Even critics who find the price exorbitant and resist what they consider dangerous trends in contemporary criticism seem inclined to join in the study of readers and reading. Witness some recent titles: Wayne Booth’s Critical Understanding, Walter Davis’s The Act of Interpretation. E. D. Hirsch’s The Aims of Interpretation, John Reichert’s Making Sense of Literature, Geoffrey Strickland’s Structuralism or Criticism: Some Thoughts on How We Read. These theorists for whom criticism is essentially an elucidation of an author’s purposes have felt compelled to provide their own accounts of reading so as to challenge those that make the reader an anti-hero, a fall guy, an unabashed hedonist, a prisoner of an identity theme or of an unconscious, or a willful inventor of meanings. Seeking to eliminate such nonsense with, as Reichert puts it, a criticism that “cuts through the plethora of competing critical languages to recover and redignify the simple procedures of reading, understanding, and assessing,” they have thrown themselves into the critical competition for the rights to “the reader” (Making Sense of Literature, p. x). If, as Barthes says, the reader can live in contradiction without shame, this is doubtless a good thing, for on this disputed figure converge the contradictory claims and descriptions of current critical debate. “Reader and audience,” writes Susan Suleiman, introducing a reader-centered anthology, “once relegated to the status of the unproblematic and the obvious, have acceded to a starring role” (The Reader in the Text, p. 3). Why should this be?

One reason for interest in readers and reading is the orientation encouraged by structuralism and semiotics. The attempt to describe structures and codes responsible for the production of meaning focuses attention on the reading process and its conditions of possibility. A structuralist poetics or science de la littérature, Barthes writes, “will not teach us what meaning must definitively be attributed to a work; it will not provide or even discover a meaning but will describe the logic according to which meanings are engendered” (Critique et vérité, p. 63). Taking the intelligibility of the work as its point of departure, a poetics would try to account for the ways in which the work has been understood by readers, and basic concepts of this poetics, such as Barthes’s distinction between the lisible and the scriptible, would refer to reading: the lisible is that which accords with the codes and which we know how to read, the scriptible that which resists reading and can only be written.

A structuralist pursuit of codes leads critics to treat the work as an intertextual construct—a product of various cultural discourses on which it relies for its intelligibility—and thus consolidates the central role of the reader as a centering role. “We now know,” writes Barthes with that assurance that comes upon some writers in Paris, “that the text is not a line of words releasing a single ‘theological’ meaning (the ‘message’ of an Author-God) but a multi-dimensional space in which a variety of writings, none of them original, blend and clash. The text is a tissue of quotations drawn from the innumerable centers of culture.” But, he continues, “there is one place where this multiplicity is focused and that place is the reader, not, as was hitherto said, the author. The reader is the space on which all the quotations that make up a writing are inscribed…. A text’s unity lies not in its origin but in its destination” (Image, Music, Text, pp. 146, 148). To be sure, emphasis falls on the reader as a function rather than as a person, as the destinataire or place where the codes on which the unity and intelligibility of the text depends are said to be inscribed. This dissolution of the reader into codes is a critique of the phenomenological account of reading; but even if the reader is conceived as the product of codes—a product whose subjectivity, Barthes writes, is an assemblage of stereotypes—this would still make possible a differentiation of stereotypes, as in Barthes’s typology of “pleasures of reading or readers of pleasure,” which “links the reading neurosis to the hallucinated form of the text” and distinguishes four readers or reading pleasures: the fetishist, the obsessional, the paranoiac, and the hysteric (Le Plaisir du texte, p. 99/63).

Discrimination of readers might be a fruitful line of research—or speculation—but is seldom pursued by structuralists themselves, who focus on the codes and conventions responsible for the work’s lisibilité or intelligibility. In S/Z Barthes describes reading as a process of relating elements of the text to five codes, each of which is a series of stereotyped models and “perspective of citations,” “the wake of what has always already been read, seen, done, lived” (pp. 27–28/20). In a later essay, “Analyse textuelle d’un conte d’Edgar Poe,” he increases the number of codes by dividing what he had previously called “the cultural code”; and doubtless further additions are necessary. Michael Riffaterre argues in his Semiotics of Poetry that codes of poetic stereotypes serve as the basis for the production of poetic texts and that recognizing the transformations of these codes is a decisive moment in reading. One must also add to the list a code generally neglected in S/Z but extensively studied in other contributions to poetics: the code of narration, which enables readers to construe the text as the communication of a narrator to a narrative audience or narratee.

Work on the audience of narration, an important branch of the poetics of reading, investigates what discriminations are necessary to account for narrative effects. The narratee, defined by Gerald Prince as someone a narrator addresses, must be distinguished from the ideal reader an author might imagine (who would appreciate and admire every word and device of the work) and from what Wolfgang Iser calls “the implied reader,” a textual structure incorporating “those predispositions necessary for the literary work to exercise its effect” (Prince, “Introduction à l’étude du narrataire,” p. 178/7; Iser, The Act of Reading, p. 34). Peter Rabinowitz, in a series of excellent discussions, distinguishes four audiences: the actual audience, the authorial audience (which takes the work as a fictional communication from an author), the narrative audience (which takes the work as a communication from the narrator), and an ideal narrative audience (which interprets the narrator’s communication as the narrator appears to wish). “Thus, in John Barth’s End of the Road the authorial audience knows that Jacob Horner [the narrator and principal character] has never existed; the narrative audience believes he has existed but does not entirely accept his analyses; and the ideal narrative audience accepts uncritically what he has to say” (“Truth in Fiction: A Reexamination of Audiences,” p. 134).

Two things should be emphasized here. First, one proposes these distinctions in order to account for what happens in reading: Rabinowitz is particularly interested in radical disagreements about Nabokov’s Pale Fire, which can be traced to disagreements about what the narrative audience and authorial audience are supposed to believe. Second, these “audiences” are in fact roles that readers posit and partially assume in reading. Someone who reads Swift’s “A Modest Proposal” as a masterpiece of irony first postulates an audience that the narrator appears to think he is addressing: an audience entertaining specific assumptions, inclined to formulate certain objections, but likely to find the narrator’s arguments cogent and compelling. The second role the reader postulates is that of an audience attending to a serious proposal for relieving famine in Ireland but finding the values and assumptions of the proposal (and of the “ideal narrative audience”) singularly skewed. Finally, the reader participates in an audience that reads the work not as a narrator’s proposal but as an author’s ingenious construction, and appreciates its power and skill. Actual readers will combine the roles of authorial, narrative, and even ideal narrative audiences in varying proportions—without embarrassment living in contradiction. One ought perhaps to avoid speaking of “the implied reader” as a single role that the reader is called upon to play, since the reader’s pleasure may well come, as Barthes says, from the interaction of contradictory engagements.

Focus on the conventions and operations of reading leads critics to treat literary works as a succession of actions on the understanding of the reader. An interpretation of a work thus comes to be an account of what happens to the reader: how various conventions and expectations are brought in to play, where particular connections or hypotheses are posited, how expectations are defeated or confirmed. To speak of the meaning of the work is to tell a story of reading. This is to some extent the mode of Barthes’s S/Z but is more pronounced in works such as Stanley Fish’s Surprised by Sin: The Reader in Paradise Lost, Wolfgang Iser’s The Implied Reader, Stephen Booth’s An Essay on Shakespeare’s Sonnets, Michael Riffaterre’s Semiotics of Poetry, and my Flaubert: The Uses of Uncertainty.¹ Each of these critical accounts describes the reader’s attempt to bring to bear on the text the codes and conventions deemed relevant and the text’s resistance to or compliance with particular interpretive operations. The structure and meaning of the work emerge through an account of the reader’s activity.

This use of the reader and reading is not, of course, new. Long before Barthes, the response of the reader was often essential to accounts of literary structure. In Aristotle’s Poetics the reader’s or spectator’s experience of pity and terror, at certain moments and under certain conditions, is what makes possible an account of tragic plots: the types of tragic plot are correlated with differences in effects on the reader. In Renaissance criticism too, as Bernard Weinberg notes, the qualities of a poem were to be sought through a study of its effects upon an audience.²

Even the New Critics of our own day, now reviled for banning talk of readers as an instance of the affective fallacy (“confusion of what a poem is with what it does”), often show considerable interest in what a poem does when they describe its dramatic structure or praise the complex balance of attitudes it produces. The moments when New Critics do specifically acknowledge the role of the reader suggest a connection between reader-oriented criticism and modernism. In “Poetry since The Waste Land” Cleanth Brooks argues that a basic technique of modernist poetry is the deployment of unanalyzed juxtapositions, where “the interconnections are left to the reader’s imagination.” In The Waste Land Eliot declines to develop the implications of a juxtaposition of scenes but “has thrown this burden upon the reader himself, demanding that he relate the two scenes in his own imagination.” Once this modernist technique is identified, the critic can recognize its importance in earlier poems: Wordsworth’s Lucy poems, Brooks notes, “reveal gaps in logic that the reader is forced to cross with a leap of the imagination—they hint at analogies that cry out to be completed—and yet which can only be completed by the reader himself” (A Shaping Joy, p. 58).

Criticism must acknowledge the role of the reader when literary works, in Henry James’s phrase, “once more and yet once more glory in a gap” (Selected Literary Criticism, p. 332). But such acknowledgment does not basically alter the role that notions of reader and audience have played in descriptions of literary structure. When discussing many modernist works, one can stress the activity of the reader while treating it as the accomplishment of a determinate task: the reader must “work out for himself” the relation between two images, must complete analogies that “cry out to be completed,” or must piece together from disparate clues what must “really” have happened, bringing to the surface a pattern or design that the work conceals. This is the general role that Roman Ingarden and Wolfgang Iser have assigned the reader: to fill in gaps, to render concrete and determinate the Unbestimmtheitsstellen or places of indeterminacy of a work.³

If the activity of the reader has recently become decisive for criticism, it may be because some works—those Umberto Eco describes in L’Opera aperta as “open works”—provoke a general revaluation of the status of reading by inviting the reader or performer to play a more fundamental role as constructor of the work. Music provides striking examples, such as Pierre Boulez’s Third Sonata for Piano, whose first section consists of ten different pieces on ten sheets of music paper that can be arranged in various sequences (Eco, The Role of the Reader, p. 48). Works presented as a series of components that readers or performers put together in different ways often seem rather obvious experiments, whose primary interest may well lie in their impact on notions of art and of reading. By foregrounding reading as writing—as construction of the text—they provide a new model of reading that can describe the reading of other texts as well. One can maintain, for example, that to read Finnegans Wake is not so much to recognize or work out for oneself connections inscribed in the text as to produce a text: through the associations followed up and the connections established, each reader constructs a different text. In the case of more traditional works, this model invites one to account for resemblances among readers’ productions by investigating the productive influence of textual codes and institutionalized conventions. In this perspective, other accounts of reading—reading as recognizing a meaning or a pattern—are not eliminated but become particular and limited cases of reading as production. Although, as we shall see later, there are disadvantages to the view of reader as producer, theorists such as Booth, Hirsch, and Reichert, who combat this view of reading, in fact offer proposals that can be inscribed within it, as rules for particular, restricted sorts of rewriting.

In this perspective where, as Barthes says, “the stakes of literary work (of literature as work) are to make the reader no longer the consumer but the producer of the text,” variations in readers’ construction are no longer regarded as accidents but treated as normal effects of the activity of reading (S/Z, p. 10/4). This has implications even for critics who reject notions of readers constructing texts, for emphasis on the variability of reading and its dependence on conventional procedures makes it easier to raise political and ideological issues. If the reader always rewrites the text and if the attempt to reconstruct an author’s intentions is only a particular, highly restricted case of rewriting, then a Marxist reading, for example, is not an illegitimate distortion, but one species of production. This revised conception of the status of reading may thus subtend criticism that takes no interest in the avant-garde texts that provide the leverage for the change in perspective.

Contemporary literature also encourages concentration on the reader because many of the difficulties and discontinuities of recent works become amenable to critical discussion only when the reader serves as protagonist. To analyze one of John Ashbery’s poems is first of all to describe the reader’s difficulties in making sense. In France interest in the reader seems to have arisen at the moment when it became impossible to discuss the nouveau roman as a purely objective, nonanthropocentric presentation of reality. The problematizing of plot and character in works such as Robbe-Grillet’s Le Voyeur and Dans le labyrinthe encouraged critics to locate the force and interest of these novels in their violent engagement with the conventional novelistic expectations of readers and their disruption of habitual processes of sense-making. Outside the French tradition we find further evidence that analysis of difficult modern works requires reference to readers and reading. To take just one example, Veronica Forrest-Thomson’s energetic and inventive Poetic Artifice: A Theory of Twentieth-Century Poetry displays no interest in the behavior of individual readers. Concerned with poems as artifice or artifact, and with what they mean, Forrest-Thomson describes two processes, “external expansion and limitation” and “internal expansion and limitation,” by which difficult modern poems produce effects of pastoral and parody. But to explain these effects and to show how formal feature...