Universal Grammar and Narrative Form. - book reviews

Two words in the title of David Herman's book signal it as a maverick on the contemporary critical scene: "universal" and "grammar." (This is offset by "narrative," a hot item on the current market.) "Grammar" attempts to revive the once-flourishing field of linguistic approaches to literature, while "universal" counters the current emphasis on differences and relativity by making a commitment to structures that transcend cultural boundaries. The term "grammar" is taken in its widest sense, as a theory of language embracing all three components of the semiotic trivium: syntax, semantics, and pragmatics. This ambitious program mobilizes an impressive array of theoretical sources towards the analysis of narrative discourse: linguistics, logic, philosophy (especially phenomenology), structuralism, and even psychoanalysis and deconstruction. The list of references alone occupies an amazing twenty-seven pages.

The introduction provides a useful history of the concept of universal grammar. In Antiquity and the Middle Ages, the utopian project of a grammatical description valid for all human languages was grounded in the postulation of a correspondence of cognitive, ontological, and linguistic structures: the study of the grammar of natural language was regarded as an investigation of the "laws of thought" or "nature of things" (9). In modern logic (Leibniz, Frege), the role of universal grammar is entrusted to a formal calculus representing the mental operations which are expressed in discourse. For Chomsky, universal grammar is the particular "wiring" of the human mind that predisposes it to the acquisition of language. Transposed to the narrative domain by French Structuralists (Barthes, Todorov, Greimas), the idea of universal grammar expresses the belief that narrative is a cognitive structure independent of language and culture. The rules that define this structure arc seen as more general than those of particular languages, but less general than logic (30).

The main body of the book consists of three long chapters (forty to sixty pages each) featuring a literary text discussed in terms of one of the three components of the semiotic trivium: the "Sirens" episode from Joyce's Ulysses is paired with syntax in chapter 1, Kafka's The Trial with semantics in chapter 2, and Woolf's Between the Acts with pragmatics in chapter 3. The unifying theme of these chapters (and in fact, of the entire book) resides in their insistence on the dependency of meaning on frames of reference such as linguistic context, conceptual models, pragmatic situation, intertextual relations, knowledge of codes and conventions, and rules of inference. Each of the readings focuses on how a change of frame of reference affects the interpretation of verbal and physical phenomena.

In chapter 1, "The Modeling of Syntactic Structures: Sirens, Schonberg, and the Acceptation of Syntax," the frame of reference is not only intertextual but interdisciplinary. Musical comparisons are common practice in the critical literature devoted to the "Sirens" episode, but whereas Joyce claimed to have used the fugue as model, Herman proposes to read the text in the light of Schonberg's experiments with a twelve-tone scale. In contrast to the fugue analogy, which inspired critics to dissect "Sirens" into contrapuntal voices, the Schonberg comparison leads to no precise mapping of textual units onto musical ones. (Herman's reading of Joyce reverts to the scale-independent concept of polyphony.) The main point of the Schonberg analogy resides in a break with tradition: "Joyces's text disrupts classical narrative form in the same way that Schonberg's twelve-tone technique wreaks change on classical conceptions of music" (51). I wish here that Herman had considered the consequences of the Schonberg comparison for the claim of the universality of narrative structures. Is the project of a "universal grammar" of narrative forms viable, if these forms are as diverse as musical scales? Is there a meta-scale that contains all scales? No less problematic is the relevance of Herman's analysis of the "syntactic form" of "Sirens" to the question of narrrativity. The reading focuses on stylistic features such as sequencing techniques, combination of motifs, transformational operations on words, and distortions of English syntax, but it never addresses the question of the contribution of the text to the advancement of the plot of Ulysses. The unanswered question is whether all the discourse features of a globally narrative text contribute to its narrativity. In its second half, the chapter offers a fascinating discussion of the concept of syntax in logic, linguistics, and narratology, but Herman seems almost afraid to use his shiny analytical tools to dissect the flesh of the text. The theoretical build-up is simply out of proportion to the insights gained into the "Sirens" episode.

The second chapter, "Semantic Dimensions: Objects and Models in Kafka's Der Prozess," presents an opposite problem: it offers a convincing and original reading of the text, but its use of theory is questionable. The focus of Herman's reading of The Trial is the interpretive failure that befalls Josef K. in the "world" of the Court. This failure does not concern the linguistic code per se (K. is able to talk, if not communicate, with magistrates and Court employees), but the indexical meaning that people customarily attach to phenomena on the basis of world-knowledge: for instance, smoke means fire because our experience tells us that fire causes smoke. In the world of the Court, phenomena are the same as in the world of K.'s experience (his "model world" or "reference world," in Herman's terminology), but the rules of pragmatic inference that link them to meanings are entirely different. Herman follows the theme of interpretive failure in sentences such as this one: "It could only be a sign of deep humiliation, or it must at least be interpreted as such, that the Examining Magistrate now took up the notebook where it had fallen, tried to put it to rights again, and made as if to read in it once more" (130). The "must" and "as if" introduce inferences that would be valid in K.'s reference world, but not in the world of the Court. K.'s reference world, once regarded as the condition of all meaning, has thus been "virtualized." In support of his reading, Herman invokes a definition of meaning allegedly borrowed from "model-theoretic semantics" (another name for possible-worlds theory): meaning (intension) is "a function from some possible world to the truth-conditions defining the world stipulated to be actual for the purposes of a given (semantic) analysis" (124). What this formula is supposed to mean is explained by an example: the truth conditions of the word "red" can vary from possible world to possible world. In the world I regard as actual, "red" refers to a color; in another world, "red" could be used to denote "a town in Lithuania"; hence "red" has a different meaning in these two worlds. This account of meaning which Herman applies to both linguistic and indexical signs is intended to explain how certain phenomena can mean one thing in K.'s reference-world, and another in the world of the Court. As far as I know, however, the formula is nowhere to be found in the philosophical literature. For possible-worlds theory, what is variable across possible worlds is not the truth-conditions of expressions, but rather their truth-value. The proposition "Elvis is dead" can be true in one world and false in another, but in both worlds it can only be true if a certain individual singled out by the name of "Elvis" happens to be no longer alive at the time of the utterance. In Herman's version, by contrast, "Elvis is dead" could be true in a world just in case snow is white. This is only possible if in the language of that world "Elvis" refers to snow and "dead" to the color white. Herman has thus introduced an element of code-switching into a theory that was designed to explain meaning within a given code exclusively. Model-theoretic semantics is not concerned with how "terms and propositions can mean different things to different people at different times" (125) (an enterprise that involves a far fuzzier conception of "meaning" than the strictly denotative sense favored by logicians), but with whether two different expressions, such as "the Unabomber" and "Ted Kaczynski," mean the same thing for the same individual within a given code at a specific time. (The answer to my example is no, because in some possible worlds "Ted Kaczynski" and "the Unabomber" pick different referents.)

The question of represented discourse in narrative has received so much attention in the past twenty years that it would seem that all forms have been by now thoroughly catalogued and analyzed; yet the third chapter, "Toward a Metapragmatic of Represented Discourse," argues that existing typologies need further refinement to account for the almost mind-numbing modulations of voice in Virginia Woolf's Between the Acts. To resolve the question "who speaks" in Woolf's novel, Herman establishes a typology of nine modes, adding to the well-known trilogy of DD (direct discourse), ID (indirect discourse), and FID (free indirect discourse) such innovations as two types of indefDD (direct discourse issuing from unspecified speakers who can be part of either a restricted (type 1) or unrestricted group (type 2), two types of the parallel indefID, PDD/TIM (pseudo-direct discourse functioning as interior monologue), and embedded forms of these categories. An example of IndefDD embedded in IndefDD is 'Ping-ping-ping' in this passage from Between the Acts:

Feet crunched the gravel. Voices chattered. The inner voice, the other voice was saying: How can we deny that this brave music, wafted from the bushes, is expressive of some inner harmony? "When we wake" (some were thinking) "the day breaks us with its hard mallet blows." "The office" (some were thinking) "compels disparity. Scattered, shattered hither thither summoned by the bell. 'Ping-ping-ping' that's the phone. 'Forward!' 'Serving' that's the shop." (Woolf 119; quoted 179).

While the cataloging of the modes can be thought of as a pragmatics of represented discourse, the "metapragmatics" promised by the title consists of a ranking of the categories on a continuum ranging from "backgrounding" to "foregrounding" of the context. The degree of backgrounding or foregrounding reflects the importance of assessing the parameters of the speech situation (speaker, hearer, time, and place of utterance) to the intelligibility of the discourse. Herman ranks the forms in the order (from backrounding to foregrounding): narratorial discourse, attributive discourse, DD, ID, interior monologue, FID, indefDD, indefID, PDD/TIM, and embedded forms. Narratorial discourse occupies the backgrounding pole because its context is inherently indeterminate, at least in the case of third-person impersonal narration. This makes it unnecessary for the reader to reconstrue a specific situation of utterance. On the other end of the continuum, embedded forms require the reconstruction of several contexts - one for every mediating speaker in the chain of quotations. As I read Herman's ranking, however, it seems that another factor contributes to increased foregrounding: the difficulty of fixing the parameters of the speech situation. In the most foregrounding forms, the question "who speaks" (to whom) cannot be answered univocally: either because the utterance expresses a "dual voice" blending narrator and character (FID); or because the speaking source remains unidentified (indefDD and ID).

A second section contains two much shorter essays. The last chapter, "Modes of Meaning in Film" offers a reading of Peter Greenaway's The Cook, the Thief, His Wife and Her Lover (1989) inspired by possible-worlds semantics, here proposed as an alternative to the predominantly Lacanian approaches to film studies of the past fifteen years. The other essay, "Postmodernism as Secondary Grammar," flirts with postmodern aesthetics, as it relaxes the scholarly tone of the book into a mild form of "performance text." Adopting as metaphor the vocabulary of the computer field (scanning, codes, programs, interface), Herman reads postmodernism as "a set of discourse technologies" for constructing webs of citations (190). The "secondary grammar" proposed by the title would "model intuitions about how to decompose, permute, reassemble, and more generally manipulate text-sized structures" (191). This grammar, one must assume, would be a descriptive and not a prescriptive one: in the literary/textual domain, in contrast to the linguistic-sentential one, there are no prohibited combinations of elements but only as yet unrealized possibilities. Though Herman does not discuss electronic textuality per se, those interested in the theory of hypertext will encounter in this chapter an unlikely precursor in Walter Benjamin and his discussion of the newspaper. A mosaic of headlines, accounts, captions, cartoons, and images, text formatted in various ways and broken up into segments, the newspaper, like hypertext, presents a "matrix of items" out of which readers create a text customized to their own interests through selection, combination, and the following of links (194).

I find this next-to-last chapter the most successful because it takes the walk into the theoretical woods for its own sake and does not gloss over a literary text. Too often in the book the object-text is merely a pretext for the discussion of theory; here, however, Herman gives himself license to play the game he is best at, a dexterous juggling with ideas unhampered by a mission to "get somewhere": "After all, facing what other horizon [than postmodernism], standing in what other framework, could I cite so many bits of text over so short a period of time, and not be dismissed out of hand before reaching so inconclusive a conclusion?" (202). I do not mean here to endorse the postmodern promotion of literary criticism as a self-conscious textual play with nibbles of theory at the expense of the traditional scholarly criteria of building knowledge through linear argumentation, but I believe that each of the two rhetorics can be enlivened by a sprinkling of the other.

To summarize the value of Universal Grammar and Narrative Form I will parody the title of James McCawley's introduction to logic (itself, of course, a parody of the title of a well-known sex manual): Everything that Linguists have Always Wanted to Know about Logic (*but were ashamed to ask). To continue this playful chain of citations, Herman's book could be renamed "Many Things that Literary Critics Ought to Know about Logic and Linguistics (but usually don't want to ask, because the fields involve hard work and are not very fashionable)."

Marie-Laure Ryan Independent Scholar

Other Works Cited

McCawley, James. Everything that Linguists have Always Wanted to Know about Logic (*but were ashamed to ask). Chicago: U of Chicago P, 1981.

Woolf, Virginia. Between the Acts. Orlando: Harcourt Brace Jovanovich, 1941.

COPYRIGHT 1996 Northern Illinois University
COPYRIGHT 2004 Gale Group