Speech Act Theory and the Concept of Intention in Literary Criticism more

SPEECH ACT THEORY AND THE CONCEPT OF INTENTION IN LITERARY CRITICISM Jose Angel Garcia Landa Universidad de Zaragoza The aim of this paper is to trace the outline of a speech act theory of literature, taking into account the work of critics who react against the prevailing anti-intentionalist schools of criticism (such as the New Criticism, some versions of structuralism, and deconstruction). These critics prepare the ground for a theory of literary discourse considered as a speech act, since it is known that the concept of intention is central to the analysis of speech acts. Such a theory of literary discourse should not overlook the role played by criticism in expanding the meaning of a text, and should go beyond a narrow definition of interpretation as the identification of the author's intention. 1. SPEECH ACTS: LOCUTIONS In J. L. Austin's philosophy of language, "locution" is a loose term which includes all those aspects of utterances which were studied by traditional grammar: the phonetic, morphological, syntactic and semantic characterization of an utterance. The study of illocutions and perlocutions, on the other hand, would be the competence of pragmatics, according to the conceptual divisions of Peirce and Morris (cf. Lyons 1977, 114ff.). Here we shall take for granted the physical and syntactic nature of the utterance, and whenever we speak of locutions we shall concentrate on their semantics, understanding by locutionary act the transmission of a series of meanings from the speaker to the hearer by means of that physical and syntactic enunciation (Bach and Harnish, 1-10). It must be kept in mind that in speaking of communication as a transmission of meanings we are referring to the locutionary level of communication. Locutionary meaning is an intentional phenomenon which requires the assumption of shared semiotic types (phonemes, words, sentence patterns) in both speaker and listener. That is, it presupposes the existence of a linguistic convention and the intentional patterns which are to enable the use of that convention. It is easy Revista Canaria de Estudios Ingleses, ns 24, 1992, pags. 89-104 90 Jose A. GarcIa Landa to mistake this conventionalized intentionality for the psychical act of actually intending. Actually, once the conventions apply, a piece of language may signify and be interpreted as intentional quite apart from the precise ideas in the speaker's mind: in language, intentionality grounds the rules of the game, not just individual performances. Searle (1983, 167-168) distinguishes in the analysis of meaning an intention to represent and a communication intention. I understand this communication intention present in locutionary meaning to be distinct from the communicative intention involved in the performance of a speech act as a whole. The first is a constituent part of a locutionary act; the second pertains to an illocutionary act. Just as the enunciation of syntagms is instrumental to the production of a locutionary act, the locutionary act is a means of producing an illocutionary act. This act consists in a socially codified act of communication (e.g. a statement, a promise, etc.) which takes place only if certsanfelicity conditions are satisfied: for instance, only future actions can be promised. The felicity conditions vary from one kind of act to another and serve to identify them. Illocutionary acts, unlike the propositions of locutionary acts, do not have a truth value: they have what Austin calls an illocutionary force. There is a certain relationship, albeit a flexible one, between the forms of locutionary and illocutionary speech acts. According to Lyons, "There is no one-to-one relationship between grammatical structure ... and illocutionary force; but we cannot employ just any kind of sentence in order to perform any kind of illocutionary act" (1978,733). Or, rather, we cannot use it in any circumstance with the same facility. A sentence carries along with it a built-in, ideal context which defines its meaning, and which interacts with the actual context in which the sentence is uttered. Alston (1964) says as much when he defines locutionary meaning with respect to illocutionary force. In his view, the meaning of a sentence is its "illocutionary-act potential." We can think of a sentence as being designed to perform one particular kind of illocutionary act, even if this original force can be redirected in a context different from the potential context implied in the locution. According to Harris (1988, 21), literal meaning goes beyond dictionary meaning. In the case of the word, literal meaning takes into account the associative possibilities of the word, the contexts where it is usually found. The notion of literal meaning at sentence-level is more difficult to define. It already depends on a context, which must be assumed (this is what Harris calls a "wonted context"). Only the equal probability of two or more wonted contexts brings about ambiguity in the literal meaning of the sentence. Accordingly, Harris defines the literal meaning of a sentence as "that which the native speaker of the language would be most likely to assign to it if assuming it to occur in the most usual context" (1988, 23). This may do in many cases as an interpretative instrument. We should not forget, however, that "the native speaker of the language" is an abstraction without age, sex, race or class. Different (types of) speakers may in some cases differ as to which is the most usual context in which the sentence would, appear, since they move in different contexts. Speech Act Theory and the Concept .. 91 The relationship between literal meaning (or locutionary meaning) and illocutionary force creates a difference between direct and indirect speech acts. An illocutionary act is direct if its meaning is not in conflict with its illoctionary force (e.g., a question couched as a question). In an indirect illocutionary act, there is a difference between the overt illocutionary force and the real illocutionary force (e.g., a polite request formulated as a question). Some indirect acts are highly standardized and may be used in any context; others rely on specific knowledge (often of a situational kind) shared by the speaker and the hearer. When defined at the illocutionary level, communication is no longer a transmission of semantic meaning: it is an exchange of speech acts.1 This illocutionary communication consists in a recognition of the kind of speech act which is being produced. Indeed, as Austin noted, "the performance of an illocutionary act involves the securing of uptake" (1980, 117). The sender must make sure that the receiver identifies the speech act in question. This involves the recognition on the part of the hearer of the speaker's intention to make him recognize the speech act he recognizes.2 Therefore, the performance of a speech act is inherently tied to intentionality: a standardized intentionality which is communicated together with the locutionary content. It makes all the difference in the world to speech act theory if this brand of intentionality is involved in a speech act, because this allows us to tell an illocutionary from a perlocutionary speech act. Those speech acts or those aspects of a speech act which fall outside this communicative intentionality are called by Austin perlocutionary acts. By means of his locutionary and his illocutionary speech act, the sender influences the receiver in some way; he causes in the receiver a perlocutionary effect, or perlocution. This process may also involve intentionality on the part of the sender, but, unlike the intentionality present in illocutions, the perlocutionary intent of the speaker need not be recognized by the hearer in order for the perlocutionary act to take place. For instance, when we lie we have the perlocutionary intention of deceiving the hearer, and the lie is a lie whether this intention is recognized or not.3 That is, the perlocutionary intent is not structurally standardized. The locutionary element of a literary utterance has a more problematic status than the simple proposition which can be isolated in an isolated sentence, and the neatness of Austin and Searle's categories is considerably compromised as the ' different levels of verbal actions become hierarchically structured and instrumentalised within more comprehensive discoursive acts. A sentence is an abstraction, while a text cannot help being somewhat more concrete: links appear between the sentences; a virtual communicative situation, with a speaker and a hearer, takes a more definite shape. Therefore, the study of locutions at textual level cannot be isolated from the study of illocutions. Indeed, both locutions and illocutions are abstractions, analytical phases in the study of verbal action; the study of locutions simply implies a further step in abstraction, since locutions become manifest only in and through illocutions. It is obvious that a speech act theory of literature needs this adaptation of speech act concepts to the level of discourse. 92 Jose A. GarcIa Landa Austin and Searle's original theory is unable to cope with literature, as should be clear from their own discussions of fictionality and literary speech acts. 4 Whole areas of the textual structure, such as the fictional narrative in a story, may seem to be primarily locutionary in character: after all, the narrative is a series of related meanings communicated by the narrator to the narratee, a macro-Jocution which can be regarded as a means to accomplish the macro-illocutionary act of telling him a story. But the comprehension of this macro-locutionary act already presupposes many illocutionary maneuvers: for instance, the utterances of the characters must be understood in their full communicative value for the text to have a meaning at all. However, it is useful to consider the different layers of the narrative structure as being the discourse-level counterparts of the locutionary, illocutionary and perlocutionary phases of the basic speech acts described by Austin. The narrative, then, can be thought of as a locution, with big imaginary scare-quotes. Its transmission from the narrator to the narratee is a fictional illocution, and the composition of a novel or a story by the author is a real illocution. Most of the work on literature and speech acts done so far is an attempt to determine the precise status of the illocutionary acts of the narrator and the author, and the role of the reader/interpreter in the literary process. Needless to say, most of this discussion is carried on without using the terminology of speech act theory. But whenever there is an effort at precision, the critical debates on the status of interpretation begin to look like a textual extension of speech act theory. For instance, E. D. Hirsch's theory of interpretation is based on the sharability of verbal meaning. This view is in principle consonant with speech act theory. However, speech act theory further specifies the notions of sharability and of verbal meaning. According to Hirsch's definition, "verbal meaning is whatever someone has willed to convey by a particular sequence of linguistic signs and which can be conveyed (shared) by means of those linguistic signs" (1967, 31). This is loose enough to include both locutions and illocutions, and maybe even perlocutions. More specifically, verbal meaning is "a willed type which the author expresses by linguistic smbols and which can be understood by another through these symbols" (1967,49). From the point of view of speech act theory, this is on the right tack, but still not specific enough. What characterizes locutions and illocutions and sets them apart from perlocutions is that the linguistic type in question is willed to be conventionally sharable —that intentions must be publicly specified. Hirsch gets close to defining something like that: "The willed type must be a shared type in order for communication to occur" (1967, 66); but still the illocutionary level of analysis is not clearly defined in his theory; it occupies an indeterminate area between the locutionary meaning he is primarily thinking of when he defines verbal meaning, and the "intrinsic genre," which is something like a macro-illocution. Speech Act Theory and the Concept ., 93 2. ILLOCUTIONS An opposition between an intrinsic and an extrinsic meaning of the concept of "authorial intention" can be found in hermeneutic theories as apposed as Wimsatt and Beardsley's (1967) or Hirsch's. The development of an mtrinsic theory of intention in literary analysis can benefit from the insights of speech act theory. Many of the early definitions of fiction, or literature, use a concept of intention which could be formulated in pragmatic terms. Let us remember, for instance, Sir Philip Sidney's famous distinction between the poet and the liar: The poet, he nothing affirms, and therefore never lieth. For, as I take it, to lie is to affirm to be true which is false ... But the poet (as I said before) never affirmeth. The poet never maketh any circles about your imagination, to conjure you to believe for true what he writes.... And therefore, though he recount things not true, yet because he telleth them not for true, he lieth not. (1973,124) In Speech Act Theory parlance, we should say that lying is not an illocutionary speech act, since a he does not require the hearer's uptake of the speaker's intention of lying in order to count as a lie. But literary fiction is a kind of (derived) illocutionary act, since in order to count as fiction we must recognize it as such —we must recognize the intentional design of the author, the genre to which a work belongs. As Jon-K. Adams puts it, "fiction is defined by its pragmatic structure, and, in turn, this structure is a necessary part of the interpretation of fiction" (1985,2). Other classical accounts of the receiver's basic interpretive assumptions, such as Coleridge's "willing suspension of disbelief (1956,168) follow much the same lines.5 Even a critic like I. A. Richards, who usually blurs the difference between illocutions and perlocutions, preserves an implied version of the generic illocution when he differentiates scientific language from poetic language. Scientific language is characterized by a predominance of referential meaning, while in poetic language this meaning is conventionally neutralized, and only feelings remain significant: When this happens, the statements which appear in the poetry are there for the sake of their effects upon feelings, and not for their own sake. Hence to challenge their truth or to question whether they deserve serious attention as statements claiming truth, is to mistake their function. 6 Like Richards's theory of the literary "pseudo-statements," many modern accounts of literary fiction presuppose such shared conventions. Among them are Ingarden's theory of the "quasy-judgmental" status of literary works (1973, 131), Northrop Frye's account of the hypothetical nature of literary works (1957, 79, 84f.), Felix Martinez Bonati's description of literary statements as "pseudo-phrases" (1972, 216) and Carlos Castilla del Pino's conception of a presupposition of fictionality on the author's part which is subsequently inferred by the reader (1983, 321). 94 Jose A. GARCfA Landa Hirsch's notion of "intrinsic genre" takes the quasi-illocutionary definition of discursive activities one step further. The philosophical basis of Hirsch's intrinsic genres is Wittgenstein's notion of "language game" (1958, section 7). An intrinsic genre is "that sense of the whole by means of which an interpreter can correctly understand any part in its determinacy" (Hirsch 1967, 83). "Correctly".must be understood in the light of Hirsch's notion of communication as sharability of types: intrinsic genres are not purely heuristic constructions, because they are also constitutive of the author's activity —of writing, not only of reading. Hirsch correctly assumes that these genres do not have any permanent essence, that they may be created, be transformed or disappear (cf. Wittgenstein 1958, section 23), so that they must be studied from a historicist viewpoint. Also, intrinsic genres cut across Saussure's distinction between langue and parole, and cannot be properly subsumed under any of these categories (Hirsch 1967, 111). The same is true of illocutionary acts. And yet, as we have noted, Hirsch's intrinsic genres are more vaguely defined than Austin's speech acts, and above all they have a macrostructural import. They are macro-illocutionary acts avant la lettre. Explicit attempts to develop speech act theories of literature have been unequal and wavering. At first these attempts were very simple analyses of performativity in dialogue, a direct application of Austin's initial thrust in How To Do Things with Words. The best instance of this kind of criticism might be Stanley Fish's analysis of Coriolanus as a "speech act play" (Fish 1976). More interestingly, there have been attempts to formulate the conception of textual illocution from a linguistic perspective, extending towards the realm of literature the analyses of oral communication made by Austin, Grice and Searle. These definitions are a continuation of the ones we have mentioned, rather than a completely new outlook on the matter. They usually rely on the notion of the "hypothetical" or "Active" speech act.7 Some of the early definitions are defective because of a simplistic conception of the intentional moves required to perform an illocutionary act. Among these are Ohmann's (1987a) and Searle's (1975). In spite of some useful insights, Ohmann's (1971) account of the illocutionary nature of narrative and fiction is blurred; above all, it does away with the notion of uptake (cf. Fish's criticism, 1976). This same defect is present in the early attempts at synthesis by Close (1972), Skinner (1972), and Hough (1973). Close notes that "a major cause of confusion is to see intention as a 'mental objective,' connected with action contingently or causally but not logically" (1976, 176). In his view, there is another, more important sense in which authorial intention is internal to the work: illocutionary force.8 Close favours a common basis for dealing with intention in both literature and ordinary language; literary utterances, just like the conversational utterances studied by Austin or Strawson, are speech acts (1976, 183). Close's conception, however, is still vague. It is not clear whether the literary speech acts he mentions ("allegorizing, parodying, using dialogue") are meant to be speech acts in the same sense, or at the same level, as Austin's directives or behavitives; in Close's view the varieties of speech acts are "so diffuse that one Speech Act Theory and the Concept .. 95 cannot envisage them all falling under descriptive identifiable levels ...; yet they display diverse family-likeness and are consequently capable of description by a process of analogy and contrast" (1976, 187). Hough's use of speech act theory is similar to Close's, and it shares its limitations. He distinguishes the approach which studies the illocutionary acts inside the work from that which considers the work as a whole as an illocution: "But can anything as complex as a whole poem constitute an illocutionary act in Austin's sense? If so, it ought to be possible to say what it is doing —asking questions, warning, admonishing, describing, or what not" (1976, 226). What Hough and Close do not see is that writing a poem is the illocutionary act performed by the author in writing a poem. This failure to see literary discourse and literary genres as illocutionary types of activity is common to other early attempts to translate speech act theory into the realm of literary study, such as Ohmann's (1971; 1987a) and Searle's (1975). They define fictional discourse as a sum of "hypothetical" or "fictive" speech acts, rather than as a comprehensive speech act with its own characteristics. Ohmann's later versions of his theory are somewhat more satisfactory. Although he does not solve the problems set by his earlier defective definitions of the literary speech acts, he proposes the idea of "literature" as a specific type of illocutionary act (1987b, 47). But his account is still vitiated by his persistence in an uncritical identification of literature with fiction. Searle's speech act theory of fiction also fails to see in it a specific kind of illocutionary act. According to Searle, the author of a work of fiction feigns that he is performing speech acts, within generic conventions which suspend the illocutionary rules which could normally link his speech acts to reality. In the case of first-person narration, the author feigns he is a character who performs (real) illocutionary speech acts (1975,326). Searle holds that there are no formal traces of this fictionality: the difference between fiction and non-fiction is purely intentional. Searle's contention, however, is not valid. It relies on an angelic conception of literary form, one which ignores the actual contexts in which forms are recognized. The formal marks of fiction include such obvious and down-to-earth reminders as the words "story" or "novel" on the cover of the book, not to mention the conventions of subject matter and style, and the difference in the contexts (both physical and discursive) in which fiction and non-fiction are read. These "marks" provide us with the essentials of interpretation. A theory such as Searle's leaves unanswered the question of how, without any formal marks, the author manages to convey the idea that he is the only feigning the performance of an illocutionary act. Searle does propose a solution, but an absurd one: the author conveys his meaning by performing a locutionary speech act. This does not explain the difference between real and feigned illocutionary acts, since the speaker must always perform a locutionary act in order to perform an illocutionary act, feigned or real. Moreover, in Searle's theory (as in the early Ohmann) the author is not performing any real speech act, apart from those which are performed in an implied way, by means of the fictive speech acts of which the 96 jose A. GARCfA Landa work is made. These (the moral, the message, etc) are, however, not specifically literary in nature. Searle's theory does not recognize that the very act of performing fictional speech acts is a real speech act generically defined —writing fiction; more specifically, writing a novel or a play.9 The conclusion is therefore that an author can be said to be performing Active speech acts at the level of the story or of the fictional narrative, but these are only the means to perform a real speech act at the level of the discursive interaction between the textual author and the textual reader. 10 In the unmarked case of literary communication that we are considering, this speech act is a conscious literary performance. Pratt (1977) has argued this most forcefully, pointing out the similarities between conversational and literary narratives in their ultimate status as speech acts and in their constitutive elements. Searle's rejection of the idea that "telling a story" or "writing a novel" are illocutionary speech acts is understandable, since his conception of illocution is too rigid and abstract to describe real discoursive activity (in spite of its pretensions to do so). Searle's theory, like that of Austin, from which it derives, is based on a sentence grammar, and narrative or literature can only be understood in the terms of a text grammar. A sentence grammar is a highly abstract construct which in general ignores all contextual information save that which is inscribed in the locutionary meaning of the sentence. Therefore, Austin's and Searle's analyses of oral conversation, in spite of appearances, do not describe actual speech activity as such. They describe speech activity under highly standardized circumstances. Speech acts are always performed in a discursive situation: not by means of a sentence as such, but by means of a text uttered in specific circumstances. Therefore, Austin and Searle describe not speech acts as they are actually performed, but the Platonic models of speech acts —not real illocutions, but the atomic constituents of illocutions, the abstract illocutionary primitives which are instrumentalized to produce concrete illocutions in actual discourse. Concrete illocutions differ from primitive illocutions in two respects: — They must be defined at both a macrostructural and microstructural level. An extended, discursive or macrostructural speech act can be analyzed into partial, instrumental, sometimes sentential micro-speech acts. A novel, insofar as it is a narrative (macro-speech act), consists of a multitude of questions, exclamations, and above all assertions (all of them micro-speech acts). — They may be highly indirect. Austin and Searle only discuss those indirect speech acts which are so highly standardized as to be almost direct. 11 But a fictional assertion presupposes an analytical phase in which its hypothetical assertive nature is identified, just as the assertive nature of a real assertion is identified. That is, the uptake of the (virtually present) real-assertion-element is a previous condition for the uptake of the actual speech act, the fictional assertion. A whole scale of priority could be developed in this way. A novel presupposes (and goes beyond) the conventions of factual narrative, just like factual narrative presupposes (and goes beyond) the conventions of the simple assertion. Under the simple illocutionary definition of literary genres there is a whole underlying Speech Act Theory and the Concept .. 97 structure of petrified, standardized or conventionalized intentional attitudes which make it possible and give each of these discourse activities its distinctive character. Skinner (1976, 218) identifies the author's meaning of the work to the illocutionary intention of the writer. Given the conception of genre as illocutionary type which we intend to develop, Skinner's conception is too restrictive. That is, unless we understand that the illocution includes the earlier locutionary phases of interpretation —the locutionary meaning of the sentence, and not merely the identification of its illocutionary force. We need, moreover, a theory which has a place for the macro-illocutions of discourse as well as for the micro-illocutions inside the story, such as are performed by the characters or the narrators. 12 It is to be noted that the two major works which develop a speech act theory of literature, Pratt's (1977) and Petrey's (1990) ignore the whole problem of literary voice as a fictional construct —a promising area for future literary pragmatics. And so far the attempts to couch the debate on interpretation in terms of speech act theories are merely tentative, although Petrey's contribution (1990, 131 ff.) leaves it clear that speech act theory will lay emphasis on how conventions create meaning in specific interpretive situations, rather than on the undecidability or the endless deferral of meaning favoured by current deconstruction. Still, closer pragmatic analyses of the concepts of uptake, understanding and interpretation as applied to literature must be developed. It is clear that there is a gap between the identification of the bare illocutionary intention and the understanding of the work. This is pointed out by Close and Hough (1976, 228). And there is another gap between understanding and criticism which is often overlooked —Hirsch's concern with significance as a crucial part of criticism should be remembered when some of these theories identify criticism and the hunt for illocutions. In Close's view, the essential aim of literary criticism is discovering the meaning of a work: "here 'meaning' has a fuller sense than either Austin's concept of locutionary meaning or the notion of paraphrase. It is equivalent to 'illocutionary force' (with the comprehension of locutionary meaning being taken for granted) and to what Grice has recently called «utterer's occasion-meaning.»" 13 This account does not include the determination of significance and, in spite of Close's precautions, establishes a rigid connection beween the recognition of intention and the response, between aesthetic response and communication. In his view, literary art... comes under the general category of 'communication' and ... the responses which literature seeks are securely anchored in utterer's intention. By this I do not mean that we find farces, for example, funny just because their creators intend them to be so. Rather I mean that our finding them funny is normally dependent on our prior recognition that they are intended as such. (1976,193) Close's account is only partially right. A reader may be in tune with the authorial intention and accept the role of the implied reader. For instance, an early reviewer 98 Jose A. GarcJa Landa of Stephen Crane's "The Monster" finds that "the comedy of the Dutch barber shop and of the negro dandy's call upon his sweetheart is irresistible" (Bridges 1898:166). Not many readers seem to have found these scenes funny enough to call attention to them; some readers do not consider them to be funny at all due to their racist overtones, even though they recognize that the author intended to create a comic scene (e. g. M. Solomon, 1956 n, 39). So far, so good. But the reverse case is also possible: I may find very funny a work (say, a Barbara Cartland novel) in which I do not recognize an intention to be funny on the part of the author. Close's speech act theory of criticism ignores that response is not just interpretive response —that I may successfully identify the illocutionary intent of an artistic message and yet respond to it in a way the author did not intend. It neglects perlocutions, which are a vital part of the critical enterprise. 3. PERLOCUTIONS Hough complains that the speech act analysis of literature does not account for all the elements of literary understanding. The intended meaning it is concerned with is easily recoverable, but this is only a diagram of what we need. A full interpretation cannot be accounted for in terms of intended illocutionary meaning. Recognizing that a convention is being used "can tell us nothing about the intention with which [the writer] employs [the convention]" (1976, 230). Hough seems close to Roger Fry when he affirms that the most important part of composition is not that which can be intentionally accounted for —it is a symptom, not a full deliberation: "mounted on that basic intentional structure is the whole inspectable surface of the poem in all its varied detail, much of which cannot properly be described as intentional" (1976, 233). The author's intention is merely one more element in the intertextuality of the work. The analysis of unintended elements is just as significant and pertaining to the work as that of intentions. Unintended elements, as well as the indefinite area which shades off into the intentional structures of locution and illocution, may be related to the perlocutionary phase of speech acts. Perlocution is left out of the conventional analysis of speech acts by both Austin and Searle: "Illocutionary acts are conventional acts: perlocutionary acts are not conventional" (Austin 1980: 121); "The perlocutionary effects of our utterances cannot be included in the conventions for the use of the device uttered, because an effect which is achieved by convention cannot include the subsequent responses and behavior of our audiences" (Searle 1983,178). Which should be the status of perlocutions in critical theory? The notion might be usefully developed in various senses. First of all, one related to the macro-illocutionary status of literary communication. Skinner is too rash when he affirms that "a writer's perlocutionary intentions (what he may have intended to do by writing in a certain way) do not need to be further considered" (1976, 217). Criticism is often concerned with perlocutionary intentions. That is why we have a concept of literary hoax, for instance: many of the intentions which critics of Speech Act Theory and the Concept .. 99 Macpherson, Chatterton or Defoe find relevant are not communicative intentions, but rather the literary equivalent of a lie. But these are nevertheless exceptional cases. We also need to identify the perlocutionary element in any kind of literary discourse. Close seems to deny the existence of such an element. Like Bateson (1953,16) before him, Close identifies the task of the critic with the interpretation of authorial meaning. Interpretation of meaning at large, evaluation and the interpretation of significance are secondary for him (Close 1976, 188). 14 It is not surprising that at one given moment of his account of literary communication, dealing with the effects of and reactions to literature, he divides his analysis from Austin's: "The favourable or unfavourable impact that art makes upon us is not, like Austin's 'perlocutionary' force, a set of causal effects conceptually separable from the identification of illocutionary force, but is part-and-parcel of it" (1976, 193). However, Austin's perlocution is subsequent to illocution, and is its consequence. The difference Close draws here between literature and ordinary communication results from an inadequate conception of the communicative role of illocutionary force. The effects literature has on us are not confined to the uptake of illocutions —and this is the case for the aesthetic response Close is concerned with as well as other kinds of response. Perlocution, like illocution, is a notion devised at a highly abstract level in Austin's analysis of speech acts. Its nearness must be expected to be somewhat blurred when we transport it from a sentence grammar to discourse analysis. Literary analysis makes particularly clear a phenomenon which is present in some measure in any text by virtue of its rhetorical structure: that the purity of macro-illocutions is compromised at discourse level because of the elasticity of the conventions which must be posited at this level, and consequently of the notion of uptake. Far from being neatly codified, as it is at the idealized and normative level of primitive illocutions, the uptake of macro-illocutions is fuzzy-edged, probabilistic and approximative. Micro-illocutionary uptake is consciously regulated and only becomes subliminal by virtue of the automatism of habit; it can always be brought to consciousness and stated explicitly. The rules of macro-illocutionary uptake, on the contrary, remain unconscious for the subject who applies them. Even in critical theory, which aims at formulating conscious theories of literary response, they remain an object of debate. Therefore, there is no sharp line between illocutions and perlocutions at discourse level; instead, their relationship is constantly being negotiated. A text, and most of all a literary text, is always redefining the codes that allow us to understand it, escaping automatism and convention, and therefore redefining the play of illocution and perlocution. Each phase of the sender's utterance has a corresponding activity in the reader if communication or understanding is to take place. The author's speech act must be complemented by the reader's interpretive act (cf. Harris 1988, x). Through its work on the cultural codes which constitute it, the text redefines the implied image of the reader, and stimulates the reader to transform his own codes in order to make the most of the text. The superior cultural intuition of a good writer uses the socially codified modes of intention and takes them one step r 100 Jose A. GarcIa Landa beyond, towards the unknown, playing them against each other so that a new significant relationship can be perceived between them. The activity of the writer is to a great extent one of hypercodification, of using already existing codes to create new codes, and of stimulating the same activity in the reader. 15 Such activity could be described at any level of analysis of the literary work. The most common instance found in practical analyses of literary works is the transmission of implied authorial attitudes through the use of symbolism. Whenever two sets of elements of the work are seen to maintain a relationship of analogy to each other, a provisional semiotic code is constituted: the symbolic elements signify the authorial dianoia, and the semiotic density of the work is increased. Of course, not everything has to be accounted for in terms of a more or less conscious perlocutionary intent. Perlocutionary effect is not always the result of perlocutionary intent. We may even recognize the perlocutionary intent of a writer (although this recognition supposes that his strategies have become conventional, and hence un-perlocutionary to some extent) and fail to respond to the work according to that intent. The critic is also a shifter of rules and a hyper-codifier —criticism is also creative writing, which is what makes it criticism, and not just retrieval of meaning. The creative element may be dominant, submitting the text to extrinsic codes and literally rewriting it. It may even overpower the moment of meaning retrieval, as can be easily seen in Wimsatt and Beardsley, the critics who denounce the "intentional fallacy." 16 But it is also inevitably present at that level of objectivist interpretation which is concerned with the boundaries between illocution and perlocution, taking the form of what is to count as the shift in meaning operated by the text —whether an effect must be "institutionahzed" as an illocution, or declared to be beyond the pale of convention. Criticism is a supplementation of literature because its main aim is the further conventionalization of the creation of meaning achieved in literature. Of course, it is bound to achieve this only in the way supplements do 17 —by subverting the conventionalization it strives to achieve, through ideological contention and the conflict of interpretations. Horton (1979, 24) makes a difference between reading and interpreting. In her account, reading is a temporal, sequential activity, and as such is opposed to interpretation, which consists in building a "spatial" pattern of relationships out of the matter furnished by this activity. This is a useful distinction, because it accounts for the difference in levels of hypercodification in the semiotic structure of the text, and relates it in a very immediate way to the reader's experience. In our own terms, we will conceive of the difference between reading and interpreting as two poles of the reader's activity. Semiotic phenomena have a structural nature, that is, they exist only in a system of related phenomena. A code is what allows us to relate a semiotic phenomenon to the system which gives it its identity, a system which is a code in its turn. That is, some semiotic phenomena are instrumental in allowing an access to other semiotic phenomena. The structures of signs are variable and are constantly being modified by their usage. This involves a necessary strategy: we must rely on certain codes which are (taken at this moment to be) relatively stable in order to have a fulcrum to exert a force on other codes. When an interpreter uses a semiotic Speech Act Theory and the Concept .. 101 system to retrieve meaning, without thereby modifying the system, we have an instance of wholly codified meaning. Reading, in a (conveniently abstract) technical sense should refer to the retrieval of codified meaning; interpretation is concerned with the retrieval of partially codified meaning. It will be observed that, since the process of retrieving partially codified meaning involves the transformation of existing codes, it is inherently open, and "retrieving" ceases to be an adequate term at this level. That is, the line between the construction of those meanings intended by the author and its logical continuation by the activity of the reader often resists precise definition, and is better conceived as a matter of of critical convenience. Horton despairs of tracing this line: "How do we ever know if what we have is only an instance of form speaking and not content or intent?" (1979, 119). The answer is that there are some cases clearer, or less undisputed, than others. There is no need of tracing the precise line of demarcation (although we will have to learn to live with a reasonable percentage of interpretive cruxes); what is fundamental is recognizing that there are two poles, that of the undisputably intentional and that of the undisputably unintentional, and that this situation seems to be the right one for both literature and criticism to thrive on. The lack of a clear border between intentional and unintentional features will not invalidate the relevance of the authorial intention to the creative or interpretive process (cf. Raval 1981,52). Speech act theory presents itself as a bridge between the extremes of individual creativity of use (the Saussurean parole) and the social givenness of language {langue). However, the gulf between the two main currents of speech act theory reproduces the original distance between langue and parole. Searle's version of speech act theory tends towards objectivism, and Grice's towards intuitionism (cf. Hirsch 1976, 25f.). According to Hirsch, Strawson has had the last word in this debate, because Searle's version cannot account for the growth of the system through use (1976, 67ff). An initial guess is in this view always an essential part of interpretation. Still, I would argue that a guess does not take place in a void. It occurs in a context which is already grasped through conventions. The guess itself will not be absolutely original; it will no doubt assume one form or another derived from a typology of possible kinds of guesses. Use of language in general, and interpretation in particular, is not possible apart form a complex system of varieties of discourse and relevant situational features. The final integration of both theories is possible only at a textual-discursive level, which watches the workings of the interpretive movements taking place in the specific uses of language in specific situations —such as academic literary criticism, for the concerns of this paper. According to Foucault (1976: 222), all critical commentary serves a "principle of limitation": its function is not to reveal all the meanings of the text, but rather to restrict the possibility of meanings of the text. Each interpretation tries in some way to arrest and determine the meaning of the work, but when it is confronted to the other interpretations it quickly becomes evident that criticism makes meaning proliferate, instead of arresting it; that the move of each interpretation towards a determinate meaning is in a way self-contradictory, since 102 Jose A. GarcIa Landa the interpretive activity by definition defeats its own purpose. In a way. We could apply Mallarme's line, "Jamais un coup de des n'abolira le hasard," to the inevitable limitations of the objectivist critical act which tries to determine the meaning and significance of a work. But it is convenient to remember also Jan Kott's riposte, "Et le hasard n'abolira jamais le coup de des." Criticism- is not a post-lapsarian activity or an essential failure in the sense that it (often) fulfils its immediate goals and has its own institutional validity. Moreover, interpretations are not simply divergent or contradictory. Later interpretations tend to reinforce some elements of earlier interpretations and discard others, and usually some clear patterns of agreement emerge. Some meanings are privileged, and together with them the modes of interpretation which have contructed them. Through this dialectical process of contention and accumulation, the status of the work in question as an instrument of communication or ideological sign, and the status of the interpretive practices which make it such, are constantly redefined and kept in touch with the transformation of the dominant social discourses. To return to an example mentioned earlier: the interpretation of Stephen Crane's story "The Monster" as a fable on racial relations became generalized in the sixties.18 The reasons for this change in interpretation are not dependent on the internal dynamic of literary interpretation alone: the new interpretation was only a symptom of a wider change in the ideology of American society. It could be argued, of course, that the process works both ways: criticism also has an effect on social discourses. But the changes in the discourse of literary criticism are more a product than an agent of social transformations; it is therefore of the utmost importance that the critic be aware of social reality and of the institutional constraints of his own activity. An intentional theory of literary discourse should go beyond the individual and microstructural level in analysing intention, and move towards more complex, composite and communal forms of intention. Literary genres and interpretive conventions as traditionally understood are only two instances of such higher-level intentional structures. Notes 1 Cf. Lozano, Pefia-Marin and Abril, 1982,41. 2 Cf. also Searle, 1980,52; Lyons, 733; Lozano, Pena-Marin and Abril, 194ff. 3 On perlocutions, see Searle 1980,51ff. 4 See Austin 1980, 22; Searle 1975. For discussion, see Petrey 1990, esp. 51-69. 5 Cf. also Hegel 1985,49; Mill 1971,538. 6 Richards 1973,186. Of course, the notion that the referential meanings of the sentence are discarded or annulled is absurd; see Fish's criticism of Richards (Fish 1980) 7 Cf. Ohmanh 1971; 1987a; Searle 1975; Pratt 1977,173; Lanser 1981,280; Ruthrof 1981, 53. 8 Cf. Skinner 1976,216. 9 The same shortcoming is also found in Beardsley's attempt at a speech act theory of literature (1970, 59), and in J-K. Adams's pragmatic theory of fiction (1985, 10). 10 On the textual author and reader, cf. the more common terms of "implied author" and "implied reader" as defined by Booth (1961) and Iser (1972). Speech Act Theory and the Concept ., 103 11 In fact, Bach and Harnish (1979) describe this kind of indirect speech acts as those which no longer require an indirect interpretation on the part of the listener. "Can you pass me the salt?" is so highly standardized that usually its status as a question does not even enter the mind of the hearer. 12 For the concept of macro-speech acts or macro-illocutionary force, see Van Dijk, 1980, 325ff. 13 Close 1976, 187. For the notion of 'utterer's occasion-meaning' see Grice, 1989, 89-91. 14 For a more detailed discussion of these concepts and their differences see my article "Authorial Intention in Literary Hermeneutics: On Two American Theories." 15 On hypercodification, see Eco, 1977,239ff. 16 Wimsatt and Beardsley, 1967. On this subject, see my article "Authorial Intention in Literary Hermeneutics." 17 On the logic of supplementation, see Derrida, 1967. 18 See my article "Reading-Racism: The Interpretation of Authorial Intentions in Stephen Crane's "The Monster'." Works Cited Adams, Jon-K, 1985. Pragmatics and Fiction. Amsterdam: Benjamins. Alston, William, 1964. Philosophy of Language. Englewood Cliffs: Prentice-Hall. Austin, John R. !9S0.HowToDo Things withWords. 1962, Oxford: Oxford UP. Bach, Kent, and Robert M. Harnish, 1979. Linguistic Communication and Speech Acts. London: M.I.T. Press. Bateson, F. W. 1953. "The Function of Criticism, at the Present Time." Essays in Criticism, 3 (1953), 1-27. Beardsley, Monroe C. 1970. The Possibility of Criticism. Detroit: Wayne State UP. Booth, Wayne C. 1961. The Rhetoric of Fiction. Chicago: U of Chicago P. [Bridges, Robert]. 1898. "Stephen Crane's 'The Monster.'" Life, 1 Sept.: 166. Castilla del Pino, Carlos. 1983. "Psicoanalisis y critica literaria." In Introduction a la crltica literaria actual. Ed. Pedro Aullon de Haro. Madrid: Playor. Close, A. J. 1976. "Don Quixote and the 'Intentionalist Fallacy'." 1972. Rpt. in Newton-De Molina 1976,174-193. Coleridge, Samuel Taylor. 1956. Biographia Literaria. 1817. London: Dent. Derrida, Jacques. 1967. De la grammatologie. Paris: Minuit. Eco, Umberto. 1977. Tratado de semiotica general. Barcelona: Lumen. Trans, of Trattatto di semiotica generate. 1975. Fish, Stanley. 1976. "How to Do Things With Austin and Searle: Speech Act Theory and Literary Criticism." MLN, 91, 983-1025. _. 1980. "Literature in the Reader: Affective Stylistics." 1970. Rpt. in Reader Response Criticism. Ed. Jane P. Tompkins. Baltimore: Johns Hopkins UP, 70-100. Foucault, Michel. 1976. The Archaeology of Knowledge and the Discourse on Language, trans. A.M. Sheridan. New York; Harper and Row. Frye, Northrop. 1957. Anatomy of Criticism. Princeton: Princeton UP. Garcia Landa, Jos6 Angel. 1989. "Reading Racism: The Interpretation of Authorial Intentions in Stephen Crane's «The Monster»". Misceldnea, 10,63-80. _. 1991. "Authorial Intention in Literary Hermeneutics: On Two American Theories." Misceldnea, 12, 61-92. Grice, Paul. 1989. Studies in the Ways ofWords. Cambridge (MA): Harvard UP. Harris, Wendell V. 1988. Interpretive Acts: In Search of Meaning. Oxford: Clarendon. Hegel, G.W.F. 1985. Introduction a la estetica. 1835. Barcelona: Peninsula. Hirsch, E.D., Jr. 1967. Validity in Interpretation. New Haven; Yale UP. 104 Jos£ A. Garcia Landa _. 1976. The Aims of Interpretation. Chicago: U of Chicago P. Horton, Susan R. 1979. Interpreting Interpreting. Baltimore: Johns Hopkins UP. Hough, Graham. 1976. "An Eighth Type of Ambiguity." 1973. Rpt. in Newton-De Molina 1976,222-241. Iser, Wolfgang. 1972. Der Implizite Leser. Munich: Fink. Ingarden, Roman. 1973. The Literary Work of Art. Trans. George G. Grabowicz. Evanston: Northwestern UP. Trans, of Das Literarische Kunstwerk. 1931. 3rd e<3. 1965. Lanser, Susan Sniader. 1981. The Narrative Act. Princeton UP. Lozano, Jorge, Cristina Pefia-Marfn and Gonzalo Abril. 1982. Andlisis del discurso: Hacia una semiotica de la interaction textual. Madrid: Catedra. Lyons, John. 1977-1978. Semantics. 2 vols. Cambridge: Cambridge UP. Martinez Bonati, Felix. 1972. La estructura de la obra literaria. 1960. Barcelona: Seix-Barral. Mayoral, Jose Antonio, ed. 1987. Pragmdtica de la comunicacion literaria. Madrid: Arco/Libros. Mill, John Stuart. 1971. "What is Poetry?" 1833. Rpt. in Critical Theory Since Plato. Ed. H. Adams. San Diego: Harcourt, 537-543. Newton-De Molina, David. 1976. On Literary Intention. Edinburgh: Edinburgh UP. (Introduction, vii-xvii). Ohmann, Richard. 1971. "Speech, Action, and Style." In Literary Style: A Symposium. Ed. Seymour Chatman. London: Oxford UP, 241-254. _. 1987a. "Los actos de habla y la definition de literatura." In Mayoral 1987, 11-34. Trans, of "Speech Acts and the Definition of literature." 1971. _. 1987b. "El habla, la literatura y el espacio que media entre ambas." In Mayoral, 35-57. Trans, of "Speech, Literature, and the Space Between." 1972. Petrey, Sandy. 1990. Speech Acts and Literary Theory. London: Routledge. Pratt, Mary Louise. 1977. Toward a Speech Act Theory of Literary Discourse. Bloomington: Indiana UP. Raval, Suresh. 1981. Metacriticism. Athens (GA): U of Georgia P. Richards, I. A. 1973. Practical Criticism. 1929. London: Routledge. Ruthrof, Horst. 1981. The Reader's Construction of Narrative. London: Routledge. Searle JohnR. 1975. "The Logical Status of Fictional Discourse." New Literary History, 6.2, 319-332. _. 1980. Actos de habla: Ensayo defilosofla del lenguaje. Madrid: Catedra. Trans, of Speech Acts: An Essay in the Philosophy of Language. 1969 _. 1983. Intentionality: An Essay in the Philosophy of Mind. Cambridge: Cambridge UP. Sidney, Philip (Sir). 1973. An Apology for Poetry. 1959. Ed. Geoffrey Shepherd. Manchester: Manchester UP. Skinner, Quentin. 1976. "Motives, Intentions and the Interpretation of Texts." 1972 Rpt. in Newton-De Molina 1976,210-221. , Solomon, M. 1956. "Stephen Crane: A Critical Study" (I and U) Masses and Mainstream 9:1,25-42, and 9:2, 31-47. Van Dijk, Teun A. 1972. Some Aspects of Texts Grammars. The Hague: Mouton. _. 1980. Texto y contexto. Madrid: Catedra. Trans, of Text and Context. 1977. Wimsatt, W.K., and Monroe C. Beardsley. 1967. "The Intentional Fallacy." 1946. Rpt. in WK. Wimsatt, TheVerbal Icon. 1954. Lexington; U of Kentucky P, 2-18. Wittgenstein, Ludwig. 1958. Philosophical Investigations. Trans. G. E. M. Anscombe. New York: Macmillan. NOTAS