Langage, sujet, lien social || Complete Utterances

Complete UtterancesAuthor(s): Paul RastallSource: La Linguistique, Vol. 30, Fasc. 2, Langage, sujet, lien social (1994), pp. 81-91Published by: Presses Universitaires de FranceStable URL: http://www.jstor.org/stable/30249054 .

Accessed: 14/06/2014 04:39

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

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

.

Presses Universitaires de France is collaborating with JSTOR to digitize, preserve and extend access to LaLinguistique.

http://www.jstor.org

This content downloaded from 188.72.96.102 on Sat, 14 Jun 2014 04:39:09 AMAll use subject to JSTOR Terms and Conditions

http://www.jstor.org/action/showPublisher?publisherCode=puf

http://www.jstor.org/stable/30249054?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp


COMPLETE UTTERANCES

par Paul RASTALL

University of Portsmouth, U.K.

Traditional grammarians and philosophers have often included in their definitions of the 'sentence' some reference to the expression of a 'complete thought'. Other components of the traditional definition have been, of course, the subject and the predicate. Functionalist linguists and others have avoided such definitions out of a reluctance to be committed to any psychological or phi- losophical presuppositions about the nature or existence of 'thoughts', complete or otherwise. It has been considered prefe- rable to concentrate on the relation between phonetic events and the messages associated with them. Sandor Hervey has recently pointed out that the 'constellation' of linguistic data modelled

by sentences must also include pragmatic features'. Functiona- list linguists have been particularly concerned to draw attention to the variety of grammatical functions and have rightly been sceptical about the supposed universality of the subject-predicate relation.

Any serious linguist would wish to escape a vicious circularity in which sentences supposedly expressed 'complete thoughts' which were themselves recognisable by virtue of being expressed by 'sentences'. Furthermore, even at a very simple level one can observe that more than one 'thought' seems to be expressed by a single sentence (most obviously by means of conjunctions such as and or but and, more subtly, that a single train of thought may require more than one sentence for its expression. In: - My wife is very nervous. She doesn't want to drive

there is an obvious logical connection and continuity of thought between the first and second sentences. We can note, however,

1. See S. Hervey, "Sentences and Linguistic Data", La Linguistique, 1991 / 1.

La Linguistique, Vol. 30, fasc. 2/1994



82 Paul Rastall

that the 'suppressed premise', nervous people are loath to drive, is

plainly present as a 'thought' albeit an unexpressed one. A similar situation is found in,

- I thought you were going to clean up the kitchen. There's still a mess in there

where there is a clear implication, - Obviously, you have not kept your promise

which exists as a thought somewhere associated with the two sentences but unexpressed2

Functionalist linguists scarcely need to be reminded that the other component of the traditional definition, the subject- predicate relation, supposedly present in all 'sentences' would exclude numerous utterances in many, if not all, languages which should not be excluded from any definition of the sentence. The alternative is to force the facts into a preconceived mould after the fashion of transformationalist approaches. Examples include the well known.

--The bigger, the better. - Down with the government! - Hier wird getanzt. -U m'en'a masina (Russ. lit. 'with me-car', 'I have a car') - Emu pov'ezlo (Russ. lit. 'to him was lucky', 'he was lucky') --N'ebo zavoloklo tucami (Russ. lit. 'covered the sky with clouds', 'The sky clouded over').

One may remember also utterances containing 'actualisers' or ergative expressions discussed by Andre Martinet in this

regard 3. We can note, furthermore, the commonly encountered ellip-

tical expressions in speech, such as that oficer over there or outside in response to questions such as MWho is in charge here? and Where's the cat?

In such cases, the answers supply only the requested information. The utterances show every sign of being 'complete' without their being 'sentences' in the traditional sense. Without presup- posing an ellipsed structure for such utterances, it is plain that

2. This was pointed out to me by my former colleague at Sunderland University, Dr. Arnold Spector. 3. See A. Martinet, Studies in Functional Syntax.



Complete utterances 83

their interpretation is context-dependent and that they relate to stored, previously mentioned structures. One must assume that it is an application of the law of minimum effort that stored information is not repeated when new information is introduced in relation to what is stored.

However little we may be inclined to deal in 'complete thoughts' or in the traditional view of the grammatical structure of 'sentences', we must recognise that there is an important sense of 'completeness' which we require. The utterances we have mentioned so far have been in some sense 'whole' or 'complete'. Furthermore, while we are in no way committed to an analysis of all sentences into subject and predicate, we can see that the traditional view at least points to the element of 'cross-reference' in complete utterances. That is, the completeness of an utterance involves the expression of more than one way looking at the same experience.

If it is not easy to give a positive identification of the factors which make an utterance complete, it is easy to see when one is not complete. In:

- Her over there is out of action. - He shouldn't go around saying. - My jewels have stolen.

there are missing elements conveying grammatical and lexical information. Part of the explanation of completeness is then a lexical or grammatical one. The particular lexical or grammatical information required depends on the utterance in question.

Another part of completeness is intonation. Intonation patterns may be "final" or "non- final" ("suspensive"). A suspensive intonation pattern (in English expressed by a maintained pitch level) indicates that the speaker will continue (or sometimes expects the interlocutor to continue) the utterance. The suspensive intonation is found for example in lists (with a final, rise-fall, pattern on the final element of the list),

-A book, a pen, a ruler, ..., and a notepad.

- I've cleaned the car, done the washing, cooked the

dinner, ..., and fed the cat.



84 Paul Rastall

It is also found in other clauses or expressions where a

completing clause or expression is intended,

--I bought that book ... because I liked it.

--Have you seen the cat ... with the black tail?

The completing expression will contain a final intonation

pattern (with a rise and fall as indicated). As far as we know, final (and suspensive) intonation patterns

may be found with signs of varying grammatical status-simple signs, complex signs, grammatically incomplete complexes, cons-

tructionally unrelated signs and complete predicative syntagms to form perfectly acceptable utterances which appear to be "com-

plete ". Thus, we find,

-John (As an assertion in answer to, for example, Who toldyou?).

- Our new fridge (In response to, What's that?).

- I bought a car and John a bike (where the expression and John a bike contains grammatically unrelated elements).

--I bought that book (With final intonation by comparison with the same syntagm above with a suspensive intonation).

It is because of the possibility of combining complete intonation patterns with such a range of sign types to produce "complete" utterances that Mulder distinguishes a syntagm from a sentence (consisting of the sentential base, which-as we have

seen-may be of varying grammatical types, and an intonation

pattern). In particular, it is important to distinguish the predicative syntagm from the sentence. Mulder's formal definition of the sentence as "a signum with such features that it cannot be a feature (constituent or other feature) of another signum" is

noteworthy because it plainly expresses, in a negative way, the idea that a sentence is somehow communicationally complete or "whole". Mulder's definition is instructive also in that it leads us to distinguish grammatical form, as a means of modelling sign complexity, from the modelling of utterances. Mulder is clearly




free to draw on ideas of grammatical form to model utterances, but he is not uniquely dependent on them. The modelling of utterances is a good deal closer to communicational tokens in

speech. Grammatical form is a more abstract concept4 We may, however, wish to add something to Mulder's view

of the sentence. Hervey has rightly pointed out that in addition to the complex of phonetic, grammatical and semantic features which we normally associate with "sentences" we must take into account pragmatic features found in utterances as part of the

linguistic data to be analysed. One might want to add also that attitudinal and affective information, which is often associated with the degree of loudness, relative duration and intonation pattern, should be included in the linguistic data. Again, while sentences

may not be part of other signa, many utterances which appear to be complete are so because of their intonation pattern. But it is also true that their acceptability as complete utterances depends on the particular linguistic context in which they are found. This is clear in the case of pro-forms, e. g. - Mary does (In response to Who teaches music here?). - Her? Oh, yes (In response to Is that Mary?).

The expressions containing the pro-forms, does and her, are complete only in the defined context (or other such context). A linguistic context may determine other linguistic features for some sentences to come as in reported speech or conditionals, - He said he was coming. He would bring his wife. They were travelling

from Cardif... and they would be on time. - If I had a million pounds, I would retire to the Bahamas. I could

have a house with a swimming pool. I would have three sports cars. I could...

What we take to be "complete" is, then, often context dependent. This reinforces the point made earlier that "completeness" in speech is a feature of utterances. It should be modelled at the level of sentences. The other side of the story is that grammatical completeness, however important, does not of itself confer communicational completeness. Grammatical completeness is deter- mined at the level of the syntagm and it is found just as much with the nominal syntagm (e. g. that beautiful cat) or verbal syntagm

4. See J. W. F. Mulder, "Postulates for Axiomatic Functionalism ", def. 20 ff.



86 Paul Rastall

(e. g. may have been sunbathing) as it is with the predicative syntagm (that beautfiul cat may have been sunbathing). No amount of grammatical completeness, however, confers communicational completeness which requires also a final intonation pattern and the complex of pragmatic, attitudinal and affective features in addition to con- textual appropriacy that have been mentioned.

It may be that some expressions (with appropriate intonation

patterns) can be interpreted without reference to the linguistic context, - Mr Major is the Prime Minister of the United Kingdom. - To every action there exists an equal and opposite reaction, etc.

Such utterances contain all the grammatical and lexical information for an interpretation without reference to other linguistic expressions. They are beloved of traditional grammarians and

philosophers. One suspects that their occurrence is relatively res- tricted and infrequent. One could think of them and other

"context-setting" expressions such as, - Oh, you know that car we saw yesterday... - With reference to your recent letter I wish to inform you that...

--Here is the news for today...

as "independent". Given the purpose of such expressions it is clear that they cannot be elliptical. They must be complete and

non-elliptical in order to serve their pragmatic purpose. Where utterances can be regarded as complete only in a given

linguistic context, we could call them "autonomous". Expressions containing pro-forms or ellipsis are plainly autonomous. In both cases, we have instances of what Martinet would call the application of the law of minimum effort in communication. On the one hand, shorter pro-forms are used to avoid repetition and, on the other, known or stored information is omitted and only "new" information is conveyed.

It is noticeable that, in English, the ellipsed element is often the nuclear one (or a pro-form represents the nuclear element). This is to be expected as the nuclear element is obligatory and must be established in the "independent" utterance(s) which create the linguistic context. Such elements also determine the structure

by reference to which a pro-form or an elliptical expression or both may be interpreted. It is precisely the nuclear element which,




once established in the independent utterance, can be omitted or represented by a pro-form in a following autonomous utterance. This happens when, for instance, the first (subordinate) auxiliary is used as a kind of pro-form representing the entire verb phrase in English and the nuclear main verb is ellipsed either in a succeeding utterance or a succeeding part of the same utterance, as in, - John has (In response to Who has been working in here.). -John will (In response to Who will do the job.). -john is (In response to Who is going.). -John will be going, but I won't.

-John has eaten and so has Magy. --I am going and John is too.

A dummy pro-form occurs when there is no auxiliary in the

independent utterance.

-John does (In response to Who knows?). -Johhn goes, but I don't.

-John went and so did Magy. - I go and so does John.

It is the nuclear verb, bought, which is ellipsed in the earlier example, and John a bike, which follows, I bought a car. In the nominal syntagm, the nuclear element may be represented by the dummy element, one(s), - I bought the green pullover and May the red one. - I like the blue flowers and Maty likes the red ones, etc.5

It is important to remember, as Martinet has shown, that there can be cases of conflict between minimum effort and communicational efficiency. In such cases, the "completeness" of an utterance may require an increase in the number of elements in a structure. Such is the case when components of an utterance appear to be abnormally long, or when the components would disrupt a structure by divorcing elements in normally close proxi- mity, or when there would be a combination which would appear to be ungrammatical. Then, dummy elements may be inserted. In English, the dummy, it, is used to mark a grammatical posi-

5. Of course there are also cases of 'unrecoverable ellipsis' where the lexical elements cannot be known, although usually the grammatical structure can.



88 Paul Rastall

tion which will be occupied by a structure appearing later in the temporal sequence. The dummy, it, is used for all of the three above-named purposes. - It is clear that he doesn't want to do the washing up (The that clause

is the subject which is displaced to the end to avoid a long "wait" for the predicate).

- He took it for granted that I would go (Take for granted is a fixed

phrase whose elements cannot be divorced too far). - I will see to it that he goes (The preposition, to, cannot be fol-

lowed by a that clause).

We can think of the expressions involved as discontinuous elements, it... that... The purpose of dummy elements is plainly to help maintain grammatical clarity at the expense of an extra element and the displacement of major elements in a less usual "word order", and so an increase in effort. A similar pheno- menon occurs where there is "displacement" of relatively long adjectivals in English. In English, adjectivals normally precede the nominal element and follow the article and/or numeral (if any). Complex adjectivals, however, may be displaced so that

they follow the noun or may be split into discontinuous structures with one or both parts being "displaced" (occasionally with the first part of the discontinuity before the article), as in:

-an apple too bitter to eat, - too bitter an apple to eat,

--a solution more difficult than the last, -a more diffcult solution than the last, - more difficult a solution than the last,

where too bitter to eat and more difficult than the last cannot occur in the "normal" adjectival position. I. e. we cannot have

-a too bitter to eat apple, - a more difficult than the last solution,

probably because the nuclear element would be delayed too long and it would be excessively divorced, for that reason, from the article. Again the cost of clarity in a complete structure is a less usual word order and a corresponding increase in effort.

Now, it should be clear that an utterance will be regarded as complete if its intonation pattern is final and if we can relate




it to some grammatical model (either directly or within the lin-

guistic context). While completeness of sentences involves practical issues of communicational efficiency and the social role of the speaker in the discourse, completeness in the grammatical sense involves containing all necessary grammatical relations for instantiating a syntagm. Thus, the green bottle, with the green bottle, and may have been working are "complete" in this sense, and green bottle, with green bottle, and may have work are not. There may be completeness of nominal, prepositional and verbal syntagms, etc. It is only the predicative syntagm which the traditional gramma- rian, with his confusion of sentence and syntagm, would accept as complete (i. e. of the sort the man may have been working, etc.).

With all its drawbacks, the traditional view is, to some extent, justified. The subject is said to introduce that which is talked about and the predicate states something about the subject. It is clear that the "completeness" of an utterance requires that the components of the utterance be connected and that connected components constitute a "cross-classification". That is, we must view the experience from at least two different angles. This feature of "cross-classification" is particularly important at the level of predicative syntagms or syntagms of an equivalent level (where no "predication" is involved, as in the earlier examples). Jes- persen spoke of this as the "nexus" and we might borrow further from his terminology to revise the traditional view by saying that the "cross-referencing" of experience may take place in any form of syntagm (i.e. not just in predicative syntagms or the equivalent), but that any complete utterance contains or implies at least one "primary" cross-classification of experience (which may or may not be contained within that utterance)6.

The cross-classification of experience by linguistic means takes place within any syntagm. The cross-classification is partly a matter of the speaker's choice and it is partly a matter of the "choices" which are imposed on the speaker by the communication system. If we imagine an experience to be communicated, such as a dog lying asleep, these points become clearer. Clearly, any speaker is faced with an immense number of possible utterances. The speaker may make a statement, ask a question, make an excla- mation, and so on. He may focus on the dog's position, its colour,

6. See O.Jespersen, Essentials of English Grammar, p. 91 ff.



90 Paul Rastall

or any one of innumerable possibilities. Whichever choices the

speaker makes, it is always necessary for him to view the experience from at least two different points of view. In: - The dog is asleep, the dog is identified and the fact of its sleeping. These two factors are first connected and, second, cross-classified. This is, of course, an analysis of experience in Martinet's sense. The dog and its

sleeping are, in experience, part of one totality of impressions. Similarly, in questions and exclamations such as:

--Is that your dog? - What a lazy hound!

two aspects of experience are connected and cross-classified, viz. the

dog and the request for information about its ownership and the

dog and an expression of personal surprise (Of course, the ownership of the dog or the reaction of the speaker are part of the wider context of the experience not directly related to the sleeping dog.)

While the speaker has a number of free lexical and grammatical choices, numerous "choices" are also imposed upon him. Some article (your, the, a, my, etc.) must be chosen with a count noun such as dog. Some tense (past or non-past) must be chosen. Some intonation pattern must be chosen, etc. These "choices" are enforced

by the communication system. Often they relate to "ways of

viewing" the experience rather than to the experience "itself" (defi- nacy, for example, is not part of the "dog experience ", but some-

thing imposed upon it). In any case, however, there is always at least a double way of looking at the experience or part of the

experience and this applies at different "levels" of syntax. It is important to note that where there is no connectedness

there can be no cross-classification of experience. The connectedness of an utterance appears to depend on the array of com-

binatory possibilities available. Thus, any article may be in construc- tion with dog, but a verb, for instance, could not relate to dog at the same level of analysis. - Keep dog is asleep, is therefore uninterpretable. Similarly, there can be no primary cross- classification of a prepositional syntagm and a verbal syntagm in English. - In the garden is asleep. There is no connectedness in such a case.




It is important to remember that the cross-classification of experience may not be present within the utterance. In elliptical utterances, it may involve the stored linguistic context, as we have seen. It is also important to remember that it is not necessary that there should be two discrete signs in the utterance. A discrete sign and a significant intonation pattern constitute a kind of minimum

complete communication act. Such a case might be calling someone

by name with an emphatic intonation pattern (and increased loudness) or in questions/uncertainty with the interrogative intonation,

A - John!

- John?

The finiteness of utterances has a certain practical importance in communication. No utterance can be of infinite or even of inde- finite length. Furthermore, the hearer must recognise when an utterance is finished and that it is finished in order to respond appro- priately and in a socially acceptable manner (in terms of conversational turn-taking, for example). Completeness can be identified as a property of utterances and can be accounted for in terms of the properties of sentences (in Mulder's sense). There must be a "base" of grammatical structure and lexical elements providing the necessary connectedness, a non-discrete intonation pattern and, in the case of autonomous sentences, discourse relations with other utterances. Most importantly of all, however, complete utterances contain at least one primary cross-classification of experience.

REFERENCES

Hervey S. G.J., "Sentences and Linguistic Data, La Linguistique, 1991/1. Jespersen 0., Essentials of English Grammar, London, 1979 (1st Ed., 1933). Martinet A., A Functional View of Language, Oxford, 1962. Martinet A., Studies in Functional Syntax, Munich, 1975. Mulder J. W. F., "Postulates for Axiomatic functionalism", in J. W. F. Mulder

and S. G.J. Hervey, The Strategy of Linguistics, Edinburgh, 1980.



Documents

Langage, sujet, lien social || Complete Utterances