From: Cervantes: Bulletin of the Cervantes Society of America 10.1 (1990): 35-50.
Copyright © 1990, The Cervantes Society of America

raditional criticism has repeatedly described Persiles as a “split” text, with Part 1 set in the Arctic zones, and Part 2 along the Mediterranean. This schizoid view of Cervantes's last romance survives among its most recent critics, as Alban Forcione has rightly observed, “even in the one study [Casalduero's] which argues for its thematic unity.”¹ Indeed, excepting Forcione himself, the majority of serious commentators since Friedrich Bouterwek (1804) have shared a view of Persiles as a work split into two halves. Casalduero shared that view, having likened the two halves of Cervantes's book to the “dos mundos que encontramos constantemente en el Barroco, de cuya tensa oposición surge la

     ⁺ This is a paper from a symposium on Los Trabajos de Persiles y Sigismunda, as is explained in the Foreward of this issue of the journal.
     ^*This essay is a revised version of portions of Wilson's forthcoming book, Allegories of Love: Cervantes's “Persiles and Sigismunda” (Princeton University Press, 1990).
     ¹ Alban K. Forcione, Cervantes' Christian Romance: A Study of “Persiles and Sigismunda” (Princeton: Princeton UP, 1972), 11. Later in the text Forcione argues that “far too much emphasis” has been placed on the differences between the two halves of the novel: “the traditional distinction between a symbolical and a realistic half of the Persiles . . . ignores the fact that the second half of the work continues to reactualize the Christian mythology and employs the symbolic methods of the first half to do so” (46-47n).

unidad.”² Avalle-Arce shares that view, or at least he has written of the “diferencia tan tajante entre las dos mitades del Persiles.”³ I share that view and, in this essay, I hope even to problematize it.
     In a study entitled “The Enigma of the Persiles,” William Atkinson, who stresses the “obvious concern with symmetry in the architecture of the novel,” sees Cervantes as not only balancing the action “between North and South, unknown and known, symbolical and material,” but also introducing in the first half of the book one significant national representative of each of the four countries that would be traversed in the second half.⁴ A less theorized but far more schematic reading by Alberto Navarro González distinguishes between the “idealistic” and “realistic” halves of Persiles.⁵ To wind up this little catalogue of critics on the “split,” Rafael Osuna not only regards Cervantes's last obra as “escendida” (cut, divided, separated), but he also declares this “escisión” as “sin duda de lo más misterioso del Persiles.”⁶
     A text divided against itself cannot stand consensus. The enigmatic notion of Persiles's two textual halves has led certain critics to promote one half of the text over the other, not without an interesting clash of epistemologies. Atkinson, for instance, sees only the first half of Persiles as the “vital” half, where “everything is relevant.”⁷ This contests Menéndez y Pelayo's earlier claim, at the other pole, that the Persiles “contiene en su segunda mitad algunas de las mejores páginas que escribió su autor.”⁸ Noting that the lack of unity in the novel is a problem yet to be studied, Osuna concludes his useful agenda for Persiles scholarship

     ² Joaquin Casalduero, Sentido y forma de “Los trabajos de Persiles y Sigismunda” (Madrid: Editorial Gredos, 1975), 14.
     ³ See the “Introducción” to his edition of Los trabajos de Persiles y Sigismunda (Madrid: Clásicos Castalia, 1969), 20. Avalle-Arce dates Part 1 of the Persiles between 1599 and 1605, and Part 2 between 1612 and 1616, with Don Quixote interpolated between the two parts (18-20).
     ⁴ William C. Atkinson, “The Enigma of the Persiles,” Bulletin of Spanish Studies 24 (1947), 248.
     ⁵ Alberto Navarro González, Cervantes entre el “Persiles” y el “Quijote” (Salamanca: Universidad de Salamanca, 1981), 56.
     ⁶ Rafael Osuna, “El olvido del Persiles,” Boletín de la Real Academia Española 48 (1968), 61.
     ⁷ Atkinson, “The Enigma of the Persiles,” 248.
     ⁸ Marcelino Menéndez y Pelayo, “Cultura literaria de Miguel de Cervantes y elaboración del ‘Quijote,’” Discursos, ed. J. M. Cossío (Madrid, 1956); cited by Emilio Carilla in the introduction to his edition of the Persiles (Madrid: Biblioteca Anaya, 1971), 14.

with the crucial critical question: “si las dos partes son tan diferentes, ¿qué es lo que las agrupa en un todo?”⁹ What is it that joins two different halves together? If the “enigma” of Persiles is to be articulated on the level of theory, such a theory might begin with the signifying potential of that mysterious “escisión.” How does scission or division work as a force of signification? A split need not be a flaw if it can be shown to be a strategy. As I read it, Persiles is a strategically split discourse, a cunning analogue of the ancient split between the sexes it aims to explore. Escisión, as it turns out, may be a natural fault.
     Centuries before Cervantes, the author of a very different kind of romance, the Romance of the Rose, tried to describe the production of meaning through the scission or division of contraries:

Thus things go by contraries; one is the gloss of the other. If one wants to define one of the pair, he must remember the other, or he will never, by any intention, assign a definition to it; he who has no understanding of the two will never understand the difference between them, and without this difference no definition that one may make can come to anything (vv. 21573-82).¹⁰

This vivacious little sermon on difference as the prelude to definition is fleshed out for the Renaissance in Il cortegiano, where Castiglione himself —speaking of good and evil as contraries— observes that “the one must always sustain and reinforce the other, and if the one diminishes or increases, the other as its necessary counter-force, must do the same.” Later in the text, the same argument is gendered:

Thus male and female always go naturally together, and one cannot exist without the other. So by very definition we cannot call anything male unless it has its female counterpart, or anything female if it has no male counterpart.

These observations by Giuliano de' Medici are cut off by Castiglione's stage misogynist, signor Gaspare: “I do not wish us to go into such subtleties because [the] ladies would not understand

     ⁹ Osuna, “El olvido del Persiles,” 67.
     ¹⁰ This passage is cited and translated by Kevin Brownlee in his fine study of “Jean de Meun and the Limits of Romance: Genius as Rewriter of Guillaume de Lorris,” in Romance: Generic Transformations from Chrétien de Troyes to Cervantes, eds. Kevin and Marina Brownlee (Hanover, N.H.: UP of New England, 1985), 129.

them.”¹¹ This remark bears out Ian Maclean's recent claim that, for the Renaissance mind, “concepts of difference, division, definition and opposition” remained problematic —and not only for the ladies.¹²
     Renaissance scholars were familiar with the two-term system for organizing discourse found in the Metaphysics (1.5.6.986a), where Aristotle attributes his ten pairs of contraries to the Pythagoreans. A more sophisticated presentation of contraries was available to the age in Aristotle's Categories (10.llb 15ff), where oppositions were handily divided into correlatives, contraries, privatives to positives, and affirmatives to negatives. But Renaissance humanists perceived these abstract categories and speculative subtleties as mechanically arid: they were the intellectual apparatus, indeed, the detested baggage, of a scholasticism that Luis Vives, for one, denounced as a “pestilencia” that had infested the minds of men for over five-hundred years.¹³ Intimately involved with this Aristotelian area of speculation deplored by the humanists, however, was the concept of female difference —a notion that nicely reflects “the hesitancies and incoherences inherent in Renaissance modes of thought.”¹⁴
     The concept of difference repressed by the humanists, however, would return again, ironically, in that genre “de quien nunca se acordó Aristóteles” (DQ I, Prologue) —Renaissance romance. When it returned, like Ovid, it was moralized. And it has remained that way. Even modern critics who have focused on Otherness as an archaic character of romance tend to privilege moral oppositions as the genre's organizational categories. Thus Northrop Frye writes that “every typical character in romance tends to have his moral opposite confronting him, like black and white pieces in a chess game.” Elsewhere Frye speaks

     ¹¹ Baldesar Castiglione, The Book of the Courtier, trans. George Bull (New York: Penguin Books, 1967), 109 and 220.
     ¹² Ian Maclean, The Renaissance Notion of Woman: A Study in the Fortunes of Scholasticism and Medical Science in European Intellectual Life (Cambridge: Cambridge University Press, 1980), 2-4. As the source of his own speculations on polarities, Maclean cites G. E. R. Lloyd's Polarity and analogy: two types of argumentation in early Greek thought (Cambridge, 1971).
     ¹³ Juan Luis Vives, Adversus Pseudodialecticos in Juan Luis Vives: Obras completas, trans. and ed. Lorenzo Riber (Madrid, 1948), 2:310. Binaries were to receive devious illustration in the work of those Renaissance humanists who detested Aristotle but valued Lucian's irony, e.g., Erasmus and More.
     ¹⁴ Maclean, The Renaissance Notion of Woman, 4.

of “the curious polarized characterization of romance, its tendency to split into heroes and villains, [into] . . . good and bad . . . virtue or vice.”¹⁵ Fredric Jameson, although bent on modifying Frye's transhistorical notions of romance, similarly stresses the semantic code of moral opposites as primary to romance. Indeed, Jameson would suggest “that the most important of those organizational categories is the conceptual opposition between good and evil, under which all the other types of attributes and images (light and darkness, high and low, etc.) are clearly subsumed.” Stressing that belief in good and evil “is a very old form of thought,” Jameson then goes on to conflate evil with Otherness: “In the shrinking world of today,” he laments, “with its gradual leveling of class and national and racial differences, it is becoming increasingly clear that the concept of evil is at one with the category of Otherness itself: evil characterizes whatever is radically different from me.”¹⁶
     But James never mentions, in his study of romance, the one difference that has not been leveled, and that can never be leveled no matter how much our world shrinks: sexual difference. Aristotle's remark in the Categories —that “whatsoever is better, more honourable, is said to be naturally prior” (12.14b)— colors his own prioritizing of the male/female over the good/evil opposition in his Metaphysics (1.5.6.986a). Although I would scarcely wish to rest my case on Aristotle (whose contribution to gender studies is the view of an entire sex as morally undeveloped)¹⁷, I question Jameson's need to enshrine the opposition of good and evil as the “most important” organizational category in romance. It seems hard to think of good or evil behavior without first thinking of the agents of such behavior, and agency, as we are beginning to discover, comes in two genders. “When you meet a human being,” Freud writes, “the first distinction you

     ¹⁵ Northrop Frye, Anatomy of Criticism (Princeton: Princeton University Press, 1957), 195, and The Secular Scripture: A Study of the Structure of Romance (Cambridge: Harvard UP, 1976), 50.
     ¹⁶ Fredric Jameson, “Magical Narratives: Romance as a Genre,” New Literary History 7 (1975), 140.
     ¹⁷ Barbara Johnson dryly calls Aristotle “the founder of the law of gender as well as of the law of genre.” See A World of Difference (Baltimore: Johns Hopkins University Press, 1987), 33. One code of that Aristotelian “law of gender” represents the deliberative capacity of women as akuron, that is, as “lacking in authority” or as “easily overruled” (The Politics, 1260a13).

make is ‘male or female?’”¹⁸ Hélène Cixous speculates, further, that all hierarchized oppositions ultimately derive from gender difference, that most ancient and fundamental of distinctions. For Cixous, the Man/Woman ratio is the basis for many other oppositions, such as “Sun/Moon, Culture/Nature, Day/ Night . . . ,” a thread of invidious binarism that she sees as running down through centuries of representation, through literature and criticism, philosophy and reflection.¹⁹ Jameson's privileged opposition of good and evil, in any case, would seem to be untenable in the Persiles, where it is pointedly destabilized by a narrator loudly skeptical about overly schematic oppositions: “Parece que el bien y el mal distan tan poco el uno del otro, que son como dos líneas concurrentes, que aunque parten de apartados y diferentes principios, acaban en un punto” (4.12).
     This convergence of “el bien y el mal” —the “punto” where Cervantes' opposites meet— would appear to be missing in the romances of chivalry, where, according to Dan Eisenberg, “Black is black and white is white,” and where “heroes and villains are clearly distinguished.”²⁰ In Persiles, however, where black is sometimes white —or where evil is given a dangerous and seditious clarity (as, e.g., in the Barbaric Isle narrative)— it seems unproductive to come to terms, as Jameson urges, with the relationship between romance as a mode and the “deep-rooted ideology of good and evil.”²¹ It is the ideology of sexual difference —whose roots are as deep and as ancient— that tends to generate issues of good and evil in Cervantes's last romance (as it does in most of Shakespeare's romances).
     Coming to terms with this ideology sheds a useful light on Cervantes's remotivation of Greek romance —on his avowed decision “to compete with Heliodorus.” That Cervantes was not attracted to Heliodorus for his epic perfections —for his admired unity of plot or for his exemplary use of verisimilitude— Persiles as an agonistic imitation makes clear: suspense and plausibility,

     ¹⁸ Sigmund Freud, The Standard Edition of the Complete Psychological Works of Sigmund Freud, 24 vols., trans. James Strachey (London: Hogarth, 195-374), 22:113; emphasis added.
     ¹⁹ Hélène Cixous' extract reprinted in New French Feminisms: An Anthology, eds. Elaine Marks and Isabelle de Courtivron (Amherst: University of Massachusetts Press, 1980), 90-91.
     ²⁰ Daniel Eisenberg, Romances of Chivalry in the Spanish Golden Age (Newark, Delaware: Juan de la Cuesta Press, 1982), 74.
     ²¹ Jameson, “Magical Narratives,” 140.

for Cervantes, were hardly the call to emulation. Other motives, however, all clustered around gender as a category, seem more likely candidates for his choice of Greek romance, a subgenre that appears to embrace what chivalric romance, los libros de caballerías, seemed so stubbornly to resist: the notion of sexual Otherness as both different and equal, the notion of woman as not fundamentally inferior, not divinely superior, but simply different.
     It is a prosaic fact, but one that greatly enhances our understanding of Cervantes's imitation, that he kidnapped from Heliodorus a Greek romance —what Hellenist Thomas Hägg characterizes as the “first great literary form to have had its main support among women.” Not only did the Greek romances of the Second Sophistic appeal “especially to women,” but Hägg further speculates on whether some of their authors may have been women.²² This last speculation remains an unproven hypothesis, but the gynocentricity of the Greek romances is in any event not our concern here. What does interest us, rather, is the projection of sexual difference, of gendered alterity, in some forms of romance, whatever the sex of the projector. The development of a literary heroine as a resourceful, albeit suffering, woman may be the most significant legacy of Greek romance to later fiction —and not only to Cervantes.
     For openers, woman is given positional significance by being ritually advertised (along with her man) on the title page of Greek romances. Her “entitlement” may be found in virtually all of the surviving romances and, again, in the twelfth-century Greek romances of the Byzantine revival (novelas bizantinas such as Hysmine and Hysminias) with their obsessively heterosexual titles. Even in the fourteenth century, during the interplay of Byzantine and Western institutions, we meet a postscript of Greek romances whose titles are similarly dual-gendered: Callimachus and Chrysorrhoe, Libystrus and Rhodamne, Phlorius and Platzia Phlore, and so forth.²³ In imitation of Heliodorus' Aethiopica,

     ²² Hägg ultimately resists his own speculation on the grounds that the image of Woman idealized in these romances may have been “a typically male product.” See The Novel in Antiquity (Berkeley: U of CA P, 1983), 96.
     ²³ Three of the four twelfth-century novels, all written in learned literary Greek, were verse imitations of Heliodorus: Hysmine and Hysminias (by Eustathius Macrembolitis); Rhodanthe and Dosicles (by Theodore Prodomus); Aristandros and Callithea (by Constantine Manasses); and Drosilla and Charicles (by Nicetas Eugenianus). On the fourteenth-century Byzantine [p. 42] romances, with their newly westernized folk-tale motifs, see Hägg, The Novel in Antiquity, 80.

whose official title is The Loves of Theagenes and Chariclea, the copulative conjunction of Persiles and Sigismunda on Cervantes's title page links together a man and a woman as joint owners of an unspecified sum of trabajos.
     The significance of this sex-coded detail is made evident only by the clamour of its absence in the Spanish chivalric romances —generally named after heroic males such as Amadís, Belianís, Clarián, Lisuarte, Primaleón, Esplandián, Olivante, Cristalián, Felixmarte, Florisel, Florismarte, Palmerín, Tirante, the Caballero del Febo, et al.—and even in the Italian romance epics, where only the male hero is entitled to be maggiore or innamorato or furioso. Indeed, of the many romances scrutinized in Don Quixote's library (the reading that “dried up” his brain), only one little-known work is heterosexual in its conjunction: the pastoral Nimphas y pastores de Henares by Bernardo Gonzáles de Bovadilla (1587).²⁴ Apart from Paris e Viana (c. 1494) included by Henry Thomas in his antecedents (pre-1500) to the Spanish romances of chivalry, and Curial y Guelpha, part of Catalonian chivalric matter, no libros de caballerías entitled after two lovers seem to exist. Cervantes swerves from the chivalric —even from those “good” books of chivalry saved from the inquisitorial bonfire of Don Quixote's library— to choose a model with a more egalitarian title: Heliodorus' Historia etiópica de los amores de Teágenes y Cariclea.
     But titles are only the most superficial symptom of what Greek romance offered Cervantes toward a poetics of sexual difference. An analysis of sex roles in Greek romance will show these narratives to be remarkable in eschewing a double standard —in setting “the same high moral requirements for the hero as for the heroine.”²⁵ This is in stark contrast to most of the literature of Cervantes's age, which both reflected and perpetrated the kind of double standard described by Cardinal Bibbiena (Bernardo Dovizi) in Il Cortegiano:

     ²⁴ The sixteenth-century Spanish pastoral romances were distinguished by such monogendered female titles as La Diana, La Galatea, etc. On the enormous popularity and subsequent erasure of La Diana, the prototype of these pastoral works, see Elizabeth Rhodes, “Skirting the Men: Gender Roles in Sixteenth-Century Pastoral Books,” Journal of Hispanic Philology 11 (1987), 131-149.
     ²⁵ Hägg, The Novel in Antiquity, 95-96.

we ourselves, as men, have made it a rule that a dissolute way of life is not to be thought evil or blameworthy or disgraceful, whereas in women it leads to such complete opprobrium and shame that once a woman has been spoken ill of, whether the accusation be true or false, she is utterly disgraced for ever.”²⁶

The single standard for both sexes found in Greek romance, an atypical morality for the Renaissance, was not lost on Cervantes. Persiles easily competes with Heliodorus in setting a single moral standard, on the sexual front, for both sexes.
     This “two-in-one” sexual program, visibly “embodied” in the Persiles, is a retrospective construct. Cervantes had sketched out a rough draft of this gendered structure in La Galatea, his first romance, where he began by “splitting the difference.” The remarkable division of that whole work into masculine and feminine halves was long ago pointed out by Ruth El Saffar: “what is probably central to any deep analysis of the work,” she argues, “is that the shepherds are kept quite distinct from the shepherdesses almost throughout the entire novel.”²⁷ Sexual “distinction” (difference) is also central to any deep analysis of Persiles, a text in which duality is continuously expressed through tropes. To “compete with Heliodorus” —Cervantes's avowed intention in the Persiles— was to compete with the chief rhetorical figures in Heliodorus —figures such as antithesis, oxymora, paradox, and syllepsis. One Hellenist regards syllepsis (a.k.a. zeugma) as the master trope of Greek romance, the rhetorical figure that would seem to dominate over others. In this figure, two different meanings of the same word are invoked without repeating it. (The classic example in Cervantes is Dorotea's language of her deflowering: “y con volverse a salir del aposento mi doncella, yo dejé de serlo” [DQ I.28]). More intriguingly to our point, this rhetorical scheme, with all its delirium and incompatibility, engenders the topos of transvestism (as it does in Dorotea's story).²⁸ The master trope of Persiles, however,

     ²⁶ Castiglione, Courtier, 195.
     ²⁷ “The whole work pivots around four major interpolated tales, two told to men by men, and two told to women by women” (Ruth El Saffar, “La Galatea: The Integrity of the Unintegrated Text,” Dispositio 3 (1979): 340, 346.
     ²⁸ This complicated and uncommon scheme —the use of a single word that is grammatically and idiomatically compatible with two other words that it modifies or governs— provides readers with a double logic: suspended between “two divergent codes,” the Greek romance text becomes [p. 44] “a delirious seam edging incompatible systems of order which will never manage to address each other face to face.” Daniel L. Selden, “Genre of Genre: Theorizing Ancient Fiction” (Paper delivered at the Second International Conference on The Ancient Novel, Dartmouth College, 24 July 1989), 9-10.

yet another classical figure expressing duality, is, to my thinking, syneciosis. In this figure, as Susenbrotus explained it to the Renaissance, a contrary is joined to its opposite or two different things are conjoined closely.²⁹ I would suggest that this scheme of joined contraries motivates the formal, thematic, and intertextual features of Persiles. This figure operates in the text as a formula that resembles an architect's model: from its lineaments and dimensions, in other words, we may deduce the proportion of the whole work.
     The most notorious use of the scheme of syneciosis in world literature is the Androgyne in Plato's Symposium. Although it is a figure of ancient pedigree, the Androgyne is a classically debased figure that many Renaissance thinkers regarded as a “category of monster.”³⁰ As is well known, the Androgyne has its Western literary inaugural as a primordial myth in Plato's Symposium, where Aristophanes explains how the original nature of mankind embraced not two but three sexes: man, woman, and Androgyne (man/woman), all circular beings with eight hands and feet, four ears, and two privy members apiece. Zeus punished the insolence of these octopods by cutting them all in half, “like a sorb apple which is halved for pickling, or as you might divide an egg with a hair” —amusing kitchen similes that betray the Platonic denial of all difference between the two halves of the heterosexual Androgyne. (Two halves of an apple or an egg are not terribly different.) Plato's disarmingly ludicrous account of the genesis of two yearning, incomplete sexes is aptly articulated by Aristophanes, represented in Plato's dialogue as a buffoon given to prolonged seizures of the hiccups. His account of the degeneration of the happy Androgynes, nonetheless, includes a passage of considerable resonance in Western culture. Its language of division and reunion, with its metaphysics of nostalgia, is amazingly resonant across Persiles:

     ²⁹ Syneciosis is also known as contrapositum, conjunctio, or commistio. See Joannes Susenbrotus, Epitome troporum ac schematum et grammaticorum et rhetorum (Antwerp 1566), 82; cited by Lee A. Sonnino, A Handbook to Sixteenth-Century Rhetoric (London: Routledge & Kegan Paul, 1968), 61.
     ³⁰ See Maclean, Renaissance Notion of Woman, 39.

After the division the two parts of man, each desiring his other half, came together . . . longing to grow into one . . .: so ancient is the desire of one another which is implanted in us, reuniting our original nature, making one of two, and healing the state of man. Each of us when separated . . . is always looking for his other half.

Cervantes's imitation of this scene of splitting and fusion, however, although overpacked with redundant echoes of mathematical language, cannot be regarded as simply another “footnote to Plato.” In the Platonic text, it bears recalling, the androgynous sex is roundly disdained: the term androgynous there signifies “adulterous” men and women and is preserved, as Aristophanes is careful to explain, only as “a term of reproach.” In the same context, the “manly” sex is teasingly exalted —the sex represented by “the best of boys and youth.”³¹ Elaine Pagels reminds us that, in the Symposium, Plato showed “a group of men fighting hangovers from the night before by raising the glories of erotic— and especially homoerotic— love.”³² And Suzanne Lilar, along the same lines, laments that “it is to the love of boys that we owe the one and only great Western philosophy of love.”³³ The Androgyne that informs Persiles, however, is a syncretic Renaissance creation that corrects and criticizes its Platonic model. By way of formal endorsement of the sexuality it proposes, Cervantes's text is itself metaphorically “an” androgyne —a fiction structured in the shape and by the conjunction of two different halves. Let us look closer at Cervantes's uses of the overarching figure of the Androgyne, made visible for us through rhetorical, emblematic, and onomastic devices across Cervantes's text.
     The rhetoric in Persiles perhaps most blatantly figures forth the Androgyne by iterating variants of the formula of splitting followed by imaginative, blissful fusion. When contraries are joined to their opposites, as I mentioned above, the figure is called syneciosis. When contraries become one —when they fuse— the figure is called anagogy. Cervantes's use of anagogy —mystical mathematics of two-in-one that transcends all syntax

     ³¹ Plato, Symposium, trans. Benjamin Jowett (New York: Bobbs-Merrill, 1956), 30-32; emphasis added.
     ³² Elaine Pagels, Adam, Eve, and the Serpent (New York: Random House), 85.
     ³³ Suzanne Lilar, Aspects of Love in Western Society, trans. Jonathan Griffin (New York: McGraw-Hill, 1965), 68.

and logic— verges on abuse. In response to Auristela's long disease of jealousy, for example, Periandro speaks of producing a compound out of their conjoined souls: “un compuesto tan uno y tan solo . . . que tendrá mucho que hacer la muerte en dividirle” (185). The use of anagogy recurs even when the lovers are engaged in a divisive argument: “he dicho mal en partir estas dos almas, que no son más que una,” the heroine protests; and the hero, in turn, replies that no contentment can equal the experience of “dos almas que son una” (414-15).
     This two-in-one figure reappears throughout the text in an ostentatious partitive construction, one that focuses rhetorically on the concept of halves. The hero's first address to the heroine, whispered during their transvestite embrace foregrounds this figure: “¡O querida mitad de mi alma!” (67).³⁴ As a classical topos, this same phrase appears in Horace's ode addressing his friend Virgil —in the Latin pars animae meae (Ode 3.8)— as Avalle-Arce has usefully noted (57n). We also find the phrase in Ovid's Metamorphoses (8.406), Augustine's Confessions (4.6), and Petrarch's Epistola Metrica (1.1) —to note only a sample of Latin writers who employ variants of the vocative pars (or dimidium) animae meae to address their male friends. But the formula need not be consigned to a literary tradition of male friendship, which Cervantes himself tapped into at the beginnings but not the ends of his career. In La Galatea (1585), Timbrio laments that both his female love and his male rival for this love are the “dos verdaderas y mejores mitades de su alma.”³⁵ The transgressive logic of precisely this kind of metaphor may have moved Aristotle to remark, centuries earlier in the Politics, that the idea of Aristophanes' Androgyne appears to be somewhat watery when it embraces more than two persons (2.1. [1262b]). Cervantes must have outgrown such triangulated figurations, however, because they never occur in Persiles, where the “other half” is always and only the other gender. As such, the phrase gestures back to origins more remote than Horace and to differences more radical than those found between men. It is the myth of the two different

     ³⁴ To cite some other instances of this formula: “Auristela, mitad de su alma sin la cual no puede vivir” (57); “donde iba la mitad de su alma, o la mejor parte della (251); “no la mitad, sino toda su alma que se le ausentaba” (252); “¡Qué mudanza es ésta, mitad de mi alma?” (364); and “la mitad que le falta que es la del marido” (399).
     ³⁵ La Galatea, ed. Juan Bautista Avalle-Arce, 2 vols. (Madrid: Espasa-Calpe, 1968), 1:158.

sexual halves that underwrites Persiles —not like the split apple or egg in Plato, nor like the split masculine souls in Horace.
     All these “two-in-one” figurations engender, even more insistently than in Don Quixote, the topos of transvestism in Persiles. Emblematic representations of sexual difference surface through Cervantes's frequent use of cross-dress, a traditionally androgynous hallmark, in different episodes of Persiles. The emblem of the Androgyne, indeed, is programmatically installed at the threshold of his narrative, when the protagonists of Persiles first come together in the text in spectacular cross-dress (60). This duplex “metamorfosis” saves the protagonists' lives, for their androgynous embrace triggers the holocaust on the Barbaric Isle which allows them to escape to freedom. At the end of Persiles, the narrator sums up its beginning episodes in one memorably androgynous image: “la Isla Bárbara, cuando se vieron [Auristela] y Periandro en los trocados trajes, ella en el de varón, y él en el de hembra: metamorfosis bien estraño” (341).
     Persiles includes a notable population of transvestites, whose cross-dress augments the gender reversal of Cervantes's protagonists in their opening embrace. There is the unremarkable Ambrosia Agustina, who serves as a drummer's boy in Philip II's army while chasing after her new husband (361-66). A more militant cross-dresser is the young Lithuanian widow Sulpicia, who becomes an avenging pirate after hanging some forty would-be rapists from the tackle and yards of her ship. Although Sulpicia is “dressed to kill,” her transvestism is less bizarre than that of Tozuelo (3.8), the rustic who impersonates his pregnant girlfriend in order to spare her a possible miscarriage during some country dances in Toledo: “ella está encinta, y no está para danzar ni bailar” (329). Although Tozuelo's gender disguise signifies, in fact, a rare literary act of male nurturing, all the conventional phobias to male cross-dress are here articulated by the mayor of Toledo, who berates Tozuelo for the “delinquimiento” of his dress. Ultimately, however, the civic and paternal authorities in the text are persuaded to marry the couple, on the spot, with the groom still in female garb. An ad hoc “pregnant male,” Tozuelo's liminality seems pointedly daring in the light of Rosa Rossi's reminder, in Escuchar a Cervantes, of how transvestism was used, in Cervantes's Spain, as part of the death sentence for homosexuality. The suspect who confessed under torture to such practices, and was sentenced to

burn for them, could expect to approach his death disguised as a woman: in his last walk to the stake, the condemned man would be exhibited “trasvestido y rizado.”³⁶
     Despite these rebarbative associations of male transvestism —which has canonically represented humiliation and, in the context above, even death— Persiles asks its readers to respond to its blurring of sexual categories with a suspension of the old aversions: cross-dress is represented as an enterprise vital to its practitioners, vexing to its critics, and fatal to its opponents. Its representation gestures toward a different ordering of sexual relations, toward an economy founded on an ethical exchange of gendered subjectivities. Such an exploration does not require transvestism. There are characters in Persiles who experience the reality of the other gender without cross-dressing: the polyglot Transila, based upon the model of the Italian “virago,” is celebrated for her “varonil brío” (69); the putative “adúltero” Rosanio is dubbed “el caballero de la criatura” because he inherits the care of his son when the infant's unwed mother, Feliciana de la Voz, must flee for her life (290); and Isabela Castrucha, perhaps the most protean figure in the text, is addressed as “malino” and a “viejo” —as an obscene old man— while she successfully feigns demonic possession (409).
     Apart from these gender crossings, Cervantes's alertness to sexual difference is visible at the onomastic level. Persiles, the male protagonist of this work, travels across the text under a pseudonym that begs for interpretation. When we recall how Cervantes's earlier hero, Alonso Quijano, agonized about choosing a name that was “sonoro y significativo” for himself, his lady, and his horse (DQ I.1), we are moved to examine the compound pseudonym of Periandro for its own resonant signification. From the Greek prefix peri- (meaning round, around, about) and the noun andros (man), we infer a gender yoked to its own qualifier: Periandro, a region or place lying around the fixed gender of masculinity, much as the region about the heart is called the pericardium. Cervantes may, in fact, have borrowed the name of Periandro from Greek romance, from the characters called Periandro/Periandra in Achilles Tatius's Leucippe and Clitophon.³⁷ In response to repeated criticism of his narrative

     ³⁶ Rosa Rossi, Escuchar a Cervantes: Un ensayo biográfico (Valladolid: Ambito Ediciones, S.A., 1988), 7; see the instructive footnote on this page.
     ³⁷ The debt of the Persiles to Leucippe and Clitophon may be an indirect one, via the Castilian version of Núñez de Reinoso, the Historia de los amores de Clareo y Florisea (1552).

techniques, Cervantes's hero likens himself to some anagogic region containing all time and space, a narrative container “donde todas las cosas caben y no hay ninguna sin lugar” (227). Speaking from a peripheral reality, this ex-centric protagonist insists on incorporating —both into himself and his narratives— all differences, sexual and otherwise. This refusal to limit or confine either gender or speech is characteristic not only of Periandro —whose individual consciousness is notably capacious— but also of the narrative that contains him, Persiles itself.
     Ruth El Saffar warns that “as long as we are caught in the dichotomies that entangled the Cervantes of 1605, we will not be able to read his late romances.³⁸ One of the critical tasks of Persiles criticism in the 1990s will be to rethink, among other dichotomies, the male/female “split.” For centuries, it would seem, we have talked about “reconciling” or “transcending” dualisms, an unproblematic act, it would seem, for “the lunatic, the lover, and the poet” —to borrow Shakespeare's triad of imaginative souls. The lunatic's approach to dualisms may be instanced by Don Quixote, whose endearing “locura” is of the kind “que las más veces . . . juzga to blanco por negro y lo negro por blanco” (DQ II.10). The lover's approach may be exemplified by Castiglione, who movingly proclaims that “Love gives likeness to the unlike” (Courtier, 341-42). The poet's approach is taken (ironically) by Plato, who articulates the figure of the Androgyne, that visionary merger of male and female. What we celebrate in all these dreamers, however, their ability to synthesize opposites, we question in literary critics. The unabashed reductionism of a Casalduero, for example —for whom unity simply “rises out of” opposition (“de cuya tensa oposición surge la unidad”)— now strikes us as a repression of the knowledge of difference.
     Those of us who are (at least for the moment) neither crazy, nor lovesick, nor writing Petrarchan poetry may wish to transfer the analysis of sexual difference out of the aesthetic transcendental and into the “real world” of human politics. Will it help us to think of male and female the way Cervantes thinks of good and evil, as not merely juxtaposed, but like “two concurrent lines” that end up in one place? What would this “place” be like? Would it resemble Periandro's “self”—that hospitable “lugar” where “todas las cosas caben”? Would it resemble Julia Kristeva's “new theoretical and scientific space,” where the very notion

     ³⁸ Ruth El Saffar, “The Truth of the Matter: The Place of Romance in the Works of Cervantes,” in Romance: Generic Transformations, 251.

of sexual identity is challenged?³⁹ Like Freud's notion of femininity in the 1930s, this new space still remains a “terra incognita.” Our task is to chart it for the “new Hispanism” of the 1990s.

     ³⁹ Julia Kristeva rejects the dichotomy between masculine and feminine as an attitude belonging to metaphysics. This approach, which necessarily challenges the very notion of sexual identity, is a refusal of both biologism and essentialism: you are not your biological sex, Kristeva would argue, but the subject position you wish to take up. See “Women's Time,” trans. Alice Jardine and Harry Blake, Signs 7, 1 (1981), 33-34.