the practical side of linguistics rather than to
linguistic theory.
Actually, the three-way primary division of linguistic structure, and with it the
. a
Louis Hjelmslev
I consider a language to be a system of relationships. It may be analyzed
into a series of subsystems, called STRATAL SYST EMS, each of which has a
syntax or tactics and certain other characteristic patterns of relationships.
The elementary relationships of which these patterns are composed are of
a very small number of types.
In analyzing and describing linguistic relationships it is a matter of
practical necessity to employ some system of precise and simple notation.
Some linguists have been saddened by the increasing use of algebraic
notation in linguistics and have longed for the "good old days" in which
.
linguistic descriptions were written in ordinary words. Unfortunately, in
those good old days linguistic descriptions were neither very complete nor
very precise. Languages are such complicated systems that none but the
most gifted mentalities, if even they, can cope with their structural intrica
cies unless aided by notational devices. To engage in linguistic description
without the use of a notational system well suited to the purpose makes no
more sense than to try to do long division with Roman numerals. It has
sometimes been objected that if linguistic descriptions are written in a
special notation rather than in ordinary words they are no longer accessible
to the untutored general public. But it is better to have precise descriptions,
even if they can be read only after training, than not to have precise
descriptions at all. One wonders how far physics and chemistry would
have progressed if similar objections had dissuaded them from using nota
tional devices.
A two-dimensional graphic notation appears to be superior to algebraic
notation for some types of work with linguistic structure, while for other
purposes algebraic notation is more convenient. It is therefore desirable
to employ notational systems of both types, designed so that they are
mechanically interconvertible (as is done in electronics). The diagrammatic
notation used here provides a more direct portrayal of structural relation
ships than the algebraic, and it may be considered the primary notation,
while the algebraic notation may be regarded as a means of describing
structural diagrams, and thus only indirectly a means of describing linguistic
structures. In addition, it is convenient to employ a tabular notation for
some descriptive purposes, as illustrated by Newell in the Appendix to
this Outline.
L INGU ISTIC GRAPHS are made up of LINES and NODES, i.e. places where
lines intersect. Three fundamental dichotomies provide eight types of
nodes, of which at least seven play a very important role in characterizing
Linguistic Structure
UNORDERED
AND
ORDERED
OR
AND
OR
DOWNWARD
UPWARD
Figure
1.
lbc
d j e,f
gh 1i
j,k 11
m lno
pjq+r
s t 1u
v+w 1x
The directions UPWARD and DOWNWARD are in keeping with the diagram
ming convention according to which meaning is at the "top" and speech at
the "bottom" of linguistic structure. Thus UPWARD means "towards mean
ing" while DOWNWARD means "towards expression." Note that in th e
algebraic notation the diagonal clearly indicates the direction since it has
10
Linguistic Structure
an upper side (at left) and a lower side (at right). Where convenient,
the diagonal of opposite direction may be used. Thus
x\ y . z
jx
fb
1 e, f
g. h
li
j, k I 1
m
jno
s t; u
v
I x Dv OR Dw goes to Dx
Note that the impulse moves upward from a downward ordered AND
(e.g. up m ) only after the second (or last) impulse from below (e.g. up o)
has reached the element; the preceding impulse(s) from below (e.g. n)
wait at the element until the last impulse from below arrives there.
Patterns
g
c
f Ig, (h n
led
bide
I k (l,
m)
nI
(p, q)
v Iw,z
Figure 2.
vx
v. y
11
12
Linguistic Structure
Patterns
Patterns
13
LEXEMIC
-n
MORPHEMIC
PHONEMIC
14
Linguistic Structure
au
le
(See Figure 1 0, r11ge 24, and the Appendix for additional examples of tactic
configurations.)
Figure 5 introduces the ZERO element, represented by a small circle at
the end of a line. It is found in tactic patterns, alternation patterns, and
sign patterns. Impulses moving to a zero element disappear; and an impulse
may move from a zero element to the connecting line at any time. In the
algebraic notation, the zero element is represented as 0; xI y,jiJ may be
written xI[y] and x,0 Iy may be written [x] Iy, where x andy are any
symbols representing lines.
As Figure 5 illustrates, tactic patterns are in general made up largely
of downward ANDs and upward and downward unordered ORS. Upward
ORS in tactic patterns represent situations in which units have two or more
alternative tactic functions. For example, Figure 5 shows that a glottal
stop in Monachi may function either as the initial consonant or as coda
of a syllable. Figure 1 0 below has an upward OR for nominals indicating
that a nominal in English lexotactics may function as a predicate nomina
tive, as object of a verb, as subject of a clause, or in construction with 's
(e.g. Mary's ) . Downward ANDS in tactic patterns are often called CON
STRUCTIONS. Semotactic patterns differ from tactic patterns for lower strata
in having considerable numbers of upward ANos; occasional upward ANDS
are also found in lexotactic patterns.
A tactic pattern has at its top only the zero element. The connections
to the next higher stratum are not through the top of the tactics, but
through the alternation pattern ( see below) . Thus the syllable, a phono
tactic unit, does not connect upwards to anything in the morphemic system;
and the morphological word, a morphotactic unit, does not lead up to
anything in the lexemic system. At the bottom of the tactic pattern of a
given stratal system are connections to the "emes" of that system.
Patterns
15
16
Linguistic Structure
TACTIC
PATTERN
Morphemes
Morphemic
Signs
KNOT
PATTERN
----+
SIGN
PATTERN
Morphons
---+
knots, and, below them, occasional upward ORS. The upward ANDS of the
knot pattern are the "emes" of the stratal system. Thus an "erne" has a
function in the tactics, a connection to the next higher stratum through the
alternation pattern, and a connection to the next lower stratum through the
sign pattern. In addition, a knot pattern commonly has one or more lines
going upward directly to the tactic pattern rather than to upward ANDS.
Such lines represent nondistinctive units, whose occurrence is determined
by the tactics and which have no connection to the next higher stratum.
Figure 6 also specifies the meanings of the terms MORPHON, MORPHEMIC
SIGN, MORPHEME, and LEXON. As a characterization of actual morphemic
systems it is deficient in that it shows only eight morphemes, i.e. only a
fragmentary sample of the hundreds of morphemes actually present in
morphemic systems. (Linguistic graphs are, of course, too cumbersome
to describe large quantities of relationships, and this consideration is the
Patterns
17
ALTERNATION
PATTERN
KNOT
PATTERN
beT
gud
.,
primary reason for the use of an algebraic notation and a tabular notation
in addition to the diagrams. )
The alternation pattern, knot pattern, and sign pattern, taken together,
may be said to constitute the REALIZATIONAL PORTION of the stratal system.
Any node in the realizational portion represents a discrepancy from a simple
one-to-one relationship between units of neighboring strata. Terms for
types of interstratal discrepancy are indicated in Figure 6 by capital letter
abbreviations, whose meanings are as follows (cf. Lamb, 1 964b ) : D,
DIVERSIFICATION; Z, ZERO REALIZATION; N, NEUTRALIZATION; E, EMPTY
REA-LIZATION; P, PORTMANTEAU REALIZATION; C, COMPOSITE REALIZATION.
Of the operations of realizational analysis mentioned above (Introduc
tion ) , horizontal grouping involves recognizing instances of composite
realization, while horizontal splitting is concerned with portmanteau realiza
tion. Vertical splitting is the discovery of instances of neutralization ( in
cluding those involving empty realization ) and the determination in each
case of just where the associated upward OR is to be placed for maximum
simplicity of the description; and vertical grouping is associated with
diversification (including the special type called zero realization) .
Figure 7 provides an illustration of diversification in the alternation
pattern of the English morphemic system, together with a closely associated
18
Li nguistic Structure
portion of the morphotactics. The tactic construction shown is that for the
forms consisting of an adjective followed by the comparative suffix or the
superlative suffix. The morpheme Ml beTI is shown as occurring only in
this construction, while MI gudl is shown as an ordinary adjective, entering
into a class ( cf. the downward OR ) which has a variety of morphotactic
functions ( cf. the upward oR ) . (The morphon MNITI of MlbeTI is a
special "t" which is realized as zero ( i.e. has no phonemic realization) in
best. ) When the lexon LNigoodl occurs with LNI-er11, the tactics would
allow either better or gooder. but the downward oRfor Lliigoodl is ordered
and therefore MlbeTI takes precedence. When neither LNI-er11 nor
L81-estl is present, the realization MlbeTI cannot occur, since the impulse
from LNigoodl to MlbeTI cannot get past the upward AND unless the other
line leading into it also has a downward impulse. An alternative analysis of
the tactic pattern would specify that MIgudl cannot occur in the compara
tive construction; but this analysis would require an unnecessarily compli
cated graph, since the occurrence of MIgudl in this construction is ruled
out by the ordering in the downward OR. Moreover, the analysis shown
accounts for the English speaker's ability to recognize gooder when it is said
by a child who has not yet learned this diversification (whereas, on the
other hand, *betness would not be understood as meaning the same as
goodness, and this form is not allowed by the morphotactics).
Figure 7 also shows an example of neutralization in a knot pattern,
involving the comparative -er1 and the agentive -er2 ( as in swimmer) . Note
that, as the diagram indicates, these homophonous morphemes differ both in
their meanings and in their morphotactic functions. Also present in typical
knot patterns is a type of neutralization in which one line leads from the
upward OR directly into the tactic pattern. Such cases have particularly
attracted the attention of linguists in hypophonemic systems, and the terms
NON-DISTINCTIVE FEATURE and DETERMINED FEATURE have often been
used in this connection ( cf. Lamb 1966b).
Strata and Terms
19
20
Linguistic Structure
Units
e m
Hypremon r
/re.
Hyperseme
/s
Hypersememic
Sememic
0 0
Semon
I ooooooSeme.
/xeme
Lexemic
on
Lex
I
MorphemeLe.x
Morphemic
Morphon/1
I -Morph
Phonemic
/"Te
Phonon
:
I
Hypophonemic Hypophoneme
Phonei
o o -
o o o
21
A stratum differs from its neighbors with regard to its elements and their
combinations. The elements of a stratum, i.e. the upward ANDS of its knot
pattern, are obviously unique to that stratum. Between them and the emes
of a neighboring stratum there is a sign pattern and an alternation pattern,
in addition to neutralization in the knot pattern. Perhaps the interstratal
discrepancies occasioned by the sign patterns are the most apparent ( cf.
Figure 3 above ) , since these are obvious differences in size: between
phonological components, phonological segments, morphemes, lexemes,
semological components. In fact this difference has been so obvious rela
tive to the other discrepancies that the latter have often been overlooked;
linguists have often supposed that some of these different elements merely
represented different tactic levels within a single tactic pattern, i.e. different
RANKS ( cf. Halliday 196 1). Thus the division between morphology and
syntax was based upon a distinction of rank by those who held it: morp
ology dealt with combinations of morphemes up to the rank of the word,
while syntax was supposed to deal with combinations of words. The com
monly held opposing view was that such a distinction was unnecessary and
should not be maintaint:d; rather, it was held, there should be a single
mp.rphotactics dealing with combinations of morphemes all the way to
the rank of the sentence. Both views were incorrect, since both were
attempting to deal with the tactic phenomena of grammar in tern1s of a
single stratum. The units dealt with in syntax, properly treated, are
lexemes, not words or morphemes, and they differ from morphemes in
their arrangements as well as by virtue of an intervening sign pattern
and an alternation pattern. Morphotactics is a tactics of morphemes,
whereas syntax -i.e. lexotactics -is a tactics of lexemes. (To confuse
words with lexemes, as has been quite commonly done, is to make the
same mistake as that involved in confusing syllables with morphemes ; the
word is a tactic unit of the stratum below the lexemic, which is often but
by no means always the realization of a lexeme. )
Every node in the realizational portion of a stratal system represents
a difference between two strata. That is, every AND in a sign pattern
involves a difference in size of units of neighboring strata; every down
ward OR represents an alternation among two or more units of the lower
stratum realizing a single unit of the upper; every upward OR represents
a neutralization of units distinct on the next higher stratum.
Less apparent at first glance is the type of difference represented by the
upward ANDS of a knot pattern, since these nodes each have a single line
22
Linguistic Structure
going to the next higher stratum and a single line leading to the next lower
one. But the third line, leading to the tactic pattern, represents an im
portant interstratal discrepancy, since that tactic pattern differs from those
of neighboring strata. Thus it is not in general the case that the emes of a
stratum and their realizations on the next lower one occur in the same
arrangements. For example, sememic arrangements generally and lexemic
arrangements in part are non-linear, while combinations of morphemes
appear to be generally linear ( at least in many languages ) . The phenome
non of difference in the ordering of units ( i.e. difference in the sequencing
of impulses ) from one level to another may be called ANATAXIS. For a
simple example of a difference between lexemic and morphemic ordering,
compare LJ-z have-en be-en take/ with Mjhave -z # be -en # take -en/
(the word boundaries are provided by the morphotactics ) .
Differences between the ordering of morphons and that of (basic)
phonemes have often been observed, and have commonly been referred to
by the term metathesis. As an example, consider the following forms of
Zoque (from Nida 1 949 : 67-68, 1 7 1 - 1 74 ) , written phonemically at the
left and morphonically ( with morpheme boundaries indicated ) at the right :
kenu
kenpa
kyenhayu
kyenhapya
kento?yu
kento?pya
looked
he looks, will look
he saw it for him
he sees it for him
he wanted to look
he wants to look
ken u
ken pa
y ken hay u
y ken hay pa
ken to?y u
ken fo?y pa
23
MOR PHEM IC
SIGN
PATTERN
PHONEMIC
KNOT
PATTERN
accord with the lower tactics, despite what the higher stratal system speci
fies. In the case of has been taking, the lexemic units L/have-en/, L/be-ing/
are single lexemes, hence elementary units of the lexotactics. The lexemic
sign pattern specifies that each has two components; but the lexonic order
ing Lj-z have -en be -ing take/ is unacceptable to the morphotactics,
which allows verbal suffixes to occur only after verb stems ; in general, the
ordering specified to a tactics from above is of importance to that tactics
only for sp e cifying the succession of choices at individual choice-points
( i.e. downward ORS ) ; but when a lexonic ordering is of elements belonging
to different morphotactic classes, e.g. verb stem and verb suffix, it is up
to the morphotactics to determine their relative order. Thus M I -z/ is the
first verbal suffix to be selected, while 31/-en/ is the second; and lljhave/
is the first verb stem, while 31/be/ is the second. So the morphotactics puts
out, under specification of this combination of lexons, li /have -z # be -en/.
Intrastratal anataxis, on the other hand, is that which involves a differ
ence in ordering from one level to another of a single tactic pattern. Two
examples are shown in Figure 1 0, a simplified tentative description of a
fragment of English lexotactics. In the lower right hand portion, which
concerns the verb phrase ( VP ) , the line leading down to the object ( Obi )
of a verb from the verb-object ( VO ) construction encounters an ordered
upward AND. An impulse coming down this line to this AND will be delayed
if the line labeled d (which is optionally active if Av occurs ) is activated,
since it is not possible for an impulse to proceed below the AND node until
the impulse from line d reaches it. Thus the graph generates such pairs as
Ljtake away the cat/ and Ljtake the cat away/, whose difference is due to
whether or not the optional line d is taken. The other example shown in
24
carry
run
25
Nom
Johnny
pla y
can
lhe
saxo- p hone
mz
VP
can
Johnny
play
Fig u re 1 1 .
lhe
saxo-phone
mz
26
process, the impulse going from the clause construction to the subject
phrase is delayed until the impulse comes from the predicate construction
to the other line connecting to this upward AND ; the resulting order is that
seen in Can Johnny play the saxophone? ( Figure 1 1 ) . The lexeme LJQ/
is itself realized as M.ff/J/ in this example (because Mjcan/ is present) , but
in some environments ( as specified in the morphotactics ) it is realized as
do, as in Does Johnny know any other songs?
Figure 10 introduces a new type of node, the COORDINATION element,
represented by a half circle. A downward impulse meeting this element
may optionally result in two or more separate impulses continuing down
ward from it, each of which is treated by the morphotactics as if it were
the only one. In the algebraic notation the coordination element may be
represented by an asterisk, together with identification in parentheses of
any units which may occur between the last two of the coordinate units ;
for example, the coordination element above the Pred construction in Figure
1 0 might be represented in the formula for CLAUSE as follows :
CLAUSE I Subj * ( and,or,but) Pred
27
Distr.
Pref.
tactics. This is the special property of the reduplication element. (It would
be possible to account for reduplication without the use of a special ele
ment, but only at considerable expense of ordinary AND and OR nodes;
whether the reduplication element is to be regarded as a shorthand device
for the more complex structure of AND and OR nodes, an element of struc
tures of individual languages whose analysis in terms of ordinary lines and
nodes belongs to linguistic theory, or a separate element in general linguistic
theory as well, remains an open question for the present. ) Reduplication
occurs not only in phonology, but also in morphemic and lexemic systems.
Also, there may be distinguished from each other the type of reduplication
which realizes a unit of the upper stratum, as in the example of Figure 1 2,
and the type which exists within a single tactic pattern; i.e., which is
determined by the tactics itself rather than by the upper stratum. The
latter type is present in the lexotactics of various Indo-European Ian-
28
Linguistic Structure
Grammar
29
G ra m mar
The distinction between the morphemic stratum and the lexemic has
likewise been overlooked or else haphazardly treated in the past. Figure 3
above, showing part of the lexemic sign pattern of English, indicates one
of the types of evidence for the separation. The two strata also have quite
different tactic patterns, as illustrated in the Appendix. This difference in
tactics is the source of the traditional morphology-syntax division, but it
has usually been supposed that the distinction was based merely on the
size of units dealt with, and the existence of a sign pattern and an alterna
tion pattern between the two tactic patterns generally went unrecognized.
The unit whose combinations are defined by the syntax (i.e. lexotactics ) is
actually not the word but the lexeme.. To suppose that syntax should deal
with combinations of words ( or that the tactics of syllable structure should
be based upon the classical phoneme ) would be to make the same mistake
as that involved in supposing that morphological structure should deal with
combinations of syllables.
The common view put forth in opposition to supporters of a morphology
syntax distinction has been that there is a single morphotactics, concerned
with combinations of morphemes up to the tactic level of the sentence ( e.g.
Hjelmslev 1 943, Harris 1 9 5 1 ) . The structural unit taken as the basis of
this confl.ation of tactic patterns, called the morpheme by American struc
turalists, commonly approximated that which is here called the lexon, a
unit as appropriate as could be expected for such a compromise since it
is intermediate between the morpheme and the lexeme.
Properties of morphemic and lexemic structure are exemplified in the
Appendix. A special feature of the lexotactics of English and other Ian-
30
Linguistic Structure
Figure 1 3 .
Semolog y
31
the-kid-pl-re-have-en-be-ing
/ a d
for-hour-pi
dace /
the lexemes Ljsing/ and Lfdance/ each occur with preceding Ljbe-ing/
and following Ljfor hour pl/. Such a non-linear combination, however,
is unacceptable to the morphotactics, which arranges the realizations of the
coordinate elements sequentially ; but each of these realizations occurs in
the morphemic realization of its environment, up to the limits of a morpho
logical construction. Thus the lexon 1:-/ing/, a component of Ljbe-ing/,
is multiply realized on the morphemic stratum :
M.f
Examples of various other features of the lexemic and morphemic systerns of English are given in the Appendix.
Semology
32
Linguistic Structure
33
prevalent even into the decade of the sixties. This concept was often held
side by side with several others, so that the term morpheme was used for a
wide variety of different units, sometimes within the pages of a single book.
A very widely held concept of the morpheme has been that it is a class
of combinations of classical phonemes, called its allomorphs. But this con
cept too is untenable. For a class can be completely specified by a listing
(or other precise identification) of its members; this is not true of any of
the units that linguists have wanted to call morphemes, since their specifica
tion requires not only an identification of their realizations but also the
information as to the environments that provide the conditions for the
occurrence of the realizations. This concept also runs into logical difficulties
because of the phenomena of empty realization (cf. Hockett 1 947c ) , port
manteau realization ( cf. Hockett 1 947c, Nida 1 948 ) , "link phonemes"
(Wells 1 949 ) , and the type of problem exemplified by the past tense of
English take (Bloch 1 947, Nida 1 948, Hockett 1 9 54 ) . To take account
of empty realization with the morpheme as a class would mean recognizing
entities which are members of nonexistent classes ; no version of set theory
accommodates such a notion. The opposite relationship, zero realization,
would involve classes with no members, and this is a part of set theory
(i.e. the empty set ) ; but empty realization violates the concept of the class.
The same is true of the portmanteau, since it would involve an entity being
a member simultaneously of two classes, but only when they are concate
nated. This is not the same as the relationship of membership in two or
more separate classes . That concept, which is allowable in set theory,
would be the relationship of neutralization ( i.e. upward oR ) , if the class
concept of the morpheme were tenable.
These difficulties led Hockett to the position that phonemes and
morphemes are on different strata and that morphemes are REPRESENTED
by combinations of phonemes, rather than having these combinations as
members (Hockett 19 54 ) . Hockett's morpheme, during this period, was
approximately the same as the unit here called the lexon. This improved
conception was still deficient in that it assigned no structural status to
morphophonemes or to alternations among phonemes independently of
morphemes, and in that, as Hockett pointed out ( 1 9 54 ) , it failed to pro
vide a thoroughly realistic and economical account of the forms like English
took.
A further development. of Hockett's two stratum view ( 1 9 6 1 ) replaced
34
Linguistic Structure
35
36
37
38
Linguistic Structure
level of sophistication is the approach which would ( for the sake of a simple
description ) allow basic forms of some stems to be written so that they
coincide with neither of the actually occurring forms, even though they are
still written in terms of "phonemes" ( cf. Nida 1 949 : 34-3 5 ) . Allowing
himself this freedom the linguist sets up each basic form in such a way that
the description of the processes yielding the actually occurring forms ( i . e .
the rewriting rules ) will b e a s simple a s possible. Even more advanced i s
the approach which allows some basic forms t o contain phonological units
that never actually occur as ( classical ) phonemes . Thus Panini's grammar
of Sanskrit uses, for the sake of economy, a long vocalic r in some basic
forms, even though only short vocalic r occurs in actual hypophonemic
forms. Similarly, one means of treating English knife:knives ( in which
there is alternation in voicing of the spirant) along with fife:fifes and
hive:hives is to make use of three separate phonological units underlying
labial spirants in basic forms, even though only two are distinguished
hypophonemically.
Finally, the highest degree of sophistication of process description may
be called quasi-stratificational. Here the linguist realizes that he is dealing
with different levels of structure, but he continues to use the method of
description that is appropriate only to the process conception or to dia
chronic description. In classical morphophonemics the advance to this
degree of sophistication was achieved sporadically but never became in
corporated into general doctrine. Instead many neo-Bloomfieldians, recog
nizing the lack of psychological reality of process description, concluded
during the forties that one must abandon economy in this area altogether,
since it did not occur to them that any other means than process description
was possible for dealing directly with morphophonemic alternation . Some
linguists continued to use process morphophonemics at one of the degrees
of sophistication, sometimes apologetically, while others took the view that
morphophonemic alternation had no structural significance and instead was
merely a matter of alternate phonemic shapes of morphemes ; where a
morphophonemic rule could be used in a description it was regarded as
merely a means of summarizing or abbreviating several statements about
allomorphs but was in itself considered to be without structural significance .
In using the quasi-stratificational process approach the linguist recog
nizes that there are separate strata but he states his rules in a way that would
be appropriate to a pure process conception ; i . e . , as rewrite rules, which
replace certain units under certain conditions . He thus fails to give explicit
recognition to the third, underlying entity, except in occasional cases like
that of the long vocalic r of Sanskrit. Given a particular alternation in a
language between two units /x/ and /y/, where /y/ occurs, let us say, in
some restricted environment and /x/ elsewhere (but suppose that the
language also has forms with /yI not participating in the alternation ) ,
39
9 in the context : - 1
40
Linguistic
Structure
1 This
in Lamb 1 9 64b.
2 To
properly relate to one another my papers of the past few years one must
Figure 1 4.
41
42
Linguistic Description
Effective Information
43
Figure
1 5.
Two graphs G and H which are alternatives to each other (i.e. have the
same bottom lines and top lines ) CONVEY THE S A M E E F F ECTIVE I N FORMA
TION if and only if ( 1 ) for any combination T of downward impulses along
top lines of G and H, the downward set of T through G is the same as the
downward set of T through H, and ( 2 ) for any combination B of upward
impulses along bottom lines of G and H, the upward set of B through G i s
the same as the upward set of B through H. If G conveys the same effec
tive information as H, it may be said to be EQUIVA L E N T to H, and the rela-
Fig ure
1 6.
44
(1}
(2)
b
d
(3)
(4)
(5)
(6)
(8)
(7)
figure 1 7.
Effective Information
45
(3)
(4 )
(5)
(6 )
(7)
(8)
a l b,b = a l b
c I d, ( e,f ) = c I ( d,e) ,f = c I d,e,f
g I h ( i j ) := g I ( h i ) j = g I h i j
k I I ( m n) = k I (I m ) n = k I I m n
p I qr,qs = p I q ( r,s )
t l uv,v = t l [u]v
w ly,z and x l y,z = w,x l y,z
a l cd and b l cd := a,b l cd
(9)
( 10)
(11)
(12)
(13)
c I d+ ( e+f )
c I ( d+e ) +f = c I d+e+f
p I qs,rs = p I ( q,r) s
p I q r,q s = p I q (r, s )
p I q s,r s = p I ( q,r) s
t I u,uv = t I u[v]
-
46
Figure
1 8.
the first. The situation shown is typical of alternations ; the section at right
is that part of the tactic pattern which specifies the conditions for the occu r
rence of the alternants . Of the alternants x and y, realizations of Y, x
occurs in a restricted environment.
S u rface I nformation
Surface I nformation
47
Fi g u r e 1 9.
refined analysis it too fails, for it turns out that not all lines express the
same amount of information . In Figure 20 the alternatives have the same
number of lines as well as the same number of nodes ( and they convey
the same effective information ) . But the graph at left is repetitive while
that at right is economical in treating the fact that a leads down to x. The
algebraic representation of the two graphs is :
a / x,xyz
a I x[yz]
Note also that if the situation is altered very slightly to that of Figure 2 1 ,
the simpler graph has more lines as well as more nodes .
Thus a more refined measure of surface information is needed. The
information that is actually relevant is just that which is needed to specify
an individual linguistic structure, i .e. to distinguish it from all other lin
g uistic structures . A brief discussion of just what kind of information this
48
z
X
Figure 2 1 .
Surface Information
49
its
P LURAL SIDE.
(1 )
(2)
(3)
(4)
(6)
(5)
(7)
l 1 1 I ft\ TI 1
1\
I I
I
I
I I
I \
I
I
4
II
Figure 22.
I I
I I
fl
,,
"
I I
,
I
p
I
I
I
as additional
50
F i g u re
23.
51
Surface Information
Figure 24.
52
linguistic Description
F i g u re 2 5 .
more nections ( i.e. one more ) among which each terminal selects. On
the other hand , if the diagram is taken as representing part of a sign pat
tern, say the morphemic sign pattern, then the terminals of the graph at the
right have considerably more information than those at the left ; for if the
sign pattern is organized in the manner of the diagram at the left, then
each terminal leading downward from a morphemic sign represents a choice
among a very small number of possibilities ; but if the type of organization
shown at the right is allowed, the possibility would exist for hundreds of
additional nections like that at the right to be present in the sign pattern, and
the surface information value of every terminal leading down from a
morphemic sign would be greatly increased. Therefore, for example, we
refrain from identifying a common component ar in car and art, it would
represent a spurious generalization. Moreover, the organization of a sign
pattern according to the graph at the left is apparently to be arrived at as a
matter of general linguistic theory rather than separately in the analysis
of individual languages.
Of the numerous principles that have been put forth at various times to
guide the conduct of linguistic analysis (e.g. in Nida 1 949 ) , some are
invalid or superfluous because ( 1 ) they are based upon unsound theoreti
cal foundations, or ( 2 ) they involve relatively aimless classifying without
apparent theoretical j ustification, or ( 3 ) they were of use only in their
roles as steps in unsuccessful procedures of analysis ; others, however, are
entirely or largely valid and remain useful. The validity of these latter does
not rest upon their own inherent virtues nor on some other independent
foundations ; rather, they are derivable from the simplicity principle, as
imple"mented through the concepts of surface information and effective
information . This is true, for example, of the principles of biuniqueness
and complementary distribution in phonology ( cf. Lamb 1 966b ) , and of
several (but not all ) of the criteria that have been put forth for defermin
ing immediate constituents ( Wells 1 947 , Nida 1 949 : 86-9 5 ) .
The determination of the immediate constituents of sentences, phrases,
and words is of course not important as an end in itself; rather, such de
termination is an approach to the description of the tactic pattern of a
Surface Information
(1)
wise
true
un
ly
(2)
true
un
wise
ly
(3)
true
Fig ure
wise
26.
ly
53
54
stratal system (e.g. the lexotactics of some language ) . Every valid "cut"
into immediate constituents represents a construction (i.e. a downward
AND ) in the tactic pattern. Alternative immediate constituent analyses
mean alternatives in the tactic description ; that alternative is to be pre
ferred which is simpler. Consider, for example, the English word untruly,
which offers these possibilities for analysis : ( un ) ( true ly ) , ( un true ) ( ly ) ,
( un ) ( true ) ( ly ) . The choice among the alternatives depends upon how
they fit into the tactic description, which must of course account for other
forms as well, in particular such forms as truly and untrue, and for the
fact that the distribution of true, untrue, wise, unwise, etc., is different from
that of truly, untruly, wisely, unwisely, etc. The three analyses are shown
in Figure 26 . Diagram ( 1 ) , for the analysis ( un ) ( true ly ) , has quite
obvious excess surface information whose elimination results in diagram
( 2 ) , representing the three-way cut. But ( 2 ) can be further simplified
to ( 3 ) . Thus the analysis ( un true ) ( ly ) fits the simplest tactic description,
and is for that reason to be considered the correct immediate constituent
analysis . This is also the analysis that would be arrived at by applying the
principles of Wells or Nida.
S u perfi c i a l I n formation
Fig ure
27.
Su perfi cial
Information
55
b I c,d
clgh
dIel
e I i,f
fljk
(2)
b I gh, (i,jk) l
(3)
B = A I C,D.
C=BIGH
D = B I (I,E ) L
E=DIJK
56
Linguistic Description
ing linguistic structure. Thus the diversification for the English lexon
Igood/ depicted in Figure 7 could be described as
LN
if MbeT and Mgud are used as symbols for the morphemes to which LN Igood/
leads; and the conditions for the occurrence of the one realization or the
other are specified in the tactics, to which the reader of the description may
refer. The reader will be spared the trouble of referring to the tactics,
however, if the formula for the diversification includes the relevant tactic
information :
L:\'good I I - (erhest) I MbeT
+ I I - I Mgud
A formula of this type may be called a REALIZATION RULE (cf. Lamb
1 964a ) . It states that the structural relationships associated with Lx I good/
are such that the realization lijbeT/ occurs only when the comparative
suffix or the superlative suffix follows ; otherwise, i.e. in any other environ
ment, the realization MI gud/ occurs .
Some Practica l Considerations Concer n i n g Description
Hj
HSj
HNj
s;
ss;
s:\';
.
.
I
I
I
/
. I
Lj
I
LSj . . . I
I
LN j
.
hypersememes
hypersememic signs
hypersemons
sememes
sememic signs
semons
lexemes
lexemic signs
lexons
57
Mj
I morphemes
MSj
I morphemic signs
MN j
I morphons
P f . I ( basic) phonemes
P Sf . . . I ( basic) phonemic signs
PNf . . . I ( basic) phonons
Hf
I hypophonemes
S
I hypophonemic signs
Hf
Fr
Do
Pd
Ph
<Jl
Cl
Sp
Ns
Uv
Vo
Hi
Mi
Lo
Labial
Apical
Retracted (a component of English
P Sfr/, P Sjz/ , and P S/s / )
Frontal
Dorsal
Postdorsal
Pharyngeal
<Jlottal
Closed (i.e. stop )
Spirant
Nasal
Unvoiced
Vocalic
High
Mid
Low
58
Description
59
BI BLIOGRAP H Y
BARKER, M. A. R.
1 964
Klamath grammar. University of California Publications in
Linguistics 32.
BILIGIRI, H. S.
1 959
Kannada verb : Two models . Indian Linguistics 2 : 66-89.
BLO CH , BERNARD
1 947
English verb inflection. Language 23 : 399-4 1 8 ; RIL 243-254.
1 948
A set of postulates for phonemic analysis. Language 24 : 3-46.
1950
Studies in colloquial Japanese IV : Phonemics . Language
26 : 86- 1 2 5 ; RIL 329-348.
CHOMSKY, NOAM
1 9 57
1 9 64
1965
CONKLIN,
1962
1 964
HAROLD C.
Lexicographical treatment of folk taxonomies. In F. W.
Householder and S. Saporta ( eds. ) , Problems in Lexicography.
Indiana University Research Center in Anthropology, Folk
lore, and Linguistics, Publication 2 1 .
Ethnogenealogical method. In Ward H. Goodenough (ed. )
Explorations in cultural anthropology : Essays in honor of
George Peter Murdock. New York, McGraw-Hill.
FRAKE, CHARLES 0.
1 96 1
The diagnosis of disease among the Subanun of Mindanao.
American Anthropologist 63 : 1 1 3- 1 32.
GLEASON, H. A., Jr.
1 964
The organization of language : A stratificational view. In
Monograph Series on Languages and Linguistics, No. 1 7,
Report on the 1 5th Annual R. T. M. on Linguistics and
Language Studies, Georgetown University.
GOODENOUGH, WARD H.
Componential analysis and the study of meaning. Language
1956
32 : 195-2 1 6.
HALLE,
1 959
MORRIS
HALLIDAY, M. A . K .
1 96 1
Categories o f the theory o f grammar. Word 1 7 : 24 1-292.
60
Bibliography
HAMMEL,
1 965
61
E. A. ( ed. )
Formal semantic analysis. American Anthropologist 67, No.
5, Pt. 2 .
HARRIS, ZELLIG S.
1 942
1951
1 954
1 957
1 965
HJELMSLEV, LOUIS
1 943
HOCKETT, CHARLES F.
1 942
1 947a
1 947b
1 947c
1 954
1955
1961
LAMB, S YDNEY M .
1961
M T research a t the University of California, Berkeley. In
H. P. Edmundson (ed. ) , Proceedings of the National Sym
posium on Machine Translation, Englewood Cliffs, Prentice
Hall.
Outline of stratificational grammar. Berkeley, ASUC.
1 962a
On the mechanization of syntactic analysis. In 1 96 1 Inter
1 962b
national Conference on Machine Translation of Languages
and Applied Language Analysis, London, Her Majesty's
Stationery Office.
On alternation, transformation, realization, and stratification.
1 964a
In Monograph Series on Languages and Linguistics, No . 1 7 ,
Report o f the 1 5th Annual R . T. M. on Linguistics and Lan
guage Studies, Georgetown University.
62
Bibliography
1 9 64b
In A. Kimball
American Anthropologist
1 9 65b
66,
No.
3,
2 : 57-7 8 .
Journal of
4 : 1 9 6-2 10.
In E. A. Ham
1 966a
67, No.
5, Pt.
2 : 37-64.
Romance Philology
1 9 : 5 3 1 -57 3 .
1 966b
Language
42 : 53 6-
573 .
1 9 66c
1 966d
In press
Stratificational linguistics as
1956
Language
3 2 : 1 5 8- 1 94.
1 964a
Proceedings of
The Hague,
Mouton.
1964b
terminologies.
1 965
In E . A .
67,
No
5,
Pt.
2 : 1 42- 1 8 5 .
NIDA, EUGENE A .
1 948
1. 949
24 : 4 14-44 1 .
Ann Arbor,
Bibliog ra p hy
63
NOREEN, A.
1903- 1 8 Vart Spn'ik (Lund ) . Selections translated into German by
H. W. Pollak, Einfiihrung in die wissenschaftliche Betrachtung
der Sprache. Halle, 1 9 2 3 .
PITTMAN, RI CH ARD S.
1 948
SHAUMJAN, S. K.
1965
1 964
Maidu grammar.
Linguistics 4 1 .
S WADESH, MORRIS
1934
TRAGER, GEORGE L.
1955
French morphology : Verb inflection. Language 3 1 : 5 1 1 -529 .
TRUBETZKOY, N. S.
1939
Grundziige der phonologie. TCLP 7.
TWADDELL, W. FREEMAN
On defining the phoneme. Language Monograph No . 1 6 ;
1935
RIL 55-80.
WALLACE, ANTHONY F. C. , and JOHN ATKINS
1 9 60
American Anthropologist
WELLS , RULON S.
Immediate constituents. Language 23 .8 1 - 1 1 7 ; RIL 1 86-207 .
1 947
Automatic alternation. Language 25 : 99- 1 1 6 .
1 949
EXERC ISES
1 . Maidu
1 . wepa
2 . pano
3 . heme
4. koj6
5 . kyle
' ( the )
' (the )
' ( the )
' (the )
' (the)
6 . pybe
coyote'
grizzly'
gopher'
valley'
woman'
wepa na
pano na
heme na
koj6 na
kyle n a
pybe na
'toward
'toward
'toward
'toward
'toward
'toward
( the ) coyote'
( the ) grizzly'
( the ) gopher'
( the ) valley'
( the ) woman'
( the ) boy'
Exercises
65
TACTIC
PATTERN
lo.phons
----.
d
Exercise
66
Exercises
4. Latin
The following forms are written in a classical phonemic transcription
( i.e. each character represents a combination of one or more hypophonemes ) .
morphemes.
NOMINATIVE
5.
GENITIVE
reks
regis
'king'
duks
dukis
'leader'
arks
arkis
'citadel'
ops
opis
'power'
pleps
plebis
'the people'
urps
urbis
'the city'
Latin
The following forms are written in traditional Latin orthography,
which serves as an accurate classical phonemic representation of these
forms.
Account for
NoM.
GEN.
GLOSS
NoM.
GEN.
GLOSS
fel
fellis
bile
mel
mellis
honey
cor
cordis
heart
lac
lactis
milk
OS
bone
glls
ossis
far
farris
sp . grain
gliris
dormouse
OS
oris
mouth
mos
moris
custom
flos
floris
flower
criis
cruris
leg
mus
muris
mouse
aes
aeris
copper
ma:s
ma:ris
male
cinis
cineris
ashes
pulvis
pulveris
dust
vomis
vomeris
plowshare
tempus
temporis
time
corpus
corporis
body
leporis
charm
telliis
telluris
earth
lepus
pignus
plgnoris
pledge
ver
veris
spring
fiir
fiiris
thief
whip
uber
'ii b eris
udder
verber
verberis
mulier
mulieris
woman
ebur
eboris
ivory
femur
femoris
thigh
robur
roboris
oak
marmor
marmoris
marble
murmur
murmuris
murmur
vultur
vulturis
vulture
fulgur
fulguris
lightning
calcar
calcris
spur
soror
sororis
sister
honor
honoris
honor
amor
amoris
love
Exercises
Additional
material :
pro dTs
you g o forth
prOdis
he goes forth
pro dit
he publishes, reports
vincis
vincis
vi n cit
you conquer
vincit
pro dit
6.
67
he conquers
Natchez
Account for the phonological alternation exhibited by the following
data.
'one to
'several to
__,
th i nk
?ayi ho ?is
?ayho ?is
tu ho ? is
tuhuho ?is
pi Lhal?is
beat
pi lihal?is
plaster
tie
kihelu ?is
ki helu ?is
paYheLsu ?is
payaheLsu ?is
stand on
hemhal?is
hemi hal?is
destroy
polohal?is
polo hal?is
la t he
tolhal?ill
toll hal?i s
twist
toMhal? is
tomohal?is
lend
kawhelku s
kaw i helku s
we Lhaki s
we lehaki s
wanhetahnu ?is
puLchaLsi s
pu l cuha Lsi s
dip
ta l k
quarrel
write
w acho ?is
pi ho ? is
wacaho ?is
scat t er
pi hiho ? is
whip
pataheNci s
pata heNci s
ti Mhaki s
ti mihaki s
tuku hesku s
get drunk
rub
tukuhuhesku s
me hal?is
mehal?is
tisa hekti s
ex t in gu ish
tisahahekti s
sneeze
peyi hohsi s
peyhohsi s
7.
GLOSS
__ ,
store
effective information.
8.
Draw the diag ram of the s implest possible capotactics which will
generate all and only the following combinations of capemes
capeme i s any of the following : A , B , C , D, E , F, G ) :
A
A B C
A B D E
A B D F
A B D G
(a
68
Exercises
Exercise
Exercises
10.
69
Exercise 1 0
11.
A D
A E
A G J
A H J
B C
B
B
B
B
c
C
D
E
G J
H J
F J
D
D F J
E
E F J
I J
1 2 . English
70
Exercises
forms to any set if you find it helpful to do so. What are the appro
pr i ate considerations in arriving at a decision? Give reasons for each
decision.
Set A :
cra nberry
r aspberry
huckleberry
strawberry
ba rb erry
elderberry
Set B :
mushroom
Set
cr a wdad
C:
otter
t i g er
beaver
Set
lobster
b adger
gopher
fisher
D:
patter
clatter
tatter
spatter
shatter
batter
Set E :
gleam
gl are
glitter
Set F :
slip
flip
drip
trip
glow
glance
gloaming
slop
flop
d rop
Set G :
confer confuse
refer refuse
infer infuse
p erfuse
defer
s l ide
glide
ride
stride
sc atter
s platter
ch a tter
glimmer
gleed
gloze
slash
bash
d ash
c ras h
s mash
conform confine
reform refine
i nf o rm
perform
deform define
glis ten
glint
glimpse
s l urp
burp
urp
slime
grime
c o n tain
ret ai n
concei v e con du ce
receive reduce
induce
perceive
deceive deduce
pertain
det ai n
slush
mush
APPENDIX
STRATI FICAT I ONAL ANALYSIS O F A N E N GLISH TEXT
by
Leonard
E.
71
Newell
I . I NTRODUCTIO N
G:
M:
G:
M:
G:
F:
G:
F:
G:
F:
G:
F:
G:
F:
G:
F:
M:
G:
M:
G:
M:
G:
M:
Appendix
74
K:
F:
Hmh?
K:
G:
K:
G:
F:
G:
M:
G:
Down
M:
fire escape.
G:
Mhm.
M:
G:
M:
G:
M:
U' 'u.
G:
F:
G:
a great big tube and you just slide down and you get down faster.
2 . LEXEMIC SYSTEM
The lexemic description includes a sketch of the lexemic tactic pattern
for all the well-formed sentences . Some utterances, considered not to be
well formed, were excluded from the description as, for example, "Today
we had a . . . the fire alarm went."
accounts for the sentence "Today we had a fire drill" and "Today the fire
alarm went. "
2. 1
Analysis of an English
Text
75
M:
G:
M:
G:
F:
G:
F:
G:
F:
G:
F:
G:
F:
G:
F:
M:
G:
M:
G:
M:
G:
M:
K:
F:
K:
76
Appendix
G:
K:
G:
F:
G:
M:
G:
M:
G:
M:
G:
M:
G:
M:
G:
F:
G:
They -re not have any slide -m2 huh? You -re know some -s school
-s have a great big tube -m2 and you -re just slide down and
you -re get down fast-er.
We -re have to walk down.
You -re can -ould not run huh?
No, we -re will -ould be-en hang a t the break -m2 of day -m2.
Sentence:
S I (MajSn,ExclSn,Itj ) Inton
A sentence consists of one of three sentence types plus intonation. Of
these, one is a major sentence type and is the only one of the three patterns
containing a subject and a predicate. This preliminary presentation does
not include a description of the intonation contours. ( Word juncture and
some features of stress and pitch would be accounted for in a more com
plete description of the morphemic system than is given here. ) The form
huh, closely associated with intonation, is not included in this description.
Major Sentence:
MajSn I [Conj] [how-come] Sn
Conj I and,but,or
'Yes but you didn't strike out then. '
Sentence:
Sn I [AttP] [SubClause] [Int] Clause [com Voc]
A ttributive Phrase:
An attributive phrase may optionally occur as a modifier of any one of
three constructions within the lexemic tactic pattern : it may occur as the
Analysis
of
77
InfP l to2 VP
Verb phrases are described below under Verb Phrase.
'To get out of the school we had to run down the steps.'
Locative Phrase:
LocP I outside2,there2
'Outside the boys were playing ball.'
Temporal Phrase:
TempP I then,today,tonight,yesterday,NP
'Today the fire alarm went.'
A noun phrase occurs as a temporal phrase in sentences such as the
following :
'The day before yesterday the fire alarm went.'
Subordinate Clause:
SubClause I (if,
) Clause
Interrogative:
78
Appendix
Pr I PrPers,PrDem,Prlmp
PrPers I I,youhhe,she, ( they,we,you2 ) [all]
'I got to hit it once. '
PrDem I that2,this1
'that2 means that there were three throws'
Primp I everyone,it,something,there1
'everyone was scared'
Noun Phrase:
Text
79
-S
Figure 1
Determiner1 :
Det1 I Det2 =Nmb
Det2 I a,any,some,the,DemDt,IntDet
Nmb I [-s]
80
Appendix
'I jumped
A djective1 :
Adh I great-big, .
'some schools have a great big tube'
.
A djective2 :
Adh I fast, . . .
Noun:
N I (B,N1 ) Nmb
B I ball,base,bell,break,call,class,do,dog,draw,for;>t,get,go,ground,
hang,help,hit,jump,like,make,man,picture,play,ring,run,say,
scare,school,score,see,show ,side,slide,start,step,story,strike,
throw,tube,walk,work
N1 I afternoon,any, baseball,chute,day ,door ,fire-alarm, fire-escape,
first-base,homework,movie,once,running,second
Bases ( B ) occur both as nouns as specified here and as verbs ( see Verb
Phrase ) .
Post-Noun Modifier:
81
Figure 2
thab.
82
Appendix
Pred I
Con
Concord:
Con I
-mh-re,-z1
-m1 indicates concord with a first person singular subject, -z1 with a
third person singular subject and -re with either a first person plural, a
realization of
have-en
Figure
83
sec. 3.2 under the headings A uxiliary and Verb, and sec. 3 . 3 where port
manteaus involving concord elements are discussed.
A uxiliary:
A uxiliary consists of either a past tense lexeme -ed, or can or will with
an optional subjunctive lexeme -ould. The order of occurrence of the
suffix -ed within a word is specified by the morphemic tactic pattern.
Subject Delay:
84
Appendix
Figure 4
85
co-occurred with its first occurrence and which the morphotactics specifies
as an affix to a verb phrase constituent is also repeated, as though the repe
tition were the first occurrence of the verb phrase. Thus, in the following
sentence the lexon -ed is repeated with each successive occurrence of the
verb phrase : 'Everyone was scared and jumped up and started running out
the door'.
Verb Phrase8 :
VP2 I PrepVP,LiVP,AcVP2
Verb Phrase2 indicates verb phrases which may be optionally followed
Pas VP I be-en V
V I B ,have,know,mean ,put,write
Base may occur as a verb or a noun ( see under Noun ) . A verb may
also consist of various lexemes which do not occur as nouns (have ,know,
etc. ) .
'Everyone was scared. '
A ctive Verb Phrase1 :
AcVP1 I V [DirObj]
AcVP1 [ObjDel] I DirObj I Obj
An active verb phrase (AcVPh A c VP2 ) is characterized by an optional
(A cVP1 ) or obligatory (AcVP2 ) occurrence of a direct object.
The second line of an active verb phrase1 occurring with an optional
object delay1 in an upward AND construction is realized as object. This
specifies the optional delay of the direct object until following a preposi
tion! or an indirect object, as described below.
Prepositional A dverb:
86
Appendix
Indirect Object:
Comparative A dverb:
* (Conj ) CompAdv I (fast, . . . ) -er than Obj
Anal ysis
of
87
A ttributive Phrase:
A ttributive phrase is described above, where it occurs as the first con-
stituent of a cl ause.
'I have something to do in that tonight for homework. '
' I was out o n first.'
Purpose A dverb:
PurAdv I to -see-if Clause
'They were going to show the pictures to the other schools to see if they
DegAdv I so-much-that Sn
'It scared me so much that I jumped about six feet off the ground.'
Quotative Verb Phrase:
Voc I NPers
Personal nouns a re described above in the treatment of nominal.
' What did you do at school this afternoon, Gordy?'
Exclamatory Sentence:
88
Appendix
HEADING
VALENCE
Det
NumAdj
Prep
about1
about2
after-noon
all
and
any
at
ball
base
base-ball
be
be-en
be-ing
be-fore
be-go-ing-to
bell
break
but
call
can
-ce
chute
class
day
do
dog
Donello
door
down
draw
else
-ed
-er
every-one
fast
fire-alarm
fire-escape
first-base
foot
for
SEMON
N1
PrPers
Conj
Det,N1
Prep
B
B
N1
VPLi
PasVP
VP
Prep
PreV
B
B
Conj
B,AcVP2
Aux
NumAdv
N1
B
fiNFut
N1
B
B
NPers2
N1
Prep,Prep1
B
RIP2
Aux
CompAdv
Primp
Adj2,CompAdv
N1
N1
N1
B
Prep,Prep2
sNPast
Ana lysis of
HEADING
VALENCE
from
S:\'Fut I will,be-go-ing-to
get
go
Gordy
great-big
ground
hang
Harriet
have
have-en
he
help
hit
hmh
home-work
how-come
I
if
in
it
jump
just
Kathy
know
like
-m1
-m2
make
man
mean
mhm
Mister
movie
m' 'm
no
nope
not
of
off
oh
Prep
an
SEMON
B
B
NPers2
Adh
B
B
NPers2
v
VP , PreV
PrPers
B
B
Itj
Nt
MajSn
PrPers
SubClause
Prep,Prep1
Primp
B
Pred
NPers2
V,QuotV
B
Con
Obj
B
B
V,QuotV
ltj
NPers
Nt
ltj
Excl
ltj
Pred
PostNMod
Prep,Prep1
Excl
sNPast
89
90
Appendix
HEADING
VA LENCE
on
Prep,Prep1
N1
Num,NumAdv
Conj
NP
Aux
Prep,Prep1
Prep
LocP
once
one
or
other
-auld
out
out-side1
out-side2
SNPast I -ed,have-en
picture
play
put
Q
-re
ring
run
run-ing
-s
-'s
say
scare
school
score
second
see
she
show
side
six
slide
so-much-that
some
some-thing
start
step
story
strike
-th
than
that1
that2
B
B
v
SubjDel
Con
B
B
N1
Nmb
PosNom
B,QuotV
B
B
B
N1
B,QuotV
PrPers
B
B
Num,NumAdv
B
DegAdv
Det
Primp
B
B
B
B
NumOrd
CompAdv
DemDet
PrDem
SEMON
HEADING
VA LENCE
thata a
thata b
the
then
there1
there2
they
this1
this2
three
throw
to1
to2
to-day
to-night
to-see-if
tube
Tubman
two
up
u' 'u
walk
we
whatt
what2
which1
which2
who
will
work
write
yea
yester-day
you1
you2
-z1
RelPra
QuotVP
Det
TempP
Primp
LocP
PrPers
PrDem
DemDet
Num,NumAdv
91
SEMON
Prep,Prep2
InfP
TempP
TempP
PurAdv
B
NPers2
Num,NumAdv
Prep,Prep1
ltj
B
PrPers
RelPr1
IntDet
Re1Pr3
IntDet
RelPr2
Aux
B
V,QuotV
SNFut
Excl
TempP
PrPers
PrPers
Con
3. MORP H EMIC SYSTEM
92
Appendix
M:
G:
M:
G:
M:
G:
F:
G:
F:
G:
F:
G:
F:
G:
F:
G:
F:
M:
G:
M:
G:
M:
G:
M:
K:
F:
K:
G:
K:
G:
F:
G:
M:
G:
M:
G:
M:
G:
M:
G:
M:
G:
F:
G:
93
-r
94
Appendix
Pronoun1 is any one of the following : ours theirs, hers, yours or the
possessive pronouns our, their, her and your.
,
Pronoun2 :
Pr2 I I -'s [-ne]
Pronoun2 is either my (portmanteau realization of the two morphemes
I and -'s) or mine.
Object Pronoun:
ObjPr I ( d'e-,hi--, I,she,we,who ) -m2
I, she and we with -m2 have portmanteau realizations me, her and us.
Substantive:
Subst I (Pr.,NPers,Nimp,NDet,NumOrd ) [-'s]
A substantive may be either a simple pronoun, a personal noun, an
impersonal noun or a determiner noun, followed by an optional possessive
suffix.
Simple Pronoun:
Prs I he,I, it, she,that2,they,thist.we,whicht.who,you
Personal Noun:
NPers I Donello,Gordy, Harriet,Kathy,Tubman, . . .
Impersonal Noun:
Nlmp I Nlmpt. NJmp2
Impersonal Noun1 :
Nlmp1 I (B,N.,NPrep ) [-s]
Base:
B I Num,alarm, ball, base, bell, break,call,class,do 2 ,dog,draw,escape,
fire,foot,get,go,ground,hang,help,hit,home,jump,know,like,
make,man,picture,play,put,ring,run,say,scare,school,score,see,
show ,side,slide,start,step,story,strike, throw ,tube,walk, work
Num I one,two, . . .
N. I chute,day,door,Mister,movie,night,noon,other,they
Prepositional Noun:
NPrep I Prep (B, N. ) , before
Prep I at,about,after,down,for,from,in,of,off,on,out,to,up,before
before I be - -fore
Analysis
of
95
Impersonal Noun2 :
Nlmp2 I (foot,man ) [Ad
lmpersonal noun2 represents a class of nouns which are pluralized by an
ablaut morpheme. The phonemic shape of the vowel of an impersonal
noun so pluralized is specified in the phonemic system.
Determiner Nouns:
NDet I Det1 ( body,thing )
Det1 I any,every,no,some
Ordinal Numeral:
h ave
be
Figure 5
96
Appendix
A uxiliary:
Aux I Aux1,Aux2,be -en
A uxiliary1 is characterized by an optional constituent not auxiliary2 by
an adverb constituent just or always.
,
A uxiliary1 :
Aux1 I Auxs not
A uxiliary3 :
Auxa I ( [Emph] Aux4 ) + f1
A uxiliary3 specifies an ordered OR, the first line realized at a lower level
of the morphotactics as various auxiliaries and the second line realized
as zero . This rule requires not to be suffixed to an auxiliary if not occurs
as part of a predicate within the lexemic tactic pattern where auxiliaries
may occur. In other environments, not occurs as a separate word .
A uxiliary4:
Aux4 I Aux5 + doAux
This rule specifies an ordered OR with do auxiliary as the second line.
The ordering involves two features ; the occurrence of auxiliaries with not
and their occurrence in the formation of question sentences. If not is
activated and is specified by the lexotactics as a predicate constituent and
an auxiliary5 morpheme is also activated, not is suffixed to that auxiliary as
specified by the ordering indicated in the above rule.
'You couldn't run, huh?'
However, if an auxiliary5 morpheme is not activated, then the second line
of the ordered OR is chosen and the morphotactic pattern supplies an aux
ilary do1.
'You didn't hit it?'
The formation of an interrogative sentence involves the activation of the
lexeme Q. This lexeme is realized in the lexonic alternation pattern as
follows : LN Q I do1 + f1. do1 is chosen if the morphotactic pattern will
permit its occurrence. A uxiliary4 rule, however, specifies the choice of an
auxiliaryrs morpheme if one occurs, in which case LNQ is realized as zero.
'What's that about?'
If an auxiliary5 morpheme does not occur, Q is realized as do1.
'Did you all go outside?'
do A uxiliary:
doAux I do1 Suf1
Suf1 I Con1oSuf2
Con1 I -m1,-re,-z1
Suf2 I Con2 -ed1
Con2 I -m1o-re, f1 1
-m1 and -re when affixed to do1 are realized as zero as specified by the
morphemic sign pattern ( 3 . 3 . 1 ) . The phonemic shape of the special vowel
Analysis
of
an Eng lish
Text
97
morphon of do with the various suffixes described here (do, does, did ) is
specified in the phonemic system.
A uxiliary5 :
A ux6 I canAux,Auxo
A uxiliaryo consists of any of the inflected forms of be and have1 A ux
iliaryo may be inflected by the third singular concord morpheme -z1 ( has, is )
whereas a can auxiliary ( can, will, etc. ) may only be inflected by the third
singular concord morpheme 0 t .
can A uxiliary :
canAux I (can,will ) Suf8
Sufa I Con2 [-edtJ l
The phonemic shape of can and will with the past morpheme -ed1 is
specified by the phonemic system.
A uxiliaryo :
Auxo I ( have1obe ) [Suft]
For a description of Sufo see above under do A uxiliary.
A uxiliary2 :
Aux2 I (Aux5 + 0 ) just
A uxiliary2 has, as one constituent, an auxiliary6 or a zero in an ordered
OR. This rule requires just to be suffixed to an auxiliary5 (various inflected
forms of can, will, be or have ) if just is activated as part of a predicate
within the lexemic tactic pattern and if an auxiliary5 is also activated. If an
auxiliary5 is not activated, the first constituent of auxiliary2 is realized
as zero.
'I just ran down the steps. '
Verb:
98
Appendix
PresPart I Vs -ing
Adverb:
Adv I Advhall,how,therehthen,whatt.when
Adv1 I ( Qne,twi-, thri- ) -ce
All adverbs occur with contractions of be, will, etc . as single morpho
logical words, e.g., all's 'all is', how'd 'how would', etc.
Adjective:
Adj I AdjComp,Adj.
Comparative Adjective:
Adj. I big,fast,great, . . .
99
Determiner:
Det I Dett;Det2
For a description of determiner1 see above under Determiner Nouns.
Det2 I ( a,thaththis2 ) [-s]
The determiners a, that1 and this2 occur with an optional plural mor
pheme -s. a with plural is realized in the morphemic sign pattern as zero,
and the latter two determiners with plural are realized as portmanteaus
those and these. The determiner lexemes any, the, some, what2 and which2
are specified by the lexotactic pattern as optionally occurring with -s. The
morphotactic pattern, however, does not combine these determiners with
-s; it thus forces the choice of 0 ( second line of an ordered OR ) as the
realization of -s within the lexonic alternation pattern.
Simple Word:
Wa I a, and,but,else,if,mhm,m' 'm,nope,oh,or,so,that8,the, there2, u' 'u,
well,what2,which,yea,yes, . . .
that8 represents the lexemes thataa and thatab
1 00
Appendix
Determiner:
a -s 113
this -s /these
that -s /those
Ordinal Numerals:
one -th /first
two -th /second
three -th /third
3 . 4 Morphicon.
HEADING
VALENCE
a
A1
A2
Aa
about
A- ed
A-en
after
all
Det2
Nlmp2
PastV1 ( ordered AND )
P astPartV 1 ( ordered AND )
Prep
PastV2 (ordered AND )
PastPartV2 ( ordered AND )
Prep
B
Adv
and
any
Det
alarm
at
aw
ball
ba se
be
LEXON
L N. S
LN. ed
LN. en
LN. ed
L N .en
Wa
Prep
PosPr
B
B
Aux,Auxo
LNwe
Analysis
HEADING
of
VALENCE
before
beB
bell
Adj .
big
NDet
body
B,V1,Va
break
w.
but
B
call
canAux
can
Adv1
-ce
N.
chute
B
class
v1ov.
com e
N.
day
PosPr,ObjPr
ledoAux
do 1
B
do2
B
dog
NPers
Donello
door
N.
Prep
down
B,V 1 oV4
draw
L-ed I (A2, A-ed, 0) + -ed1
Suf2 ,Sufs,PastV4
-ed1
PastPartV5 ( ordered AND )
-ed2
w.
else
LN _en I A3,A-en,-en, I 0s ,-ed 2
Aux,PastPartV3
-en
AdjCo mp
-er
B
escape
Det
every
Adj .
fast
B
fire
B,Nimp2
foot
Prep
for
before
-fore
Prep
from
B,V1oVa
get
B ,Vs
go
NPers
Gordy
Adj.
great
1 01
LEXON
LN_ ed
L-en
LN_ en
1 02
Appendix
HEADING
VALENCE
ground
hang
Harriet
have1
have2
LN he I hi,he
he
help
B
B,V1> V2
NPers
h iht-
ObjPr
PosPr
B,PastV3 ,PastPartV4
B
Adv
Pr2,0bjPr,Pr8
if
w.
hit
home
how
in
-ing
it
jump
just
Kathy
know
like
-m 1
LN-m2 I m2 + 0
-m2
make
m an
mean
mhm
Mister
movie
m' 'm
-ne
night
no
noon
nope
not
of
off
oh
LEXON
Auxo
v.
Pr.
B
LN he
LN he
LNshe
Prep
PresPart (ordered AND )
P r.
B
Aux2
NPers
B ,V1o V4
B
Con1oCon2 (ordered AND )
ObjPr
B
B,Nimp2
V.,PastV2
w.
N.
N.
w.
Pr2
N.
Det
Na
We
Au x 1
Prep
Prep
w.
LN_m2
HEADING
VALENCE
on
one
or
other
out
pi ctu re
play
Prep
Num, Adv1
Ws
-r
d o1
+ JJ
-re
ring
run
LN I A1.-s
+ ftJ
-s
LN_ s I -r,-'s
- 's
say
scare
school
s c o re
see
LNshe I hi,she
she
show
s ide
six
s lide
so
some
- st
start
step
story
strike
-th
that1
that2
thata
the
then
there1
LEXON
Ns
put
LNQ I
1 03
Prep
B
B
B,PastVa,PastPartV4
PosPr
Con1.Con2 ( ordered AND )
B,V1 .V2
B ,V 1.PastPartV4
LN
Nlmp1,Det2
LN_g
Pr2,Subst
B,P astV2
L N_ '
B
B
B,VhV4
ObjPr,Pr8
B,V4
B
Num
B ,V1>V2
Wa
Det
AdjComp
B
B
B
B ,V1.V2
NumOrd
D et2
Pra
w.
w.
Adv2
Adv
L N sh
1 04
Appendix
HEADING
VALENCE
there2
LNthey
w.
cfe-,they
they
Pr.
thing
N.,NDet
this 1
Pr.
this2
L Nth ree
Det2
LN they
thri-,three
three
Nu m
thri-
Adv 1
LNthre
e
LN t hree
B,V1oV4
throw
to
Prep
tube
Tubman
twiLNtwo
NPers
Adv1
LN two
Num
LN two
twi-,two
two
up
Prep
u' 'u
walk
LNwe
LEXON
w.
aw-,we
we
well
w.
what1
Adv
what2
w.
when
Adv
which 1
which2
Pr.
w.
ObjPr,Pr.
who
will
canAux
work
v.,vbva
w.
w.
write
yea
yes
yi-LNyou
LNwe
ObjPr,Pr.
PosPr
LNyou
you
Pr.
LNyou
Z1
-z2
Con1 ( ordered
yf;you
LN.zl I 01oz1
01
02
0a
AND)
PosPr
Con2
AND)
( ordered AND)
PastY4 (ordered
PastPartV4
LN. z l
LN.zl
LN. ed
LN. en
Ac
Adj
Adv
Att
Aux
B
com
Comp
Con
Conj
D, Det
Deg
Del
Dem
Dir
Emph
Excl
Imp
Ind
Inf
Int
Itj
Li
Loc
Maj
Mod
N
Nom
Nuc
Num
Nmb
Obj
Ord
p
Part
Pas
Past
Pers
Pos
Post
Pr
PreV
active
adjective
adverb
attributive
auxiliary
base
comma
comparative
concord
conjunction
determiner
degree
delay
demonstrative
direct
emphasis
exclamatory
impersonal
indirect
infinitive
interrogative
interjection
linking
locative
major
modifier
noun
nominal
nucleus
numeral
number
object
ordinal
phrase
participle
passive
past
personal
possessive
post
pronoun
preverb
1 05
Appendix
1 06
Pres
Pred
Prep
Pur
Quot
Rei
RIP
present
predicate
preposition
simple
purpose
quotative
relative
relative interrogative pronoun
sentence
Sn
Sub
Subj
Subst
Suf
subordinate
subject
su bs tantive
suffix
temporal
verb
vocative
word
Temp
Voc
w
Footnotes :
1
analyses, obtained from his writings, published and mimeographed materi als, class
room lectures and from oral discussions with him regarding particular problems as
they have arisen . His help has been so extensive that only a general acknowledgement
is possible here.
3A
Cfe-
quotative verb phrase, however, has a clause as one of its constituents which
may contain one of these modifiers : 'He said that the fire alarm went today at
school'.
4
The Iexemes -ould and -ed are neutralized and are realized as the Jexon -ed.
I ndex
1 07
I NDEX
actualized chains 34
affixation 37
algebraic notation 8-9, 1 1 , 1 4, 17, 26,
45, 55-56
1 9 , 3 3 , 34, 3 8
allomorph
alternation 2 1 , 3 1 , 4 6
alternation pattern 1 1 , 1 4 - 1 7 , 1 8-20, 2 1 ,
55, 5 8
anataxis 2 2
interstratal 22
intrastratal 22-23
anticipatory reduplication 2 6
apical
1 1 , 57
assimilation 3 6
Atkins
31
Barker 3 4
basic forms
37-3 8
Biligiri 3 4
biuniqueness 28, 52
B loch 2 8 , 3 3
Bloomfieldian linguistics
bottom line 4 1 -43
brackets
56-57
28
chains
actualized 3 4
ideal 3 4
Chomsky 3 4, 3 6, 3 9-40
clause construction 25-26
closed
1 1 , 57
complementary distribution 2 8 , 52
componential ana lysis 3 1
1 1, 17, 3 1
composite realization
concord 28
Conklin 3 1 , 32
connective information 49-50
constructions
14
content 3 4
coordination element 26, 3 0
decoding
10, 55
deep structure 3 4 - 3 5 , 40
determined feature
18
d iachronic linguistics 3 6 , 3 8
dictionary o f the stratal system
59
dissimilation 3 6
d istinctive features 40
distinctiveness 28
diversification
17, 5 5-56, 58 -59
dors al 57
downward 9
impulse 4 1 -43
output 4 1 -43
set 43
1 9 , 58-
economy of description 3 8, 40
effective information 4 1 , 43 -47, 52, 5455
empty realization
17, 33
encoding 2 5 , 5 5
English
1 , 5 , 6, 1 1 , 1 4 , 1 7 , 2 2 , 2 3 , 28,
29, 3 1 , 3 2 , 3 8 , 54, 5 6
equivalence 4 5
expression 3 4
external line 5 0
Frake 3 2
French 5, 6 , 1 1
frontal 57
3, 45, 46, 52
generalization
Gleason 3 5
glossematics 2 , 3 4
glottal 57
Goodenough 3 1
grammar
1 , 20, 2 1 , 29
grammatical process 3 6
graphic notation 8 , 5 5
grouping
horizontal 5 , 1 7
vertical 5-6, 1 7
Halle 3 6
Halliday 2 1
Hammel
1 1, 3 1
Harris 29, 3 2 , 3 4
Hebrew 3 6
high 57
Hjelmslev 2, 29, 34, 4 8
Hockett 2, 2 8 , 3 3 , 3 4 , 3 6
homophonous morphemes
horizontal groupin g 5 , 1 7
horizontal splitting 5, 1 7
hypersememes
56
hypersememic
signs 56
system
1 8 , 20
hypersemon 3 2 , 5 6
hypophonemes 57
hypophonemi c
signs 57
stratum 2 8
system
1 8- 1 9
hypophonotactics 28
18
ideal chains
34
idioms 34
5, 3 2 , 52, 54
immediate constituents
impulse
10-1 1 , 4 1
Indo-European languages 27
infixation .3 6
1 08
I ndex
internal change 36
internal line 49-50
interrogative lexeme 25-26
interstratal anataxis 22
i nterstratal discrepancy
17,
59
intrastratal anataxis 22-23
2 1 -22,
26,
kinship terminologies 1 1 , 3 1 - 3 2
knot pattern
1 1 , 1 5-2 1 , 5 8 , 5 9
l abial 1 1 , 5 7
17, 1 8 , 19, 28, 29, 32, 3 4, 35, 3 6,
Lamb
40, 52, 56
lexeme 28, 29, 56, 58
lexemic
signs 56, 58
stratum 29, 3 1 , 35
system
1 8, 20, 5 6
lexicon
19, 58
lexon
1 6, 29, 3 3 , 5 6 , 5 8
lexotactics 2 1 , 2 3 , 2 9
line-activation 4 1 -43
lines 8 , 1 1 , 46-47, 55
internal 49-50
external
50
linguistic graphs
8, 1 6, 4 1 -45, 5 1 -52, 5 5
linguistic structure
1-4, 6 , 8, 1 9-2 1 , 29,
3 2 , 3 5- 3 7 , 40, 4 1 , 48-49, 5 1 , 56-58
link phonemes 33
Lounsbury 3 1 , 3 2
low 5 7
manifestation o f hypophonemic units
1 9, 29
metathesis 22
mid 57
onachi
1 1 , 1 4, 26, 3 7
monostratal v iew 3 2
1 6, 2 1 , 29, 32-34, 57, 58
morpheme
morphemic
signs
1 6 , 52, 57, 5 8
stratum 2 9 , 3 5
system
1 4- 1 7 , 20, 56, 58
morphicon
1 9 , 58
morphological process 3 6-37
morphology 2 1 , 29, 3 7
morphon
1 6, 29, 57, 58
morphophoneme 29, 3 3-34
morphophonemic
alternation 2 8 , 3 5 , 3 8
process 3 6, 3 8
stratum 3 4, 3 5
morphophonemics 29, 3 6
morphotactics 1 7 , 2 1 -2 3 , 26, 2 8 , 2 9' 3 1 ,
36
multiple realization 26
nasal
1 1. 57
nection 50-5 2, 5 5 , 59
neo-Bloomfieldian linguistics 28 ' 3 8
neutralization
1 7- 1 8, 2 1 , 26, 3 3 , 34, 5 9
neutralized realization 6
Newell 8
Nida
5, 22, 3 3 , 3 7 , 3 8 , 52, 54
nodes 8-9, 1 1 , 46-47
non-distinctive feature
18
Noreen 7
notational system 2, 8
ordering
1 0, 3 9-40
Panini 3 8
pattern
11
pharyngeal 57
phoneme
basic 28-29, 57-58
classical 2 8 , 32-34, 3 8
phonemic
signs
57, 58
stratum
28, 3 5
system
1 8, 20, 28, 5 6
phoneteme
18
phonetic system
18
phonological components 40
phonological systems 2
phonology
1 , 20, 2 8
phonon 5 7
phonotactics 22
Pittman 37
plural side of a node 49-50
portm anteau realization
1 1 , 17, 3 1 - 3 3 ,
48, 59
postdorsal 57
prefixation 3 6-37
pre-stratificational theory 3 2ff.
procedures of analysis 6-7
process
approach 3 4 , 3 6 , 40
description 3 5-40
production process
10
proto-stratificational 2
quasi-stratificational
3 2 , 3 8, 40
ranks
21
realization
1 9 , 2 1 -22, 26, 3 1 , 3 3 , 56-59
rules 39, 56, 5 8
realizationa l
analysis 4 - 5 , 1 7
description 5 8
portion
1 7, 1 9 , 2 1 , 5 8-59
reduplication 26-27
anticipatory 26
repetitional 2 6
repetition 3 , 5 , 6, 45, 5 1 , 55
repetitional reduplication 2 6
Index
replacement 3 5-3 6, 3 9
retracted 5 7
rewrite rules
3, 3 8-40
roots 3 6-37
Russian 39
Sanskrit 3 8
semanteme
18
semantic system
18
sememe 32, 56
sememic
signs 3 1 -32
stratum 3 1 , 35
system
1 8, 20, 3 1 -32, 56, 58
semological systems 2
semology
1 , 20, 3 1
semomorphemic stratum 3 5
semon 3 1 , 56
semotactics
32
set theory 3 3
Shaumyan 3 4
Shipley 3 4
sign pattern
1 1 , 1 5-2 1 , 52, 58
simplicity 3 , 5, 1 7 , 46, 52
singular side of a node 49-50
spirantal
1 1 , 57
splitting
horizontal 5, 1 7
vertical 5 , 1 7
stratal systems
1-2, 5, 8, 1 4 , 1 6- 1 9 , 2 1 ,
3 1 , 3 2, 3 5 , 3 6 , 48-49, 54, 56, 5 8-59
stratificational
framework 40
theory
1 -2, 28, 32, 3 5-3 9
stratum
1 -2, 6, 1 6, 2 1
structural linguistics 2, 3 5
suffixation 3 6-37
superficial information 54-55
suppletion 3 6
5, 4 1 , 43 , 45-49,
surface information
52-55
measure of 46-5 1
surface structure 34-3 5
Swadesh 28
1 09
syllable
14, 2 8 , 29
symbols 1 1 , 5 6-59
syntax 2 1 , 29, 3 7
tabular notation 8 , 17, 56
tactic
analysis 4-5
construction
17
description 54, 58
pattern
1 1 , 1 4- 1 6, 1 8 , 2 1 -2 3 , 32, 5 1 5 3 , 58, 59
system 5
taxonomic
hierarchies 3 2
structure 3 1 -32
terminals 50-52, 5 5
top line 4 1 -43
Trager 34
transformational analysis 3 4
transformations
34, 3 7 , 4 0
tree diagrams
32
Trubetzkoy 3 9
Twaddell 2 8
unvoiced 57
upward 9
impulse 4 1 -43
output 4 1 -43
set 43
valence 59
vertical grouping 5-6, 1 7
vertical splitting 5, 1 7
vocalic 57
Wallace 3 1
Wells 5 , 3 3 , 52, 54
word
1 4 , 2 1 , 28, 29
zero element
1 4, 43 , 50
zero realization
17, 32, 3 3 , 3 6
Zoque 22