The problem of Cervantes in Bakhtin's poetics is initially a simple one: his inconspicuousness. Neither Cervantes nor his novel Don Quixote figure at all prominently in Bakhtin's voluminous theorizing about the novel as a literary form. Bakhtin wrote (and later revised) a whole book on Dostoevsky. He devoted a book and a large section of an important essay to Rabelais. And he wrote a third book, on the novel of education and the history of realism, that paid considerable attention to Goethe, although only part of this has survived.48 But about Don Quixote, a novel for which Dostoevsky claimed literally apocalyptic significance, Bakhtin has curiously little to say. In the roughly 1200 pages of his major writings on the novel currently available in English, there are some twelve pages worth of discussion of this «grandest and saddest book conceived by the genius of man», in Dostoevsky's assessment of Cervantes' masterpiece.49 Given this one per-cent —30→ solution, it would be easy to conclude that contemporary Cervantes scholars concerned with modern literary theory should ignore Bakhtin and concentrate on other more Cervantes-centered theorists -Ortega y Gasset, for example, or René Girard, or even Viktor Shklovsky, who devotes considerable space to Don Quixote in his Theory of Prose.50
Although it would be easy to come to this conclusion, it would also be a mistake. Even in the brief glimpses we get of Cervantes in Bakhtin's writings on the novel, there are significant insights into Don Quixote. And in the relative absence of Cervantes from Bakhtin's novelistic pantheon, we learn important things about the contours and limits of this contribution to modern literary theory that has aroused such interest in the West in recent years. Most of all, the initial noncongruence of beholder and beheld allows us to reflect on the relationship between the theory of the novel and the history of its practice and to question whether it is theory or practice that should have the final word.
In Problems of Dostoevsky's Poetics, a book which Bakhtin seems to have completed a version of as early as 1922 but which he revised for republication in 1963, Cervantes and Don Quixote appear mainly as prototypes for the distinctive novelistic formation of «polyphony» which Dostoevsky is credited with originating. «A plurality of independent and unmerged voices and consciousnesses, a genuine polyphony of fully valid voices is in fact the chief characteristic of Dostoevsky's novels», Bakhtin writes. «What unfolds in his works is not a multitude of characters and fates in a single objective world, illuminated by a single authorial consciousness; rather a plurality of consciousnesses, with equal rights and each with its own world, combine but are not merged in the unity of the event» (p. 6). Initially, Bakhtin is unwilling to attribute any of this energetic alterity and independence of character to earlier novelists. Dostoevsky «created a fundamentally new novelistic genre [and] his work does not fit any of the preconceived frameworks or historico-literary schemes that we usually apply to various species of the European novel» (p. 7). Later on, however, Bakhtin allows that «embryonic rudiments» or «early buddings of polyphony» can be detected in Cervantes, as well as in Shakespeare, Rabelais and Grimmelshausen (p. 33). And in material added to the 1963 edition, Cervantes achieves a somewhat fuller measure of this polyphonic potential under the rubric of «carnival». This affords a significant insight into the peculiar inwardness of Don —31→ Quixote when Bakhtin compares the tonality of laughter in Cervantes and Rabelais. «In Cervantes there is no longer that public-square intensity of sound, although in the first book of Don Quixote laughter is still quite loud, and in the second it is significantly (when compared with the first) reduced. This reduction is also linked with certain changes in the structure of the major hero's image, and with changes in the plot» (p. 165).
This distinction between the two parts of the Quixote is something that Auerbach, for all his close reading of the novel in Mimesis, overlooks. Nevertheless, the momentary eminence Cervantes achieves here is flattened out in the sweeping historical vista that concludes this discussion of Dostoevsky's precursors. «Dostoevsky is the creator of authentic polyphony, which, of course, did not and could not have existed in the Socratic dialogue, the ancient Menippean satire, the medieval mystery play, in Shakespeare and Cervantes, Voltaire and Diderot, Balzac and Hugo. But polyphony was prepared for in a fundamental way by this line of development in European literature» (p. 178).
The phenomenology of persons, of author and characters, in this first phase of Bakhtin's theory of the novel is succeeded by a sociolinguistics of voices in his second major contribution to the subject. In «Discourse in the Novel», a long essay written in 1934-35, Cervantes achieves a new prominence. In the 160 pages of this essay in The Dialogic Imagination, Dostoevsky is only mentioned in passing while there are some six pages worth of discussion of Don Quixote. Much of this discussion is of real significance, both for understanding Bakhtin's protean concepts of language and literature and for understanding the place of Cervantes in a historical poetics of the novel. The main concept in this essay, one which Don Quixote exemplifies particularly well, is «heteroglossia», the condition of many different discourses struggling to be heard in any concrete utterance. On one level, heteroglossia is the general condition of all communication in language for Bakhtin, but it is something that is realized on a higher level most fully and self-consciously in the literary form of the novel. After considering a number of different strategies for «incorporating and organizing heteroglossia in the novel», Bakhtin notes the possibility of a still higher-order combination of combinations. «Of such a sort is the classic and purest model of the novel as a genre -Cervantes' Don Quixote-», Bakhtin observes, «which realizes in itself, in extraordinary depth and breadth, all the possibilities of heteroglot and internally dialogized novelistic discourse».51 Many critics, of course, have —32→ commented on this multi-generic, encyclopedic quality of Don Quixote, but Bakhtin focuses on the novelistic typicality of this form in a way that has yet to be fully appreciated. He also expands his sense of the heteroglossia of the Quixote beyond the merely literary. Beginning with Don Quixote, he argues, «the novel must represent all the social and ideological voices of its era, that is, all the era's languages that have any claim to being significant» (p. 411). This is a challenge to the literary formalism of much Cervantes criticism.
Don Quixote figures significantly as well in this essay in Bakhtin's attempt to develop a more intrinsic poetics of the novel. Cervantes' text becomes the epitome of the «Second Stylistic Line» of the novel's development, a line that is more radically dialogic or heteroglossial than the First Stylistic Line, which opens dialogic possibilities only to foreclose them. The contrast between these two stylistic lines is a more sophisticated version of the traditional distinction between novel and romance. Furthermore, within the Second Stylistic Line, Don Quixote turns out to embody both of the two basic types of testing that purely literary discourse is subjected to: the testing that centers on a hero trying to live according to the books he has read and the testing that centers on an author trying to live by writing a book of his own. «Both these types of testing literary discourse [are] blended into one... as early as Don Quixote», Bakhtin says, noting the importance of Cide Hamete as well as of Quixote himself (p. 413).
Nevertheless, after «Discourse in the Novel», Bakhtin's attention to Cervantes as exemplary novelist begins to decline. In «Forms of Time and of the Chronotope in the Novel», the other major essay on the novel from the 1930s, Rabelais begins to overshadow Cervantes as the major embodiment of novelistic energy. Some fifty pages of this 174-page essay are devoted to «The Rabelaisian Chronotope» and «The Folkloric Bases of the Rabelaisian Chronotope», while Cervantes is only mentioned three times. Bakhtin does claim «enormous significance» for Don Quixote in «the long history of literature's assimilation of historical time», but apologizes that «in this essay... we cannot undertake an analysis of Cervantes' novel».52 This eclipse of Cervantes in Bakhtin's poetics continues in Bakhtin's separate book on Rabelais, completed as a dissertation in the 1940s but only published in 1965. Here the sociolinguistics of voices gives way to —33→ a folklorics of festivity, and while Cervantes is mentioned often enough in the company of Shakespeare, Boccaccio, and others, the only concrete discussion of the Quixote centers on the «narrowing down» of the exuberant energies and appetites to which the carnivalization of Gargantua and Pantagruel had given expression. Although Bakhtin admits that this process of decline is only in its «initial stage' in Don Quixote, he argues for an increasing privateness and morbidity of physical experience in Cervantes' novel. «Bodies and objects begin to acquire a private, individual nature; they are rendered petty and homely and become immoveable parts of private life, the goal of egoistic lust and possession. This is no longer the positive, regenerating and renewing lower stratum [of Rabelais], but a blunt and deathly obstacle to ideal aspirations».53
Thus the concrete, explicit view of Cervantes and Don Quixote afforded by Bakhtin's literary theory is a tantalizing but a fleeting one. It is of course quite possible to ignore the limits of the actual and to extrapolate from one or another of Bakhtin's generative concepts of literature and language and discuss what a full developed Bakhtinian view of Don Quixote might be. I am aware of essays by Manuel Durán and Donald Fanger which offer preliminary forms of such an extrapolation,54 and unless a manuscript like Nabokov's Harvard lectures on Don Quixote turns up (it is known that Bakhtin did lecture on Cervantes at the Pedagogical Institute of Saransk), we will have to resort to the borrowed illumination of «dialogism» or «the chronotope» or the «extralocality» of the author to the hero to see Cervantes and see him whole from Bakhtin's point of view. There is also the possibility of borrowing from Bakhtin at strategic points in one's own theoretical analysis, an approach that I take in An Exemplary History of the Novel and that Alban Forcione takes in Cervantes and the Mystery of Lawlessness.55 It would certainly be perverse to hide the light of this brilliant theorist under the bushel of his own preferred examples.—34→
It is also possible to inquire whether there is some reason or rationale for Bakhtin's relative silence on Cervantes. We know that Don Quixote was a favorite text of Shklovsky and the Russian Formalists and that Bakhtin, publishing under the name of his friend Medvedev, was critical both of Formalist literary theory in general and of Shklovsky's interpretation of Don Quixote in particular.56 Michael Holquist and Katerina Clark in their indispensable biography of Bakhtin remark how certain literary topics like «the epic» functioned in part as veiled references to issues in Bakhtin's own ideological environment like «socialist realism».57 It is arguable that in the politics of literary discourse in the increasingly oppressive Stalinist era, Don Quixote was neither sufficiently acceptable on an official level nor sufficiently subversive on an unofficial level to engage Bakhtin's full critical attention. It is also arguable that the peculiar ironies concerning the status of the reader in Cervantes' fiction partially eluded Bakhtin's critical ear, so attuned to the more powerful resonances of the «idea» in Dostoevsky, the «body» in Rabelais, or «emergence» in Goethe. As Caryl Emerson has shown in her analysis of Bakhtin's view of Tolstoy, there are important novelists whom Bakhtin's theory of the novel (and philosophy of existence) are simply incapable of hearing clearly and seeing sympathetically.58 Nevertheless, it is certainly less a case of Bakhtin polemically «closing down» Cervantes -as Emerson argues he does with Tolstoy- than with his seeing over him, looking with a farsighted astigmatism beyond the «long, thin graphism» of Don Quixote, to use Foucault's arresting phrase,59 into the farther reaches of the novelistic terrain.
While these are all promising avenues of reflection on the problem of Cervantes in Bakhtin's poetics, I would like to open a rather different perspective at the close of this brief discussion, passing through the looking glass of Bakhtin's image of Cervantes, as it were, and considering the image of Bakhtin in Cervantes' poetics. Ortega y Gasset claims in Meditations on Quixote that we need a book «showing in detail that every novel bears Quixote within it like an inner filigree, in the same way as every epic poem contains the Iliad within it like the —35→ fruit its core».60 What Bakhtin's concept of dialogism suggests is that we also need a book showing that every theory of the novel contains the Quixote within it, like an organism its chromosomes. The book that is a prototype for so much of the subsequent practice of fiction is also a prototype for much that passes for the theory of the novel, particularly in our century. This is surely a plausible perspective with a text that so confounds one's easy assumptions about the difference between literary creation and literary criticism.
Within the purview of Cervantes' poetics, then, however implicit or «unwritten» this system of literary norms and types may be, we can consider the place of Bakhtin in relation to the positions of other twentieth-century theorists. What is interesting is that most modern theories of the novel seem to inherit their characteristics from the opponents and understudies that Don Quixote encounters in the course of his adventures. They are theories inimical to Quixote himself. Ian Watt's theory of the novel as «formal realism» for example, is a lineal descendent of Sancho Panza's attempt to counter-balance Quixote's romancing idealism. René Girard's theory of the novel as «triangular desire» derives from the antagonistic mimesis of the Bachelor Sampson Carrasco. Wayne Booth's prescriptive «rhetoric of fiction» can be traced back to the inquisitorial proceedings of the priest and the barber at the beginning of the First Part of Don Quixote, while Lukacs' more «historico-philosophical» analysis finds its prototype in the pronouncements on fiction by the Canon of Toledo at the end of the First Part.61 In the case of Bakhtin's theory of the novel, however, we have an understanding of the art of fiction that can only be traced back to Don Quixote himself. There is an uncanny family resemblance between Quixote's theory of his practice and Bakhtin's practice of his theory, a resemblance that suggests another explanation of why the novel Don Quixote is so little visible in Bakhtin's writings.—36→
Consider the following similarities. Both Don Quixote and Mikhail Bakhtin imagine life and literature as a matter of combat. Emerson notes in her preface to Problems of Dostoevsky's Poetics Bakhtin's «fondness for military metaphors. In his texts words are always competing, doing battle, winning and losing territory» (p. xxxvii). Like Quixote, Bakhtin envisions «letters» in terms of «arms». There is then Bakhtin's tendency to treat the novel as the noblest of literary forms, the «leading hero in the drama of literary development», in Bakhtin's phrase or the «Robin Hood of texts», as Holquist puts it.62 The novel in Bakhtin's imagination not only exercises its own freedom; it also liberates other genres from the fetters of official culture, like Don Quixote the galley slaves or any of the other «captives» he encounters. Furthermore, the novel as Bakhtin presents it expands as alarmingly as the chivalric romance in Quixote's conception of the genre. For Bakhtin «the novel» ends up including any number of literary manifestations that a more empirically minded critic would see as separate and distinct: Greek romance, Socratic dialogue, classical biography, Menippean satire, the mystery play, and innumerable forms of folkloric discourse and performance. Like the ideal of chivalric romance in Quixote's imagination, the idea of the novel in Bakhtin assimilates other genres to itself as its proper nourishment. Finally, and most importantly, Bakhtin reveals himself as a theorist of Don Quixote's party in the way he recasts issues of representation and truth as issues of aesthetic wholeness and ethical responsibility. Paul de Man's criticism of Bakhtin for ignoring the «fundamental question of the compatibility between the descriptive discourse of poetics and the normative discourse of hermeneutics»63 is a twentieth-century echo of the Canon of Toledo's amazement at Quixote's «so mingling truth and fiction» in spite of all his attempts to make Quixote separate these realms.
In his important and as yet officially untranslated essay entitled «The Author and Hero in Aesthetic Activity», Bakhtin describes the way in which an author «loses his evaluating point of extralocality to the hero» -that is, the way an author may fail to establish his imaginative authority over his protagonist. In one version of such a loss of control, Bakhtin says, «the hero takes possession of the author. The hero's emotional and volitional situation among other subjects, his —37→ cognitive and ethical position in the world, is so authoritative for the author that he can only see the world of subjects with the eyes of the hero and can only experience it within the events of the hero's life. The author cannot find a convincing and persistent evaluative support point outside the hero».64 This analysis seems appropriate, mutatis mutandis, to Bakhtin's position as critical author vis-a-vis Cervantes' hero Don Quixote. Bakhtin's theory of the novel cannot hold the novel Don Quixote clearly in its gaze because it contains a project analogous to the character Don Quixote's own so deeply within itself. Like that other eccentric cervantista of the twentieth century, Borges' Pierre Menard, Bakhtin demonstrates the centrality of Cervantes' great fiction more by existential homage than by essential reflection.
I acknowledge the Quixotic eccentricity of this final assessment of the problem of Cervantes in Bakhtin's poetics. But it seems to me that Cervantes himself mounts powerful arguments against our tendency to assume that, in the dialogue of theory and practice, literary theory can and must have the last word. Both Cervantes and Bakhtin demonstrate, from opposite sides of the critical fence, that the privilege and authority of all theory are on loan from the creative imagination but also that the escapism and delight of the literary imagination are vehicles for serious philosophical investigation.