1—
The Concept of the Canon

I.

The concept of a literary canon is generally understood in either of two ways, each very much opposed to the other. Let me state them in their most extreme form: on the one hand, the canon is conceived of as a body of texts having the authority of perennial classics. These texts, "the great books" (as at least one American college has institutionalized them in a course of instruction), are, as they always have been and always will be, nothing less than the very best that has been thought and said. To understand their content—to have isolated for further meditation their themes or ideas—is to gain or make some nearer approach to timeless wisdom; to apprehend their form is to experience the beautiful or at the least to perceive a significant order. Sympathetic contact with these texts cannot help but make one a better person, or—the phrase is a curious one on inspection, to be sure—more human.

On the other hand, however, the canon is taken simply as the name for that body of texts which best performs in the sphere of culture the work of legitimating the prevailing social order: canonical texts are, as they always have been, the most useful for such a purpose (although the modality of their usefulness may, of course, alter with time). To understand their content is largely to accept the world view of the socially dominant class; to apprehend their form is to fail to perceive that acceptance as such. Sympathetic contact with these texts tends mostly—although not always or exclusively—to con-

tribute to that ideological conditioning, the production of that consciousness, necessary to conform one willingly to one's—usually subordinate—class position in society.

The latter conception of the canon derives from what Paul Ricoeur and others have called the hermeneutics of suspicion. Rather than presenting the best that has been thought and said, or the "great tradition," for this perspective on the matter, the canon is simply what Raymond Williams has called "the selective tradition " (1980, 39), where "tradition" may be understood as "an aspect of contemporary social and cultural organization in the interest of the domination of a specific class . . . ," as "a version of the past which is intended to connect with and ratify the present" (1977, 116). For the hermeneutically suspicious, the texts of the canon have no inherent authority or value; rather, they are socially authorized and institutionally legitimated. Nor do they make their own way into our hearts by appealing unmistakably to an intrinsically human longing for truth, wisdom, beauty, order, and such; instead, they are regularly and insistently urged upon us so that they may promote and sustain by cultural means perspectives and values that are not necessarily consistent with the fullest conceptions of human possibility. This view need not abandon entirely the question of greatness, nor must it deny that ideologically objectionable works may be constructions of great power; rather, it simply points out that, in actual practice, a considerable number of the items that continue to be named among the master texts seem less and less able to sustain their position by any appeal to greatness alone.

In one form or another, hermeneutical suspicion arises, naturally enough, among those who have been impressed by varieties of Marxian, psychoanalytic, and anthropological thought, leading them—here it seems not only awkward but

false for me not to say us —leading us to consider ourselves oppositional intellectuals (neither traditional nor yet organic in the Gramscian sense). In this country, most of those interested in these matters, no matter what their view, are affiliated with the academy—the University, more precisely—and the disciplinary allegiance of the oppositional intellectuals tends to be to the Geisteswissenschaften, or sciences humaines: most of us would not, today, object, I think, to being described as social scientific workers, although our institutional and professional allegiance most usually places us in departments of English, language, and literature. As members of the professoriate, even those suspicious of the canon's claims to eternal truth and beauty are nonetheless culturally conservative, for one cannot, after all, not teach some of the traditional canon. And, in any case, there is much in the traditional canonical texts—in many of them—which, after the appropriate work of historical referencing and demystification has been performed, still appears even to the hermeneutically suspicious as beautiful and/or somehow true.

If this instrumental or pragmatic view of the canon (to call it that) I have so sketchily presented derives from suspicion and generates opposition, the traditional view of the canon—transcendental and abstractly theoretical—might be said to derive from a hermeneutics of faith that generates validation of the society whose noblest cultural productions the great canonical texts are taken to be. It is only fair to note here, however, that the canonical validators would—again, so far as these broad generalizations may hold—object to descriptions of their position that make reference to such things as cultural opposition and validation as being too politicized to explain that position. Such descriptions, they would insist—and in the present instance they would be entirely correct—

could only come from someone with a view very different from theirs. They are interested in literature, not politics; in Man, not individual men (it is only recently that some small opening to women has been worked into the system); in what is enduring and unchangeable in our putatively common human destiny, not in the vicissitudes of epochal accident or of cultural variation.

Oppositional intellectuals, for all that they may appreciate and linger over the internal workings of canonical or even noncanonical texts, must always, as I have noted, engage in a certain historical referencing as they read, relating the literary text to a world that is indeed hors texte, "real," and prior to the text.^[1] Traditional, validating intellectuals, however, are not, in their reading, particularly concerned with what is socially and materially beyond the text itself (except as "background"), the "real," in the Mallarmean sense, existing mostly to provide materials for a book—which, as they can often demonstrate in a sophisticated manner, performs a complex imaginative transformation of context or reality: this is the artwork that makes art. (And there is, of course, nothing to demystify in the work of art.) For them, literary study has little to do with science (except perhaps as defined by Northrop Frye). Rather, it is either concerned with that compound of ethics and esthetics perhaps best summed up in a phrase such as "the moral imagination," or it is concerned with something now called simply "reading," a purely technical and so wholly disinterested operation. The validators in the mold of Arnold, Trilling, or, indeed, Clifford Geertz tend to consider their professional allegiance to be to the humanities; and these are, to be sure, kept alive most

diligently, once again, in university departments of English (although Geertz is professionally an anthropologist).^[2] Curiously enough, the most recent formalists—let me call them for these purposes, with reference to the New Critics, the New Rhetoricians—for all that they are also conservative vis-à-vis the canon, tend to be antihumanist: their allegiance is not to a universalized, ahistorical concept of Man, but to an equivalently ahistorical, systematized concept of Language. Canonical texts, then, are those that show us either Man struggling to be fully Himself (hopelessly but nobly) or else Language trying to say fully what it means (hopelessly but—in the master texts—interestingly).

There are difficulties with both these views of the canon, as there are attractions in both as well. The transcendental—essentialist view, for example, like all full-blown idealisms, can only be accepted as an act of faith; it requires that one engage in what Edward Said has called "religious" criticism in contradistinction to the secular criticism he so eloquently affirms. Appeals to inherent greatness, to unseen yet dimly felt orders of great power, and the like are not conducive to this-worldly understanding—unless one chooses to say with Saint Augustine credo ut intellegam . The traditional view of canonical

authority can be maintained only by rigidly separating literary value from value of other kinds—and by positing literary judgment as a first-order judgment, something that can be formed prior to or independently of our judgments about other kinds of value. Once the introduction of empirical (not to say statistical) evidence is permitted and the condition of cognitive responsibility is accepted,^[3] the transcendental-theoretical conception of the canon tends to lose a good deal of its force.

Yet a purely instrumental or pragmatic view of the canon, for all its attractions, has its problems, too. For, at least in its extreme version, it tends to see the canon as formed exclusively by power relations: the canonical texts are the surviving victors on the battlefield calling for due praise. If they are an obnoxious group, we can never join them and so we must get out there and beat them. But this is to adopt a Thrasymachean perspective, the world view of primitive capitalism, a crude form of social Darwinism, or indeed an equally crude version of Marxist class struggle. It is to adopt Foucault's bleak vision (somewhat modified in the last work) of discourse as power and power as everywhere, so that even to fight and win is only to become oppressor in one's turn—to force people to read our books now, not theirs (and, of course, they will fight back, conflict unending). And this sort of Hobbesian war of all against all forever is very far from what most instrumental pragmatists desire.

To avoid it, three positions in relation to the canon have, thus far, seemed open to the instrumentalist. The first is simply to declare a universal peace, as it were, one in which everyone is conceded equal greatness—or, rather, one in which any claim at all to greatness or superiority is automatically suspect. This means, in effect, that one simply dispenses entirely with the concept of the canon—of any canon of authorized texts—so that what one teaches and writes about is simply books that happen to be interesting, or useful for one sort of demonstration or another. Not evaluative but only functional criteria are admitted as determining text selection. Cognitive responsibility here consists in a willingness to provide arguments for justifying the functional criteria as generally reasonable ones.

According to this view, one could not claim that Uncle Tom's Cabin is either better or worse or even as good as The Scarlet Letter; one chooses it for presentation as it fits a particular context of concern in the interest of fully acknowledged extraliterary values; those with other concerns and other values will choose other books—perhaps The Scarlet Letter, perhaps the novels of Lydia Maria Child and William Gilmore Simms and the poems of Fanny Fern. Or, if one is still interested in such a category as "nineteenth-century American literature," one may as well found one's view of it upon reading Kit Carson's autobiography and the journals of frontier women as upon novels and poems of any kind. Whether these latter texts are literary texts, indeed, whether it matters whether they are or not, is something I shall take up in a moment.^[4] In any case, the series of texts in question has no

authoritative status beyond its functional instantiation in fact; other texts might do as well; at other times, no doubt, other texts will do as well. The body of texts actually taught in any course of instruction or referred to in any work of criticism, far from constituting a canon, now becomes merely a pedagogical (some might say polemical) series; others may teach and comment upon these texts, if they wish—or they may equally well ignore them.

This solution certainly clears the air of a lot of cant about greatness and inevitability and unquestionable authority, but it also works in the interest of a social vision of pure atomized fragmentation. Everyone does her own thing and we can only hope that our thing will hold up at least well enough to be visible; even should our thing attract the attention of many, there are no grounds for proposing it generally to all, for it remains "ours" only. Barbara Foley's phrase, "the fetishization of the refusal of mastery" (129), comes to mind for this strong form of cultural relativism as one way to describe literary life after the canon. Or perhaps an analogy might be made to the poststructuralist infinity of signification: just as there is no decidable meaning to a text, so, too, here, would there be no central or major texts, no ground on which a canon might stand, only the free intertextual play of books taught or read. Of course, in the postmodern world, if one dispenses with belief in the continuity and commonality of a tradition of excellence (moral, political, esthetic) embodied in some form of canon, one then acquiesces in what Fredric Jameson has described, in "Periodizing the '60's," as a "dispos-

able" culture of the "metabook, " a culture made up of those texts that (perhaps against the functional intentions of those who compose and propose them) fill the residual (but dying) book function in our society. Rather "than the ambition to express a proposition, a position or a system with greater 'truth' value"—Jameson is speaking of philosophy—or, for literature, to read or to write a great or central book, one engages in expression of a "radically occasional" nature (in Sayres, 193). This, too, is something most of the hermeneutically suspicious do not desire.

A second way to avoid endless ideological struggle for the canon of great or central books is not to jettison the canon entirely but to propose a canon whose authority is merely statistical. Wayne Charles Miller, for example, has suggested revising the canon of American literature on principles—so it seems to me—of virtually statistical representation.^[5] This position on the canon I shall call the ethical-ontological reform. It claims that we should, in fairness, read a proportionate number of authors who are actually available. Since it is indisputable that America is not almost exclusively made up of white, male, eastern WASPS, courses in American literature should not almost exclusively be made up of texts by white, male, eastern WASPS. With this observation I agree wholeheartedly. But even if one were to accept the principle of proportional revision, the ethical-ontological reform has no way of deciding which authors should finally be taught, or even which authors should represent their particular ethnic and/or racial groups. If it is easy to agree that Black Elk might well be taught instead of William Cullen Bryant (but maybe it isn't so easy to agree after

all, and maybe the most famous Native American autobiographer should yield to some other Native American autobiographer) or that Whittier (if he still shows up) could be dropped to make room for Frederick Douglass (but why not Linda Brent or Harriet Wilson?) or Maxine Hong Kingston, it is less easy to know whether Kingston should be supplemented by Frank Chin, or whether (for the semester has only so many weeks) Ferris Takahashi should be taught instead of Piri Thomas. (And we have still left out immigrant Jewish writers, Chicanos, Italian Americans, Scandinavians, and a great many others as yet unmelted into general Americanness.) This inability or refusal to imagine a principle that might merge valorized difference into some collective identity (a dialectical principle, to be sure, so that common identifications do not obliterate historical differentiations) finally works against the hope of a common culture; its canon reflects and can reflect nothing but the changing demographics of each historical moment. We never approximate an American literature but remain instead at the level of Miller's multiethnic literatures of the United States.

It is perhaps in recognition of this truth that the third way comes about for hermeneutically suspicious pragmatists to avoid a view of the canon as inevitably oppressive and endlessly contested. Rather than an abandonment of any canon whatever or the acceptance of a demographically authorized one, the position I turn to now reconstructs the canon on the basis of a strictly experiential authority. Recognizing that the traditional canonical texts have been pretty exclusively the phallogocentric texts, in Jacques Derrida's term, those representing the experiences and values of a very small group of elite Western males, this position decides that the books worth promoting are not just any books, not even those that offer a kind of proportional representation of what is actually there,

but, instead, those which are directly relevant to their audience's experience. This reform of the canon leads increasingly to the sense that any attempt to consider American or English literature as a whole is just too hopelessly broad, and it is bound to urge a shift more and more to courses in and books about ethnic literature, working-class literature, third world women's literature, regional literature, and so on, and more nearly in isolation from one another rather than in relation to one another.

Here, one gives up the proposal to a wide range of American students and readers even of a multiethnic canon. Instead, one offers to readers of the Plains region, Willa Cather's novels and the epics of John Neihardt; to readers of the southwest, perhaps Paul Horgan and Witter Bynner; New Yorkers may choose from Emma Goldman, Mike Gold, and Abraham Cahan, Langston Hughes and early James Baldwin, Oscar Hijuelos and Piri Thomas, among others. In New England, at least to predominantly male readers, Hawthorne, Thoreau, Melville, even Longfellow and Whittier and Bryant will continue to be taught. Thus Paul Lauter's experience in trying to present Faulkner's The Bear to a group of working-class students in New York would rarely be repeated (they thought the story silly; they laughed),^[6] for everyone will find his personal concerns more or less closely reflected in what he or she is asked to read.

The latter two of these reactions against the essentialist view of the canon and against the instrumentalism of endless conflict do retain a measure of the authority that traditional views accorded to canonical texts, for the books they propose to us are, after all, deemed to be more important or somehow

more necessary than others—demographically, or at least in regard to their audiences' own experience. But what they give up is enormous. For they surrender the possibility of attaining just that sort of perspective on our individual experience and our historical moment that a broad acquaintance with literature can provide. Instead of the attempt to define a canon that might become what Elizabeth Fox-Genovese has called a "collective autobiography" and "to introduce some notion of collective standards," they "settle for education as personal autobiography or identity," and thus "accept the worst forms of political domination" (133). In their desire to acknowledge difference and accord it its due, they give up the very possibility of a common culture, one that may at least imply a common society. An American literary canon, I should say, is worth fighting for as the complex record of possible national identities; Crèvecoeur's question, What is an American? is unanswerable, of course. But it needs again and again to be posed.

In these regards, then, the most appealing aspect of the traditional view of the canon is its determination to believe in a common culture, a body of work that defines some part of what we all (whoever "we" are) can believe in and share. To quote Elizabeth Fox-Genovese once more," . . . however narrow and exclusive the canon we have inherited, the existence of some canon offers our best guarantee of some common culture" (132; my emphasis), one which, as I would gloss her remarks, in its heterodoxy expresses some part of our selves as collective selves, so that we see ourselves neither as simply accommodating or opposing but, rather, informed by others dialogically . The idea of the collective self as it is presented in the collective autobiography we call the canon, and the social vision implied by such a collective self, is something I will return to. These are

of especial importance in any attempt to keep the concept of an American literature meaningful.

II.

I began this discussion of the canon with the concept of the canon itself; but at several points it may have seemed that I should, instead, have begun with the concept of literature: that is, to have spoken first not of the literary canon but of the canon of literature and thus to define what is at stake in the various understandings of that term before speaking of the makeup of the literary canon. Even if one abandons not only the traditional literary canon but any canon, choosing, instead, only a pedagogical series constituted by no authority other than the intention of the one who instantiates it, one constituted by a purely demographic authority, or only by the authority of its audience's experience, still the question persists whether these texts, whatever the principle governing their instantiation, are literary texts. Fox-Genovese, for example, whose work I have cited with approval, does not concern herself with distinguishing between those texts of the canon that may reasonably be designated literary texts and those that may not—as others who would revise the canon, myself among them, on occasion do not. If we would keep the distinction between literary and other texts, whether it is one of kind or of degree, we must not only offer grounds for it but justify it as well: to ask, that is, whether the distinction between literature and other types of written discourse is possible to maintain and also whether it is worth maintaining.

The order I have adopted—canon first, then literature—seems preferable not so much for logical as for strictly empirical reasons inasmuch as one can more or less point to the literary canon, at least "as we have inherited it," whereas it is more difficult, probably impossible, to point to "litera-

ture" itself. The canon of Western literature, as this is traditionally conceived in England and America, has consisted of the texts of Homer, Sophocles, and Dante, Chaucer and Spenser and Shakespeare, Milton, Donne, Dryden, Wordsworth, Austen, Flaubert, George Eliot, Dickinson, Yeats, James, Joyce, Proust, Lawrence, Mann, Woolf, and T.S. Eliot, probably Stevens and John Berryman, along with the texts of many unnamed others—which texts, whatever their authors' names, would not significantly alter our sense of the kind and quality of the traditional literary canon. For traditionalists, to be sure, "literature"—"great" literature—is simply what these men and some few women wrote. But just as there is a challenge to the canonicity of these ostensibly "great" authors, so, too, is there a challenge to the view that their writing defines literature.

For "literature" obviously does not include everything that even these authors wrote, probably not their letters, for example. Or would it be that some of these authors wrote "literary" letters while others did not? We might possibly admit Milton's Areopagitica to literature, although that would likely not be the case with Eliot's Idea of a Christian Society, nor with Berryman's critical biography of Stephen Crane. Henry James's plays would have to be admitted as compositions of a literary kind—although these last might be considered bad literature. Do James's novels alone then help us define literature (but what of the prefaces for the New York edition?)—or is it hopelessly circular reasoning to say so? Meanwhile, we know today that all "texts," even menus, ads, and TV programs, can be given literary readings, can be read, that is, as literature or in ways consistent with the sorts of readings that have traditionally been reserved for literature. Are we approaching the time therefore when the canon of literature should open itself to the best-written ads, or is it, instead, the case that literary readings do

not literature make? And if not, what happens to such apparently nearer-to-literary texts (nearer than ads or menus, anyway) as Rousseau's or Augustine's Confessions ? It may be that we should indeed speak of the canon without qualification as to canonical type; still, if only for practical reasons (I refer to the traditional ways of mapping subject matter areas in American universities and colleges and the classification systems in our libraries and bookstores), perhaps we would do well, first, to consider what may be meant by literature. (There are also good theoretical reasons for doing so.)

In this matter there are also two opposed views that I will again state in extreme fashion; they parallel the essentialist and the instrumentalist views of the canon. On the one hand, it is held that there is certainly such a thing as literature, a type of discourse qualitatively distinct from other types, with discernible, distinctive traits and unique functions, a thing that changes, to be sure, but only within certain fixed limits (beyond those limits literature would not be literature), so that its lineaments may be traced over time with some assurance. "The existing monuments," as T. S. Eliot put it, "form an ideal order among themselves," an order "modified by the introduction of the new . . . work of art among them," but modified, it appears, only "ever so slightly" (50): for literature, plus ça change plus ça reste le même . No wonder, then, that, as a Canadian doctoral candidate in literature complained to me, according to her department head, fur-trade journals are simply not literature nor may they be studied as such for the degree. And as Christopher Clausen has recently insisted, "It Is Not Elitist to Place Major Literature at the Center of the

English Curriculum";^[7] so to place it, of course, one needs a certain steady faith that he knows the exact nature of Literature, Majority, and Centrality.

On the other hand, literature is taken as just another socially determined category of discourse: like "greatness," what is or is not "literature" is whatever those empowered to define it say it is. Literature has no essence nor any distinctiveness of its own; it is not inherently recognizable, nor can we know it except insofar as it is institutionally defined. Just as capital gains are what the IRS says they are, so are novels or poems—so is literature—only what the publishers, reviewers, and professors say they are, only what the librarians class under one group of numerals rather than another, or the bookstores place in one section or another. Consider, for example, the opinion of those very differently flamboyant personalities, Leslie Fiedler and Roland Barthes. Fiedler writes, "Literature is effectively what we teach in departments of English; or, conversely, what we teach in departments of English is literature" (73). Or, in Barthes's words, "The 'teaching of literature,' is for me almost tautological. Literature is what is taught, that's all" (1971, 170).

For there is, after all, no single trait that can be isolated in the texts traditionally treated as literature which could unequivocally distingush them from their nonliterary relatives: neither their fictionality (lies and sensational journalism, not to mention Ronald Reagan's accounts of Libyan terrorist activities, are fictional but not usually considered literature); nor their form (greeting card verse has rhyme and meter as do the lyrics of some popular music, but the first of these is never claimed for literature, whereas the second, when claimed, meets vociferous opposition); and assuredly not the figuralism of their language or any excess of signification over semantic communication. William Labov's studies of black

American speech on the one hand, and Paul de Man's studies of philosophical writing on the other, uncover a considerable element of figuralism in ordinary or scientific discourse, but neither of these has as yet—the situation is perhaps in the process of change—been claimed specifically as literature.^[8] And as Barbara Herrnstein Smith remarks, there simply are "no functions performed by art works that may be specified as unique to them" (14). This is to say that apart from "literature's" existence as an abstract and idealized object of veneration or theorization, what we have to deal with is simply those texts that for one reason or another have been treated as "literature," literature as a strictly social category of discourse. As Terry Eagleton has claimed, a time might even come when Shakespeare himself would not be seen as a producer of literature; nor is there any fixed and intrinsic manner of determining just what sort of thing Gulliver's Travels, Our Lady of the Flowers, Glas, or, indeed, Canadian fur-trade journals might best be considered.

This is a strong position, I believe. But just as one may adopt a generally pragmatic and instrumental view of the canon without denying the usefulness of such a category as the canon, so, too, one may adopt a hermeneutically suspicious view of literature without denying the usefulness of the category itself. Though I agree that there is —ontologically, essentially—no such thing as literature, the term nonetheless embodies a very specific history and, as well, an aspiration that should not too quickly be abandoned.

We might begin by noting that historically literature was

simply the term for whatever language Western cultures deemed important enough to preserve by means of the technology of writing. Literature, from Latin littera, letter, as is well known, served broadly to indicate anything that had been written down and—to achieve a measure of social circulation—copied over. (For oral societies without alphabetic "letters," literature is whatever language is deemed worthy of sufficient repetition to assure that it will be remembered and passed along. From a quantitative point of view, oral cultures will inevitably have less literature than chirographic and print cultures.) Now, although everything written down might have at least an etymological claim to be taken as "littera-ture," poetry, the term that precedes our modern term literature, was always distinguished as a specific kind of writing different from lettered discourse in general because of its curious capacity to appeal either to the emotions, the senses, the imagination, or the fancy—the faculty posited varies with each commentator's psychological understanding and social vision—all of which were contrasted in some greater or lesser degree with the rational or analytical faculties to which other written discourse makes its primary appeal. Poetry, by its rhyme and meter, or—this is the case in literary prose—by its figures and structure, delights us; it is pleasurable beyond what can be accounted for rationally. Whether it stands therefore in opposition to rationality is the great question.

In general it may be said that all littera-ture, including the poetic part, was required to be pedagogical to the extent at least that it traditionally was expected to play a role in instructing our reason in how to be the kind of citizens valued by our particular societies. (If littera-ture did not do this it would be against the self-interest of society to preserve it in the first place.) But a type of littera-ture—poetry—which

delights us and appeals to our emotions, imagination, and so forth, may well be something, as Plato felt, that works against rationality and so against social formations presumed to rest upon rational foundations. The problem is to reclaim poetry for the state, or, less grimly, to show its social value. Aristotle's cathartic notion of what happens to the emotions aroused by tragic poetry had exactly this aim; perhaps I. A. Richards's psychoanalytic revision of Aristotle did, too. As Sir Philip Sidney summarized Renaissance thinking on the matter, poetry might be defended by emphasizing its epistemological status, its fictive nature, which permits a more abundant freedom to mix types of discourse: the philosopher must teach abstractly, by precept, the historian concretely, by example. But the poet can do both, and without being tied, in the latter case, to what actually occurred.^[9] The delight poetry produced was a kind of bonus, one that contributed to the effectiveness of its teaching—which was, to be sure, the chief reason for preserving and valuing it. Put another way, the surplus of signification, the excess that pleased, had cognitive value.

In general, defenders of poetry from Aristotle to Sidney, Percy Shelley, Walt Whitman, Matthew Arnold, I. A. Richards, T. S. Eliot, Leon Trotsky, and Herbert Marcuse—let me take them as examples of some of the best-known spokesmen for (what at this point we must call) literature—have affirmed the social utility of poetry, the importance of its pedagogical function, and the compatibility of its extrarational appeal with right reason, regardless of their political

differences (and so their stance in relation to an existing society), regardless of their understanding of the component makeup of its delightfulness.

It is only recently—dating perhaps from around the mid- or late nineteenth century—that anyone has attempted to divorce literature from social utility. The argument is that poetry is good in itself, presenting beauty that needs no equation to truth; it is purposeless purpose, and so on. In the art-for-art's-sake line of reasoning (to call it that), poetry's delightfulness is deemed sufficient to justify it without any need to claim a teaching function—or any function whatever. In America, it is not so much the opinion of Verlaine, J. K. Huysmans, Walter Pater, or Oscar Wilde, mutatis mutandis, that has weighed in the attempt to isolate literature from the social, but that of Flaubert as passed through Stephen Dedalus's esthetic and expressed by the New Critics, whose poetic theory—for all their abundant concern for social questions—insisted on the literary relevance only of the tensions and ironies internally generated by or in a poem. This separation of the esthetic and the socially useful was reinforced by Roman Jakobson's specification of the poetic function of language as self-reflexiveness, and, most recently, by New Rhetorical deconstruction's insistence on linguistically generated moments of textual aporia as the ultimate horizon of literary language. Recent valorization of "the pleasures of the text,"^[10] a turn to jouissance and a literary erotics which brackets if it does not overtly scorn questions of poetic functionality, may be seen as only the culmination of a determination to uphold the delighting function of literature in spite of, not as complementary to, its teaching func-

tion. In America this view animates a range of criticism from Van Wyck Brook's distinction between "art" and "expression," to Lionel Trilling's separation of "genuine" literature from fictionalized propaganda, and Richard Poirier's valorization of a stylistic "world elsewhere." Historically, however, this remains a minority view born of a dissatisfaction with the actualities of social life and a despair of political change in the interest of a transformation of social life.

The foregoing comments are intended to suggest that there are some historically urgent criteria for claiming literature as a distinct category of discourse. Granted, as Saussure said of language, literature is not a positive entity; it has no essence, no body. But it is distinguishable differentially, if only in degree, from other uses of language. The question, What is literature? then, may be answered by positing a range of texts in which the conjoint attempt to teach and delight appears, in the old, Russian formalist sense, as the "dominant." In this way we may distinguish literature (although always as the result of an act of interpretation and evaluation that is open to question) from other uses of written language that have aimed at, though certainly not achieved, instruction (philosophy, theology) or pure delight.

But is the effort to sustain the difference between literature and other forms of discourse to any point? If, as I would readily acknowledge, one can delight in philosophical discourse that is primarily oriented toward instruction; if one is always inevitably instructed by discourses of delight of whatever type,^[11] why

then labor to maintain a special category called literature, a category, as I have claimed, that would be constituted precisely by the dominance of its conjunction of the pleasurable and the pedagogical? What answer I can give depends upon an analogy to the canon: just as it is worthwhile to retain the notion of a body of texts centrally significant to a common culture, an approximation to a collective autobiography, the story of our integrated selves as participants in a common society, so too is it worthwhile to retain the notion of literature as comprised of those texts that provide instances of and occasions for the integrated self. Literature is where the affective and rational coexist and complement each other in language, where fancy and imagination press for freedom beyond the bounds of a material constraint to which proper due is given. Literature is that mode of discourse which foremost seeks to enact and perform its insights, insisting that we understand with affect, feel with comprehension. In this sense, T. S. Eliot was wrong: there is no literature of dissociated sensibilities; in this sense, Roy Harvey Pearce was right: literature is always an instance of the search for wholeness.^[12]

Let us then optimistically assume we have staked out a claim for some properly literary territory; let us say we know differentially what literature is. This only tells us which texts are candidates for the canon of literature. How to choose from all of them those that are—what? Great? best? most useful, relevant, cogent? From a purely formal point of view, considering what I will call the techné of delight, there probably are intrinsic criteria for literary excellence. The only problem is that decisions as to which intrinsic criteria to privilege turn out themselves to be dependent upon prior extrinsic or ideological determinations. And, to be sure, even if we agree upon which texts meet our particular standard of literary "excellence," our further choice from among them must rest upon the kinds of imaginative and affective appeals they make, a function of our phronetic desire (phrónesis^[*], or "knowing how to act . . . knowing what a situation calls for in the way of action" [Bruns, 16]), or the kind of social pedagogy we wish to further. It is the image of individual wholeness and collective cohesiveness ("content") we approve as presented by means of those techniques ("form") we enjoy that determines our choices for the literary canon. However much we may, with Christopher Clausen, dream of value-free literary judgments, these exist—alas?—only in dreams. (And as Roy Harvey Pearce is fond of noting, in history, not dreams, begin responsibilities.) To acknowledge that in literary studies all is ideology is not suddenly to endorse the Thrasymachean-Hobbesian-Foucauldian view of the endless war of all against all I was at pains to criticize above; rather, it is to say quite simply that literature is not the first thing (nor is it the last and only thing) anybody ever cares about. It is to take my epigraph for this book to heart.

Consider first the question of form, and so the techné of

delight. At least from the modernist period up until the current postmodern reaction, subtlety and complexity of verbal texture were the intrinsic criteria for literary excellence: connotation rather than denotation, implication rather than direct statement, irony rather than the metaphoric and metonymic tropes of romanticists and realists. Degrees of subtlety and complexity probably can be determined by intrinsic tests; what kinds of subtlety and complexity—and, whether subtlety and complexity of technique necessarily mark the "best" texts—cannot be determined intrinsically. That the triumph of modernist taste represents the triumph of a masculinist bias—once more, that these intrinsic criteria rest upon extrinsic grounds—has been noted by feminist critics; what has been noticed less is that this bias represents a privileging of writing over the voice, something readily apparent to students of orality and literacy, if not to deconstructionists.

Thus, current attempts to take women's diaries seriously as types of literature or as cultural documents of more than marginal significance, for example, need to be aware that the values of these texts are not only gender- but technologically-inflected. That is to say that their techniques are not only those assumed to be appropriate to women rather than men, but as well those deemed appropriate to cultures marked by oral rather than chirographic or print modes of information storage and retrieval. In oral cultures, to be sure, these techniques are used not only by women but (even predominantly) by men.

In England—I offer a very broad generalization here—Romantic and Victorian writers, male and female, were thoroughly committed to the values of presence, immediacy, and full communicative intentionality—parole pleine —associated with oral forms, as was Whitman in America (but certainly

not Thoreau, Hawthorne, and the other "major" eastern WASPS). The sexist part of the modernist shift has to do with the fact that in leaving these values to women, and to men who write like women (curiously, D. H. Lawrence, for one!), it left them stigmatized as outmoded and subordinate. Textuality was new, advanced, and male; orality was old, backward, and female. Not only female, of course, for the backwardness of the oral encompassed the "primitive" as well: the unlettered red savage or black slave. This aspect of modernism is carried forward quite unselfconsciously by Jacques Derrida and Paul de Man: Yale School is Male School, as Barbara Johnson put it. For, although Derrida has been a strong critic of phallogocentrism, he has not extended his critique to the subject of textual complexity. To the contrary, he has reveled in every sort of textual complication as if that were not itself implicated in the Western phallogos. And, to point up the obvious, although he has been tireless in deconstructing phallogocentric/metaphysical texts, he has been completely uninterested in seeking out texts that might be Different.

In America, the new standard of technical excellence, by subordinating not only writing by women but all writing deriving from types of oral presentation, also subordinated the west to the east; Mark Twain's reputation as a major American author, for example, was not securely established until the 1950s, when the first serious critical challenges to the New Critical/modernist program began to make their way. What all this means for Native American literature should be obvious. Its earliest proponents were in fact mostly women and westerners (e.g., Natalie Curtis Burlin, Mary Austin, and Yvor Winters) or those interested in spoken models for verse

(e.g., William Carlos Williams), and their strongest case for it often invoked its primitive charm. (Lawrence's sense of the powers of indigenousness is a particular, complex version of similar notions.) At a time when institutional canon-making power was solidly eastern and male and the standard for excellence was based on capabilities apparently possible only to certain kinds of writing, clearly claims for the importance of Native American literature had to go against the grain.

Without endorsing subtlety and complexity as inevitably the stuff of literary excellence so far as the techné of delight is concerned, I should still like to take a moment to point out what may not be so well known to many students of written American literature: oral literatures, while they do not privilege ironic tropes, structures, or attitudes, are capable of a wide range of complex and subtle effects in their own right. I will let two brief examples stand here for the very many that might be cited. Consider the much-quoted remark of Maria Chona, Papago woman and autobiographer, in commenting on the fact that Papago songs are very brief: "The song is very short because we understand so much" (23). Chona is usually taken to be referring to the extraordinary degree of condensation in Papago songs and pointing out the very complex interaction between a given singer's words, the culturally defined functional context for those words, and the audience's active awareness of all this at every moment. A different instance of extreme condensation is presented by Keith Basso in relation to the Apache practice of "speaking with names," a technique used in "socially taut" situations, and one in which the enunciation of toponyms invokes "ancient wisdom" as this is contained in stories commonly known to all. As Basso summarizes it,

the expressive force of an Apache utterance seems to be roughly proportionate to the number of separate but complementary functions it accomplishes simultaneously, or . . . to the number of distinguishable subject matters it successfully communicates "about." (1988, 121)

Condensation thus turns out—and this seems to be the case for some Apache storytelling or "literature," as well—to be hand in glove with overdetermination, all of it most subtle and complex indeed.^[13] We are very far, here, from the fantasies of primitive simplicity Americans have long attributed to Native American oral expressiveness. But let us go on.

As with the matter of formal criteria for excellence, in matters of content, decisions as to which books provide the instruction we approve can only be made in reference to a reader's prior values. For example, in The World, the Text, and the Critic, Edward Said writes that

criticism must think of itself as life-enhancing and constitutively opposed to every form of tyranny, domination, and abuse: its social goals are noncoercive knowledge produced in the interests of human freedom. (1983, 29)

To produce such criticism, Said may choose to study virtually any literary or nonliterary text; were he to propose a canon of the "best" literature,^[14] however, it would have to be in line

with his stated values. One further example. Terry Eagleton conceives the tasks of the revolutionary culture worker to be consistent with socialist aims. And, he writes, "Ultimately,"

the only reason for being a socialist is that one objects to the fact that the great majority of men and women have lived lives of suffering and degradation, and believes that this may conceivably be altered in the future. (1981, 112)

Eagleton has argued powerfully for no-such-thing-as-literature, and so it appears unlikely that he would ever propose a literary canon of "greats." Eagleton's position is that one's choice of texts to work with, whatever their discursive type, must and can only be determined by one's intention in working with them—culture criticism, then, in the interest of mitigating human suffering and degradation. What should be clear is that any claim to present a canon of literary excellence is inevitably a function of one's prior values and always implies a social vision; the canon proposed by William Bennett or suggested by Allan Bloom and Christopher Clausen is as fully determined by extrinsic factors as any canon Said or Eagleton or I might construct. The difference, let me put it baldly, is in the willingness of the former to produce knowledge coercively in the interest of perpetuating an order whose material effect—doubtless undesired, doubtless inevitable, and resolutely denied—is further suffering and degradation for the many. This is what is at stake in the argument over literature and the literary canon.

III.

As we move to proposing principles for a canon of American literature, the example of the twentieth-century's fore-

most canon-maker, T. S. Eliot, is again worth consideration. John Guillory has recently shown how Eliot came to realize what I have been arguing, that one's choice of a canon could not, finally, be defended on literary grounds alone (however much one may value the category of literature). Rather, Eliot saw that the choice had to be a reflection of "a more fundamental evaluative norm, extrinsic to literature (Guillory, 348). Thus, according to Guillory, "Eliot tells us in After Strange Gods [1933], that he is rewriting 'Tradition and the Individual Talent' [1919] by substituting [the term] 'orthodoxy' for 'tradition,' and this is unquestionably an ideological correction" (348). "Orthodoxy," "true opinion," and "right tradition" are made up of the beliefs inculcated by "a Christian education," one that would not so much attempt to produce believers as it would "primarily train people to be able to think in Christian categories." Those poets whose teaching was consistent with orthodoxy had claims to being the "best" poets, the best, at least in the sense that reading them might help promote the Christian Society that Eliot envisioned. Eliot did not say much about the kinds of delight consistent with such teaching, although delight along with teaching there would have to be for Eliot to attempt to overcome his—mistakenly theorized—notion of the "dissociation of sensibility" that presumably set in some time after John Donne.

In selecting his orthodox canon, Eliot had the problem all of us have when we acknowledge that the literature we most approve pedagogically is not always or all of it the literature we most admire esthetically, and vice versa. Eliot's own solution to this problem, as Guillory's discussion makes clear, was to make a distinction between major and minor poets and to suggest that the minor poets could turn out to be more useful—and thus in this regard actually "better"—than the

majors. But Eliot's pleasure in techné, in certain kinds of formal prowess, was such that he could not demote to the category of minor author any technical genius whose teaching alone he deemed counterinstructive, nor promote to majority the technically lesser author merely because he took the "right" line.

In time, then, Eliot gave up the claim for an essential, ahistorical (his phrase for this was "timeless"), and universal greatness as intrinsic to the works that make up the "monuments" of the tradition. Rather, it was the extrinsic principle of orthodoxy that came to authorize his choices of the canonical texts (the "best" if not always the "major" texts)—as it was that principle which animated Eliot's own religious criticism. The term Eliot offered for the principle opposed to orthodoxy was heresy; thus, to support the orthodox and oppose the heretical would be the aims of a religious criticism deployed in the interest of a religious canon authorizing a Christian Society. A secular criticism in the interest of a secular canon and a pluralistic democratic society need not, however, oppose orthodoxy in the name of heresy. To the contrary, it is necessary to reject heresy along with orthodoxy as equivalently the conceptualizations of religious thought. Rather than heresy as the principle (extrinsic, ideological, apart from esthetic valuation), informing my own selections for a canon of American literature, I would suggest what Guillory, in the final sentences of his article names heterodoxy, a term that, for him, is best suited to "consider what it means that 'difference' has become our central critical category . . . a teaching that will enact discursively the struggle of difference" (359). That is all Guillory says on the matter. I have already indicated why I think it is a mistake to construct and endorse any category that would reify an endless struggle of discursive difference. This is—to put the matter

another way—only to valorize a radically ironic perspective, a commitment to tension or difference as an end in itself (I shall have more to say about this temptation, a strong presence in American criticism since the modernist period).^[15] Politically, as I have heard Cornel West say, it is a commitment not to revolution, which envisions an end however long-postponed, but to resistance, which is a permanent condition of existence. Or, to take Allon White's formulation of this matter,

A politics of pure difference which refuses to theorise the unity-in-difference of humanity ends by replicating the individualism of the self-sufficient bourgeois ego—a dangerous fiction if ever there was one. (233)

I will thus appropriate the term heterodoxy and use it to name exactly such a principle as "unity-in-difference," as this principle may inform an American literary canon—a canon of national literature—and an American social order. I shall also try to extend it to the international order of literature and society in taking heterodoxy as informing that cosmopolitanism my own discussion will take as its ultimate horizon.

But just here, let me turn again to Edward Said, who has movingly argued against religious criticism and in favor of an explicitly secular criticism. Although Said himself, as I have already noted, is not a canon-maker, still, his theoretical and practical insistence on "how ostrichlike and retrograde assertions about Eurocentric humanities really are" (1983, 29), has strong implications for opening the canon to difference not as a "central critical category" but as the condition of any informed critical stance. Add to this Raymond Williams's comment

that however dominant a social system may be, the very meaning of its domination involves a limitation or selection of the activities it covers, so that by definition it cannot exhaust all social experience, which therefore always potentially contains space for alternative acts and alternative intentions which are not yet articulated as a social institution or even project. (in Said, 29)

Further adduce almost any typical remark of Bakhtin's on the dialogical nature of language, literature, and society (e.g., "There is neither a first nor a last word and there are no limits to the dialogic context" [quoted in A. White, 220]). Recall "without apology" the canon reformation of the 1960s, and there is in all of this more than enough warrant for a heterodox secular canon of literature, not only in opposition to a Christian society, or, indeed to any monologic orthodoxy, but as well to an insistence upon rather than a permission of difference.

For the canon of American literature, secular heterodoxy on an empirical level means something very specific: it means that any proposed canon of American literature that does not include more than merely occasional examples of the literatures produced by red and black people as well as white people—men and women, of indigenous and African, as well as European origins—is suspect on the very face of it. The history of that national formation called the United States of America is such as to insist upon the primacy of Euramerican, Native American, and Afro-American literary expression in any attempt to define an American literature. In saying this, I hope it is clear, to repeat, that I am not calling for some kind of proportional representation for these groups, nor restricting the canon to texts associated with these groups, aprioristically denying that canonical American books might well be pro-

duced by people of Asiatic, or mixed origins, or of any background whatever. In these regards, it is worth noting that Spanish is now the second language of the United States (as it is the second most widely spoken language in the world), a fact sufficiently important to have provoked intense efforts on the part of S. I. Hayakawa and others to pass a regressive and repressive constitutional amendment making English the official language of the United States. As I write, on the eve of Election Day, 1988, proposals are on the ballot in Arizona, Colorado, and Florida to this effect. Given the increase in American Spanish speakers, there is no doubt in my mind that Latino literature will soon exert major pressure on the canon, a development I look forward to with enthusiasm. Nonetheless, to the present, the cultural expression of red, white, and black people seems to me to have a historically urgent claim to primary attention.

I mean, here, to assert that Afro-American and Native American literary production, when we pay attention to it, offers texts equivalently excellent to the traditional Euramerican great books. It is not only that these texts should be read in the interest of fairness or simply because they are available; nor is it because they provide charming examples of "primitive" survivals: they should be read because of their abundant capacity to teach and delight. But for that capacity to be experienced and thus for the excellence of these texts to be acknowledged, it will be necessary, as I have suggested above, to recognize that what they teach frequently runs counter to the teaching of the Western tradition, and that the ways in which they delight is different from the ways in which the Western tradition has given pleasure.

Still, it may be that we have entered into a period in which the prospects for Native American phrónesis^[*] and techné are

rather promising. In terms of its teaching, let me note only the fact that traditional and contemporary Native literatures tend pretty much without exception to derive from an ecosystemic, nonanthropocentric perspective on the world that we may at last be coming to see—as the ozone layer thins, as the polar ice melts, as the nonbiodegradable garbage mounts to the skies—as being centrally rather than marginally important to human survival. This is not to say that Indian literatures are explicitly "about" a particular view of "Nature," far from it; yet this is indeed the perspective with which they all, in my experience, are consistent.

In terms of technique, even the most recent and most complexly composed Native American works are still likely to have roots in or relations to oral traditions that differ considerably in their procedures from those of the dominant, text-based culture: if these works are indeed equivalently excellent, still it must be recognized that they are differently excellent. To the extent that we are perhaps already in what Father Walter Ong calls the "secondary orality of our electronic age" (305) to the extent that print culture is already receding from the importance it had for a full five hundred years, we may currently be producing just the conditions of possibility for such a recognition. That postmodernist fiction, poetry, and painting have found a substantial audience; that the disjointed, even spasmodic styles of "Miami Vice," "Crime Story," and MTV music video have proved popular, indicates that a wide public has lost interest in attempts to represent the world realistically in causally connected, continuous linear narrative. Ronald Reagan's popularity, matched with the unpopularity of his actual political positions, is only further evidence of a paradigm shift whose description is already possible to produce, although its evaluation remains somewhat more difficult. In any event, the

material situation as I can understand it, for all that I am wary of it, nonetheless seems to me encouraging for the appreciation of Native American literature.

In the chapters that follow, I first explore the relations between critical perspectives and the canon, most particularly as these relate to the inclusion or exclusion—indeed, the very visibility—of Native American work. This exploration is followed by an essentially historical study that traces attempts to place Indian literatures within the canon of American literature. I embark next upon a lengthy return to the genre of Native American literature with which I have been most concerned, Native American autobiography. I try to do some careful reading here, practical criticism, as it used proudly to be called, of texts that are only marginally literary. My aims are multiple: to call these texts to the attention of readers and make them interesting and illustrative in a variety of ways; to deduce from them some theoretical principles that may, to this point, have occupied a certain" space" but have not been even modestly "grounded." My final chapter attempts to consider the place of Native American literature as part of a heterodox American national, international, and ultimately cosmopolitan world canon of literature. I hope here, as throughout, to provide some theoretical basis for the enormous amount of practical work that needs to be done. A brief conclusion rounds out this book.