Jahn, Manfred. 1999.
"Stanley Fish and the Constructivist Basis of Postclassical Narratology".

Note: A much shorter version of this essay is due to appear in the Proceedings of the Anglistentag 1999 Mainz (Trier: WVT, 2000).


Abstract -- According to Stanley Fish, narratology is an ill-conceived theoretical project that can never succeed. Narratologists generally reciprocate by marking Fish down as counterproductive or by pointedly ignoring what he is saying. The present essay argues that both sides are wrong. Assuming that much is to be learned from Fish's denunciation of theory, it will trace the development of his philosophy of reading from the late-structuralist program of "affective stylistics" to his current position of radical relativism constrained only by the norms and values of "interpretive communities." Specifically, it will discuss Fish's test cases concerning ambiguity, situatedness, fact, and truth, and make an attempt to reorient his insights within the wider contexts of cognitive linguistics, pragmatics, and artificial intelligence. In conclusion, the essay argues that both Fish's insights and his oversights significantly impact on many branches of what has come to be termed "postclassical" narratology.


1. Ambiguity

Although the keyword is unaccountably missing from the indexes of Is There a Text in This Class (Fish 1980) and Doing What Comes Naturally (Fish 1989), Fish's oeuvre is full of references to ambiguity, and a compulsive indexer could profitably arrange them under subdescriptors such as "in literature," "in natural language," "in puns," "local," "global," "illocutionary," "pervasiveness of," etc. Ambiguity, as Fish recalls (1980, 56-58), came into the focus of critical interest in the 1930s when publication of William Empson's Seven Types of Ambiguity started what might be described as the first school of dedicated ambiguity hunting. At the time, Empson's strategy of finding rather than of resolving ambiguities fed into the New Critics' program of demonstrating the richness of (mainly poetical) texts, not -- as is the custom today -- of disclosing the faultlines at which they fissure and fall apart. Empson eventually asserted that "all good poetry" was ambiguous (1949, xv), elevating ambiguity to a criterial feature that prefigured today's notion of literary polyvalence. Paradoxically, while Empson's brilliant textual analyses were widely accepted as the cutting edge of the discipline, the "seven types" themselves were fuzzy to the point of uselessness, as Fish notes (1980, 57). Consider, for instance, the mess Empson makes of type number four:

An ambiguity of the fourth type occurs when two or more meanings of a statement do not agree among themselves, but combine to make clear a more complicated state of mind in the author. Evidently this is a vague enough definition which would cover much of the third type, and almost everything in the types which follow . . . . (1949, 133)
Actually, Fish's critique is leveled not so much at Empson's definitions (though these are surely "vague enough") but at the fact that he approaches his subject from a mixture of heterogeneous points of view -- cohesion, intention, and reading effects, in the above quote. Fish himself, in contrast, begins his own project of a "Literature in the Reader" (1980, ch. 1) by addressing small-scale and volatile ambiguities that Empson did not even consider worthy of name, let alone type. Consider the following almost unnoticeable ambiguity from Paradise Lost:

          Satan, now first inflamed with rage came down,
          The Tempter ere th' Accuser of man-kind,
          To wreck on innocent frail man his loss
          Of that first battle, and his flight to Hell.
                    (qtd Fish 1980, 3)

At issue, here, is the referent of "his" in lines 11 and 12 (= 3 and 4 in quote). Globally speaking, this is clearly Satan. Yet, Fish says, if one carefully traces one's steps and surveys the state of affairs as it presents itself at the end of line 11, "his" is evidently strongly attracted to the immediately preceding "man." It is only the wider context of line 12 and its allusion to the person in question's "flight to Hell" that the connection between "his" and "man" dissolves in favor of one between "his" and Satan. Naturally, Fish seizes on the opportunity to point out that it is an aesthetically and morally relevant fact that the passage momentarily tempts the reader to confuse man with Satan.
          For many critics, the foregoing analysis will not present too compelling a case, first because Fish's reconstruction of the reading process is handled a bit too glibly,1 second, because the ambiguity lasts only the briefest of moments, and third, because many readers will simply deny that they were ever tempted to make that false connection. Evidently, Fish's point is better served by local ambiguities that introduce more palpable textual difficulties. As a matter of fact, Fish manages to unearth a very convincing item from an early 17C sermon by Lancelot Andrewes:

          He is found of them that seeke Him
                    not
                    but
          of them that seeke Him
                    never
                    but found (qtd. Fish 1980, 184)

The structural ambiguities suffusing this passage do not resolve as easily here as in the Milton example. Indeed, a number of trial-and-error moves are needed to arrive at a reading that makes overall sense, and an impatient reader may easily give up. Ultimately, after following up and duly backtracking on a number of wrong turns, the globally consistent (correct? intended?) reading worked out by Fish is that lines 1-2 constitute a sense unit presenting a statement about people who find God even though they do not seek Him, while the rest of the passage asserts that people who do seek Him are certain to succeed (He . . . is never but found = He is always found). Although this reading resolves the textual difficulty, it is as anticlimactic as a finished puzzle, and Fish's point is precisely that the passage is valuable not for the message that it ultimately conveys but for the mental exercise that it affords as one labors over it. As Fish points out, recognition of this fact slips through the net of a globalizing formalist approach because "the only making of sense in a formalist reading is the last one, whereas everything a reader does, even if he later undoes it, is part of the 'meaning experience' and should not be discarded" (1980, 3-4).
          The counterpart of local ambiguity is persistent ambiguity, the only "true" kind of ambiguity in the view of many commentators. As Fish shows, discussing the concluding lines of Milton's sonnet 20, persistent ambiguity comes with its own set of processing effects:

          He who of these delights can judge, and spare
          To interpose them oft, is not unwise. (qtd Fish 1980, 148)

"Spare" has two likely, but contradictory, meanings ('leave time for' and 'refrain from') which cannot be easily accommodated in the ordinary naturalizations available for cases such as this, namely as paradoxes or puns. As Fish points out, Milton exegetes usually follow the impulse to resolve the ambiguity by looking for disambiguating evidence in the wider context -- the context of Milton's sonnets, Milton's works, Milton's "known attitudes," and so on. What happens, however, is that consideration of higher levels of context merely leads to an infinite regress of alternate confirmations and disconfirmations. Fish himself concludes from what he thinks is an "equal availability of both interpretations" (1980, 151) that Milton's text conveys, precisely, a "blurred judgment." Making his methodological point, Fish states that this type of semiosis is inaccessible to "formalist criticism," which assumes that "meanings can be specified independently of the activity of reading" (1980, 152). Specifically, the formalist approach reduces ambiguities like Milton's "spare" either to "an (insoluble) crux" or eliminates them "in the course of a procedure that is incapable of finding value in temporal phenomena" (1980, 155).
          In linguistics, context is usually defined as the text surrounding an item, ambiguous or otherwise. The conventional formal notation is W__Z, where W and Z denote preceding (left) and subsequent (right) contexts, respectively. More generally, the extralinguistic situation in which an utterance is embedded is also usually called "context".2 While situational context is a (more or less) holistic construct, the only verbal context available at any one point in time during processing is previous context, i.e., W__. Specifically, Charles Hockett showed that a textual segment that begins with the phrase A man eating fish is locally ambiguous because it may refer either to a piscivorous man or to a hominivorous fish. If the text continues ... on Friday is not necessarily a Catholic then the first reading is strongly confirmed; if the text continues called the piranha is found in the tropical waters of Brazil the second reading is confirmed; and if it continues has an unbalanced diet then neither reading is confirmed and further disambiguating evidence is called for (Hockett 1961, 226).
          Hockett's speculative excursions into a "grammar for the hearer" (1961) and a "grammar of silence" (1967, ch. 7.4) were counterproposals to what he fittingly labeled "marble slab grammars." As Hockett put it, the traditional grammarian regards

a sentence as an enduring structure, to be scanned at leisure and repeatedly, and as easily from right to left or upside-down or inside-out as from left to right. He can do this because he deals not directly with a sentence, but only with a representation thereof, spread out before him like a cadaver on a marble slab, to be dissected at his convenience. (Hockett 1961, 220)
Today, Hockett's deliberations are justly considered seminal prolegomena toward a project of cognitive linguistics that materializes and acquires momentum in the sixties. One of the questions that has been investigated since is whether and to what extent comprehension processes use any wait-and-see strategies (van Dijk and Kintsch 1983, 66). As Hockett (1961, 221) suspected, all present evidence indicates that readers do not suspend parsing of a sentence until they have heard or read all of it. On the contrary, resting content with the flimsiest of evidence, readings are usually construed as early and quickly as possible. Naturally, given such readerly strategies, there is a certain risk of being occasionally wrong. Being occasionally wrong in turn requires compensatory strategies of undoing prior readings and following up untried options. Processing mechanisms such as these are typically highlighted by "garden path" constructions which trap the reader in a processing failure from which it is difficult to recover. The standard example, invented by Thomas G. Bever (1970, 316), is The horse raced past the barn fell [= the horse that was raced past the barn fell]. Fish, in his discussion of local ambiguities in Milton and Andrewes, is the first critic to unearth instances of literary garden paths which are not just temporary readerly errors but have a specific (writerly?) functionality. The garden-path phenomenon has since become a favorite testing ground in psycholinguistics and artificial intelligence research. Significantly, it can also be observed in minimal stories like jokes and riddles as well as in more complex narratives like short stories and novels (see Jahn 1998 for a survey).
          While local and persistent ambiguity are grist for Fish's early program of "reader-response analysis" (Fish 1980, 148), the question of the range and spread of ambiguity gets him going on a more general philosophical and epistemological agenda. Interestingly, language philosophers are sharply divided in their assessment of ambiguity. One position is exemplified by Mey's (1993, 7) claim that "[i]n real life -- that is, among real language users -- there is no such thing as ambiguity." The opposing view is put by Sperber and Wilson, who state that
narrowly defined, almost every utterance is ambiguous. In fact, almost every utterance is multiply ambiguous, with possible semantic interactions among its individual ambiguous constructions. . . . It is thus quite typical for an utterance to have dozens, or even hundreds, of possible propositional interpretations. However, speaker and hearer are normally able to select a single one of these interpretations without even realizing that they have made a choice. It is generally agreed that this choice is a function of the context; but to define the function, as opposed to claiming that it exists, is no easy task. (Sperber and Wilson 1981, 298)
As one can see, Mey takes a "product-oriented" view that assumes that true ambiguity is rare and of little ultimate consequence in everyday communication, whereas Sperber and Wilson take a "process-oriented" view claiming not only that ambiguity is rampant but that disambiguation is an essential part of meaning construction. It is this latter view of process-oriented, pervasive ambiguity that Fish embraces; indeed, for Fish, univocity is so unreal that ambiguity lacks the true opposite that would render it distinctive: "to label a sentence 'ambiguous' will be to distinguish it only if there are sentences that always and only mean one thing, and I would contend that there are no such sentences" (1980, 281). Thus, when E. D. Hirsch suggests that The air is crisp is unambiguous, Fish points out that the sentence does not in fact disambiguate the word air ('local atmospheric condition'/'melody') (Fish 1980, 309). Similarly, when linguists argue that The suit is too light [in color or in weight?] to wear is conclusively disambiguated when placed in a context like __ on such a cold day, Fish shows that such seemingly disambiguating contexts are themselves prone to becoming reinterpreted in wider contexts (1980, 281).
          Due to his early interest in speech-act theory, Fish is also one of the first commentators to discuss the specific effects of illocutionary ambiguities. He points out, for instance, that I will leave may "in different circumstances, be a promise, a threat, a warning, or a prediction" (1980, 229, 284). Another example, Can you pass the salt?, is given particularly close and searching attention:
Consider the small example of the utterance "Can you pass the salt?" immediately construed by the vast majority of native speakers as a request for performance of a specific action rather than as a question about the hearer's physical abilities; but this is so because in the very hearing of the utterance we assume the mealtime setting populated by agents concerned with eating and drinking" . . . . If one varies the setting and reconceives it as a conversation between a doctor and a patient recovering from surgery, the utterance "Can you pass the salt" could indeed be heard as a question about the hearer's physical ability . . . . Independently of some such already assumed context (and there could be many more than two), the utterance wouldn't have any meaning at all and wouldn't be an utterance, but merely a succession of noises or marks. . . . In the example of "Can you pass the salt?" it is always possible that someone at a dinner table may hear the question as one about his abilities, or that a patient may hear his doctor asking him to pass the salt (perhaps as a preliminary to an experiment). (1989, 295-6).
Fish makes four important points here. First, he shows that Can you pass the salt? is illocutionary ambiguous because it has a literal and an indirect-speech-act reading. Second, he offers an explanation of why the indirect-speech-act reading is the normally preferred one (because, he says, the mealtime setting is the "standard story"). Third, he demonstrates that choice of reading freely varies with context. And fourth, he has the wonderful acuity to recognize that the only context that really counts is what the hearer assumes to be relevant. In exactly the same vein, Sperber and Wilson (1986, ch. 3.3) compellingly argue that context is not what is objectively given but what is actively chosen by understanders. Ultimately, Fish says, context is "a product of interpretation and as such is itself variable as a constraint" (1989, 108).
          In the final analysis, one cannot trust "self-explanatory" words like lion hunt (Gove 1966), one cannot rely on seemingly unambiguous sentences like Hirsch's the air is crisp, and, come to think of it, proper names like "Socrates" or, for that matter, "Fish" are not exactly constants either.3 If natural language is incurably ambiguous and affords nothing like a firm ground, and if context is subjective rather than objective, then, as Colomb and Turner (1989) convincingly argue, any approach relying on univocity, objective signifieds, and compositionality (the principle that the meaning of a sentence is a function of the meaning of its components) is doomed to failure. The logical question would seem to be which more adequate model or theory to put in its place. For reasons of his own (to be gone into in section 4), Fish does not ask this question, nor does he explore whether any of the conceptual tools developed in cognitive linguistics and pragmatics might contribute to shedding light on the problem. Fish's tendency to overlook common concerns and, possibly, common resources becomes even more evident when the question of context is considered in its more general epistemological dimension.

2. Shaping eyes and nontrivial machines

Even as he recognizes that context is subjective and "variable as a constraint," Fish does not for a moment doubt that its import extends far beyond the management of ambiguity. Context determines meaning, changes meaning and creates meaning out of meaninglessness. Specifically, in order to demonstrate that context is also a major factor in the assessment of truth, Fish picks up a test case first presented by John Austin:

In the penultimate chapter of How To Do Things with Words, J. L. Austin presents a sentence and asks us to consider it. The sentence is "France is hexagonal," and the question he puts to it is a very familiar one in analytical philosophy: Is it true or false? The answer, however, is not so familiar. It depends, says Austin: "I can see what you mean by saying that it is true for certain intents and purposes. It is good enough for a top-ranking general, perhaps, but not for a geographer" (p. 142). I[n] other words, the truth or falsehood of a sentence is a function of the circumstances within which it is uttered, and since it is always uttered within some set of circumstances or others, it is not in and of itself either true or false, accurate or inaccurate, precise or imprecise. (Fish 1980, 197)
Truth or falsehood, Fish says here, is a function of circumstances -- deviating slightly from Austin, whose "final answer" is that the sentence is a "rough description" and therefore neither true nor false (1962, 142). In addition, while ostensibly supporting Austin's vision of studying "not the sentence but the issuing of an utterance in a situation" (qtd Fish 1980, 231; orig. Austin 1962, 138), Fish recognizes that (just as in the case of Can you pass the salt?) it is not the situation per se that is relevant -- a mealtime setting, a general glancing at a sketch of France, a cartographer in converse with other cartographers, in short, any kind of reality seen objectively -- but the construction of the (or a) situation in the mind of the speaker or hearer passing the true-or-false judgment. Austin himself takes a step toward context-in-the-mind by granting that something can be "true for certain intents and purposes," "true for a general," etc. Fish soon makes the radical generalization that "what is normal (like what is ordinary, literal, everyday) is a function [Fish's italics] of circumstances in that it depends on the expectations and assumptions that happen to be in force" (1980, 287).
          The point is gone into in considerable detail in what is perhaps the most famous of all of Fish's essays, "How To Recognize a Poem When You See One" (1980, ch. 14). In this essay, Fish anecdotally recounts an informal experiment that he sprang on his students when he was a guest lecturer at the University of Buffalo, New York, in 1971. He taught two consecutive morning classes there, one on stylistics, and one on English poetry. One day, he left the first class's reading assignment, which consisted of a tabular list of five authors, on the blackboard. When the poetry students entered, he told them that the "text" on the blackboard was "a religious poem of the kind they had been studying" (1980, 323). Asked to "interpret" it, the students treated the "text" as a poem, and proceeded to generate a number of surprisingly plausible readings. Here is Fish's own account of what happened:
My students did not proceed from the noting of distinguishing features to the recognition that they were confronted by a poem: rather, it was the act of recognition that came first -- they knew in advance that they were dealing with a poem -- and the distinguishing features then followed. . . . As soon as my students were aware that that it was poetry they were seeing, they began to look with poetry-seeing eyes, that is, with eyes that saw everything in relation to the properties they knew poems to possess. . . . It was almost as if they were following a recipe -- if it's a poem do this, if it's a poem, see it that way -- and indeed definitions of poetry are recipes, for by directing readers as to what to look for in a poem, they instruct them in ways of looking that will produce what they expect to see. . . . Skilled reading is usually thought to be a matter of discerning what is there, but if the example of my students can be generalized, it is a matter of how to produce what can thereafter be said to be there. Interpretation is not the art of construing but the art of constructing. Interpreters do not decode poems: they make them. (1980, 326-7)
Emphasizing, in his conclusion, that "Interpretation is . . . the art of constructing," Fish is only a figura etymologica away from "constructivism," a set of influential epistemological axioms accumulating from the works of Giambattista Vico, Immanuel Kant, Wilhelm von Humboldt, William James, Charles Sanders Peirce, Edward Sapir, and Benjamin Lee Whorf (among the more recent prophets of "radical" constructivism one usually includes Humberto Maturana, Heinz von Foerster, Paul Watzlawik, Siegfried J. Schmidt, and Francisco J. Varela). On his private journey toward constructivism, Fish starts out by establishing that the original "assignment reading" of the Buffalo text is in no respect more accurate than, or logically prior to, the poetic reading that is projected on it by his poetry students. In fact, seeing the text as a list of names is just as dependent on interpretive strategies as seeing it as a poem is. Evidently, Fish feels that the outcome of the experiment is indicative not only of understanding of texts but also of "acts of recognition" and of "thinking, seeing, reading" in general (1980, 335). Humans (like all living organisms, presumably) have a "shaping eye" (1980, 333) that has no other way of seeing than to see x (a real object) as y (a percept). Specifically, Fish adds, seeing x as y is a function of a mind's culturally acquired beliefs and interpretive operations.
          Although the concept of function or product that crops up here and elsewhere in Fish's texts primarily stresses a strong dependency relation, it clearly also invokes the mathematical formula y = f(x), "y is a function of x." In addition to that, Fish's use of concepts like "interpretive program" (1980, 170), "recipe" (in block quote above), and "routines" (1995, 41) closely dovetails with the main constitutive feature of a computational function model, or "computational machine." A computational machine houses an "operation" (algorithm, program, recipe, etc.) f that turns input data x into output data y. Following a suggestion by Heinz von Foerster, Fig. 1 distinguishes between a "trivial" and a "nontrivial machine".4

(ntm3.gif)

Fig. 1. Trivial and nontrivial machines (von Foerster 1993, 357-59).

According to von Foerster, a trivial machine is one that always produces the same results, no matter how often or in what sequence values of x are fed into it, and no matter what operation f1 performs. The interesting thing here is that there are tasks that trivial machines do extremely well and others where they are wholly useless. For instance, while a trivial machine easily outperforms a human in tasks like squaring numbers, searching a text for a word, or watering the plants, it is highly unlikely that it will ever plausibly disambiguate a sentence or be able to judge whether it is true or false. Even though cognitive judgments are often so spontaneous as to seem "automatic," artificial intelligence theorists have come to recognize (a) that basic cognitive processes are extremely hard to monitor, and (b) that a computational language processor has no choice but to closely mimic human cognitive processes. In other words, there can be no understanding outside human understanding, and human understanding, as all of Fish's test cases amply show, is neither abstract nor ahistorical nor generally consistent.
          Intriguingly, von Foerster's nontrivial machine does present a design that has access to a memory of situations and the capacity to modify its states in accordance with what it is doing. Superficially, a nontrivial machine is not much different from a trivial machine; in fact, both are indistinguishable from the outside. Like a trivial machine, a nontrivial machine gets input data x, performs some function or process f2, and produces output y. The crucial differences are in internal design and behavior: in addition to acting on input x, f2 also gets input from the machine's current internal state (the box labeled S); and S is in turn affected (changed) by the outcome of f2's operation. Two consequences follow: like humans, nontrivial machines are historically conditioned (they learn, they forget, they change), and they are difficult to predict. For illustration, consider one trivial and three not-so-trivial examples.

a) Suppose we are observing the input-output behavior of a machine whose interior design and program is hidden from us in a black box. During a period of time we note the following six input-output pairs:

... M-m, x-X, G-g, d-D, P-p, C-c, ...
Apparently, the machine uses a function that turns upper-case letters into lower-case letters and vice versa. One can fairly confidently predict that if fed another "M" it will produce an "m", if fed an "e" it will produce an "E", and so on. Since output is both predictable and unaffected by temporal sequence it seems safe to assume that the machine's black box houses a trivial machine.

b) Suppose we observe another black box that generates the following sequence:

... A-a, B-b, A-b, A-b, B-a, B-b, ...
All that it seems safe to say here is that input consists exclusively of upper-case A's and B's while output consists exclusively of lower-case a's and b's. Otherwise it is hard to explain what goes on, or to predict what the output is going to be if the machine were fed another A or B. As a matter of fact, output can be described as the product of a nontrivial machine that uses four rules and interacts with two internal states (states 1 and 2):
(i) If x = "A" and S = 1, then produce y = "a" and leave machine in S = 1.
(ii) If x = "A" and S = 2 then produce y = "b" and leave machine in S = 2.
(iii) If x = "B" and S = 1 then produce y = "b" and put machine into S = 2.
(iv) If x = "B" and S = 2 then produce y = "a" and put machine into S = 1.
The reader will be relieved to learn that this particular nontrivial machine was constructed before presenting the input-output sequence. The reverse task, that is, deducing a nontrivial machine from its input-output data is often impossible (von Foerster 1993, 359).

c) Consider the familiar "Necker cube" reproduced in Fig. 2.

(cube.gif)

The cube is ambiguous: it can either be seen from right and above or from left and below. In perception, disambiguation is so instantaneous that many people doubt that it ever took place. After a while, that is, after continuing exposure to the same input, a "spontaneous" reversal takes place, and the cube is suddenly seen in its alternate orientation. Although one can consciously stick to one interpretation for a while, once attention flags, uncontrollable point of view switches follow. The reasons for the reversal effect are largely unknown. Even if the viewer is aware of the fact that the figure is ambiguous its two aspects are never seen simultaneously.5

d) Suppose there are four major mind states or "life positions" informing the psychological structure of human beings: (1) I'm OK, you're OK, (2) I'm OK, you're not OK, (3) I'm not OK, you're OK, (4) I'm not OK, you're not OK (Harris 1967, ch. 3). Suppose person A meets person B and is greeted with x = "And how are you this fine morning?" Consider what A may make of x, and what s/he might reply, given any of the four states s/he could be in. Further, consider how A's interpretation of x might modify or confirm his/her current state.

          One of the major advantages of models like those in Fig. 1 is that they allow the researcher to focus on constitutive elements and to stake out specific fields of inquiry without losing sight of crucial relations or the general picture. Methodologically, the model suggests "structured" strategies such as top-down design, modularity, and stepwise refinement. While f2, S, and their interfaces are open to design and trial-and-error testing, constructivists generally agree that perceptual input is just "a pattern of stimulations" (Churchland 1993, 46), or, as Fish puts it, a mere "succession of noises and marks" (Fish 1989, 296). As Fish's Buffalo experiment shows, even formal features (lines of verse vs. a list of names) have no objective reality independent of interpretation. As regards components S and f2 in von Foerster's nontrivial machine, Fish's descriptive vocabulary actually provides a number of useful conceptualizations. Specifically, S may be the dynamic store of a mind's "assumptions" (1980, vii), "convictions," "beliefs" (1980, 199), "expectations" (1980, 277), "tacit knowledge" (1980, 330), "interests and goals" (1980, 332), "values and norms" (1980, 334), while f2 may be taken to contain a mind's "interpretive program," "recipe," collection of "routines" (1995, 41), "constitutive rules" (1980, 241), etc.
          Evidently, there is a remarkable degree of convergence between Fish's notion of shaping eyes, the pragmalinguists' conception of chosen context, the constructivists' model of nontrivial machines, and the artificial intelligence programmers' project of designing a computational natural language processor. Regrettably, this is just an incidental agreement rather than the product of a genuine meeting of minds, theories, and disciplines. In fact, after arguing the constructivist axioms, Fish assumes an "anti-foundationalist" position that negates much of the interdisciplinary promise that has been so obvious in this section.

3. Interpretive Communities

Just as a person's shaping eyes see reality not for what it really is but for what it is made out to be on the basis of beliefs and interpretive strategies, texts are not read for what they really are but for what readers make of them. For many critics, including Fish, allowing texts to be "written" by readers raises the specter of interpretive anarchy (1980, 172). The moment Fish opens the Pandora's box of subjective relativism, however, he slams it to and seals it shut by making all individual judgment dependent on community conventions.

In this new vision both texts and readers lose the independence that would be necessary for either of them to claim the honour of being the source of interpretive authority; both are absorbed by the interpretive community which, because it is responsible for all acts interpreters can possibly perform, is finally responsible for the texts those performances bring into the world. (1989, 142).
Fish's turn of the screw ostensibly produces a function of a function -- he begins by establishing that a person's perception is a function of interpretive strategies, and now he establishes that interpretive strategies are functions of community perspectives. Further characteristics of interpretive communities follow readily: two readers who share the same beliefs and strategies belong to the same interpretive community. Members of the same interpretive community will "agree" (1980, 169) about what "counts as a fact, of what is central, peripheral, and worthy of being noticed" (1980, 337); "and conversely, members of different communities will disagree because from each of their respective positions the other 'simply' cannot see what is obviously and inescapably there" (1980, 15). Depending on the behavior of their populations, interpretive communities "grow larger and decline, and individuals move from one to another," and "ways of interpreting . . . can . . . be forgotten or supplanted, or complicated or dropped from favor ('no one reads that way any more')" (1980, 171-2). Over and above the members that constitute it, an interpretive community is defined not by a "closed set of rules" (1989, 151) but by an adaptive and self-regulating code (1989, 150) of assumptions that tends to change over time: "neither interpretive communities nor the minds of community members are stable and fixed, but are, rather, . . . engines of change . . . whose work is at the same time assimilative and self-transforming" (1989, 152). (Interpretive communities, one is tempted to say, behave like non-trivial machines.)
          Since a fact is only what counts as a fact under a community perspective, Fish concludes that the literary critic's argument, indeed any argument, is basically rhetorical in nature. Specifically, literary critics are in the business of persuading members of their own community of the validity of the community perspective, and, at the same time, they are tendering an invitation to non-members (members of other interpretive communities) to join. Ultimately, Fish argues, even an avowedly descriptive mode of analysis is in the service of a normative and prescriptive purpose. This insight has drastic consequences not least for the type of criticism practiced by Fish himself. Thus, in a characteristic move, Fish relativizes his own project of "affective stylistics," originally pitted against the erroneous ways of formalism, and concedes that affective stylistics is merely one among many competing models, one way of interpreting like any other:
What I was trying to persuade them [the formalists] from was not a fundamental or natural way but a way no less conventional than mine . . . This meant that the business of criticism was not (as I had previously thought) to determine a correct way of reading but to determine from which of a number of possible perspectives reading will proceed. (1980, 16)
Leaving no doubt about what is at issue, Fish states that "no interpretation can be said to be better or worse than any other, and in the classroom this means that we have no answer to the student who says my interpretation is as valid as yours" (1980, 317). But Fish also immediately assures us that there is no cause for alarm. "Total and debilitating relativism," he argues, cannot arise because all readings are always sanctioned by the interpretive community for and in which they are made. Because community restrictions are always in force, the "brakes" against subjective interpretations and idiosyncratic meanings (1980, 338) "are always on" (1989, 83). Obviously, too, the fact that different interpretive communities license different readings does not entail that a text can be read as anything. The question "What Makes an Interpretation Acceptable" (1980, ch. 15) can always be asked, and always be answered, positively or negatively, with reference to community conventions. Since there are many interpretive communities, however, no reading, not even seemingly "absurd or impossible" ones (1980, 342-345; 1989, 104-107, 193), can be ruled out on the strength of some set of absolute constraints. While it is hard to imagine that anyone would seriously claim that Hamlet is a forceful man of action (1989, 107), that Agatha Christie's novels are treatises on death, or that Blake's "Tyger" is a "prophetic message inspired by Aunt Tilley" (1980, 343), one cannot simply say that the line has to be drawn somewhere. "If someone says that King Lear is a little girl he is wrong," emeritus Cambridge (UK) Professor Derek Brewer has recently stated in an attempt to call (post)modern critics to order, and to shore up a minimal consensus no longer, notably, about what is right but about what is wrong (Brewer 1994, 44). Assuming a characteristic attitude of contrariousness, Fish makes it a point to demonstrate that absurd interpretations such as the foregoing have a habit of acquiring a sufficient measure of acceptability as soon as they are situated in an appropriate context. For instance, the Aunt-Tilley reading of Blake's "Tyger" becomes as good as unassailable within the community of reader-response critics who allow elicitation strategies such as free association. "No reading," Fish states categorically, "however outlandish it might appear, is inherently an impossible one" (1980, 347).
          It may be instructive to put Fish's rule to the test of a real case. In a recent study entitled The Racial Matrix of Modern Fiction, Laura Doyle reads Molly Bloom as "polyracial -- Irish, Jewish, and possibly Spanish" (Doyle 1994, 132) -- a characterization that will catch many Joyceans by surprise. Molly's Jewishness is established on the following evidence:
Molly points out specifically that her charged position within constructs of both motherhood and racialism at the time Bloom met her -- a young, sexual Jewess tending her dying mother -- was partly what drew him to her. ("I suppose on account of my being a Jewess looking after my mother" [18.1184]). (Doyle 1994, 133)
Joyceans are generally aware of the fact that very few nouns in Molly's monologue are capitalized and may take the trouble to check the original, just in case. The line quoted by Doyle actually reads:
we stood staring at one another as if we met somewhere I suppose on account of my being jewess looking after my mother (18.1183)
Minute as the differences may seem, Doyle's reading misconstrues and her rewriting distorts the text. Not only does Doyle illegitimately capitalize "jewess," projecting her own focus of interest, she also inserts a phantom indefinite article. Molly isn't "a" Jewess looking after her mother, but Jewess-looking, after her mother, the way her mother did. There is no known indication in Ulysses (the authoritative account being Raleigh's Chronicle of Leopold and Molly Bloom) that Molly's mother was dying when Molly met Bloom, or that she is dead in 1904, or that Molly tended her, or that Molly's father, Major Tweedy, is Jewish. In fact, what textual evidence there is favors Molly's Catholicism -- she remembers going to confession, and plans to buy "a bit of fish tomorrow or today is it Friday" (18.939). While none of this is conclusive -- a woman eating fish on Friday isn't necessarily a Catholic -- it certainly makes it problematic to proceed on the notion that Molly is Jewish. As it stands, Doyle's case is instructive for two reasons: on the one hand it supports Fish's assumption that the critic's shaping eye sees what it sees according to an interpretive community's decision about what "is worthy of being noticed" (in this case, the race thematic); on the other hand it throws into question his reassurance that a "text cannot be overwhelmed by an irresponsible reader and one need not worry about protecting the purity of a text" (1980, 336). There is, after all, a larger principle that says that erring is human, suggesting that Fish's rule no reading is an inherently impossible one must be supplemented by a corollary to the effect that mistaken readings are not inherently impossible, either.
          Fish would probably argue that Doyle's "error" comes out not by comparing it to the objective reality of the original text but by making her reading compete with a reading that conforms more adequately to the community's standards of evidence. Doyle's added "a" only becomes the momentous distortion that it is on the background of the more acceptable reading that it excludes. In another context, an added "a" could be a negligible error, a sense-preserving modification, or a plausible emendation. In the final analysis, Fish would argue, the appeal to Joyce's text is valid only because it invokes the direct-quote criteria and evidentiary procedures of the interpretive community within which the case is argued. Thus, in a sense, Doyle's error illustrates and supports Fish's position. The only thing left unaccounted for is that there should be disagreement within one and the same interpretive community. Indeed, Parrinder (1987, 44) has since argued that the community of literary critics is more profitably seen as a community of disagreeing voices rather than one that acquiesces in agreement.
          Of course, like the critic who says King Lear isn't a little girl, the critic who is smug about other critics' errors has his comeuppance waiting for him round the corner. It is appropriate, here, to recall Philip Johnson-Laird's The book fills a much needed gap (1981, 122), one of the most mischievous garden paths on record. There is no obvious reason why the sentence should tax one's cognitive abilities, yet most readers misconstrue it wildly. In fact, even when Johnson-Laird points out that the sentence explicitly states "that it is the gap, not the book, that is needed" (1981, 122), hence that it is not as "laudatory" as one may think, it usually takes an awkward moment before, by a Necker-cube type of reversal, the sentence's true (if idiotic) meaning snaps into place. In hindsight, one realizes, of course, that the correct reading is initially blocked by a powerful mental preference to read for good sense (the "filling" frame suppressing the "needing" frame). In Fish terms, this is a consequence of an interpretive community's horror of semantic emptiness, of preferring, by default, to virtually "make" sense (Lehnert 1979, 84-5) -- irrespective, as one can easily see, of what the text really says.
          The true heuristic gain of Fish's concept of community-relative interpretations comes to the fore when it is put to use in historical accounts of changing and developing critical perspectives. Surveying the reception histories of Blake's "The Tyger" (1980, ch. 15), Swift's "Verses on the Death of Dr. Swift" (1989, ch. 9), and Milton's Paradise Lost (1989, ch. 12), Fish draws attention to the fact that, over the period of their reception, these texts were all interpreted not only differently but in often spectacularly antithetical ways. Indeed, many readings that were put forward seem little less than absurd or impossible in the light of contemporary critical opinion. But, as Fish demonstrates in these case studies, the concept of interpretive communities effectively counterbalance the paralyzing influence of a potentially endless array of retrospective master perspectives.

4. Anti-foundationalism and anti-theory

In section 2 of this essay I drew attention to the fact that there is a remarkable convergence between the tenets of Fish's philosophy and those of constructivism. Yet rather than embrace constructivism in proper fashion, Fish prefers to turn to anti-foundationalism, the "going argument" (1989, 345), he claims, in philosophy, history, sociology, hermeneutics, history of art, legal theory, and literary theory. Anti-foundationalism, in Fish's definition, "teaches that questions of fact, truth, correctness, validity, and clarity can neither be posed nor answered in reference to some extracontextual, ahistorical, nonsituational reality, or rule, or law, or value" (1989, 344). But, properly understood, anti-foundationalism also removes any "ground" on which to base a theory in the first place. No theoretical undertaking, Fish says, can change or guide the beliefs and interpretive strategies that the members of an interpretive community are already using in their everyday practice (of "seeing, thinking, reading") without the assistance of theory. So, even before a Fishian constructivist literary studies program (or whatever one might call it) can work up a stride, Fish's anti-theoretical stance radically alienates him from constructivists, literary theorists, and narratologists alike. As Culler complains as early as 1981,

It is not a little ironic that a man who has so imperiously thrust the reader before us, announced a new age of criticism focussed on the reader, and insisted that meaning and value lie not in the text itself but in the activity of reading, should then turn and tell us that we need not enquire what that activity involves. Indeed, it is not ironic but bathetic. (Culler 1981, 126)
          Nevertheless, Fish's argument against theory does deserve closer attention. Appropriately enough, Fish begins by deliberating the question of "what theory is and is not" (1989, 315). Picking up E. D. Hirsch's distinction between "rules" and "rules of thumb," Fish explains that
A rule is formalizable: it can be programmed on a computer and, therefore, can be followed by anyone who has been equipped with explicit (noncircular) definitions and equally explicit directions for carrying out a procedure. A rule of thumb, by contrast, cannot be formalized, because the conditions of its application vary with the contextual circumstances of an ongoing practice. (1989, 317)
According to Fish, strict (programmable, formalizable) rules have their place in mathematics and linguistics, while rules of thumb underlie behavioral patterns, legal scenarios, and games -- roughly, they are strategies "to try if you want to succeed in the game" (1989, 316). Theory, in Fish's view, is definitionally predicated on strict rules, while any description that invokes rules of thumb merely reflects "the contingent practices of particular communities" (1989, 320). Since every so-called fact is historically and culturally situated, it can, at best, fall under rules of thumb, not abstract or strict rules. For this reason (though in singularly ill-chosen phrasing), "every rule is a rule of thumb" (1989, 321), meaning, presumably, every rule of interest to Fish. Theory, in contrast, amounts to imagining a "theoretical machine" that, flawlessly and consistently, produces some "desired result":
In linguistics, that result would be the assigning of correct descriptions to sentences; in literary studies the result would be the assigning of valid interpretations to works of literature. In both cases (and in any other that could be imagined) the practitioner gives himself over to the theoretical machine, surrenders his judgment to it, in order to reach conclusions that in no way depend on his education, or point of view, or cultural situation. (1989, 319)
Apart from the image of minds dominated by machines (which is always greatly effective), the rational part of Fish's argument is notably fragile. To begin with, it is oddly ahistorical (for Fish, of all people) to reduce theory to the programmability of its rules -- evidently, theories existed before the advent of computers. Second, if theory by definition excluded strategic and context-sensitive rules then there could be no such thing as "game theory" -- yet computers can play a mean game of chess, and it is not the end of civilization as we know it (well perhaps it is, for master chess players). Third, the rules of "preferences" (Jackendoff 1987), "fuzzy logic" (McNeill and Freiberger 1993), and "neural nets" (Churchland 1993) that are the order of the day for today's "intelligent" computational algorithms are ostensibly closer in nature to rules of thumbs than to strict rules. Fourth, the only type of "theoretical machine" Fish can think of is a trivial machine, a design now generally granted to be unsuited for capturing essential and basic human cognitive abilities. Interestingly, "being right" and "yielding correct results" (1989, 316) for today's computational natural language processors means going wrong on items like garden path sentences. A computational parser that cracks The horse raced past the barn fell without batting an eye is not worth the paper it has been designed on -- the trick is to design it in such a manner that it falls for the trap just like ordinary people do (Fodor and Inoue 1994). The reason for this is obvious: the cognitive mechanisms that trigger the garden path effect are precisely those that allow effective and effortless comprehension of ordinary sentences in the first place.
          Rather than leave it to Fish to decide "what theory is and is not," it is prudent to check his definition against a more traditional definition. In his Philosophy of Natural Science, Carl G. Hempel details the following "general characteristics of theories":
Theories are usually introduced when previous study of a class of phenomena has revealed a system of uniformities that can be expressed in the form of empirical laws. Theories then seek to explain those regularities and generally, to afford a deeper and more accurate understanding of the phenomena in question. To this end, a theory construes those phenomena as manifestations of entities and processes that lie beyond or beneath them, as it were. These are assumed to be governed by characteristic theoretical laws, or theoretical principles, by means of which the theory then explains the empirical uniformities that have been previously discovered, and usually also predicts "new" regularities of similar kinds. (Hempel 1966, 70)
Unlike Fish, Hempel requires his laws to be specifiable "with appropriate clarity and precision" (Hempel 1966, 71), a far shot from requiring them to be formalizable and programmable. Hempel is also not at all concerned about the fact that a theory can make do with whatever rules are at hand. Mathematically precise rules are fine, presumably, but if all the rules available turn out to be rules of thumb then that is apparently fine, too, without the project losing the name of theory.
          Over and above critiquing theory's reliance on "strict rules" and its assumption of impossibly "objective" points of view, Fish also questions the possibility of "self-conscious" reflection. As Fish argues, the beliefs that enable reflection cannot, at the moment of reflection, be the subject of reflection. Specifically, Fish says, no-one can stand "to the side of [one's] own ways of thinking" (1989, 437), no-one can "step back from one's beliefs in order to survey or reform them" (1989, 465), and hence there can be no such thing as "critical self-consciousness" (1989, 465). To the extent that theoretical thinking attempts such stepping back, it undermines the anti-foundationalist credo "that we are always and already interpretively situated" (1989, 437).
          As Fish notes, the problem of self-conscious reflection boils down to a problem of recursive position taking. When critical theorists argue that self-conscious awareness can and ought to be put to the use of interrogating one's position Fish asks, "from what position is this question about our position to be asked?" (1989, 438). For the purposes of further discussion, let P1 informally stand for a position on something, and P2 for a "metaposition" on that position. According to Fish, P2 cannot in any sense be superior or more objective than P1 because P2 is as "interpretively situated" as P1. On Fish's view, this implies that P2 is like P1 -- different in "constraints" perhaps (1989, 13), but otherwise "no different (in kind)" (1989, 448). Basically, however, the question whether a position is like a metaposition can be answered, with roughly equal plausibility, in the affirmative as well as in the negative. On the one hand, P2 is like P1 because both are positions (thus Fish); on the other hand, P2 is unlike P1 because P2 is, and P1 isn't, a metaposition. Consequently, while one can accept that P2 cannot claim any kind of "superior" objectivity (an evident absurdity in the light of both constructivism and anti-foundationalism), Fish does not convincingly establish that a (self-conscious) position on a position is impossible.6
          Significantly, in his own argumentative practice, Fish has no problem with taking a position on a position. Disputing the viability of a formalist linguistic approach, he writes, not mincing his words:
I am not saying that [Ruth] Kempson is beyond criticism simply because the context of which she is an extension prevents her from seeing certain arguments as respectable or even makable. In my very strong opinion the arguments she clings to, the arguments that underwrite the project of formal linguistics, are wrong. And it is part of my argument that I can say that despite the sympathetic analysis I make of her "epistemological condition." This does not mean that I am not in the same condition -- embedded in conviction -- but that precisely because I am embedded in conviction, my sense of the rightness of my arguments is no less strong than hers and is in no way diminished by my ability to give an account of its source. (1989, 3)
The question here is how Fish's "sense of rightness," which allows him to see Kempson's approach as "wrong," can be "in no way diminished" by his "ability to give an account of its source" -- when at the bottom of this account is the anti-foundationalist insight that there is no once-and-for-all truth, and the conviction that a statement "is not in and of itself either true or false" (1980, 197). The fact of the matter is, Fish wants to have it both ways. When he feels like it, he describes divergent or contradictory readings as naturally issuing from the beliefs and strategies of interpretive communities and contends that "the business of criticism [is] not . . . to determine a correct way of reading but to determine from which of a number of possible perspectives reading will proceed" (1980, 16). To his credit, he can take that same stance of aloofness in questions where his own interests and convictions are concerned, in Milton studies, for instance. When the mood hits him, however -- and this usually happens when an article of anti-foundationalist faith itself is at issue -- he will stand on his convictions about what is true or false, right or wrong, and can adamantly declare that an approach like Kempson's is plainly "wrong." Inevitably, at this point, the classical law of contradiction rears its ugly head and declares that, situatedness or no situatedness, let constructivists and anti-foundationalists say what they will, something cannot be both true and false.
          Ultimately, Fish's argument against theory in general and narratology in particular is only a part of an emerging larger argument for "natural" reasoning. Fish thinks theory is an impossible project not only because it builds on abstract rules but also because active beliefs "cannot be the object of my attention because they are the content of my attention" (1989, 326). "[R]e-examining the principles underlying one's practice" (1989, 332) is an option that pales to insignificance when compared to Doing What Comes Naturally (= Fish 1989). As Fish explains in the preface of that book:
I intend it [the title] to refer to the unreflective actions that follow from being embedded in a context of practice. This kind of action -- and in my argument there is no other -- is anything but natural in the sense of proceeding independently of historical and social formations; but once those formations are in place (and they always are), what you think to do will not be calculated in relation to a higher law or an overarching theory but will issue from you as naturally as breathing. (1989, ix)
If all actions are "unreflective" then awareness dwindles into insignificance regardless of the fact that, as Fish takes pains to point out, it is ever-present. "[O]ne is always aware, one always knows what one is doing, and, when challenged, one can always give reasons," Fish says (1989, 462). "To be in a situation," Fish adds, "(as one always is) is already to be equipped with an awareness of possible goals, obstacles, goals, dangers, rewards, alternatives, etc." (1989, 466). And yet the only conclusion Fish draws from this is that there is no point in raising a call for becoming aware (as critical theorists do). Ultimately, Fish thinks, both theoretical reflection and judgments based on "analytical and critical attention" can be safely ignored as immaterial to the practice of perception and action.
          As one can see, Fish's argument deteriorates in tightening spirals of paradox. Needless to say, one is not "always aware": one is not normally aware of the "dangers, rewards, and alternatives" of perceiving the current visual scene (a point Fish himself makes often enough: "the immediacy of perception . . . escapes our attention," 1995, 76); one does not know what one's mind is doing when it disambiguates a word, or when it reverses one's view of the Necker cube. But it is clearly just as pointless to go to the other extreme and claim that one is never aware or, as Fish does, that all cognitive decisions passively "issue" from "unreflective actions." The main problem here is that Fish's argument builds on an adulterated concept of consciousness (a point also noted by de Beaugrande 1983, 118). Even though it is beyond question that "unconscious processes are involved in all mental tasks" (Baars 1997, 182), activities like skill-learning, action control, and problem-solving have to be conscious to some degree -- multiplying 365 by 24 will never "come naturally." Unreflective learning by doing is an option, the option Fish prefers to the exclusion of all others, but so is reflective learning by definition and theory. As far as the present author is concerned, the total number of sentences in this essay that "issued" from him "as naturally as breathing" is exactly nil.
          What Fish's argument builds on is, essentially, a "fractal" or wheels-within-wheels model of consciousness. Fish assumes that the mechanisms of belief-grounded interpretive strategies extend, via "beliefs that are so deep as to be invisible" (1989, 328), to the lower levels of subliminality. By fitting coincidence, the paperback edition of Doing What Comes Naturally has a fractal cover picture showing two bush-like objects silhouetted against a streaked evening sky. Looking closer, one can see that the shape of these bushes creates a visual pattern that reoccurs on several levels of lower magnification. While the aesthetic appeal of this is considerable, fractalism is not a generally useful explanatory scheme. Constructivists and cognitive scientists make it a point to carefully distinguish levels of consciousness (Schmidt 1989; 1991; Jackendoff 1997, ch 8; Bickerton 1995, ch. 4) and to integrate them in their models. Above all, researchers recognize that unconscious processes operate in parallel and handle a variety of autonomous tasks in a fast and efficient manner, while conscious processes operate in laborious sequence but handle tasks that consider contexts and weigh options (Baars 1997, 182). There is little in this that relates to Fish's system of beliefs and interpretive strategies, and less still to the assumption that it's beliefs and interpretive strategies all the way down.

5. Fish's contribution to postclassical narratology

Everything Fish says against literary theory in general also applies to narratology in particular. Narratology, he claims, is a "paradigm instance of theory in the strong sense" mustering "formidable apparatuses" (1989, 567n2) and strict rules to formulate general laws that disregard what happens when the practitioner-critic "makes" a valid interpretation -- processes that, he believes he has shown, are inaccessible to abstract theoretical analysis. True enough, according to Todorov, narratology's main goal is "no longer the description of the particular work, the designation of its meaning, but the establishment of general laws of which this particular text is the product" (1981, 6-7). Todorov also defined narratology as "a science of narrative" (1969, 10), an epithet that many literary critics reared in French and Anglo-American humanities departments found somewhat unsettling. As Seymour Chatman, himself a first-generation narratologist, commented in 1990,

'Narratology' is a word that Henry James would have deplored, though he might have found merit in its objects of concern. After twenty years, I still feel something of an ironic twinge when I see it in print. A "science" of narrative seems an unlikely, even a slightly shady pursuit. But, of course, "-ology" can also mean "theory of," and who, these days, would dare fault theory?" (Chatman 1990, 1)
Well, Fish does, for one, and lest it be forgot, he is by no means the only one to do so.7
          What is worthy of note in this discussion of possible and impossible projects is that the kind of narratology condemned by Fish is a late nineteen sixties to early eighties "classical" narratology -- a period that roughly begins with Chatman's putting "Eveline" on a marble slab and dissecting it into "kernels" and "satellites" (Chatman 1969), and ends with Gerald Prince's (1982) drawing up of a narratological grammar complete with well-formedness conditions, competence and performance distinction, and transformational component. In fairness, it must be added that not all early narratological models had the abstract orientation laid down by Todorov. For instance, both Chatman (1978) and Genette (1980) presented highly plausible and influential exemplifications of an "applied narratology," and "narratological criticism" has been an option ever since. Fish does not mention these groundbreaking works, but he would probably insist that all the interpretive gestures made in them were just excuses for building a theoretical machinery of discovery procedures that dictated both what researchers were to research and, worse, what they were to find.
          Leaving narratology for defunct, Fish remains unaware of the fact that the discipline has actually weathered the worst of the poststructuralist storm and moved on in a number of (what should be, to Fish) interesting directions. Entering a "postclassical" phase (Herman 1999), narratology has become a highly diversified undertaking spanning branches as varied as postmodern narratology (Hutcheon 1988), historiographic narratology (Cohn 1990), feminist narratology (Warhol 1989; 1999; Lanser 1992, Mezei 1996), possible-worlds narratology (Ryan 1991, Ronen 1994), natural narratology (Fludernik 1993; 1996), cognitive narratology (Jahn 1997), and constructivist narratology (Herring 1998; Nünning 1999). Of course, there is considerable overlap among the approaches, and current researchers emphasize the openness of the discipline, particularly vis à vis linguistics (Fludernik 1993), cognitive science (Duchan et al. 1995), artificial intelligence (Ryan 1991), and pragmatics (Adams 1996). Generally speaking, postclassical narratological projects embrace any or all of the following positions.
          1. Characters have pragmatic identities. In early structuralist accounts, literary characters are just "roles," "functions," or as Barthes called them, "paper beings" (1975, 261). The narrator, in particular, was often reduced to a sexless "narrative instance" (Genette 1980), a "narrative function" (Hamburger 1977), or a reified "it" (Bal 1985); in one notorious account, the heterodiegetic narrator was simply abolished in toto (Banfield 1982). Under a cognitive orientation, however, narrators and characters obtain a less cloudy existence as deictic centers, gendered beings, holders of assumptions, sources of rhetoric, and situated entities driven by affects, purposes, motives and intentions. In fact, a reader must project a pragmatic identity on fictional characters in order to understand description and narrated perception, speech, and action. When it comes to judging a character's or a narrator's reliability, readers obviously also invest them with psychological identities, categorizing them into sociocognitive types such as liars, deceivers, loudmouths, neurotics, madmen, etc. (Jahn 1998). Note well, this reinstatement of a pragmatic and a psychological dimension is no invitation to repeat the mistakes made in the past; specifically, there is no suggestion here to return the Freudian practice of treating literary characters as flesh-and-blood persons with instructive neuroses.
          2. Story is an "interpretant." Classical narratology strictly distinguished between discourse as narrative's signifier and story as narrative's signified. Following the logic of the Saussurean terms, this was conceived of as a transparent duality that extended, according to some accounts, not only to the "world-creating utterances" (Ryan 1981, 530) of heterodiegetic narrators but also to unreliable homodiegetic texts and severely refractive reflector-mode narratives -- narratives that, as Chatman put it, "transmit" story-events "through" a "character-filter" (1986, 193). However, the assumption that action is something that exists "prior to" and "independently of narrative presentation" was crucially challenged by Jonathan Culler, who suggested to identify story "not as the reality reported by the discourse but as its product" (1980, 28-9). Culler's position is clearly related to Fish's argument that readers "make" texts and that signifiers have no objective correlates in a world of signifieds. Postclassical narratology today generally acknowledges that story, causality, and chronology are readerly constructs or, using C. S. Peirce's term, interpretants (Morris 1964, 2-3; Peirce 1965, 2.228).
          3. There is no relation of mimesis. The postclassical subversion of the story dimension also calls into question any of the "mimetic" functions that narrative discourse is traditionally thought to possess. Indeed, questioning mimesis is an early stratagem (Genette 1980, ch. 4; 1988, 42-43), which is mirrored also in a general tendency to prefer qualified terms such as "formal mimetics" (Glowinski 1977, 106), "mimesis I, II, III" (Ricoeur 1984-88), "mimetic illusion" (Iser 1978), and so forth. On an explicitly constructivist note, Nünning (1999) suggests replacing mimesis (the ways and means of imitating reality) by "poiesis" (the ways and means of constructing models of reality).
          4. Narratives are interpretively situated. Rather than follow Todorov's proposal to draw up a timeless inventory of abstract laws and categories, postclassical narratology goes to considerable lengths to reconsider and redefine its units in the shaping contexts of historical, cultural, and pragmatic parameters. Specifically, as was suggested in section 3 of this essay, situational factors submit to Fish's model of norms-and-conventions driven interpretive communities. One of the liberating consequences of situatedness is that texts and genres are no longer definable by specific sets of inherent qualities. It is a commonplace today that something can be read "either as literature or as history" or that "the same sentence can have different meanings in poetry and prose" (Culler 1981, 123). Among the literary recontextualizations that come close to Fish's Buffalo experiment is the genre now known as "found poetry" (Stanzel 1990; see Culler [1975, 175] on poetic/quotidian reading strategies that make/unmake William Carlos Williams's "This is Just to Say" poem). Other examples come to mind: Fish mentions the use of Milton's poetry as Allied propaganda in WW II (1995, 66); the British psychiatrist R. D. Laing published the double-bind catches of his psychopathic patients in what looks, feels, and reads like a collection of poems (Knots). A detailed discussion of specifically postmodernist transcontextualization experiments can be found in Fludernik (1996, ch. 7).
          5. Narratives have a processual dimension. Classical narratological models arrive at their categories of events and existents by considering the text as a finished product and judging all textual detail from a global and retrospective view. Process-oriented analyses such as the ones conducted by Fish access "previous" verbal context only, and focus on the stepwise integration and combination of textual information, paying due attention to backtracking and revisionary moves. The notorious garden-path effect usually arises when a set of active reading strategies initially misinterprets a local ambiguity. The test cases presented by Fish (section 1) show that literary garden paths often have an aesthetic as well as a semiotic function, and this functionality clearly also extends to higher-level garden paths such as those found in jokes, riddles, and short stories (Jahn 1999). Both Sternberg (1978) and Perry (1979) draw attention to literary uses of cognitive "primacy" and "recency" effects in story construction and characterization; Perry, in particular, presents an account that establishes "a place for rejected meanings" (1979, 355). Ultimately, of course, the main goal of the cognitive approach is to develop a combined process-and-product model, one that builds complex conceptual structures by cumulatively integrating all local interpretive decisions, including wrong turns (van Dijk and Kintsch's 1983 model of discourse comprehension presents a step in that direction). Postclassical narratology's cognitive orientation is also exemplified in Fludernik's (1996) project of a "natural narratology" (see below) and the present writer's bid for a cognitive narratology (1997; 1998).8
          6. "Mental" narratives have a place in narratological analysis. While Genette (1988, 17) suggested to restrict narratology to pertain to discourse-based texts, most narratologists accept Chatman's proposal that narrative covers all sorts of texts -- today's official catalogue includes, among others, stained-glass windows, comic strips, drama, film, computer games, and latterly, music (McClary 1997). Naturally, the widening range of narrative forms requires appropriate conceptual adjustments; for instance, discourse time has to be replaced by performance and/or reception time. Furthermore, textuality itself, that is, a narrative's textual realization for use of altero-addressees (traditionally, a constitutive feature or necessary condition), appears to recede behind the spreading insight that a person's perceptions are already informed by internal narrative "scripts" (Schank and Abelson 1977). Using a variety of metaphors, theorists from various disciplines have suggested that life plans are scripted on fairy-tale patterns (Berne 1973), that "everyone is a novelist" (Dennett 1988), and that everyone has a "narrative identity" (Ricoeur 1991). Roger Schank (1995) considers memory an indexed database of stories that are retrievable by the process of "reminding," and Mark Turner claims that "most of our experience, our knowledge, and our thinking is organized as stories" (1996, i). Mental narrativizing devices were in fact already noted by Cohn (1978, ch. 6) in her analysis of "memory monologues," and the currently surging interest in mental narratives may well provide the necessary impulse to readdress the narrato-logic of "subjective analepses," dream narratives, and hypofocalizations (Bal 1983).

          Most of the features isolated above, and indeed most of the postclassical narratological branches mentioned above, can be accommodated within the framework of a "natural" narratology as proposed by Fludernik (1996). Summed up very briefly, Fludernik's model exploits real-world cognitive frames and scripts -- specifically, schemata of experiencing, telling, and reflecting -- as units of theoretical description (1996, ch. 1.3). According to this approach, there is a prototypical narrative situation of conversational storytelling which provides a default frame for concepts like narrative communication, story and discourse, narrators and focalizers, pragmatic contracts, tellability conventions, good form, good performance, and so on. At this level of naturalization, the narrator unabashedly "speaks," the narrative "represents" a fictional world, the mimetic illusion overrides constructivist skepticism, and characters exist as pragmatically and psychologically real beings. Situatedness, in this model, locates the narrator in a discourse here-and-now, the recipient in a reception here-and-now (in the audience, in front of the text), and the reflector in the story here-and-now (due allowance has to be made for the wider senses of "story" and "discourse" alluded to above). A special set of standard deviations from these primary orientations is exemplified by what Bühler (1965) called "transpositions to the Phantasma," that is, shifts to second or third level deictic coordinates. For instance, reflectors may phase out to or return from daydreams or recollections, while narrators may imaginatively transpose to the story here-and-now (description, hypothetical focalization, Herman 1994), or adopt a reflector's view of events (reflector-mode narrative, "delegated" focalization). Readers, in turn, may imaginatively hear the narrator speak and adopt (or possibly distance themselves from) the narrator's point of view.
          Despite its reliance on real-life schemata of "experientiality" (Fludernik 1996, 28-30), natural narratology ultimately rests on a level of constructivist metareflection which views all natural frames and scripts, including those related to the narratological toolbox itself, not as representative of signifieds in a world out there but as interpretive strategies generating Peircean interpretants. Here is an example of how, in Fludernik's model, traditional narratological categories acquire their constructivist underpinnings:

[O]ne can now comprehend Stanzel's narrative situations as a direct development from natural categories. Fiction with a teller figure evokes situational real-life equivalents of telling and their characteristic constellations. If there is a personalized narrator, for example, a certain cognitive, ideological, linguistic and sometimes even spatio-temporal position may become attributed to that narrator, and she becomes a 'speaker' on the model of the standard communication script. One can thereby explain the entire communicative analysis of fiction as an (illicit) transfer of the frame of real-life conversational narrative onto literary personae and constructed entities. (1996, 47)
Apart from the "constructed entities" mentioned in the last line of the quote, the true constructivist watchword in this passage is "illicit." Like Fish, the natural narratologist wants to eat the cake of natural frames, and at the same time exercise the option to view them from a skeptical distance. But while Fish's account founders in paradox and incoherence, natural narratology maps out a heterogeneous belief space that protects its cognitive frames under an overarching constructivist purview.

Conclusion

What role can Fish play in postclassical narratology? The easy answer to this is, None. Fish himself would refuse to lend his name to the narratological project, no matter whether it styles itself classical or postclassical, and the notion of an anti-foundationalist narratology is a contradiction in terms. Moreover, reviewing Fish's argument from a narratological vantage, one must acknowledge that there are lasting and unbridgeable disagreements, especially concerning the feasibility of theory, the structure of consciousness, and the status of unreflective actions. Yet apart from these areas of disagreement, it is remarkable how fittingly Fish's philosophy of reading ties in with current narratological concerns, and how well it paves the road to a fruitful exchange of concepts and models with disciplines such as pragmatics, cognitive linguistics, and artificial intelligence. So perhaps the question should be rephrased. If one cannot let Fish have the last word, would it be feasible to allow him the first?
          This time the answer is Yes. There is no-one, to begin with, who argues the case for constructivism better than Fish does (even if, ultimately, he does not wholly embrace it himself). It is no disadvantage, either, that Fish is a jargon-free writer who can be as lucid as John Austin and as relentlessly probing as Ludwig Wittgenstein. Even his occasional bouts of casuistry can be excused as springing from a laudable rhetoric of "telling you, by way of examples" (1989, 159). Opening up the literary dimension of a field that is growing like a tree (or, some critics would say, like a disease), Fish instructs the reader in constructivist conceptions of mental contexts, situatedness, the nature of truth, fact, and reality, and interpretive communities. Hopefully, the reader will also learn from Fish's errors -- his proposal to abolish theory, his attempt to argue self-conscious awareness out of existence, and his fractalist model of consciousness. Owing him this -- the lessons of his insights and the lessons of his oversights -- the least the reader can do is drink him a toast. To Stanley Fish: he opened the window and in flew Enza.9


Notes

          1. One question is whether the reading process is actually open to this type of retrospective inspection. As Culler points out (1981, 130), it is also not clear whether Fish's account captures both an expert's and a general reader's reading experience.
          2. More technically, the distinction is one between co-text (verbal context) and context (situational context). The point made here is that the holistic formula W__Z applies to situational but not to verbal context.
          3. Consider Socrates has eight letters (Searle 1974, 73) -- true, false, or undecideable?
          4. Deviating slightly from von Foerster (1993), Fig. 1 uses a simplified version of the nontrivial machine model.
          5. See Jackendoff (1987, 116) on the perception of ambiguous visual and linguistic data, and Jahn (1997) for a discussion of literary analogues.
          6. See Bode (1996) for an extended discussion of Fish's argument against the possibility (or effectiveness) of theory and self-conscious awareness.
          7. See especially Knapp and Michaels (1982) and Parrinder (1987). Instructively, Knapp and Michaels exempt narratology from their attack against theory because it is "essentially empirical" and has "no direct bearing on the interpretation of individual works" (1982, 11) -- a view few narratologists today would be prepared to accept.
          8. The current interest in cognitive matters is also evidenced by an MLA-hosted discussion group on "Cognitive Approaches to Literature." See web site humanitas.ucsb.edu/users/steen/Culture/ for details.
          9. The pun comes from "A Litter to Mr. James Joyce" by "Vladimir Dixon" (Our Exagmination Round his Factification for Incamination of Work in Progress, ed. Sylvia Beach, New York: New Directions, 1962, p. 193).


References

Adams, Jon-K. 1996. Narrative Explanation: A Pragmatic Theory of Discourse. Frankfurt: Lang.

Austin, John L. 1962. How to Do Things with Words, edited by J. O. Urmson. Oxford: Clarendon.

Baars, Bernard J. 1997. In the Theater of Consciousness. Oxford: Oxford Univ. Press.

Bal, Mieke. 1983. "The Narrating and the Focalizing: A Theory of the Agents in Narrative." Style 17: 234-69.

---. 1985. Narratology: Introduction to the Theory of Narrative. Translated by Christine van Boheemen. Toronto: University of Toronto Press.

Banfield, Ann. 1982. Unspeakable Sentences: Narration and Representation in the Language of Fiction. London: Routledge.

Barthes, Roland. 1975 [1966]. "An Introduction to the Structural Analysis of Narrative." New Literary History 6: 237-72.

Beaugrande, Robert de. 1983. "Surprised by Syncretism: Cognition and Literary Criticism Exemplified by E. D. Hirsch, Stanley Fish, and J. Hillis Miller." Poetics 12: 83-137.

Berne, Victor. 1973. What do You Say After You Say Hello. New York: Bantam.

Bever, Thomas G. 1970. "The Cognitive Basis For Linguistic Structures." In Cognition and the Development of Language, edited by John R. Hayes, 279-361. London: Wiley.

Bickerton, Derek. 1995. Language and Human Behavior. Seattle: Univ. of Washington Press.

Bode, Christoph. 1996. "Why Theory Matters." In Why Literature Matters: Theories and Functions of Literature, edited by Rüdiger Ahrens and Laurenz Volkmann, 87-100. Heidelberg: Winter.

Brewer, Derek. 1994. "Abuses of English Studies." The European English Messenger 3.2: 40-46.

Bühler, Karl. 1965 [1934]. Sprachtheorie: Die Darstellungsfunktion der Sprache. 2nd ed. Stuttgart: G. Fischer.

Chatman, Seymour. 1969. "New Ways of Analyzing Narrative Structure with an Example from Joyce's Dubliners." Language and Style 2: 3-36.

---. 1978. Story and Discourse: Narrative Structure in Fiction and Film. London: Cornell Univ. Press.

---. 1986. "Characters and Narrators: Filter, Centre, Slant, and Interest-Focus." Poetics Today 7: 189-204.

---. 1990. Coming to Terms: The Rhetoric of Narrative in Fiction and Film. Ithaca: Cornell Univ. Press.

Cohn, Dorrit. 1978. Transparent Minds: Narrative Modes for Presenting Consciousness in Fiction. Princeton: Princeton Univ. Press.

---. 1990. "Signposts of Fictionality: A Narratological Perspective." Poetics Today 11: 775-804.

Churchland, Paul M. 1993. "On the Problem of Truth and the Immensity of Conceptual Space." In Realism and Representation, edited by George Levine, 44-69. Madison: Univ. of Wisconsin Press.

Colomb, Gregory G. and Mark Turner. 1989. "Computers, Literary Theory, and Theory of Meaning." In The Future of Literary Theory, edited by Ralph Cohen, 386-410. London: Routledge.

Culler, Jonathan. 1975. Structuralist Poetics: Structuralism, Linguistics and the Study of Literature. London: Routledge.

---. 1980. "Fabula and Sjuzhet in the Analysis of Narrative." Poetics Today 1: 27-37.

---. 1981. "Stanley Fish and the Righting of the Reader." The Pursuit of Signs: Semiotics, Literature, Deconstruction. Ithaca: Cornell Univ. Press. 119-131.

Dennett, Daniel C. 1988. "Why everyone is a novelist." Times Literary Supplement September 16-22: 1016-1028.

Dijk, Teun A. van and Walter Kintsch. 1983. Strategies of Discourse Comprehension. New York: Harcourt Brace.

Doyle, Laura. 1994. Bordering on the Body: The Racial Matrix of Modern Fiction and Culture. Oxford: Oxford Univ. Press.

Duchan, Judith F., Gail A. Bruder, and Lynne E. Hewitt, eds. 1995. Deixis in Narrative: A Cognitive Science Perspective. Hillsdale: Erlbaum.

Empson, William. 1949. Seven Types of Ambiguity. London: Chatto & Windus.

Fish, Stanley. 1980. Is There a Text in This Class? The Authority of Interpretive Communities. Cambridge, Mass.: Harvard Univ. Press.

---. 1989. Doing What Comes Naturally: Change, Rhetoric, and the Practice of Theory in Literary and Legal Studies. Oxford: Clarendon.

---. 1995. Professional Correctness: Literary Studies and Political Change. Oxford: Clarendon.

Fludernik, Monika. 1993. The Fictions of Language and the Languages of Fiction: the Linguistic Representation of Speech and Consciousness. London: Routledge.

---. 1996. Towards a 'Natural' Narratology. London: Routledge.

Fodor, Janet Dean and Atsu Inoue. 1994. "The Diagnosis and Cure of Garden Paths." Journal of Psycholinguistic Research 23: 407-34.

Foerster, Heinz von. 1993. "Mit den Augen des anderen." In Wissen und Gewissen, edited by Siegfried J. Schmidt, 350-63. Frankfurt: Suhrkamp.

Genette, G‚rard. 1980 [1972]. Narrative Discourse. Translated by Jane E. Lewin. Oxford: Blackwell.

---. 1988 [1983]. Narrative Discourse Revisited. Translated by Jane E. Lewin. Ithaca: Cornell Univ. Press.

Glowinski, Michal. 1977. "On the Fist-Person Novel." New Literary Histotry 9: 103-114.

Gove, Philip B. 1966. "Self-explanatory Words." American Speech 41: 192-98.

Hamburger, Käte. 1977 [1957]. Die Logik der Dichtung. 3rd ed. Stuttgart: Klett.

Harris, Thomas A. 1967. I'm OK -- You're OK. London: Pan.

Hempel, Carl G. 1966. Philosophy of Natural Science. Englewood Cliffs: Prentice-Hall.

Herman, David. 1994. "Hypothetical Focalization." Narrative 2: 230-53.

---. 1999. "Introduction: Towards a Post-classical Narratology." In Narratologies: New Perspectives on Narrative Analysis, edited by David Herman. Ohio: Ohio State Univ. Press.

Herring, Henry D. 1998. "Constructivist Narratology."

Hockett, Charles F. 1961. "Grammar for the Hearer." In Structure of Language and its Mathematical Aspects, 220-36. Providence: American Mathematical Society.

---. 1967. Language, Mathematics, and Linguistics. Paris: Mouton.

Hutcheon, Linda. 1988. A Poetics of Postmodernism: History, Theory, Fiction. London: Routledge.

Iser, Wolfgang. 1978. The Act of Reading: A Theory of Aersthetic Response. Baltomore: Johns Hopkins Univ. Press.

Jackendoff, Ray. 1987. Consciousness and the Computational Mind. London: M.I.T. Press.

---. 1997. The Architecture of the Language Faculty. Cambridge, Mass.: M.I.T. Press.

Jahn, Manfred. 1997. "Frames, Preferences, and the Reading of Third-Person Narratives: Towards a Cognitive Narratology." Poetics Today 18: 441-68.

---. 1998. "Package deals, Exklusionen, Randzonen: das Phänomen der Unverläßlichkeit in den Erzählsituationen." In Unreliable Narration: Studien zur Theorie und Praxis unglaubwürdigen Erzählens in der englischsprachigen Erzählliteratur, edited by Ansgar Nünning, 81-106. Trier: WVT.

---. 1999. "'Speak, friend, and enter': Garden Paths and their Narratological Implications." In Narratologies: New Perspectives on Narrative Analysis edited by David Herman, 000-000. Ohio: Ohio State Univ. Press.

Joyce, James. Ulysses. A Critical and Synoptic Edition, edited by Hans Walter Gabler. New York: Garland, 1986.

Johnson-Laird, Philip N. 1981. "Mental Models of Meaning." In Elements of Discourse Understanding, edited by Aravind K. Joshi, Bonnie L. Webber, and Ivan A. Sag, 106-26. Cambridge: Cambridge Univ. Press.

Knapp, Steven and Walter Benn Michaels. 1982. "Against Theory." In Against Theory: Literary Studies and the New Pragmatism, edited by W. J. T. Mitchell, 11-30. Chicago: Univ. of Chicago Press.

Lanser, Susan Sniader. 1981. The Narrative Act: Point of View in Prose Fiction. Princeton: Princeton Univ. Press.

---. 1992. Fictions of Authority: Women Writers and Narrative Voice. London: Cornell Univ. Press.

Lehnert, Wendy G. 1979. "The Role of Scripts in Understanding." In Frame Conceptions and Text Understanding, edited by Dieter Metzing, 79-95. Berlin: de Gruyter.

McClary, Susan. 1997. "The Impromptu That Trod on a Loaf: Or How Music Tells Stories." Narrative 5: 20-35.

McNeill, Daniel and Paul Freiberger. 1993. Fuzzy Logic. New York: Simon & Schuster.

Mey, Jacob L. 1993. Pragmatics: An Introduction. Oxford: Blackwell.

Mezei, Kathy, ed. 1996. Ambiguous Discourse: Feminist Narratology and British Women Writers. Chapel Hill: Univ. of North Carolina Press.

Minsky, Marvin. 1979 [1975]. "A Framework for Representing Knowledge." In Frame Conceptions and Text Understanding, edited by Dieter Metzing, 1-25. New York: de Gruyter.

Morris, Charles. 1964. Signification and Significance: A Study in the Relations of Signs and Values. Cambridge, Mass.: M.I.T. Press.

Nünning, Ansgar. 1990. "Bausteine einer konstruktivistischen Erzähltheorie: die erzählerische Umsetzung konstruktivistischer Konzepte in den Romanen von John Fowles." Delfin 7.1: 3-17.

---. 1992. "Mimesis und Poiesis im englischen Roman des 18. Jahrhunderts: Überlegungen zum Zusammenhang zwischen narrativer Form und Wirklichkeitserfahrung." In Neue Lesarten, neue Wirklichkeiten: Zur Wiederentdeckung des 18. Jahrhunderts, edited by Gerd Stratmann and Manfred Buschmeier, 46-69. Trier: Wissenschaftlicher Verlag.

---. 1999. "On the Perspective Structure of Narrative Texts: Steps towards a Constructivist Narratology." Papers of the 1997 Utrecht conference, edited by Seymour Chatman and Will van Peer.

Parrinder, Patrick. 1987. The Failure of Theory: Essays on Criticism and Contemporaray Fiction. Brighton: Harvester.

Peirce, Charles Sanders. 1965. Principles of Philosophy. Cambridge, Mass.: Belknap.

Perry, Menakhem. 1979. "Literary Dynamics: How the Order of a Text Creates Its Meaning." Poetics Today 1-2: 35-64; 311-361.

Prince, Gerald. 1982. Narratology: The Form and Functioning of Narrative. New York: Mouton.

Raleigh, John Henry. 1977. The Chronicle of Leopold and Molly Bloom. Berkeley: Univ. of California Press.

Ricoeur, Paul. 1984-1988. Time and Narrative. Three Volumes. Translated by Kathleen McLaughlin and David Pellauer. Chicago: Chicago Univ. Press.

---. 1991. "Life in Quest of Narrative." On Paul Ricoeur: Narrative and Interpretation, edited by David Wood. London: Routledge. 20-33.

Ronen, Ruth. 1994. Possible Worlds in Literary Theory. Cambridge: Cambridge Univ. Press.

Ryan, Marie-Laure. 1981. "The Pragmatics of Personal and Impersonal Fiction." Poetics 10: 517-539.

---. 1991. Possible Worlds, Artificial Intelligence, and Narrative Theory. Bloomington: Univ. of Bloomington & Indianapolis Press.

Schank, Roger C., 1995. Tell Me a Story: Narrative and Intelligence. Evanston: Northwestern Univ. Press.

---, and Robert P. Abelson. 1977. Scripts, Plans, Goals and Understanding: an Inquiry into Human Knowledge. Hillsdale, N.J.: Lawrence Erlbaum.

Schmidt, Siegfried J. 1989. "Past: Notes on Memory and Narration." Empirical Studies of the Arts 7: 191-203.

---. 1991. "Text Understanding -- A Self-organizing Cognitive Process." Poetics 20: 273-301.

Searle, John R. 1974. Speech Acts: An Essay in the Philosophy of Language. London: Cambridge Univ. Press.

Sperber, Dan, and Deirdre Wilson. 1981. "Irony and the Use-Mention Distinction." In Radical Pragmatics, edited by Peter Cole, 295-318. London: Academic Press.

--- and ---. 1986. Relevance: Communication and Cognition. Oxford: Blackwell.

Stanzel, Franz K. 1990. "Textual Power in the (Short) Short Story and Poem." In Modes of Narrative: Approaches to American, Canadian and British Fiction. Festschrift presented to Helmut Bonheim, edited by Reingard M. Nischik and Barabara Korte, 20-30. Würzburg: Königshausen und Neumann.

Sternberg, Meir. 1978. Expositional Modes and Temporal Ordering in Fiction. London: Johns Hopkins Univ. Press.

Todorov, Tzvetan. 1969. Grammaire du D‚cameron. Mouton: The Hague.

---. 1981. Introduction to Poetics. Brighton: Harvester.

Turner, Mark. 1996. The Literary Mind. Oxford: Oxford Univ. Press.

Warhol, Robyn R. 1989. Gendered Interventions: Narrative Discourse in the Victorian Novel. New Brunswick: Rutgers Univ. Press.

---. 1999. "Forms, Feelings, Feminisms: What Feminist Narratology Can Do for Cultural Studies". In Narratologies: New Perspectives on Narrative Analysis, edited by David Herman. Ohio: Ohio State Univ. Press.