Sunteți pe pagina 1din 6

The Workshop on Generative Grammar Author(s): Jean Malmstrom Source: College Composition and Communication, Vol. 13, No.

3, Annual Meeting, Chicago, 1962 (Oct., 1962), pp. 26-30 Published by: National Council of Teachers of English Stable URL: http://www.jstor.org/stable/354759 . Accessed: 14/11/2013 10:59
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 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 support@jstor.org.

National Council of Teachers of English is collaborating with JSTOR to digitize, preserve and extend access to College Composition and Communication.

http://www.jstor.org

This content downloaded from 195.130.124.78 on Thu, 14 Nov 2013 10:59:07 AM All use subject to JSTOR Terms and Conditions

26

COMPOSITIONAND COMMUNICATION

Mowrer, O. Hobart. Learning Theory and the Symbolic Processes. New York: Wiley, 1960. Skinner,B. F. Verbal Behavior.New York:Appleton-Century-Crofts,1957. Verplanck, William S. "The Control of the Content of Conversation: Reinforcement of Statements of Opinion,"Journalof Abnormal and Social Psychology, 51 (1955), 668-676. Watson, John B. Psychologyfrom the Standpoint of a Behaviorist. Philadelphia: Lippincott, 1919. Weiss, A. Paul. A Theoretical Basis of Human Behavior. Columbus: Adams, 1929.

on enerativejrammar Tbe 7Workshop


JEAN MALMSTROM The linguistic intuitions of the native speaker of English reveal that immediate constituent analysis is incomplete, since it can explain only some English sentences. Chomsky offered "He was eager to please." and "He was easy to please." as examples of grammatically similar sentences which native speakers intuitively recognize as "different." Since transformations can give powerful explanations to account for such differences, we may as well accept these explanations as good evidence for the necessity of transformations in our grammar. In the discussion which followed his talk, Chomsky further defined "grammatical rules" by subdividing them into "rewrite rules" and "transformational rules." The rewrite rules are an axiom system; each rule is a theorem within the system. The generative grammarian explains the axiom "Sentence" by applying a sequence of rules, to a single grammatical element at a time, in a predetermined order. His first rule S -> Nom + VP Mrs. Malmstrom,who served as recorder of the workshop on generative grammarat (to be read as "Rewrite the sentence the April meeting of CCCC in Chicago, is as nominal and verb phrase.") analyzes an associate professorof English at Western the sentence into subject and predicate. A later rule analyzes the verb phrase Michigan University. Noam Chomsky, professor of Modern Languages at Massachusetts Institute of Technology, spoke on the position of generative grammar in the tradition of English studies. He began by considering how linguistic creativity, or innovation, has been handled in the English grammatical tradition. In the nineteenth and early twentieth centuries, Humboldt, Paul, and de Saussure recognized language behavior as a creative act, but confused rule-governed behavior with rule-changing behavior, or analogic change. Today, new insights from mathematics and logic reinforce the notion that linguistic creativity is not simple verbal recall but instead involves new applications of old rules and the discovery of rules of infinite applicability. These rules are the linguistic regularities by which all and only the grammatical sentences of the language can be specified. The problem of generative grammar is to make the general rules explicit.

This content downloaded from 195.130.124.78 on Thu, 14 Nov 2013 10:59:07 AM All use subject to JSTOR Terms and Conditions

THE WORKSHOPON GENERATIVE GRAMMAR into its parts; a still later rule analyzes the nominal into its parts. The end result of the rewrite rules is a "labeled bracketing" which shows how each part of the sentence is related to each other part. [We might say that this procedure is somewhat like the operation which traditional grammar calls "sentence diagraming," and which structural linguistics calls "immediate constituent analysis."] By his rewrite rules, the generative grammarian can analyze and explain simple, declarative, active sentences. However, to analyze and explain other sentences-complex, interrogative, passive sentences, for instance-he needs more powerful rules. These rules are the "transformational rules." In using transformational rules, the generative grammarian, given a certain type of structure, changes it to another structure related to it in a fixed and stateable way. For instance, given a sentence like:

27

ing it with philosophic intuition, defined it as "knowledge that comes to you without your knowing where it comes from." He denied that all knowledge comes to us through experience; probably all really important learning requires a "flash of insight," that peculiarly human phenomenon which all of us recognize but which none of us can define or explain. This viewpoint is consistent with those of both philosophy and psychology, Chomsky said. At the second workshop session, Robert B. Lees, of the University of Illinois, spoke on the achievements and nonachievements of generative grammar. He focused first on Chomsky's public statement that scholarly traditional grammar ranks high on the significance of its goals but low on explicitness, whereas structural linguistics ranks high on explicitness but low on the significance of its goals. Generative grammar ranks high on both desiderata. He then used specific THE BOY THREW THE BALL. examples to flesh out Chomsky's abstract the passive transformational rule pro- presentation of these basic considerations. duces: THE BALL WAS THROWN BY THE BOY. Lees noted that scholarly traditional This transformation places the original grammarians characterized well-formed object in subject position and the origin- sentences by 1) classifying words into al subject at the end of the sentence after parts of speech, and 2) listing the perthe preposition by, and inserts a form of mitted sequences of these parts of be preceding the past participle of the speech. Their parts-of-speech classificaverb of the original sentence. Chomsky tions were always incomplete, however, suggested that, lacking any other satis- because they neglected to specify the fying formulations, the ordering of these many subclassifications. Among verbs, rules may well represent "psychological for example, a transitive-intransitive subreality," the order of historical process. class is necessary, as is an animate-inaniThese ideas run counter to the anti- mate subclass. Given enough subclasses, mentalism of Bloomfield and his follow- we can define all sentences, but our defiers, who limit themselves to analyzing nitions will lack insightful generality unobserved data instead of asking the truly less our analysis permits us to note how important question: "What is the set of parts of speech are variously grouped. rules that generates linguistic compe- Thus, we understand the two following sentences differently even though supertence?" In this connection, the generative ficially their grammatical structure seems grammarian is concerned with "intui- identical: tion," a term which needed definition for They shot the man in the leg. (. . gave the workshop members. Chomsky, equat- him a leg wound)

This content downloaded from 195.130.124.78 on Thu, 14 Nov 2013 10:59:07 AM All use subject to JSTOR Terms and Conditions

28

COMPOSITIONAND COMMUNICATION utterances like "He dood it." for affective reasons-to be funny, for instance. Thus, an adequate grammatical theory is one that can assign degrees of grammaticalness, rather than one that attempts to account for every sentence that any speaker may produce. At the third workshop session David DeCamp, of the University of Texas, spoke on applications of generative grammar to teaching literature and composition. He explained the term "grammatical" as a strategic norm to which we can refer in a rhetorical discussion. If the grammatical norms are violated at a high level in the derivation, something like "He terrified the wastebasket." may result. The deviances of e. e. cummings are on this level. A somewhat lower level deviance might produce "He drank the wastebasket." Dylan Thomas deviates at about this level. Lower still among the grammatical restraining rules might occur "He drank the garbage." and "He drank the beer." We unpoetic rhetoricians deviate at the level of lexical choice, choosing, for example, between "We drank the martini." and "We guzzled the martini." and thereby achieving differing rhetorical effects. Thus the grammatical restraining rules have rhetorical relevance to the study of literature. Generative grammar can also help students observe certain literary strategies of sentence combination. DeCamp suggested, for example, that the twentyfive or thirty kernel sentences underlying a long Faulknerian "sentence" can be recovered and then recombined by generative embedding devices to yield a Jamesian "sentence." Value judgments on both the process and the results of such experiments lie outside the grammar, which can offer only a terminology for discussing the basis of such value judgments. Thus the linguistic contributions of generative grammar to literary

They shot the man in the fez. (... [man] who wore a fez) Bracketing the prepositional phrase in the first sentence with the verb and that in the second sentence with the noun explains why the two sentences are understood differently. Generative grammar uses these old familiar bracketings of labelled word-groups as it also uses the equally familiar old concept that one sentence is derived from another. This later concept seems "the natural way to view sentences," and is obviously different from the structural linguistic concept of the sentence as a putting-together of elements in a sequence. The main contribution of generative grammar has been to render explicit what it means to say: "One sentence is derived from another." Lees pointed out that no generative grammar is complete. Research is being done on special topics; new findings cause revisions of older rules. Therefore Lees did not try to present a "complete" grammar, but chose examples to illustrate various kinds of useful grammatical rules. In an interesting disgression, he developed the generative grammarian's important concept: degrees of grammaticalness. Speakers of English produce certain utterances like "Birds sing." which all other users of English unhesitatingly agree are impeccably grammatical sentences of English. Speakers of English also can produce other utterances like "Red and green frighten one another." To these, other users of English deny immediate and automatic acceptance, since they feel intuitively that such sentences are not fully grammatical. Of course one may imagine a special context for any sentence. An interior decorator, for instance, might remark, "Red and green frighten one another." Indeed, metaphors require highly specialized contexts. Furthermore, people often use non-grammatical

This content downloaded from 195.130.124.78 on Thu, 14 Nov 2013 10:59:07 AM All use subject to JSTOR Terms and Conditions

THE WORKSHOPON GENERATIVE GRAMMAR understanding are admittedly trivial to date. In the teaching of oral and written composition, generative grammar can offer real help however. Students need knowledge of the complex repertory of available grammatical possibilities and an understanding of which ones are strategic at certain points. When a child begins to write, he is unable to carry over his intuitive knowledge of spoken sentences to his new technique of writing. He must learn a new activity in which his intuitions are not yet formulated. Developing generally in the sequence of the generative grammar rules, he first writes simple kernel sentences. By stimulating and positive instruction, the teacher may guide him to more complex structures, without his falling prey, for instance, to such constructional homonymity as: "She enjoyed riding, playing tennis, and exciting dates." If the student is discouraged by negative, severe grading, he will revert to simplicity. Turning to the problems of writing a pedagogical generative grammar, DeCamp stated that strategic choices must be made in order to keep the rules as simple as possible without losing sight of the grammar as a whole. Proper limitations vary according to the audience. In teacher training courses, a rather full and detailed grammar can be presented; from it, later selections can be made. The problems of the high school and college student will be largely in the area of transformations, whereas those of the elementary child and the foreigner will concern the whole grammar. In reply to a question, DeCamp emphasized the need for eclectic terminology in the grammatical training of secondary teachers because of the variety of available textbooks. Traditional terminology will be necessary as long as the College Entrance Examination

29

Board continues to use it. Certain terms from structural linguistics are useful too, especially for a few graphemic problems. The questions presented to the consultants in the final workshop session elicited two analyses of complementtype sentences, a series of illustrative adverb-embedding rules, and answers to three large questions. The first question was: "What is the distinction between generative grammar and transformational grammar?" DeCamp declared the two terms interchangeable as used in the workshop, yet he made the further point that a grammar which generates all and only the sentences of a language by transformational rules is not the only possible generative grammar. The structural linguist also assumes that his grammar is valid beyond the specific text analyzed. This extended validity is the mark of grammatical generation. The second question was: "Does generative grammar include phonology and morphology in addition to syntax?" In answer, Chomsky stated that Morris Halle's The Sound Pattern of Russian shows how the phonology is "buried" in the morphology of a language to form a single morphophonemic "linguistic level." The generative grammarian does not recognize the structural linguist's "phonemic level" as a separate entity. Instead, he asserts that, to get a rational transcription, the linguist must know both the morphemes and the syntactic derivation of the sentence. So equipped, he can state the sentence's phonetic transcription complete with its stresses. Using all this information, he can evolve generalizations like: "All vowels become i in certain stateable positions." The third question concerned the relation of a theory of meaning to a theory of generative grammar. Denying that "meaning" is to be identifed with either "informant response" or "in-

This content downloaded from 195.130.124.78 on Thu, 14 Nov 2013 10:59:07 AM All use subject to JSTOR Terms and Conditions

30

COMPOSITIONAND COMMUNICATION mar, however, includes transformations. As a result, it extends semantic horizons. For example, a word analysis is not sufficient to account for the ambiguity of
FLYING PLANES CAN BE DANGEROUS.

tuition," Chomsky stated that meaning is not a singular concept, but covers a variety of subjects. Note the differences among: 1. The spots on his face mean that he has measles. 2. This means [i.e., will result in] war. 3. What does this mean to you? ("connotation"). 4. X means [i.e., refers to] Y. ("denotation"). When a person produces or understands a sentence, he focuses upon points of alternative choice. Grammatical context determines which choice will be made at each crucial point to achieve a grammatical sentence. For instance, certain verbs (like believe, eat, hate) grammatically require an animate subject; certain adjectives (like afraid, ajar) occur only in the predicates of sentences. Such contextual constraints are fragments of the study of meaning. Furthermore, an item is meaningful precisely because it is selected by rules which have alternatives. For example, having selected a copula-type sentence, some form of be must be used, and consequently be, as a word in this context, has no meaning because no alternative choice was possible. We might as well ask the meaning of e in be. The study of meaning can progress no farther than the study of grammar has progressed. Generally speaking, in traditional grammar and structural linguistics, the study of meaning has been the study of morphemes, words, and word classes. If grammatical insights are limited to words, meaning theory is similarly limited. Generative gram-

This sentence needs a transformational analysis to explain that its ambiguity is a reflection of the constructional homonymity of its two possible underlying source sentences: X FLIES PLANES. and
PLANES FLY.

In other words, the transformational rules of the grammar will produce Flying planes can be dangerous. twice by separate transformational paths. The ambiguity is a corroboration of the adequacy of the grammar, proving the existence of the transformational level of analysis. This level was ignored in Fries's theory of "structural meaning," which concerned itself solely with observed morphemes, words, and word classes. Chomsky explained further that human beings universally impose certain meaning-categories upon their conceptual scene. For instance, a person learning a new word prefers to consider its function rather than its physical character. Thus in learning "knife," the child says, "It cuts." In addition, contiguity is preferred over separation. Such structuring, imposed by the human being upon his environment, interrelates with the grammar. All languages, Chomsky said, have categories which reflect this fact. Thus there are indeed valid connections between a theory of meaning and a theory of generative grammar.

The major speeches printed in this issue are being reproduced as a separate brochure by the National Council of Teachers of English, 508 South Sixth Street, Champaign,Illinois.

This content downloaded from 195.130.124.78 on Thu, 14 Nov 2013 10:59:07 AM All use subject to JSTOR Terms and Conditions

S-ar putea să vă placă și