A Lexical Phrase Grammar for ESL

Author(s): James R. Nattinger

Source: TESOL Quarterly , Sep., 1980, Vol. 14, No. 3 (Sep., 1980), pp. 337-344
Published by: Teachers of English to Speakers of Other Languages, Inc. (TESOL)

Stable URL: https://www.jstor.org/stable/3586598

REFERENCES
Linked references are available on JSTOR for this article:
https://www.jstor.org/stable/3586598?seq=1&cid=pdf-
reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.

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

is collaborating with JSTOR to digitize, preserve and extend access to TESOL Quarterly

This content downloaded from

61.92.158.55 on Sun, 28 Apr 2024 09:09:00 +00:00
All use subject to https://about.jstor.org/terms
 TESOL QUARTERLY
Vol. XIV, No. 3
September, 1980

A Lexical Phrase Grammar for ESL*

James R. Nattinger

Current work in lexical grammars suggests that we need to think of

lexicon for ESL classes in much broader terms and must pay more attention
to the importance of prefabricated speech routines in language behavior.
This article discusses the kind and extent of lexical phrases, using categories
from recent work in artificial intelligence, and outlines further research that
is necessary to make these speech routines an integral part of the syllabus.

Many linguists currently are developing models for language processin

or what are perhaps better known as performance models. These offer so
direct descriptions of psychological categories and processes, and attempt
describe languages in terms of how they are perceived, stored, remember
and produced. (Bever 1976; Bresnan 1978; Chafe 1976; Lakoff 1977.) Em-
phasizing language processes rather than language structures, performan
models are more realistic accounts of language behavior than former com
petence theories. And what goes on during performance must surely be t
to any theory of language learning. Many of these new models give more im-
portance to the lexicon as a separate component of the grammar than they do
to the syntax by greatly reducing the role of transformational rules and en-
larging the role of the lexical component. Some models, such as Richard Hud-
son's panlexical grammar, go so far as to eliminate the distinction betwe
syntax and lexicon altogether, and describe stretches of language as generated
from rules which are lodged only in the lexical units themselves (Hudson 1979
This way of looking at grammar suggests that we need to think of lexico
in much broader terms and that we need to pay more attention to the im
portance of prefabricated speech routines in language behavior. These routine
when treated at all in ESL texts, have been much more narrowly defined than
they are in current theory and are thus more incidental in our syllabuses tha
perhaps they ought to be. To suggest the potential of the lexicon, let me begi
with what can be described as idioms, or cliches, or as other sorts of fix
phrases, manipulated as a kind of oral formula. Idioms, for example, are usual
defined as complex bits of frozen syntax whose meanings cannot be deriv
from the meaning of their constituents, that is, whose meanings are more th
simply the sum of their individual parts: step on the gas, raining cats and do

* A version of this paper was presented to the 1979 ORTESOL conference in Corvalli
Oregon, October, 1979.
Mr. Nattinger is a professor of English and Linguistics at Portland State University
Oregon. He has published in Language Learning, TESOL Quarterly, and English Teachin
Forum.

337

This content downloaded from

61.92.158.55 on Sun, 28 Apr 2024 09:09:00 +00:00
All use subject to https://about.jstor.org/terms
 338 TESOL Quarterly

kick the bucket. Thus, these phrases ar

and are treated as such in most ESL texts.
Cliches are similar to idioms in that
relatively frozen; they are unlike th
of larger stretches of language and t
individual constituents: there's no do
Non-canonical forms are other sorts
both of the above in that their pattern
the more he likes it; waste not, want n
This last, of with his head, for exam
which can take many other forms-d
away with all bureaucrats, and so on
ADV+with+NP, into which different
with only the preposition remaining
one that defines them as non-canonical
have the typical shapes of English sent
tically from the usual pattern; where,
These forms appear to be halfway betw
they cannot be stored as invariable uni
be derived by the traditional rules of s
English sentence patterns. It is facts li
to question the sharp distinction bet
for a framework that could describe all of these data in the same manner.
Thus these fixed patterns are characterized in terms of how variable they
are and how possible it is to derive their meaning and structure from the
normal rules of syntax. But there is a third characteristic involved which
closely related to the above and which may be even more important for
language processing-the storage and recall of these structures as pattern
instead of as isolated units that have to be synthesized each time they ar
used. For quite a while many linguists have suggested that idioms, cliches, and
non-canonical forms are stored as patterns, Corder's "holophrases," for example
(Corder 1973:131), and have suggested that there is a difference in kind be
tween this sort of patterned speech and more creative speech. Some linguist
have further differentiated them by sharply distinguishing, as do Krashen and
Scarcella, among "completely fixed routines," "semi-fixed patterns" and "creative
speech," (Krashen and Scarcella 1978). But it is more likely that what con
stitutes a pattern and what does not is relative, a matter of degree instead of
kind, for one usually finds a continuum in the amount of variation involved,
from more invariable and frozen forms (such as idioms and cliches) to less
invariable (non-canonical) forms; this is one of the assumptions upon whic
the notion of lexicon can be expanded, for this view allows many other stretches
of speech to be described as patterns than could have been previously.
Whatever the description of pattern, it is generally agreed that the sequence
of words in phrases with less variation is more predictable, an extremely im-

This content downloaded from

61.92.158.55 on Sun, 28 Apr 2024 09:09:00 +00:00
All use subject to https://about.jstor.org/terms
 Lexical Grammer 339

portant fact in communication whi

language. It is a fact that Oiler exp
and Richards 1973; Oller and Stre
strain those around them, and th
going to follow certain others, are
and to create all sorts of subtle v
time was had by all, is a relatively
all, a glorious time was had by all,
variations on this basic pattern a
effect comes about because we ex
of certainty, in the ADJ or in the
since their meaning exists only a
these units cannot be manipulated
kick the bedpan but it would be d
ways and still have it make much se
With these criteria, we can speak
are different from the above three
in shape and which are also variab
whose meaning can be derived by t
I would show you; your health is at
please close the door. How then are
the language? The claim here is th
easily stored and recalled as patte
kind between them. They are only
some phrases are best described a
dictably, and some are best descr
arrangements are more unpredict
in this way these patterned speec
language than most previous lingu
large a part they have played in
boundary between habit and rule,
has always been most unclear.
The extent of patterned speech
some categories from current work
pear in the work of Joseph Becker
phrases in terms of their functiona
1975).
a) Polywords: These are short phrases with extremely low variability
whose meaning exists apart from syntax. They often are substitutes for single
words and are thus almost the same as other vocabulary items. They function
as euphemisms, slang, two-and three-part verbs, and idioms: the powder room,
my old man, to call up, for good. As suggested previously, some of these can
be varied, but only with difficulty: kick the bedpan, step on the oil.

This content downloaded from

61.92.158.55 on Sun, 28 Apr 2024 09:09:00 +00:00
All use subject to https://about.jstor.org/terms
340 TESOL Quarterly

b) Phrasal Constraints: These consi

of which constrain the variability of o
effect to show emphasis, irony, sarcas
forms: by pure coincidence, down wit
for example, the words can be varied n
a week ago, but also to show emphasis,
ago. or metaphor, a grief ago, among othe
c) Deictic Locutions: These are shor
ability which can serve as clauses or w
the flow of the conversations by mark
challenges, defenses, supports, retreat
were you; for that matter; frankly; I
d) Sentence Builders: Phrases up t
slots for parameters or arguments; th
is to provide a skeleton for the expr
song and dance about C; Not only A
ditional tense B. Often non-canonical f
tence-like ones, for they are not tradi
elements that vary freely: the faster
would only stop talking for a minute.
e) Situational Utterances: These ar
be highly variable. They are the app
stances, and may be used out of those c
as does much phatic speech. The cont
tions that have found their way in
Restaurant, Meeting Friends, etc.) th
in which language takes place: cold eno
what time is it; it only hurts when I l
f) Verbatim Texts: Memorized texts
nature have extremely low variabilit
and occasionally direct usage (like th
by piecing together chunks of memo
donata/etc."): better late than never; I
every pore, He oiled his way around th
As is apparent, all of these amount
phrases of infrequent occurrence, and
speech. What does it all imply? Becker
It implies to me that the process of spe
the information we wish to express or evok
some patterns that can provide the major el
lem is to stitch these phrases together into
in the blanks with the particulars of the
need be, and if all else fails to generate p
transitions and fill in any remaining conc
Becker is not saying that this is the

This content downloaded from

61.92.158.55 on Sun, 28 Apr 2024 09:09:00 +00:00
All use subject to https://about.jstor.org/terms
 Lexical Grammer 341

research does suggest that crea

patterns. L. Fillmore's work on
quisition, for example, suggests
semi-fixed patterns which in turn
1976). What Becker means, I th
this way after a while. If so, the
developed this strategy in the lan
If all of this is near the mark,
a way we process language, the
teacher. Perhaps we should base
great deal of the time anyway, la
the ready-made units appropria
hension relies on knowing which
Our teaching therefore would cen
pieced together, along with the
they occur. We would show that
the money I would leave for go
locution (as a matter of fact)--t
terminology-the other a polyw
a larger sentence builder (If I p
would show how these individu
they can be differently combined
same functional groups can subst
dents recognize recurring functio
tion to make predictions about m
It is fairly obvious how a view
one. For one thing, sentences in
dividual constituents, while phras
its sentences from separate bits
together. How much was put toge
guage used and the situation in
of the language, such as in forma
would be put together from scrat
It is not so obvious though how
pattern practice scheme. After al
sists of lists of similar patterns,
by groups of vocabulary items th
is typical: (Wishon and Burks (1

this is the house Mary was born

I have seen the village where the murder
Have you ever been to the place your grandm
the poet stayed

The overriding concern here is with structure; the boxe

any linear combination of constituents will produce a

This content downloaded from

61.92.158.55 on Sun, 28 Apr 2024 09:09:00 +00:00
All use subject to https://about.jstor.org/terms
 342 TESOL Quarterly

Semantics is useful only to insure th

anomolous, and functional considerat
functional analysis of "This is the h
you ever been to the place where th
very different sentences indeed. For p
would use semantic and functional crit
and would require that only phrases
classed together. The following, for ex
several semantic and pragmatic func
time is it; would you mind telling me
what time it is; and so on. Obviously
quite different structures: some will b
tences with embedded clauses; some wi
no part substitutable for another, whi
tain sentence builders or situational
unlike a structural one.
As for actual content and organization of a syllabus, the lexical model at
present offers only broad guidelines but does suggest some specific and interest-
ing research possibilities. Generally, of course, since patterned phrases are more
functionally than structurally defined, so also should be the syllabus. Thus, we
would take the desired communicative ability as a starting point, for what
people want to do with language is more important than mastery of that lan-
guage as an unapplied system, as Wilkins suggests (Wilkins 1976). In that
way, the items we select to teach would not be chosen on the basis of grammar
but on the basis of their usefulness and relevance to the learners' purpose in
learning. The more specialized the course, the more definite the communicative
function that we will have to teach. In a very limited course perhaps all there
would be time for would be some forms of essential functions the learner re-
quires, like asking directions, ordering a meal, reading specialized texts, etc.,
particularly those patterns described above as situational utterances (type (e)).
There are certain general communicative functions, though, that would be
required for almost all courses, like greetings, sympathy, gratitude, judgments
and evaluations, and concessions. Each of these would have associated with it
a set of common lexical phrases of the sorts I have been discussing, particularly
phrasal constraints, deictic locutions, sentence builders, and situational utter-
ances (groups (b) through (e)).
But as suggested earlier, we need the results of further research before
we can be more specific about the role of patterned speech in ESL classes.
First of all we need to know much more about patterned speech in two separate
areas which are often lumped together-in language learning and in language
use; that is, in what ways patterned speech is used while learning a second
language, and in what ways it is used after that second language is learned.
These are two quite different aspects of language behavior and it is likely that
the role of patterned speech in each will also be different. Another aspect, pat-

This content downloaded from

61.92.158.55 on Sun, 28 Apr 2024 09:09:00 +00:00
All use subject to https://about.jstor.org/terms
 Lexical Grammer 343

terned speech in first language ac

generated much research. As a s
differing role of patterns in the
for it is quite possible that a spe
of different prefabricated units. Oft
speaker and hearer is blurred, yet
differences in these behaviors. Ps
duction is highly variable but th
unlikely, then, that the strategie
ferent from the strategies for la
A third task is to investigate vari
and kind of patterned speech that
in very limited speech contexts s
in pieces of contexts, such as gree
gambits, but as I have suggested a
to be so restricted. A further task
differences in use of these patter
make use of different amounts of
ways. Finally, we should see if
notional syllabuses more practic
(Wilkins 1976) have been attack
mediate use for teachers. Kelly typ
work needs to be done before [W
(Kelly 1978). Perhaps one way to
in terms of patterned speech to d
routinized speech and which are m
categories of communicative funct
speech, as do his categories of mo
his semantico-grammatical categ
and would surely be taught differ

REFERENCES

Becker, J. D. 1975. The phrasal lexicon. Theoretical issues in natural

cessing 1. B. Nash-Webber and R. Schank (Eds.). Cambridge, Mas
anek and Newman, 70-73.
Bever, T. G. 1970. The cognitive basis for linguistic structures. Cogn
development of language. J. R. Hayes (Ed.). New York: Wiley, 27
1976. The influence of speech performance on linguistic structure
grated theory of linguistic ability. Thomas G. Bever, Jerrold J
Terence Langendoen (Eds.). New York: Thomas Y. Crowell, 65-88
Bresnan, J. 1978. A realistic transformational grammar. In M. Halle, et al
theory and psychological reality. Cambridge, Mass.: MIT Press, 1-5
Chafe, W. L. 1976. Givenness, contrastiveness, definiteness, subject
points of view. Subject and topic. C. Li (Ed.). New York: Acad
25-55.
Corder, S. P. 1973. Introducing applied linguistics. London: Penguin Books.
Fillmore, L. 1976. Cognitive and social strategies in language acquisition. Stanford
University: Ph.D. dissertation.

This content downloaded from

61.92.158.55 on Sun, 28 Apr 2024 09:09:00 +00:00
All use subject to https://about.jstor.org/terms
344 TESOL Quarterly

Forster, K. I. 1976. Accessing the mental

anisms. R. J. Wales and E. Walker (Ed
lishing, 257-287.
Hudson, R. 1979. Working Papers 1, 2 a
College, London, England.
Kelly, M. 1978. Review of D. A. Wilkins
of Applied Linguistics 16, 69-70.
Krashen, S. and R. Scarcella. 1978. On ro
and performance. Language Learning
Lakoff, G. 1977. Linguistic gestalts. CL
Oller, J. W., Jr. and J. C. Richards (Ed
perspectives for the language teacher
Oller, J. W., Jr. and V. Streiff. 1975. Dict
Testing language proficiency. Randal
Arlington, Virginia: Center for Applied
Wilkins, D. A. 1976. Notional syllabuses.
Wishon, G. E. and J. M. Burks. 1968. L
Book Company. As cited in W. E. Ruth
A Survey of Applied Linguistics. D. B
Arbor: University of Michigan Press, 13

This content downloaded from

61.92.158.55 on Sun, 28 Apr 2024 09:09:00 +00:00
All use subject to https://about.jstor.org/terms