Authorial Intention in Literary Hermeneutics:

On Two American Theories


José Ángel García Landa

Brown University, 1989

Web edition 2004


[Originally published in Miscelánea 12 (Zaragoza, 1991): 61-92]

1. intention and Intentionality


It will be convenient, before we broach the issue of intention in the theory of literature, to contextualize the concept of intention. In the phenomenological tradition inaugurated by Brentano, intention is a specific manifestation, among others, of the more general phenomenon of Intentionality. (1). Intentionality is a relationship between a cognitive representation and a state of affairs, in which the cognitive representation can be said to be "about" the state of affairs. Intentional states include perception, belief, desire, memory, and intention, among others. Contemporary philosophical schools differ on the status of Intentionality. For some thinkers (Ryle, Skinner, Quine), it is a concept which has no place in a rigorous theory of knowledge; for others, there is no understanding human activity without it. Let me side with the latter without further explanation, and say that the differences within this second current, e.g. between John R. Searle's mentalism and Daniel C. Dennett's "intentional stance," are largely irrelevant for the purposes of this work. As a rule, I will adopt Searle's idiom, according to which meanings are actual representations in the brain, although I think that Searle's theory could benefit from a greater degree of self-consciousness about its own heuristic status. This methodological choice has the advantage of setting in a larger frame the theory of speech acts, which I think is an important meeting ground for hermeneutics, linguistics, philosophy and literary theory.

For instance, both speech acts and Intentional states in general can be represented as a modalized propositional content. (2). My assertion "John loves Sally" (a speech act) can be paraphrased metalinguistically as "I assert that John loves Sally." Similarly, my belief that John loves Sally (an Intentional state) can be paraphrased as "I believe that John loves Sally." This paraphrase can be said to be Intentional, too. We must distinguish the original Intentionality of Intentional states themselves from the derived Intentionality of their semiotic representation (cf. Searle 1983a: 21). The Intentionality of a mental belief is original; that of language or literature is derived. Meaning is the relationship between primary and derived Intentionality, between an Intentional state and its semiotic representation. As we shall see, this definition allows for a variety of types of meaning.

intention-with-a-lowercase-i is a kind of Intentionality. Therefore, my intention to love Sally, which itself is non-linguistic, can be represented linguistically and rather trivially as "I intend to love Sally." Searle sees some Intentional states as more central, simple or primordial than others. These basic intentional states, such as perception or belief, can be a component part of several other more complex or secondary (3) Intentional phenomena. Searle hesitates about the status of intention in this respect. First, he presents intention as an elaborate instance of such secondary Intentionality, and isolates in it the components of belief and desire ("Bel" and "Des" in his notation):

If I intend to do A , I must believe it is possible for me to do A and I must in some sense want to do A . But we get only a very partial analysis of intention from the following:

Intend (I do A ) ­> Bel (<> I do A ) & Des (I do A )

The extra element derives from the special causal role of intention in producing our behavior. . . . (1983a, 34)

But immediately afterwards he reverses the priorities, and speaks of desire as an evolved or bleached-out intention (1983a, 36). Fortunately we do not have to solve this problem. What is more relevant to our purposes here is that intention-with-a-small-i is defined by Searle in its ordinary sense, preceding action as the aim before the shot. This is prior intention. There is another kind of intentionality: intention in the action. Even in those instances of human agency where there is not a distinct prior intention, we want to characterize the action as intentional-with-a-lowercase-i. In such cases, Searle argues, "the intention in action just is the Intentional content of the action; the action and the intention are inseparable . . ." (1983a, 84). Intention in action does not have to be present in the subject's consciousness: "Sometimes one performs intentional actions withouth our conscious experience of doing so" (1983a, 91). This is often the case in complex actions which demand a variety of minor, instrumental actions. I may consciously intend to drive to work, though not necessarily to turn on the starter. This would still be an intentional action. In turning on the starter, I might set off a booby trap installed by some terrorist--this would count as an unintentional action. However, not every unforeseen consequence of our basic actions is adequately described as non-intentional; for instance, on my way home I might suddenly become the only European born on the first of June 1961 who happens to be looking at a shop window. If we bracket away the privileged virtual observer of this circumstance, it does not even count as an action. In Searle's words, "we count an action as unintentional under those aspects which, though not intended, are, so to speak, within the field of possibility of the intentional actions of the agent as seen from our point of view" (1983a, 102). Or, from another perspective, "an unintentional action is an intentional action, whether successful or not, which has aspects which were not intended in it" (1983a, 108).

When we speak of the intentionality of a literary work, we do not mean simply prior intentionality, but neither are we referring simply to the Intentional nature of the work insofar as it is a semiotic phenomenon. We mean intention in the action, but in a sense which remains to be further elaborated, one which is specific to the structure of language.

The Intentionality of language is derived. This means that, even in the simplest of speech acts, there is a double layer of Intentionality: the Intentional state expressed, and the intention with which the utterance is made. Moreover, intention (and not only Intentionality) is inherent to semiotic phenomena. According to Searle, "the mind imposes Intentionality on entities that are not intrinsically Intentional by intentionally conferring the conditions of satisfaction of the expressed psychological state upon the external physical activity" (1983a, 27). In the case of language, the intentional association between signifier and signified soon becomes automatic, an unconscious intention in the act. Searle defines Intentional causation as that causation in which one element is an Intentional state and the other is its condition of satisfaction or part of its conditions of satisfaction. (1983a, 122). The peculiarity of the Intentional causation of semiotic communication is that an Intentional state is the condition of satisfaction of the Intentional state that causes it. The (highly conventionalized) Intentional causality of a red light is satisfied only if it causes in the motorist a belief that he is required to stop. Therefore, the intention inherent in the red light is not properly speaking to produce the effect of stopping the motorist, but to communicate to the motorist that he must stop. As Searle has noted, it is a mistake to assume that "the intentions that matter for meaning are the intentions to produce effects on audiences" (1983a, 161).

Complex Intentional systems are thus constituted by superposing different orders of intentionality. With the example of the red light, we are in our third layer of Intentionality: the original Intentional state of the sender of a message, the intention intrinsic to the constitution of the sign, and the intention to communicate a meaning through that sign. In the case of language, the issue becomes considerably more complicated, as we shall see later: language is so conventionalized that our communicative intention is directed at the performance of speech acts. For the moment, it is sufficient to say that intentions are inherent in the structure of language, and not merely in the psychological phenomena which precede or attend its use. The use and meaning of language (even of a term such as "intentional fallacy") rests on a series of practices, specific choices and purposes, which have become conventional to the extent of being taken for granted.



2. The Intentional Fallacy Fallacy

Authorial intention is a central concept in the classical theory of hermeneutics developed in the Romantic age. The classical accounts by Schleiermacher (1805-33, rpt. 1986) and Boeckh (1886) conceive interpretation as the reconstruction of the author's original conception. Emilio Betti voices their contemporary heritage best: "interpretation becomes a collaboration that the addresee extends to the author of the statement, inasmuch as he is called upon to reawaken in his own mind the idea conceived and expressed by the mind of the author" (1988, 32). Interpretation includes the reconstruction of the author's intentions. However, it is not restricted to the reconstruction of the conscious intent or the communicative interest of the author: "Even a 'manifestation' devoid of such interest and a behavior not in itself directed toward making a thought recognizable to others may be the object of interpretation" (1988, 33). Unintentional, implicit or symptomatic meanings resulting from such manifestations can be the object of hermeneutics.

This conception has long been challenged, often in a misguided way, by non-intentionalist theories of literary meaning. The modern discussion of the role of authorial intention in the activity of the critic is usually taken to start with the aestheticist reaction against romantic individualism, which was an essentially intentionalist current. This reaction is linked to literary modernism and can be seen taking shape in various forms during the nineteenth and the first half of the twentienth century. It is seen, for instance, in the work of Roger Fry, who once said,"I'm certain that the only meanings that are worth anything in a work of art, are those that the artist himself knows nothing about" (4) or in T. S. Eliot's strictures against "interpretation" (Eliot 1957). The debate has assumed different shapes, such as a contention between "historicist" and "critical" theories of interpretation, or between advocates and opponents of psychoanalysis, Marxism, structuralism or deconstruction--doctrines which challenge the traditional concept of individual intention. In America the most influential anti-intentionalist ideas came from a "critical" movement who opposed positivist scholarship--the New Critics.

The best known and most influential theoretical statement of the anti-intentionalist school is W. K. Wimsatt and Monroe C. Beardsley's article "The Intentional Fallacy" (1946; rpt. 1967). (5). The authors summarized their position in the statement that "the design or intention of the author is neither available nor desirable as a standard for judging the success of a work of literary art" (1967, 3). There has been considerable debate as to what was the precise scope of this critical axiom or doctrine, as well as its legitimacy. Other critics have found that it is just as easy to hold that "the design of the author is both available and desirable." (6). The extreme intentionalist position will have to be qualified, too. The meaning of a poem is not always the meaning the poet intended it to have; the poet is not always in control of everything which is at work in the poem (Redpath 1976, 17), and he is definitely not in control of an anti-intentionalist interpreter.

Let us note first that Wimsatt and Beardsley take "intention" to mean first of all prior intention, the original intention of the author, previous to the actual composition of the work, or the personal intention of the author in those cases where it is different from any intention which may be inferred from the work itself. (7). Sources of information as to the author's intention which are external to the work itself are not relevant for the judgement of the work. Rejection of intention is meant as a rejection of psychologism, in favour of intrinsic criticism--of criticism based on the analysis of the text itself. Anti-intentionalism presents itself therefore as a favouring of public conventions over private ones.

This is not the main sense in which I hold authorial intention to be decisive to the critical enterprise. Alastair Fowler has noted that "intention means different things at different stages of composition." (8). It also means different things from the point of view of the author at work and the from the point of view of the interpreter. "Intention" understood as a working project of the author's is not a concern of the interpreter, except insofar as it may have some kind of bearing on the intention assumed to be present in the work. This first kind of intention may nevertheless be a concern of the critic. Wimsatt and Beardsley's article is mainly concerned with "intention" seen from the author's pole, with intention "extrinsic" to the work. The response of intentionalist critics to "The Intentional Fallacy" has been to show that intention is, to use Wimsatt's own terms, intrinsic to the literary work, that the public sharability of language is not separable from a concept of intention, something which is neglected by anti-intentionalist critics. (9). In short, writing literature involves an intention in the action which is part of the communicative structure of the work. Intention is not merely something which precedes the work or exists apart from it; neither is intentionalism a blind submission to any meaning an author may claim for his work. It is a requirement to see the work in the right context. (10). For Gadamer, too, the opposition between intentional meaning and unintended meaning is too simplistic, unsatisfactory (1988, 58). An extended speech act theory (Pratt 1977; Harris 1988) can help us conceive of an intentionalism which is not narrowly and exclusively psychological.

As for Wimsatt and Beardsley's denunciation of psychologism, it remains a mystery how Wimsatt and Beardsley ever expected a historical psychic intention to be a threat at all for criticism if they believe that it is never available (cf. Peckham 156). It is also significant that Wimsatt and Beardsley assume that an unfulfilled intention is not recognizable without recourse to external evidence. (11). Any of a range common phenomena, such as bad verse which tries hard to be good, can serve as an example that this is not the case.

Let us note, too, that Wimsatt and Beardsley are mainly concerned with evaluation. However, the intentional fallacy is soon applied quite naturally to the realm of interpretation, since evaluation is logically linked to interpretation (Peckham 1976, 141). And in his revision of the earlier doctrine, Wimsatt extended the application of the intentional fallacy to the realm of interpretation: "The design or intention of the author is neither available nor desirable as a standard for judging either the meaning or the value of a work of literary art" (Wimsatt 1976, 136).

Wimsatt and Beardsley are, furthermore, concerned with the criticism of poetry, or more widely, with aesthetic criticism. In their account, poetry simply

is, in the sense that we have no excuse for inquiring what part is intended or meant. Poetry is a feat of style by which a complex of meaning is handled all at once. . . . In this respect poetry differs from practical messages, which are successful if and only if we correctly infer the intention. They are more abstract than poetry. (12)

Moreover, whatever personal thoughts or attitudes are inferred from the poem itself are to be attributed "to the dramatic speaker, and if to the author at all, only by an act of biographical inference" (1967, 5). The poem does not belong to the critic, but neither does it belong to the author. It rests on the publicly specifiable conventions of meaning.

By means of an unconscious sleight of hand, Wimsatt and Beardsley's discussion reduces the function of criticism to an inquiry about the value of a work of art, disregarding the interpretive moment. (13). They reject as uninteresting one form of evaluation, the inquiry as to whether an artist realized his (original) intentions. In their view, criticism must evaluate the final result, the poem and not the (original, extrinsic) intention. "The evaluation of the work of art remains public; the work is measured against something outside the author" (1967, 10).

This valuation of public meaning is illegitimately identified with intrinsic criticism: in an even more amazing sleight of hand, "internal" is identified with "public"; "external" with "private" (1967, 10). Hirsch has noted that (in its main claim) Wimsatt and Beardsley's argument is not directed against verbal intentional meaning, but against irrelevant meanings. They contend "not that the inferred meanings are private, but that they are probably not the author's meanings" (Hirsch 1967, 16). Wimsatt and Beardsley, however, would reject the notion of "the author's" meanings, and would insist on the public nature of language. This public nature is what makes the notion of intrinsic criticism possible at all. The language of the poem has to be accepted as internal evidence; semi-private meanings of words are an intermediary case although ultimately they must be accepted as public (since we know them) and intrinsic: "The meaning of words is the history of words, and the biography of an author, his use of a word, and the associations which the word had for him, are part of the word's history and meaning" (Wimsatt and Beardsley 1967, 10). The historical meaning of the work defined in this way is not "internal" to the text for Wimsatt and Beardsley's point of view. It is an intermediary type of evidence, not wholly external, but then not wholly internal either, because Wimsatt and Beardsley are considering the word as such, as a piece of langue, instead of seeing it as an instance of use in a text, a contextually defined parole. (14). The historical meaning, that is, is only "a part" of the meaning of the word even from the purely interpretive point of view. Ahistoricism rears its head in this conception.

But what really gives the whole theory away is a note appended to the text just quoted: "And the history of words after a poem is written may contribute meanings which if relevant to the original pattern should not be ruled out by a scruple about intention" (Wimsatt and Beardsley 1967, 281). Notes are always telling about the deep intention of an author. It is clear now that Wimsatt and Beardsley's concept of criticism sets aesthetic value foremost. The interpretation of the authorial meaning is secondary, even negligible if necessary. The critical act may evaluate the poem according to an accidental meaning or significance which it has acquired by virtue of the evolution of language. It is noticeable that in the note "intention" refers no longer to the sense they had started with, "whatever is or had been in the author's mind apart from the intentional acts publicly specified in the work itself." The note contradicts the apparent meaning of the text, if not its real thrust. The text gives a (qualified) historicist definition of meaning, a concern for the meaning of the word in the author's context. (15). It comes close to a dangerous point where Wimsatt and Beardsley would be hard pressed not to accept evidence external to the text: those cases where a biographical or other "extrinsic" remark throws light on the original context of composition and therefore on the meaning of the word. (16). The note wipes away this concern for the authorial / historical meaning of the word. All "patterned" meanings, whether historically justified or not, are legitimate. (17). Anti-intentionalism therefore results in anti-historicism.

This doctrine rests on a very specific conception of language, which has sometimes been called "semantic autonomism." According to this conception, once it is detached from its enunciator, language becomes autonomous; the author has no further rights over his utterance. This conception is shared by some structuralists critics. Roland Barthes speaks in this respect of "the death of the author." Barthes sees the author's enunciation, writing, as an empty process which needs no interlocutors. Conversely, the reader is free to make of the text whatever he wants. In this view, literature is not communication, and interpretation is a senseless activity: "Once the Author is removed, the claim to decipher a text becomes quite futile" (Barthes 1977a, 147). The critic's meaning is as good as the author's. The best meaning, in these critics' view, overrides the authorial meaning, especially if the authorial meaning has to be determined by means of information not readily accessible to the reader. (18).

This might imply, as I have said, a disparagement of historicist considerations, but in fact it does not, or at least Wimsatt and Beardsley do not intend this result. They do not rule out the knowledge of the author's original culture as an essential asset for the interpreter--they presuppose it, they presuppose a competent reader. A great deal of historical and general contextual knowledge is to be taken for granted if this view must make any sense. Even conceding the dictionary meaning of the words to be "intrinsic" to the poem (a major concession which is taken for granted by the New Critics) we still need far more than dictionary meaning would allow in order to make sense of a literary work. An hypothetical dictionary-reader would fall short of Wimsatt and Beardsley' expectations. They need an encyclopedic reader, who knows, for instance, that the "trepidation of the spheares" in the Donne poem they comment on in their article refers at all to astronomy. The poem does not say so, but it is nevertheless an internal element of meaning to Wimsatt and Beardsley because they assume that the reader knows. (19). Wimsatt and Beardsley, therefore, do not erase the difference between authorial meaning and other kinds of meaning as far as interpretation is concerned. Neither do they, for that matter, stress it as a meaningful issue. They are not really concerned with interpretation. (20). And criticism, in its evaluative side, should in their view ignore this difference.

The critical act would therefore consist of:

1) Interpreting the text of the poem to determine the range of possible interpretations, whether of authorial meaning, meaning or significance.

2) An evaluation of these interpretations, without privileging the authorial meaning of the poem.

But, given the intrinsically intentional nature of language, the creation of a kind of virtual author is presupposed in 2). The result is that the best meaning is regarded as if it were the authorial meaning of the poem. It is clear that with such assumptions about the aims of interpretation the difference between authorial meaning and other kinds of meaning need not be a basic critical concept. At best, as in Wimsatt and Beardsley, it is anecdotic; at worst, as in some of their followers, it is ignored, not understood. Interpretive anarchy is kept within bounds by presupposing a competent reader, who will generally give a historically plausible interpretation--even if his aesthetic aims allow him to occasionally improve the poem with the complicity of history. This kind of interpretation is not very different from the adoption of an intentional stance towards mechanical artifacts whose structure we ignore (cf. Dennett 1987, 15ff.). The intentional stance allows us to make sense of the artifact's behaviour, but at the cost of endowing it with a greater degree of consciousness than we would want. Paradoxically, the New Critical "intentional fallacy" has as its concomintant phenomenon an overestimation of the authorial intention and deliberation such as may be found for instance in Stanley Fish's book on Milton (Fish 1967). Theodore Redpath notes that an author cannot be said to have intended everything a reader may find in his work--even if the author claimed so himself. The object of Redpath's criticism is John Dewey's aesthetic theory (21) --no author is known to have claimed this. Perhaps the closest example was T.S. Eliot, who had a critical axe to grind and none the less saw fit to provide The Waste Land with notes. The concept of intention is in one sense inevitably linked to at least a potential consciousness and deliberation. For the sense in which it is not, I will speak of deep intention. (22). The two must not be confused; otherwise we will end up constructing a fictional author-figure tailored according to the critic´s needs.

It should be clear by now that the "intentional fallacy" is no such fallacy. It is a particular choice of what is to count as validity in interpretation. The New Critical anti-intentionalism is another choice, with a different kind of assumptions and aims, and which is not primarily concerned with hermeneutic validity--a perfectly legitimate enterprise if it is done in a self-conscious way and with a clear methodology and aims. (23). Wimsatt and Beardsley's doctrine in "The Intentional Fallacy," however, is itself a fallacy to the extent that the authors present their "seminal ukase" (24) as an objective critical principle and the universal rule for literary interpretation. Hermeden has noted that "the fundamental issues involved in a stand on intention are nonempirical: they concern normative questions." (25). No new data will solve the problem; the status of intention in interpretation has to be decided by the interpreter, in view of the aims he sets to his activity. In Hirsch's terms, the reader need not try to realize the author's intended meaning. The question of whether he should or not is an ethical question, which is answered affirmatively by Hirsch: "Unless there is a powerful overriding value in disregarding an author's intention (i.e. original meaning), we who interpret as a vocation should not disregard it." (26). Nevertheless, I would somewhat qualify this position. There are advances made on conceptual issues, too. Analyzing them is not a fruitless task. The issues are now much clearer than when Wimsatt and Beardsley denounced the intentional fallacy: a greater degree of shared assumptions has been discovered, and some extreme positions have been abandoned. And even if there are several correct modes of constructing or using intention, experience tells us that they are not all equally correct in all interpretive contexts.

It is an axiom of classical hermeneutics that interpretation logically precedes evaluation. But it has been noted by Newton-De Molina that "this logical condition of priority may tend to elide the full human importance of an awkward truth: that the relevant information upon which we base particular interpretations is not always prior in time to particular evaluations." (27). The consequence is that "unless evaluations are to be eternally postponed they must always be accepted, in some senses, as generically provisional" (1976, xi). However, it must be recognized that things work in a slightly different way:

a) Evaluation cannot (or should not!) precede that first step of interpretation that we call understanding. It is clear that otherwise we call it prejudice.

b) Some evaluations are more provisional than others. The evaluation of a passage during the reading process may be highly provisional, and the psychological attitudes of any person towards a work may change a great deal at different moments of his life. But the considered evaluation given by a critic in a scholarly article after several careful readings and taking into account a previous evaluative tradition is, I think, in a much more definite relationship to the critic's interpretation of the text. Ideally, it has gone through a process of suspension of judgement until an interpretation is established. But the decisive fact is that this evaluation is public and fixed, and is based on an interpretation of the work which is accessible to other critics. By definition (because of the logical priority of interpretation) Critic no. 2 assumes that Critic no. 1's evaluation is grounded on that particular concretization of the work --not on Critic no. 2's own concretization of the work. And evaluations may be more generally shared than interpretations, and rest on the more general interpretive level of understanding, because precisely one of the functions of great literature in our culture is to generate diverse interpretations, which therefore can affect the basic evaluation of the work only peripherally.

Interpreting with evaluation in mind leads to a confusion of aesthetic and interpretive criteria. This is evident in Wimsatt and Beardsley, and also in some of their critics. In spite of his misgivings about the rejection of biographical evidence, Emilio Roma shares with Wimsatt and Beardsley the assumption that

there is at least one reason which counts for and against interpretations, namely, the reading fails because it does not account for certain significant passages of the poem, and hence does not bring out the richness of the poem. This reason is essentially evaluative, and at the same time it does not go "outside" of the poem. (28)

But we might well ask for whom are those passages significant. The answer seems to be: for a critic with Roma's criteria of validity--which may or may not be "in" the poem, but which surely need not be in the author's creative intention. It is not surprising, therefore, that in the last analysis Roma endorses Wimsatt and Beardsley's concept of interpretation and their definition of meaning as something which is decided by the critic on the basis of criteria which are ultimately aesthetic: "the way a poet and his contemporaries understood a word or passage is relevant evidence for making a decision about the meaning of a poetic utterance. I do not say that this is always relevant evidence" (1976, 85-86). Translated into our own terms, this would read: "The present-day significance of a work resulting from the conventions of academic interpretation sometimes has more aesthetic relevance than its historical meaning." This is true if our notion of aesthetic value allows itself, as it does for Roma or Wimsatt and Beardsley, a measure of independence from historical and cognitive considerations. My own view is that the two notions of aesthetic relevance have to be kept apart. We may very well rewrite a poem in our imagination and then evaluate it, but a historicist aesthetics can hardly afford to do this. Its object is not so simple; it involves a determination of the conditions of artistic production and the assumptions about art in the original context of the work, an evaluation of the work in those terms, and then a study of the "life" of the work and its transformations in the uses to which it is put by different readers and critical projects. (29). The anti-intentionalist, aestheticist project is just one more of these historical attitudes to art, and it does not fare very well from an objectivist perspective. Aesthetic value is usually grounded on the coherence of the text, a coherence which usually is the product of the collaboration between writer and critic. Coherent intepretations are usually reached in this way, but, as Hirsch has noted, coherence is not a sufficiently objective criterion of validity--an interpretation may be both coherent and wrong. (30).

An interpretation, we may conclude, must strive towards coherence and completeness, but in doing so it must not build a perfect structure with the data internal to the work while it conflicts with historical, psychological or cultural assumptions that we might wish to hold. (31). Its hypotheses must remain translatable into other areas of knowledge, with the aim of increasing communication in culture at large.


3. Objectivist Hermeneutics and Criticism: E. D. Hirsch

The best discipline to keep our own aesthetic or ideological concerns from giving an obvious bias to the interpretive activity, as far as this is possible, is to consider the phases of the critical activity as logically and chronologically sequential activities. In Hirsch's words, "Understanding (and therefore interpretation, in the strict sense of the word) is both logically and psychologically prior to what is generally called criticism" (1967, 209). It has been rightly said that there can be no pure interpretation, that description will always contain a measure of evaluation. (32). This is true. It is none the less the case that description as such is not evaluation as such. Refusing to concede the possibility of this conceptual distinction is the first step towards the collapse of the distinction between understanding and inventing--throw the thrust towards objective description away and communication is likely to follow suit (cf. Hirsch 1967, 26).

Hirsch is the main theorizer of objective interpretation in the English-speaking academy. (33). Hirsch's basic interpretive rule is the distinction between "meaning" and significance. In the first version of his theory (1967), "meaning" is "meaning in the author's context"; significance is "meaning for us, today." "Meaning" is the result of interpretation, which is logically prior to the inquiry for significance:

Meaning is that which is represented by a text; it is what the author meant by his use of a particular sign sequence; it is what the signs represent. Significance, on the other hand, names a relationship between that meaning, and a person, or a conception, or a situation, or indeed anything imaginable. (1967, 8)

Hirsch relates this distinction to similar differences established by other theorists: Boeckh's "interpretation" and "criticism," Frege's "sense" and "reference," and Husserl's "inner" and "outer horizons of meaning." It is one of Hirsch's main tenets that significance is variable while "meaning" is fixed. Significance changes as each critic relates the work to his own interests or to his own knowledge of the subject-matter the original writer was dealing with (cf. Hirsch 1967, 58f., 63); this accounts for concepts such as the life of a work of art, which is relative to the changing significance of the work. "Meaning", on the other hand, is a fixed historical fact: "an author's original meaning cannot change" (1967, 9. Cf. Fowler 1976, 252). It does not change through the life of the text--for Hirsch, history is already written and cannot be unwritten. (34). Later, Hirsch enlarges the sense of the word "meaning" to include other meanings besides the authorial. This comes nearer to the sense in which I use the terms. Meaning is now "that which a text is taken to represent" (1976, 79). However, the determinateness of the author's meaning still plays the same fundamental role in his theory. And Hirsch's 1976-meaning is still a principle of fixity as opposed to a principle of change, significance, or "meaning-related-to-something-else" (1976, 80).

Hirsch takes great care to distinguish this relation between authorial meaning and the author from any psychologistic conception of intention or meaning: it is the difference defined by Husserl between an intentional act and a psychical act (Hirsch 1967, 217-218). The authorial verbal meaning is, in Husserlian terms, an intentional object; it is "that aspect of a speaker's 'intention' which, under linguistic conventions may be shared by others" (Hirsch 1967, 218). This meaning is Intentional in the sense Husserl gives to the word. It may not have been wholly conscious for the author himself at any given moment, but it must belong to his "horizon of expectations": "The interpreter's aim, then, is to posit the author's horizon and carefully exclude his own accidental associations" (1967, 222). The author's Intentionality does not necessarily warrant "consciousness of meaning": "there are usually components of an author's intended meaning that he is not conscious of" (1967, 21). Hirsch explains this apparent contradiction by means of the notion of typification. The author's Intentional acts (in the Husserlian sense) by which he wills a particular meaning into being are not directed towards each aspect of his intended [Intentional] meaning, but to a typical whole: "the acceptability of a submeaning depends upon the author's notion of the subsuming type whenever this notion is sharable in the particular linguistic circumstances" (1967, 49). Verbal meaning as such is nothing other than "a willed type which an author expresses by linguistic symbols and which can be understood by another through these symbols." (35). The author's use of this type must be learned, and this is possible because types are eminently sharable (1967, 66ff.). The consequence for the definition of "conscious intention" is that very often we cannot be sure whether a meaning was conscious or unconscious. In these cases, according to Hirsch, the distinction is irrelevant (1967, 51).

Moreover, the authorial meaning does not consist solely of representational elements: "Defined in Husserl's terms, 'meaning' embraces not only intentional objects but also the species of intentional acts which sponsor those intentional objects. . . . Subjective feeling, tone, mood, and value, are constitutive of meaning in its fullest sense" (1976, 8). This we might relate to Searle's analysis of intentional acts into a proposition and a modal element (1983a, 5ff.).

There is a difference between Hirsch's concern for the public specifiability of meaning and a similar concern as it was voiced by Wimsatt and Beardsley. For Hirsch, meaning is public, but never completely so. The text cannot be regarded as a piece of language, a system of possibilities or a verbal icon; it is tied to the author's original meaning (1967, 24). Wimsatt explicitly upholds the interpretation of a work as a piece of langue. Apparently, if we interpreted a work as an instance of parole, the words "would never. . . make sense to anybody but the author himself" (Wimsatt 1976, 138). The absurdity of this statement hardly needs to be pointed out after the development of discourse analysis during the past twenty years. For Hirsch, the text is not a segment of Saussurean langue, but an instance of parole (1967, 232), a historically limited phenomenon which must be understood in its context. The (present-day) communal context as such is not a sufficient criterion:

It is therefore not only sound but necessary for the interpreter to inquire, "What in all probability did the author mean? Is the pattern of emphases I construe the author's pattern?" But it is both incorrect and futile to inquire, "What does the language of the text say?" That question can have no determinate answer. (1967, 235)

Hirsch's conception of a historicized meaning also makes irrelevant Raval's contention that "Hirsch's authorial intention does not entail a biographical person but rather a 'speaking subject'" which is "not really distinguishable form the New Critical persona" (1981, 64). The New Critical persona is more or less freely constructed by the reader; Hirsch's persona is the result of the author adopting discursive conventions which must be understood in the terms of the culture and context where they originated. The ethics of interpretation starts with the respect towards the author's meaning and our efforts to grasp it objectively (cf. Scholes 1990, 50).

Unless we regard things in this way, Hirsch argues, there would be no criterion to determine a correct interpretation. This claim is the polar opposite of Barthes's doctrine that "a text's unity lies not in its origin, but in its destination" (1977a, 148). Focusing on the reader's activity without a reference to the author's meaning does not unify the text; rather, it disintegrates it completely. In a vein similar to Barthes', Horton (1979, x) argues that although meaning is ultimately context-bound, the boundlessness of the context prevents its determination in practice. Hirsch would accept this claim for meaning in general, but not for the concrete authorial meaning. Its historical typicality is what makes interpretation possible. For Hirsch, the text is only fixed at one point, the author's meaning. That is why this is an indispensable criterion for the validation of interpretations. We have seen,though, that there might be other criteria--an aesthetic criterion can determine that the "best interpretation" is the correct one. Hirsch simply has different assumptions as to the aims of criticism. The validity of a reading is not to be equated with its aesthetic excellence (1967, 5). Criticism must resist the temptation to mix evaluative and interpretive criteria. It must speak with the voice of reason, and distinguish the text as it is before the critical activity form the text as it is after the critical activity. Meaning cannot exist in a void. If a critic rejects the author's meaning, he will inevitably substitute his own (1967, 4). But doing this is indulging in a kind of vicarious authorship. (36). This would not be "intepretation" in Hirsch's sense, and so we can concede his point that there is a specific cognitive activity which consists in the identification of the authorial intention. Creative writing and criticism are two different enterprises, even if they share some common elements. In the last analysis, the interpretive theories of Wimsatt and Beardsley and those of Stanley Fish (1980) rest on a shared assumption: that there is no essential difference between interpretation and creation. Hirsch's is radically opposed: interpretation and creation are essentially different activities and they must be carefully distinguished from each other: "Interpretation is the construction of another's meaning" (1967, 244). This meaning cannot be constructed partially. Either it is ours or it is the author's: for Hirsch, Gadamer's theory of Horizontverschmelzung is a logical contradiction, since it presupposes a contact with something--the author's perspective--which is at the same time declared to be inaccessible. In Hirsch's view, we do not understand the meaning of others in our own terms, though we do value it in our own terms. (37).

Some phenomenological critics have strongly stressed the related notion that reading involves a contact with another mind. For Georges Poulet, reading is an activity with a peculiar ontology of its own--it is an immersion in a conscience different from our own, and it requires total submission to the author's consciousness as it emerges from the work. "I myself, although conscious of whatever [the consciousness inherent in the work] may be conscious of, play a much more humble role content to record passively all that is going on in me" (Poulet 1980, 47). A similar definition of interpretation is found in Maurice Blanchot. (38). These conceptions sound sometimes nearly mystical. But not all theories of interpersonal communication are drawn along these lines. Few people would agree that the reading process itself involves such passivity on the part of the reader as Poulet would have it. Dennett's version of the way the gap to the other's meaning is bridged is more satisfactory: it is relativistic to a degree, and stresses the activity of the intepreter: "when we interpret others we do so not so much by theorizing about them as by using ourselves as analog computers that produce a result. Wanting to know more about your frame of mind, I put myself in it, or as close to being in it as I can muster, and see what I thereupon think (want, do...)" (1987, 100). The hermeneutical tradition since Schleiermacher already stressed that even the simple retrieval of meaning is an active process resting on the play of hypotheses and data, not an act of "reception" (Schleiermacher 1986, 113-17).

If the interpretation of the other in his own terms already requires a degree of activity, criticising him is an eminently active and assertive activity. In Hirsch's model, the critical phase succeeds the interpretive one. Hirsch, I think, would not contest the claim that "conceding authorial privilege means giving the author the first word, not the last" (Fowler 1976, 250). The study of meaning is only a necessary preliminary to that part of the critic's activity which has direct public value, the application of meaning, significance (1976, 19). This is an essential, not merely legitimate, function of the critic. Only, the study of significance is an activity different from the interpretation of meaning, and must be carefully distinguished from it. (39).

A last, important characteristic of Hirsch's theory is his probabilism. Authorial meaning is fixed and determinate, but we never know whether we have construed it in a correct way. On this matter there are only various degrees of probability. (40). Nevertheless, knowledge is possible: "It is a logical mistake to confuse the possibility of certainty in understanding with the impossibility of understanding." (41).

Beardsley's attempt to refute Hirsch's theory of interpretation is in my view unsuccessful. He is not addressing the issues Hirsch is concerned with. His three arguments against Hirsch's thesis only prove that the reader of a text may construct verbal meanings which were not intended by the author--something which I think is self-evident to Hirsch. The real difference lies in the fact that Beardsley counts these constructions as valid interpretations, whereas Hirsch does not. Just like in "The Intentional Fallacy," Beardsley is concerned with the aesthetic value of a reading, not with its successful retrieval of the authorial meaning. Again we find the same ahistorical aestheticism at work: in pushing the authorial will out of his consideration, Beardsley is in fact pushing out the historical context of the utterance. Let me show this through a re-cycling of one of Beardsley's examples:

An ambiguous text does not become any less ambiguous because its author wills one of the possible meanings. Will as he will, he cannot will away ambiguity. There is something odd about the notion of "willing" a meaning. It is as though we ordered someone, "Say 'cat' and mean dog." Can one do that? How does one do it? True I can say, "Vote for Senator Kennedy!" and think of Edward Kennedy. Do I thereby make the word "Kennedy" in that utterance mean Edward Kennedy ? That is quite impossible. (1970, 29)

The perspective is all wrong. An ambiguous text does become less ambiguous if we find out that the author had willed one of the possible meanings. We do not usually need to will away ambiguity because our utterances are calculated to be unambiguous in the context in which they are used. Suppose a dog is barking, and someone observes, "The cat is barking." This is a puzzling behaviour, and a number of intepretations could be offered. No doubt some of them would afford a great deal of aesthetic satisfaction. But if a moment before we had heard Beardsley order our man "Say 'cat' and mean dog," there would be nothing much left to interpret. In the example of the Kennedys, Beardsley is thinking of the ambiguous reference of the sentence at the time he wrote his essay--it could have meant Edward or Robert Kennedy. But by virtue of changing contexts, the phrase no longer has an ambiguous reference in a 1986 campaign. "Kennedy" does not indeed have the same dictionary meaning as "Edward Kennedy," but if we can discover the reference by identifying the context where the sentence was uttered, we will have interpreted its authorial meaning beyond the "textual meaning" defined in Beardsley's terms. Again, the phrase is ambiguous only if it is taken as an instance of langue.

Indeed, Beardsley admits that the discovery of the "textual meaning" and the discovery of the authorial meaning are two distinct inquiries. Only, "the proper task of the literary interpreter is to interpret textual meaning" (1977, 32). He opposes the aesthetic to the historical approach (1970, 34), instead of integrating them in a historicist aesthetics. And his conclusion is inevitably contradictory: on one hand, a carte blanche for the critic to draw the limits of a poem in an act which is aesthetically motivated (42); on the other, a claim for the complete autonomy of the poem and the passivity of the critic. (43). That literary discourse is highly conventionalized does not mean that it ceases to be a historical parole. Its contexts are standardized to a degree, but not to just any degree, as the New Critical aesthetics would have it.

A historically conscious aesthetics is richer than one which ignores historical considerations. Indeed, it includes the latter as a particular historical attitude to art. It is significant that Beardsley cannot help assuming the historically and genetically conscious viewpoint when he tries to refute Hirsch's theory. His three theses against the identity of meaning and authorial meaning are, of course, correct, but they do not prove what they set out to prove, that any construction of meaning justified by "the text itself" is a valid interpretation. Argument number one runs thus: "Some texts that have been formed without the agency of an author, and hence without authorial meaning, nevertheless have a meaning and can be interpreted" (1970, 18). Misprints and poems composed by computers are examples of this. What Beardsley does not conclude is that if we learn that a verbal phenomenon is the result of a misprint or has been generated by a computer, we have gained an insight into the nature of that phenomenon and the kind of "language game" of which it is an instance. "When Hart Crane wrote 'Thy Nazarene and tender eyes,' a printer's error transformed it into 'Thy Nazarene and tinder eyes,' but Crane let the accidental version stand" (1970, 18). The difference between a willed meaning and a printer's error is not irrelevant, as Beardsley seems to suggest. Printer's errors are suppressed in revised editions, and what Hart Crane did in this case was to transform a printer's error into an authorial meaning by means of an act of will which is known and recognized.

The second of Beardsley's objections is that the meanings of words change historically, and therefore the authorial meaning becomes distinct from the textual meaning. In his example, Mark Akenside writes in 1744 of how God "rais'd his plastic arm." Beardsley comments that the line has acquired a new meaning in our century, but apparently he does not rule out the interpretations which ignore this fact because they are concerned with today's textual meaning. It is not clear to me in which context a critic can deliberately ignore this difference and still lay a claim to be interpreting Akenside's poem. Hirsch would call this activity "rewriting," and I agree with him.

The third objection is that a text can have meanings that its author is not aware of (Beardsley 1970, 20). So much the better for the interpreter. But this claim presupposes that we can tell apart what an author is aware of and what he is not aware of--that we can isolate the interpretation of the authorial meaning as a distinct phase of the critical activity. Hirsch, I think, does not ask for more. After this, it is only to be wished that an evaluation of the text does not ignore this conceptual difference we have all started with. I am afraid Beardsley is all too ready to do this.

Since my interpretive assumptions are dangerously close to Hirsch's, maybe it is the moment to make the differences between them more explicit. Some are terminological and some are conceptual.

- Terminologically, I have tried to maintain the normal usage of the terms while carefully distinguishing the conceptual differences. My use of "meaning" and "significance" are closer to Hirsch's use in The Aims of Interpretation; whenever I wish to be more specific I will use "authorial meaning." Hirsch's "criticism" is concerned only with valuation and significance (1967, 9), and as such is opposed to "interpretation," which is concerned only with establishing (authorial?) meaning. Sometimes he draws a difference between "understanding" as the construction of meaning and "interpretation" as its explanation (1967, 136). I think it is more convenient to think of "criticism" as the sum of interpretation and valuation, and to be concerned with both meaning and significance. Interpretation is not restricted to the fixation of authorial meaning. I think that in the current usage of the word it is also concerned with meaning at large or with the significance of texts. Therefore we should speak of "interpretation of authorial meaning," "interpretation of meaning" or "interpretation of significance" whenever we wish to be more specific.

- Conceptually:

1) We can borrow Ingarden's term "concretization" (1973, 322) to refer to the whole construction of meaning to which the interpretation may ultimately be assumed to refer to, as opposed to those aspects of the construction it actually refers to. This distinction draws a necessary wedge into the first term in the opposition drawn by Hirsch between "the construction of meaning to which the intepretation refers" and "the meaning of an interpretation" (1967, 129). Sometimes Hirsch does speak of "the whole meaning to which [different compatible interpretations] refer." (44).

2) Hirsch's notion of the historical fixity of a text is linked to his theory in a deficient way. There is one sense in which the author's meaning is not a desideratum--every critic reaches it, or measures his distance to it. To this extent, Hirsch's rejection of the concept of Horizontverschmelzung used in Gadamer's historicist hermeneutics (Gadamer 1977) is justified. (45). Through this concept, Gadamer rejects the idea of a fixed sense in texts, and stresses the fact that our constructon of an author's meaning is already gided by our position and aims as interpreters. But this concept cannot have more than a purely metatheorical value Its role is justified in Gadamer's philosophical study. In actual interpretation, however, we cannot have an access to the author's pole to verify the extent of the difference between our horizon and his; and in one way or another we will have to posit our own conception of authorial meaning. This cannot be done with respect to an unknown pole, a noumenon outside our reach. There is then no ultimate contradiction between Gadamer and Hirsch, since Gadamer's "relativism" cannot have any possible bearing on actual interpretation.

Hirsch is right in saying that ultimately the possibility of academic discussion rests on the conceptual difference between (authorial) meaning and significance. But every critic articulates in a different way the proportion alloted in his interpretation to (authorial) meaning and to significance. "If a Marxist critic construes a text differently from a formalist critic," Hirsch argues, "that is an irrelevant accident. No perspectival necessity requires him to do so. Marxist critics and formalist critics may be equally able to understand what a text means" (1976, 44). This is of course too optimistic. The most significant differences in interpretation do not derive from the critics drawing different significances from the same work, but from the fact that they can hardly said to be evaluating and interpreting the significance of "the same" work--their assumptions about the authorial meaning are widely divergent. Hirsch's argument seems to place the main critical contention in the determination of significance, which is variable, while authorial meaning, being historically fixed, could be approached objectively. This account ignores the fact that history is continually being revised and rewritten, and that this rewriting is itself a matter of ideological contention--hardly a basis on which to ground a critical consensus. Preserving the conceptual difference between authorial meaning, meaning and significance is fundamental, but only as a methodological principle. This conceptual difference will never be, as such, a basis for critical unanimity. The author's meaning did exist as a historical fact, but this does not have the slightest theoretical importance unless it is recognized to be a relevant intepretive element in the theoretical assumptions of the readers and critics. (Authorial) meaning and significance are critical constructs, and the subject of critical debate. Interpretations do have elements in common--but usually, different intepretations share just some assumptions about the work's total sum of (authorial) meaning and significance, not the whole of it. And here once again Gadamer's (1977) reflections on the pre-understanding which directs the direction of interpretation are relevant. This does not mean that Hirsch's probabilistic notion of the historical existence of authorial meaning is not necessary. On the contrary, it is indispensable as a regulative concept. (46). But it does not work exactly the way he puts it, nor does it work the way radical historicism would put it. In the last analysis historical meaning is always determined with respect to the critic's own historical position; his object is partly built for him by his own age and culture. The result of the scholar's investigation of the author's meaning is not authorial meaning in se, which is a historical or cultural noumenon, but rather the way in which we must conceive of that meaning in our own interpretive situation, taking into account the ways in which we conceive of other aspects of the author's cultural context. That is, the scholar's function is to make our assumptions about the past (or areas of knowledge in general) fit with each other, to insure that the past remains accessible as an object of knowledge. The critic's interpretation of authorial meaning is not a textual time machine, because we never relinquish a holistic conception about the past which is inevitably grounded in the present. What the critic's work seeks ultimately is not to uncover a noumenon, but to ensure the translatability between the concepts in his own area of knowledge and those of the larger cultural context of his own age. Our understanding of the relevant features of the past is constantly changing, and an author's activity must be reinterpreted in the light of this new understanding of the past and of other cultures.

However, Hirsch is right in observing that there is nothing intrinsically different about "the past" as an object of scholarly inquiry--another culture, another present-day conception require just the same kind of interpretive work. These objections against Hirsch's notion of objectivity, then, are not the radical historicist objections he seeks to refute. (47). The meaning of the other must be understood in the other's own terms. But we are unlikely to agree with all of our fellow critics on the precise nature of those terms. My contention against Hirsch's notion of objectivism is that different critics may share Hirsch's interpretive assumptions, be perfectly right in doing so, and yet fail to reach agreement: the heterogeneity of present-day perspectives adduced by Hirsch as a support to his argument against radical historicism (1976, 41) can draw a line right across the middle of English departments. Two scholars may validate quite different interpretations of authorial meaning. This is possible because the objectively sharable interpretive assumptions are not an algorythm for interpretation. Hirsch knows this: "The notion that a reliable methodology of intrerpretation can be built upon a set of canons is. . . a mirage" (1967, 203). Or: "There can be no canons of construction, but only canons which help us to choose between alternative [authorial] meanings that have already been construed from the text" (1967, 204). And yet he fails to draw the implications, drawn by a misguided notion of what objectivity and knowledge are. Knowledge is not ideologically neutral--it is, willy-nilly, at the service of ideological positions which make it possible and relevant in the first place. An age's conception of the past is not a monolithic, neutral construct. There is no agreement about the past--it is used as a metaphor for the present. A necessary metaphor, not one which is deliberate or wilfully distorting. The past is one of the languages of the present, and very often the interpretation carries the marks of its origin, of the project that made it possible For Horton, "interpretation will always be colored, or even determined, by present needs, and will always be as much re-creation as it is retrieval" (1979, 3). The post-structuralist enthusiasm of this affirmation needs to be toned down with another post-structuralist claim: the interpretation will only carry those signs if a further interpreter identifies them there. This is not always the case, but it is always a conceptual possibility. Interpretation takes place in the space between. Hirsch claims that "ideology is far more likely to determine the results of inquiry when the inquirer assumes that it must do so" (1976, 149). It is useful to keep this in mind. Not merely as a good piece of advice (which it is, in a way) but also as a cautionary instance of the opposite danger: believing in the possibility of neutrality is already an ideological claim; it is, moreover, a naive one, one that may make us assume that only "the others" have an ideological axe to grind. Instead, objectivity must be used as a purely regulative concept. Hirsch's objectivity is not purely regulative, since he believes in it as the final, practical result of the interpreter's activity.

Interpreting a text is an ethical question, but not in the way Hirsch would have it (cf. 1976, 90). There are ethical choices relevant to the interpretive activity which are previous to it and wider than a local compromise to be a faithful interpreter. Interpretation is of a a piece with the interpreter's overall ethical, political and generally cultural situation, the one which gives him his view of the text, which makes him capable of having an attitude towards it in the first place. The utmost respect for the text is necessary to all kinds of intepreters, if the interpretation aims at the strongest possible engagement with culture at large, including the interpreter's own project. But this will never result in a final agreement about the (authorial) meaning of the text. The past, as well as other cultures and attitudes, are the object of ideological contention just like the present and our own culture. That is why the idea of an objective value or validation is suspect. But there is something like objective knowledge. It is the basis of agreement which is indispensable for either further agreement or disagreement. We can always agree on what it is that we disagree about, and this already provides a significant degree of conceptual sharability.

Hirsch believes in the difference between validation and interpretation of meaning. Validation is always provisional and relative to the current state of knowledge, but it nevertheless aspires to a cognitive rank superior to that of interpretation: "A validation has to show not merely that an interpretation is plausible, but that it is the most plausible one available" (1967, 172). But a validation will also have to rely on a construction of the text as a measure for the different interpretations. At one point or another, the scholar's own construction of the text must determine whether a phenomenon is relevant or not, whether it counts as evidence of authorial intention or not. And at this point, validation becomes interpretation again. Its epistemological privilege over interpretation is one of degree and circumstance. Just as there are no privileged criteria for the evaluation of literary works (Hirsch 1976, 122) there is no absolute criterion to judge the validity of an interpretation. The work must be read in order to privilege one criterion of validation against the others, and it cannot be read from a neutral stance. Hirsch himself recognizes that new evidence or an analysis of the critic's reasoning may lead to a revision of the validation. But Sparshott is far more direct when he says that "there are no critical courts of last appeal" (1976, 14). This conception of interpretation also makes allowance for a more flexible approach, one which does not completely exclude the hypothetical, the provisional and the tentative form the heuristic activity. As Dennett puts it, fiction and role-playing are a necessary element in interpretation: when I interpret, "the state I put myself in is not belief but make-believe belief" (1987, 100).

The ideal of a universally valid interpretation is then most seriously compromised by the partial relativity of the concept of authorial meaning and its dependence on ideological strife. Hirsch seems to sense this, and that is why he distinguishes between a theorical aim of criticism (to achieve truth) and a practical aim, which is "agreement that truth has probably been reached" (1967, ix, 17f.). He believes that his noumenal criterion of truth does not impair its function in the system: "we can have the truth without being certain that we have it" (1967, 173). This is not very satisfactory, because it does not leave any room for a clear difference between truth and falsity. In fact, an objective interpretation as understood by Hirsch is a contradictio in adjecto . But an objectivist definition of truth need not posit a Tomistic correspondence between the intellect and the thing-in-itself. That is, the theoretical aim of criticism need not (and of course must not) be different from its practical aim: we need a criterion of objective truth which is different from Hirsch's, since his cannot be met. Humanistic disciplines do provide knowledge (it is Hirsch's main aim to maintain the idea that they do), but it is knowledge which does not need to be universally accepted. (48). Interpretations do not need to be "objective" either in the sense of being universally acceptable, or in the sense of being grounded in the nature of things, although they do need to have a thrust towards objectivity--to be objective in the only workable sense of the word. They need this in order to be convincing in the right context, since sadly enough what passes for universal truth is in fact the result of widespread agreement in an interpretive community. And widespread agreement in an interpretive community can only be reached by meeting that community's truth requirements: Aristotle noted that truth is a quite convincing rhetorical strategy. The academy needs, and has, criteria of validity and objectivity. They are not arbitrary, since they ensure that the functions of the institution be met--in fact, they rest on the very notion of agreement and communication. (49). I propose that the most widespread and effective criterion of objective truth is the (increased) translatability of knowledge from one cognitive medium, discipline or context to another. Accordingly, interpretation is the breaking down of complex semiotic objects into others which are more at hand; showing how a new or complex code can be translated into other codes which are more fixed, accessible, or better known. This is in essence nothing new: Aristotle already defined the process by which we acquire knowledge as the translation of the unknown into the known. In the Aristotelian view, interpretation consists the subsuming of particulars under universals (cf. Raval 1981, 43); that is, it is a form of translation. Of course, Aristotle's concept of interpretation is too abstractive and essentialist for us; the earlier view is more flexible. And it is further complicated by the fact that there is no universal proof that translatability between two areas of knowledge has been increased. The interpreters must agree not only on the data, but also on the rules for the validity of translation. These may be analyzed into data, but only according to further rules. Hermeneutic demonstration is always only partial, and has to rely ultimately on shared assumptions. Translatability is not increased in the void, but only in a particular institutional context.

Academic criticism is such a context--or group of contexts. The bulk of academic interpretation of literary texts in the American tradition from the forties on follows these steps:

­A paraphrasis of the most important elements of the action. When this phase is missing, it is taken for granted that the critic and his reader share the current assumptions about which is the relevant action-scheme. Although it is rarely acknowledged in an explicit way, this summary is already a part of the interpretation.

­The establishment of semantic isotopies in the action, the narrative and the discourse levels. (50). Usually, an interpretation will call attention to details which the critic thinks may have been overlooked by the reader, and will show how they exhibit the same logic which organizes the main articulations of the action which have been foregrounded in its paraphrasis. We can see thus the history of the interpretations of a work as a gradual colonization of areas of resistance, and also a gradual uncovering of new problems. The work of all the critics is not wholly compatible, but later critics benefit from the insight of the previous ones, and often borrow the ideas of a rival interpretation for their own purposes. Practical interpretation, like savage thought (and like theory itself) is essentially bricolage.

­The author's narrative is filtered through the critic's text; or, to use another metaphor, the critic's narrative is superimposed on the author's. A clear pattern thus emerges. The initial ambiguity of the story has become tractable for the purposes of the critic, and the new coherence given the story by the isotopies traced across its structure gives authority to the interpretation. A price is paid, however (Horton 1979): the interpretation systematically excludes or leaves untouched those elements which do not fit into the critic's grid. But this becomes simply a stimulus to devise a still more comprehensive interpretation, one that includes those elements which were left aside by previous interpreters. (51). The critic's narrative is finally assumed to be the essence of the author's, to have uncovered or reinforced its central meaning. A scheme of the whole work has emerged which invariably looks more accurate and meaningful than the paraphrase of the action furnished by the critic at the beginning of his essay. The next heuristic maneuver is the substitution of this paraphrase for the author's work as the object of interpretation.

­Usually, an iconic or analogical relationship is established between the isotopical version of the work produced by the critic and a literary statement or series of statements which summarize an outlook, message or world-view present in the work. Depending on the interpretive school the critic adheres to, the producer of this literary statement is the author, the author's unconscious, the collective unconscious, the author's culture, social class, epoch, gender, etc., the language of the work, or even the critic himself (assuming that there is any difference between these two possibilities). The extent to which this statement is taken to be conscious or deliberate on the part of the author will vary accordingly. For instance, the critics of Stephen Crane's story "The Monster" almost invariably point to "Crane" as the author of the literary statements they interpret in the story. (52). This is not to assume that Crane is always supposed to have thought of the literary statement in as many words. The literary statement is precisely stated by playing on the conventions of literature; it is the function of the critic to translate this statement from the language of literature to its implications on morality, social life, or politics; to make explicit in the language of criticism the analogical or iconic significance of the story. The statements formulated by these critics can therefore be said to reflect the authorial intention of the work, even if that intention was not formulated linguistically. The semantic isotopies isolated by the critic constitute a provisional semiotic code of their own. Together with the more general conventions of literature, this code is the original language of the literary statement. Sometimes, a detailed equivalence is drawn between the action and some of the cultural myths shared by the interpreter and the author (in the case of "The Monster," mainly the Bible and American history). Although many critics hesitate to use the word "allegory," this kind of interpretation is far more common than we might suppose.


Coherence is desirable, and it will always be the aim of the intepreter. But the scope in which coherence is achieved may be more or less wide. The doctrine of the "Intentional Fallacy," aiming at defining a specifically literary kind of interpretation, conventionally restricts the scope of interpretation in a way which is no longer acceptable today. When concentrating on the fascinating task of producing coherence inside the text, we should be careful not to disrupt another kind of coherence which is just as essential: that which exists between the text and its author, or between the textual image of the author and our conception of the personality of the author, of his age and his culture. All may be considered to be textual representations if we wish--of the all-encompassing text of History. A theory of interpretation should aim at making sense of literature within this enlarged context, and construct a representation which conciliates our sense of the internal voice of the text with the voices which join in the chorus, coming from other strands of the literary tradition, from the social context in which our interpretation takes place, from history at large, which is the largest of the interpretive contexts we share. It is in the arena of history where collective intentions are shaped and internalized, thereby framing our interpretations from very premisses. It is also the locus where interpretive theories can be considered as texts and objects of interpretation subject to a generalized hermeneutics. A maximum of attention to this larger context is our best way to strive for an interpretation which rests on a minimum of dogma.









1. I follow Dennett (1978) and Searle (1983a) in capitalizing the wider, specifically phenomenological sense of "Intentionality" (and "Intentional"). Back

2. According to Searle, "Intentional states represent objects and states of affairs in the same sense of 'represent' that speech acts represent objects and states of affairs" (Searle 1983a, 4). As it stands, this statement is false, due to Searle's (deliberate?) bracketing of his own theory. But I will assume throughout this type of statements mean something like "the same metalinguistic apparatus can be used for the representation of speech acts and Intentional states." With this proviso, see Searle 1983a, 4-13 for a comparison of speech acts and Intentional acts, in terms of propositional content, direction of fit, conditions of satisfaction, etc. Back

3. Both secondary and original, not derived (in the above mentioned sense). Back

4. In Virginia Woolf, Roger Fry, qtd. in Redpath 1976, 14. Cf. also Watson 1976, 163. Back

5. There are other important statements of this position by Valéry (see Scholes 1990, 54), by Frye (1957, 86), Sontag (1966, 9), Barthes (1977a) and Derrida (1988). Back

6. Newton-De Molina 1976, xvi. See also the varying interpretations of the scope of the "intentional fallacy" by the other contributors to this volume. Back

7. It is also this kind of anti-intentionalism which is found in the aestheticians (such as Eliseo Vivas) who conceive of creation as an exploration or discovery which is not the result of a pre-existing intention. Back

8. Fowler 1976, 242. Cf. Horton's view that intention is a complex concept that works differently in each level of the work's structure (1976, 104). Back

9. In Wimsatt and Beardsley's article there is no suggestion of a relation between intention and the use of public conventions. Nor, for instance, in relatively recent work such as Horton's (1979, 106). Back

10. Cf. Bateson 1953, 14; R. S. Crane 1953, 166, 169; Sparshott 1976, 108; Peckham 1976, 148; Watson 1976, 164f.; Close 1976, 182; Skinner 1976, 213; Hirsch, 1976, 87, Raval 1981, 46, de Man 1983, 25. While repeating some of the earlier injunctions, Wimsatt's revision of his position a quarter of a century later is somewhat more careful and, above all, it recognizes the legitimity of interpretations based on an intention found "in the work itself" (1976, 128). Still, there is only moderate interest in the author's meaning­the doctrine is still essentially the same. Back

11. Redpath 1976, 15 f. Cf. Hirsch 1967, 233; Skinner 1976, 213. Back

12. Wimsatt and Beardsley 1967, 5. That this is not the case is convincingly argued in Peckham 1976. Back

13. Wimsatt and Beardsley 1967, 6. Beardsley's Aesthetics reintroduced, nevertheless, the difference between interpretation and evaluation; and later Wimsatt recognizes the convenience of recognizing such a conceptual distinction (1976, 126).

14. On langue and parole, see Saussure 1949; but also Segre 1985, 190ff. Back

15. This explains why Hirsch believes that "the intentional fallacy has no proper application whatever to verbal meaning" and that Wimsatt and Beardsley ultimately respect the author's meaning (Hirsch 1967, 12). Back

16. Cf. Fiedler 1952, 259, 273; Cioffi 1976, 60; Roma 1976, 77f. In Hirsch's view, "it is unsound to insist on deriving all inferences from the text itself" (1967, 241); for her part, Horton notes that "knowledge of the conditions of composition alters interpretation" (1979, 95). Back

17. This is also Peckham's (1976, 141), Fowler's (1976, 249) and Raval's (1981, 59) interpretation of Wimsatt and Beardsley's doctrine. Back

18. This is still Wimsatt's view in his revision of his early doctrine. See his comments on Hirsch's analysis of Blake's "London" (Wimsatt 1976, 130ff.). Back

19. Cioffi notes that Wimsatt and Beardsley's notion of what is "in" the text is deceitful, because "externals" of several kinds, often biographical, are always introduced to reach an acceptable intepretation (1976, 68). Cf. also Roma 1976, 81; Harris 1988, 30. Back

20. Some critics have taken the "intentional fallacy" to refer to interpretation, in spite of Wimsatt and Beardsley's stated definition (e.g. Cioffi 1976, 57). Of course, its "indirect" connections with interpretation finally compromise its the authors' claims to an objective standard of valuation. Sparshott (1976, 108) notes that this version of antiintentionalism, in its exclusive aesthetic concern, forgets that the work of art is a human work, and not merely an aesthetic object. Back

21. Dewey 1934; Redpath 1976, 19. Back

22. Cf. Wimsatt 1976, 128; Watson 1976, 69. Back

23. Cf. Crane 33. This kind of criticism which deliberately ignores the author's meaning need not be just an aesthetic play on the part of the critic; cf. Spivak 244-245. Back

24. Newton-De Molina 1976, x. Cf. Peckham 1976, 140. Back

25. Hermeden 1975, 81. Qtd. in Raval 1981, 265. Back

26. Hirsch 1976, 90. According to Hirsch, it is the lack of a higher institutional authority in literary criticism which explains the relevance of authorial meaning. In religious or legal texts, special interpretive conventions ensure the control of authority over the meaning of the texts. But it is wrong to extend the principle of unlimited meaning to texts which do not have these institutional constraints (1967, 123). However, what is relevant in the institutional constraints of interpretation is not the structure of authority (which is after all a check) but the uses to which multiplicity of meaning is put. The institutional function of the academic study of literature as it stood in the New Critical conception was in this sense perfectly in keeping with their interpretive practices: a controlled proliferation of meaning and the hypostatization of literature are related and mutually sympathetic conceptions. Back

27. Newton-De Molina 1976, xi. Newton-De Molina draws this notion from Robson (1966). Cf. Hirsch 1967, 133, 140; Wimsatt 1976, 126; Fowler 1976, 252. Back

28. Roma 1976, 82. For Beardsley or for Robert Graves, the richest meaning is the best (Fowler 1976, 252). Sparshott (1976, 111) also believes that the "best" and most comprehensive interpretation is the right one, while at the same time he defends the criterion of the authorial intention. Back

29. On the "life" of a literary work, see Ingarden (1973, ch. 13). Back

30. Hirsch 1967, 237. Hirsch believes, however, that is very probable that the right intepretation is "the one which makes the most elements functional" (1967, 190). Horton has criticised the assumption that an interpretation ought to take every element of the poem to be the result of a unified logic of composition, or that the best interpretation is the one that makes most elements active (1979, 4ff.. Cf. also Fowler 1976, 252). Back

31. Cf. Crane 1953, 179. Back

32. Hirsch (1976: 103) quotes Kant's Critique of Judgement, sec. XXI of the "Analytic." Back

33. Hirsch's interpretive theory is not his own in an exclusive way. It is based on the practical assumptions of many critics, as well as on the theories of Schleiermacher, Boeckh, Dilthey, Frege, Husserl, Popper and Betti. Many other theorists, such as F. W. Bateson or M. H. Abrams, have defended positions similar to Hirsch's in the English-speaking academia. Back

34. The first version of Hirsch's terminology concerning the concepts of meaning, significance, interpretation and criticism is preserved by Harris (1988). Back

35. Hirsch 1967, 49. Hirsch goes on to say that the actual words in a sentence are types. Peirce's notion of token (the particular instance or manifestation of a type) would be more accurate. Back

36. Hirsch 1976, 49. This is also Peckham's argument (1976, 143). Peckham points out some fascinating analogies between the New Critical conception of semantic autonomy and the dogma of transubstantiation. Back

37. Hirsch 1967, 254 ff.; 1976, 49, 81 ff. Cf. Sparshott 1976, 113. Back

38. Blanchot 1955, 202; see de Man 1983, 63ff. Back

39. Cf. also Skinner 1976, 219. Back

40. Hirsch 1967, 236. Cf. Sparshott 1976, 112; Fowler 1976, 255. Back

41. Hirsch 1967, 17. Cf. Searle's argument against Derrida's deconstruction of speech act theories (1983b, 78). Back

42. "Therefore whatever interest comes from without, but yet can be taken as an interesting extension of what is surely in, may be admissible. It merely makes a larger whole" (Beardsley 1970, 36). Back

43. "The literary text, in the final analysis, is the determiner of its meaning. It has a will, or at least a way, of its own. The sense it makes . . . is what it offers for our aesthetic contemplation" (Beardsley 1970, 37). Back

44. Hirsch 1967, 132; my italics. The distinction between the construction of meaning to which an interpretation refers and the concretization of a work can also be compared with Beardsley's opposition between "local" and "regional" meanings; the latter belong to "the work as a whole or some large part of it" (1970, 44). Back

45. The concept of Horizontverschmelzung is endorsed by many theorists, e.g. Palmer 1969, 120; Horton 1979, 123; Wellek 1979, 577; qtd. in Raval 1981, 265. An even more extreme relativist formulation is put forward in Michaels (1980). Back

46. See Horton 1979, 5, and Derrida himself (section 2.5.1 below). Back

47. See Hirsch 1967, 40ff., 245ff.; 1976, 38ff. Back

48. In this it is not essentially different from scientific knowledge--only in the kind of use to which it is put. Knowledge of any kind does not need universal acceptance. It works in specific contexts and is irrelevant elsewhere. Back

49. Cf. Hirsch's principle of the sharability of verbal meaning (1967, 31ff.). Back

50. On the notion of isotopy, see Greimas 1966, 88ff. On the levels of narrative and discourse, cf. Bal's "story" and "text" (Bal 1985). Back

51. It is clear that beyond a certain point this activity is made possible only by a very specific set of assumptions about the nature of literature and the proper function of criticism. The extreme version I describe is historically localized in the Anglo-American New Critical tradition, to which practically all the interpretations of "The Monster" belong. The "intentional fallacy" doctrine should perhaps be considered as a carte blanche to push this practice to its limits without any qualms about the limitations of the author or the historical status of the meanings thus "retrieved." Back

52. Cf. my article on the interpreters of Stephen Crane (1989). Back