From: Cervantes: Bulletin of the Cervantes Society of America 1.1-2 (1981): 63-81.
Copyright © 1981, The Cervantes Society of America

Cide Hamete Benengeli vs. Miguel de Cervantes: The Metafictional Dialectic of Don Quijote*


THE SELF-REFLECTIVE narrative technique of Don Quijote has fascinated and baffled the novel's readers since the very beginning. The metafictional concerns dealt with in the work were clearly of interest to Cervantes himself, and they have been recognized implicitly by novelists from Laurence Sterne to Carlos Fuentes. In recent years the novel's narrative structure has been the subject of critical studies by John J. Allen, Robert Alter, Ruth El Saffar, Colbert I. Nepaulsingh, James A. Parr, Helena Percas de Ponseti, and E. C. Riley, among others.
     In this essay I propose to examine the levels of narration in Don Quijote in such a way as to distinguish among the specific narrative voices and discuss the implications of these distinctions for how we read the book. It is my contention that some technical and stylistic clues exist that enable us to discern more precisely than ever before the role and personality of Cide Hamete Benengeli. The metafictional dialectic developed between the Arabic author and Miguel de Cervantes, his editor, is established in Part I, but only fully developed and given prominence in Part II.

     * For a response to this article, see Elias L. Rivers. “Narrators, Readers, and Other Characters in Don Quijote.” Cervantes 2.1 (1982): 96-98.



     The narrative structure of most of Part I of Don Quijote appears to be relatively straightforward and uncomplicated:

1)   after I, 8, the Moorish historian Cide Hamete Benengeli records the events in the lives of Don Quijote and Sancho Panza;
2) the historian's Arabic manuscript is then translated into Spanish by an unnamed morisco;
3) The Spanish version is edited for the reader by Miguel de Cervantes.

That the editor is in fact Cervantes is frequently questioned by many of the best readers of the book, who would have the final editor be the unnamed and never identified “segundo autor” mentioned in I, 8, 971 or perhaps an even more obscure “ultimate author.”2 I find no alternative, however, to agreeing with E. C. Riley that “for practical purposes,” the author of the final, edited text which we read is Miguel de Cervantes.3 No one, to my knowledge, doubts that the “yo” of the prologue, who relates himself to the character and text by claiming to be not the “padre” (i.e, original author) but the “padrastro” (i.e., editor) of Don Quijote (p. 12), and who tells the story of being visited by a friend while pondering the problem of writing a prologue for his book, is anyone other than the person referred to on the title page where it says “compuesto por Miguel cle Cervantes Saavedra.” Since there is no evidence to indicate an alternative, the first-person editor who appears occasionally in Part I, and very prominently in his search for Cide Hamete's manuscript in I, 9, should be considered the same person who narrated the prologue. The passage that most seriously clouds the issue is the previously cited final paragraph in I, 8, where the “autor” interrupts his narration because he has exhausted his source material, although the “segundo autor” refuses to believe that no more material is available. The passage is somewhat

     1 All references in the text are to the edition of Don Quijote prepared by Martín de Riquer (Barcelona: Planeta, 1962) and refer to Part, Chapter, and page, as required by context.
     2 See Ruth El Saffar, Distance and Control in “Don Quixote” (Chapel Hill: University of North Carolina Press, 1975), pp. 30-31. See also George Haley, “The Narrator in Don Quixote: Maese Pedro's Puppet Show,” Modern Language Notes, 80 (1965), 146-48, and F. W. Locke, “El sabio encantador: The Author of Don Quixote,” Symposium, 23 (1969), 47-50.
     3 E. C. Riley, “Three Versions of Don Quixote,” Modern Language Review, 68 (1973), 808.

1 (1981) The Metafictional Dialectic of Don Quijote 65

confusing because both “autor” and “segundo autor” could conceivably refer to the writer of the text we are reading,4 but surely the latter term is in this context little more than a euphemism for “yo,” the “yo” who is so prominent in the following chapter.5
     There are a few noteworthy indications in the text that could suggest a fourth level of narration. The most suggestive is a reference to “el propio original desta historia” (II, 44, 906; see below), which gives rise to the theory of F. W. Locke that there must have

     4 The same ambiguity and uncertainty of attribution is found in I, 20, when the “autor desta historia” who concludes that Sancho Panza is an old Christian could be either the author or the editor (p. 203), and in I, 22, where the “autor desta historia” (p. 587) most logically seems to refer to the “author” of the final text, i.e., the editor, Miguel de Cervantes. The same problem exists in the final narrative paragraphs of I, 52, where the references to “el autor desta historia” and “el fidedigno autor desta nueva y jamás vista historia” (pp. 557 and 558) probably refer to the editor rather than the original author. Riley discusses Cervantes' casual use of various terms to refer to his own narrative presence in the novel in Cervantes's Theory of the Novel (Oxford: Clarendon Press, 1962), p. 207.
     5 It is the nature of the first-person fictional narrator to reveal his identity. When the narrator is not a character in the work or an identified fictional editor he is assumed to be the person whose name is on the book's cover. This does not, of course, mean that it is a literal truth that Cervantes had a friend with whom he carried on the conversation recorded in the prologue (pp. 12-18) or that he literally searched through the streets of Toledo for the book's original manuscript (I, 9, 100-102). The Cervantes who edits and narrates Don Quijote may be fictionalized, but he most certainly is Cervantes. See John J. Allen, Don Quixote: Hero or Fool? (Gainesville: University of Florida Press, 1969), pp. 11-12. The historian-translator/editor device is, of course, common to the romances of chivalry (see Daniel Eisenberg, “The Pseudo-Historicity of the Romances of Chivalry,” Quaderni Ibero-Americani, 45-46 [1975], 253-59). Many modern novels also use this type of narrative framework. One which particularly recalls Cervantes is Alessandro Manzoni's I promessi sposi, which does not have the intermediary “translator,” but does on a small scale evoke the ironic and complex tone of Don Quijote. The “io” who discusses his manuscript sources in the author's introduction (“Ma, quando io avrò durata 1'eroica fatica di trascriver questa storia da questo dilavato e graffiato, autografo, e l'avrò data, come si suol dire, alla luce, si troverà poi chi duri la fatica di leggerla?” [Florence: Adriano Salani, 1909], p. 6) and who interrupts the narration on dozens of occasions in order to stress the work's historicity, comment on his sources, etc. (“il nome di questa, nè il casato del personaggio, non si trovan nel manoscritto, nè questo luogo nè altrove,” p. 11) is not considered by any critic with whom I am familiar as anyone other than Alessandro Manzoni. It is not surprising that this device should be most readily found in historical novels.


been an even more remote source for Cide Hamete Benengeli's manuscript.6 This passage and others which can be read to make the Moorish writer a non-omniscient reporter7 are provocative, but the possible multiplication of layers of narration does not alter the basic thrust of my presentation. While the narrative structure of Don Quijote is not always consistent or clear, the tri-level author-translator-editor relationship remains the basis for a discussion of Cervantes' narrative technique.
     There are only a few hints of the editor's tampering with his material, the best example of which is the statement that Don Quijote's squire is called both Panza and Zancas, “que con estos dos sobrenombres le llama algunas veces la historia” (I, 9, 102), after which he is only called Panza, all citations of Zancas having presumably been regularized by the editor.8 In the same passage Cervantes notes that there were “otras algunas menudencias . . . de poca importancia y que no hacen al caso a la verdadera relación de la historia” and which he has presumably edited out.
     Cervantes intervenes occasionally to comment on his work and on the author. Most of the references to Cide Hamete are complimentary; he is described as “muy curioso y muy puntual” (I, 16, 158) and “sabio y atentado” (I, 27, 297), and his text is regularly praised for its veracity, attention to detail, and historicity. But from the start Cervantes also begins to exploit some of the comic possibilities that accompany his author. When he first discovers the manuscript he laments that the author is Arabic, “siendo muy propio de los de aquella nación ser mentirosos” (I, 9, 102). Then, after noting that historians should be “puntuales, verdaderos y no nada apasionados” —essentially the qualities praised in Cide Hamete— he eschews responsibility for any error in the story: “y si algo bueno en ella faltare, para mí tengo que fue por culpa del galgo de su autor” (p. 103). But if the editor first undercuts his historian's authority and reliability by describing him as a lying dog of a Moor, he further subtly reduces his

     6 Locke, p. 54.
     7 Colbert I. Nepaulsingh, “La aventura de los narradores del Quijote,” in Actas del Sexto Congreso Internacional de Hispanistas, eds. Alan M. Gordon and Evelyn Rugg (Toronto: Department of Spanish and Portuguese of the University of Toronto, 1980), pp. 515-16.
     8 See Robert Alter, Partial Magic: The Novel as a Self-Conscious Genre (Berkeley: University of California Press, 1975), p. 9.

1 (1981) The Metafictional Dialectic of Don Quijote 67

stature when he suppresses a passage by Cide Hamete in I, 16 on the grounds that hearsay evidence has it that the supposedly illustrious historian is in some fashion related to the brutish muleteer from Arévalo with whom Don Quijote fights in that chapter: “y aun quieren decir que era algo pariente suyo” (p. 158), a comment which the editor recalls shortly afterward when he describes Cide Hamete as an “autor arábigo y manchego” (I, 22, 220).
     The longest passage in which it is clearly the editor and not the historian speaking is contained in the opening paragraph of I, 28, where the praise for the protagonist and the defensive commentary on the “verdadera historia” and the “cuentos y episodios della” (p. 298) are not contained in the original Arabic manuscript but are remarks of the editor. The shift back to the translation of the historical text is signaled by the phrase “la . . . historia . . . cuenta que . . . ,” which is in fact a variant of a standard formula (significantly of Arabic origin) for appealing to an authority by citing or paraphrasing that authority's words.9 The form of this passage, as will be seen in more detail below, is crucial. It cannot be the author who refers to “la historia” or who says “dice la historia.” Cide Hamete Benengeli, as author of “la historia” cannot logically cite his own words. Only the translator or the editor, who stand apart from the text and follow it chronologically, can refer to it or to its author. The formulaic expression “dice la historia” provides our best clue in the effort to separate the narrative voices of the author and the editor.
     After the slightly confusing and conflicting evidence sometimes presented in the earliest chapters when Cervantes has to rely for his information on various sources, from I, 9 until the end of the first part, the author, Cide Hamete Benengeli, presents a relatively straightforward and reliable account of events. The morisco translator is presumably reasonably accurate. The editor, Miguel de Cervantes, once or twice casts doubt on his author but more often praises him and his work. The distinct impression left with the reader at the end of I, 52, when Cervantes, inspired by rumor of a third sally but unaware of what is yet to come, sets out in search of more manuscript material, is one of confidence in the book's narrative framework.
     Cide Hamete Benengeli's expanded role in Part II of Don Quijote

     9 Raymond S. Willis, Jr., The Phantom Chapters of the “Quijote” (New York: Hispanic Institute, 1953), pp. 100-102.


should be readily apparent to the reader; his name alone appears over six times more frequently in the second part.10 It is not only the quantity of his appearance in the work that is important, but the quality. The Moor is a far more unreliable narrator —even to the point of lying— while, at the same time, he is oblivious to changes in his characters, especially Don Quijote. His editor, Cervantes, is quite aware of these shortcomings on the part of the author and reacts by making Cide Hamete an object of ridicule.
     John J. Allen first discussed the implications of the dialectic established between Cide Hamete Benengeli and Miguel de Cervantes.11 Helena Percas de Ponseti has recognized the importance of Allen's study,12 but James A. Parr and Ruth El Saffar specifically have expressed serious objection to the idea of separating the author and the editor.13 El Saffar specifically laments that Allen “does not show us how to distinguish Cide Hamete's voice from Cervantes'”14 I suggest that the main fault with Allen's article is that it does not go far enough. A reader's perception of an author can be directly determined by an editor. The latter can ignore, mask, eliminate, or otherwise soften any possible errors or inconsistencies on the part of the former, or he can call attention to and comment on them. Cervantes does the latter.15 In the remainder of this essay I will

     10 In the index to Riquer's edition (p. 1154), Cide Hamete Benengeli's name is listed as appearing only five times in Part I and 33 times in Part II. El Saffar estimates that he “appears at least a hundred times” in Part II and that this increased presence is due to “the need for distancing and clarification.” Distance and Control, p. 83.
     11 “The Narrators, the Reader, and Don Quijote,” MLN, 91 (1976), 201-12, subsequently incorporated into Don Quijote: Hero or Fool?, Part II (1979), 3-15.
     12 “Sobre el enigma de ‘los dos Cervantes,’” The American Hispanist, 2, xvi (1977), 10.
     13 Parr, “Aesthetic Distance in the Quixote,” in Studies in Honor of Gerald E. Wade, eds. Sylvia Bowman et al (Madrid: Porrúa Turanzas, 1979), pp. 191-97, and El Saffar, “Concerning Change, Continuity, and Other Critical Matters: A Reading of John J. Allen's Don Quixote: Hero or Fool?, Part II”, Journal of Hispanic Philology, 4 (1980), 241- 46.
     14 “Concerning Change,” p. 243.
     15 Riley has noted some of Cide Hamete's shortcomings, and then has added the following: “But our principal doubts about Benengeli's reliability as a historian are planted in us by what Cervantes says about him on certain occasions.” “Three Versions,” p. 809. Mia I. Gerhardt, a generally [p. 69] careful and discerning reader of Cervantes' novel, fails completely to see the irony in certain statements by and about Cide Hamete Benengeli in Part II. She says, for example, that he is presented “comme un historien très scrupuleux” and that after Don Quijote's comment about lying Moors in II, 3, Cide Hamete “n'est plus déprécié nulle part.” Don Quijote, la vie et les livres (Amsterdam: N. V. Noord-Hollandsche Uitgevers Maatschappij, 1955), pp. 27, 33.

1 (1981) The Metafictional Dialectic of Don Quijote 69

discuss Cide Hamete's increased role in Part II, point out many of his lies and errors in judgment, show how Cervantes directs the reader to laugh at him, and comment on how the historian's new role is related to the reader's perception of Don Quijote.
     In Part I the only direct commentary on the story that can be attributed to Cide Hamete Benengeli is the marginal note on Dulcinea which causes the morisco's laughter and reveals to Cervantes that he has discovered the manuscript for which he has been searching (9, 101). In Part II his interruptions, and concurrent involvement with the work and the characters, are found throughout: he praises Allah as Don Quijote's third sally is begun (8, 630); he would prefer not to narrate the protagonist's unbelievable madness when he sees the peasant girl Sancho calls Dulcinea (10, 642-43); he discusses his dilemma in properly describing Don Quijote's adventure with the lion (17, 702-703); he comments on the events Don Quijote claimed took place in the Cave of Montesinos (24, 761-62); he takes an oath like a Christian (27, 788); he comments on the structure of his work as he is about to separate knight and squire (44, 906-907); he digresses on the subject of poverty (44, 911-12); he promises to relate an adventure in a later chapter (47, 937); he makes a parenthetical aside during the Doña Rodríguez adventure (48, 941). In addition to these primarily comic appearances of the author (most of which are discussed in more detail below), on dozens of occasions Cervantes calls attention to the author simply by mentioning his name or by referring explicitly to “la historia” (particularly by using the formulaic “dice la historia que . . .”).
     In the first part of the novel Cide Hamete Benengeli commits only one judgmental error, a relatively minor inconsistency when, after Don Quijote drinks the magic healing balm of Fierabrás, the knight is once referred to as “aliviado y sano” and then as “molido y quebrantado” (I, 17, 168 and 171). In Part II there are at least ten times when


the increasingly unreliable historian either lies,16 makes conflicting and inconsistent statements, or clearly misjudges his characters:

     1) After defeating Sansón Carrasco dressed as the Caballero de los Espejos, Don Quijote is twice described as proud and haughty: “En estremo contento, ufano y vanaglorioso iba don Quijote por haber alcanzado vitoria de tan valiente caballero como él se imaginaba que era el de los Espejos” (II, 15, 685); “Con la alegría, contento y unfanidad que se ha dicho seguía don Quijote su jornada, imaginándose por la pasada vitoria ser el caballero andante más valiente que tenía en aquella edad el mundo” (II, 16, 687). But whereas in Part I, following his comparable victory over the Biscayan, Don Quijote's pride was evident in his words (“Pero dime por tu vida: ¿has visto más valeroso caballero que yo en todo lo descubierto de la tierra? . . . ,” I, 10, 106), now he only talks calmly with his squire about whether or not his defeated opponent was in fact his friend Sansón Carrasco. Very shortly afterward there is a reference to Don Quijote's “profunda melancolía” (p. 693), which rings somewhat truer than the statements about his pride. As we will see several times below, what is said about Don Quijote and what he actually does and says are sometimes in conflict. Cide Hamete Benengeli seems to be a better historian than psychologist: his narration of events is generally quite acceptable but his interpretation of these events or of the characters' motives or psychological states is frequently questionable.

     2) Cide Hamete Benengeli's altisonant statements in preparation for the encounter between Don Quijote and the lion are clearly inconsistent with the “niñerías” and “bravatas” (17, 702-703) which follow.17

     3) After recording Don Quijote's version of the episode in the Cave of Montesinos, Cide Hamete states that while events literally could not have occurred in the fashion described he cannot believe that the knight is lying (24, 762). What the historian never considers in this passage is the possibility (for the reader, the obvious truth) that what Don Quijote related was a dream. Cide Hamete Benengeli may be consciously misleading here, or he may simply

     16 Recall that in Part I Cervantes was worried about the fact that the historian on whom he had to depend was a lying Moor (I, 9, 102-103). In Part II the characters take up the same refrain: Don Quijote is disconsolate over the idea that his historian is Moorish and therefore a congenital liar (3, 597); Don Quijote and Sancho Panza specifically criticize the work of Cide Hamete (3, 602 and 4, 606, respectively).
     17 See Allen, Hero or Fool, II, 9-11, for relevant observations concerning items 2, 3, 5, 6, and 10 in this list.

1 (1981) The Metafictional Dialectic of Don Quijote 71

be a bewildered reporter unable to verify events through objective sources, but clearly the effect on the reader is similar to that produced by the other passages discussed in this list.

     4) Perhaps the historian's most spectacular lie is his statement that when Don Quijote entered the Duke's palace “aquél fue el primer día que de todo en todo conoció y creyó ser caballero andante vercladero, y no fantástico” (31, 813). The great majority of readers of the novel have maintained that Don Quijote is most consciously chivalric at the beginning of his career. If one has to choose between the validity of this narrative comment and the cumulative evidence of the previous 800 pages of text, there is no alternative but to opt for the latter. What this statement illustrates, as well as any in the novel, is the unreliability of the narrator, specifically of Cide Hamete Benengeli. The historian's words might mean simply that Don Quijote believes that he is being honored as a knight-errant more authentically than ever before, but accepting them as the literal truth can only lead to confusion.18 What is more, Don Quijote's own words almost immediately after this comment suggest his complete consciousness of his own non-chivalric reality. After Sancho's exchange with Doña Rodríguez, the knight takes his squire aside and counsels him to restrain his tongue: “¿No adviertes, angustiado de ti, y malaventurado de mí, que si veen que tú eres un grosero villano, o un mentecato gracioso, pensarán que yo soy algún echacuervos, o

     18 Several critics have cited these words as a high point in Don Quijote's chivalric madness, often in conflict with their own previous statements about the protagonist. The most interesting reactions come, as one might expect, from those who see Don Quijote as playing or acting throughout the novel. Arturo Serrano Plaja triumphantly stresses the literal truth of these words: “Por una vez, así como de pasada, el historiador nos dice todo: que don Quijote antes nunca se había creído caballero verdadero, sino fantástico. ¿Pero acaso eso, exactamente eso, no es lo que se llama, lisa y llanamente, jugar? ¿Se puede decir más claro que don Quijote, hasta ahora por lo menos, jugaba al caballero como jugaba al enarnorado?” Realismo “mágico” en Cervantes: “Don Quijote” visto desde “Tom Sawyer” y “El Idiota” (Madrid: Gredos, 1967), p. 86. Mark Van Doren, the original and still the best proponent of the “Don Quijote actor” reading of the novel, expresses doubt concerning the reliability of the statement: Don Quixote's Profession (New York: Columbia University Press, 1958), p. 59. Gonzalo Torrente Ballester provides a careful discussion of the implications involved in either accepting or rejecting Cide Hamete Benengeli's comment, concluding: “De todas suertes, la frase aislada o en su contexto constituye un escollo difícil de salvar para cualquier interpretación.” El “Quijote” como juego (Madrid: Guadarrama, 1975), pp. 206-207.


algún caballero de mohatra” (p. 816).19 This and Sancho's response that “nunca por él se descubriría quién ellos eran” are virtual admissions that both are quite aware of reality.

     5) One of the most obvious of the Moor's deceptions is when he discusses the name of the Countess Trifaldi: “y así dice Benengeli que fue verdad, y que de su propio apellido se llama la condesa Lobuna” (38, 868). The whole Trifaldi adventure is no more than an elaborate sham in which the Duke's majordomo plays the part of the countess.

     6) In exactly the same category is the supposed farmer from Miguel Turra who is described in narration as “de muy buena presencia, . . . bueno e buena alma” (47, 934), and who turns out to be a “bellacón” playing a role in order to make Sancho Panza look ridiculous (p. 937).

     7) When Don Quijote departs from the Duke's castle it is stated that “los espíritus se le renovaban para proseguir de nuevo el asumpto de sus caballerías” (58, 1015). In the remainder of the novel Don Quijote exhibits no more willingness to see the world in terms of chivalric standards than he had while with the Duke and Duchess. In I, 52 a statue of the Virgin could be perceived as a maiden in distress; in II, 58 the four chivalric religious figures are mere statues for Don Quijote. Interestingly, the very chivalric Santiago suffers in comparison with the pacific Saint Paul in Don Quijote's disquisition about the four saints (pp. 1017-19). Equally interesting is his admission of failure, or, at least, confusion and self-doubt, in comparison with the obvious accomplishments of the saintly knights: “Ellos conquistaron el cielo a fuerza de brazos, . . . y yo hasta agora no sé lo que conquisto a fuerza de mis trabajos.” (p. 1018).

     8) By the time he arrives in Barcelona Don Quijote acts more like a celebrity tourist than a knight-errant. He rides along the beach and through the streets of the city in the company of a

     19 From the beginning of the second part Don Quijote has been afraid of what Sancho might say. When the latter burst in upon the conversation he was sustaining with the priest and the barber, Don Quijote was “temeroso que Sancho se descosiese y desbuchase algún montón de maliciosas necedades, y tocase en puntos que no le estarían bien a su crédito” (II, 2, 592). When he first sent his squire to approach the Duchess, Don Quijote warned him, “Y mira, Sancho, cómo hablas” (II, 30, 808). Then Don Quijote apologized to the Duchess, saying that she must believe that “no tuvo caballero andante en el mundo escudero más hablador.” When the lady described Sancho as “gracioso y donairoso . . . [and] discreto,” Don Quijote could only add: “Y hablador” (p. 811).

1 (1981) The Metafictional Dialectic of Don Quijote 73

prominent local citizen, Don Antonio Moreno. A dinner and a dance are celebrated, and an exhibition of an enchanted talking head is arranged, all in the knight's honor. At the dinner Don Quijote is, according to the narration, treated like a knight-errant. His reaction, it is said, is that “hueco y pomposo, no cabía en sí de contento” (II, 62, 1054). None of Don Quijote's words or acts, however, confirms this statement. At the soirée, two ladies, described as “de gusto pícaro y burlonas,” dance repeated1y with Don Quijote, largo, “tendido, flaco, amarillo, estrecho en el vestido, desairado, y sobre todo, no nada ligero,” until he sits “en mitad de la sala, en el suelo, molido y quebrantado de tan bailador ejercicio” (II, 62, 1058). He visits a printing establishment and a ship, where he witnesses an exciting sea capture. Chivalry and knight-errantry may be referred to occasionally, but they are clearly not vital factors in Don Quijote's existence.

     9) Cide Hamete uses one of his rare similes to describe Don Quijote's dejection on the way home from Barcelona: “Si muchos pensamientos fatigaban a don Quijote antes de ser derribado, muchos más le fatigaron después de caído . . . como moscas a la miel, le acudían y picaban pensamientos” (II, 67,1091). Once again, the illustrious historian's word is not to be taken literally. There is no evidence whatsoever in anything that Don Quijote does or says that suggests that his defeat has been a crucial factor in his withdrawal from action to thoughtfulness. Don Quijote's pensiveness started early in Part II and was mentioned with increasing frequency through the stay at the ducal palace and afterwards; the actual chivalric defeat by Sansón Carrasco should not be considered the literal turning point (from “muchos” to “muchos más”) in this progression. The references to the thoughts of the knight are all significant. They appear after the enchantment of Dulcinea (“Pensativo además iba don Quijote” [II, 11, 653]), after the adventure of the enchanted boat (“sepultado en los pensamientos de sus amores” (II, 30, 807]), after discovering the holes in his stockings and hearing Altisidora's love lament (“Dejamos al gran don Quijote envuelto en los pensamientos . . . .  Acostóse con ellos y, como si fueran pulgas, no le dejaron dormir ni sosegar un punto” [II, 46, 946]), after being pinched by the Duchess and Altisidora (“doloroso y pellizcado, confuso y pensativo” [II, 48, 924]), after being trampled by bulls and while contemplating —literally?— suicide (“y déjame morir a mí a manos de mis pensamientos y a fuerzas de mis desgracias” [II, 59, 1027-28]), after he is surprised and taken prisoner by the Catalan bandits (“armado y pensativo, con la más triste y melancólica figura que pudiera formar la misma tristeza” [II, 60, 1040]), after his defeat by the Caballero de la Blanca Luna (“Seis días estuvo don Quijote en el lecho, marrido,


triste, pensativo y malacondicionado, yendo y vinendo con la imaginación en el desdichado suceso de su vencimiento” [II, 65, 1082]), and once more subsequent to Cide Hamete's statement on the way home (“daré rienda a mis pensamientos, y los desfogaré en un madrigalete” [II, 68, 1099-1100]). Admittedly, the words pensamiento or pensativo do not need to be present in narration for the reader to perceive the thoughtfulness of the character, but overall it seems clear that no textual support exists for the historian's statement.

     10) Demonstrably untrue is Cide Hamete's claim that the reason why Don Quijote sees an inn and not a castle on the way back to the village is that “después que le vencieron, con más juicio en todas las cosas discurría” (71, 1119). Don Quijote has never taken an inn for a castle in Part II, a fact which the historian himself noted on two previous occasions (24, 768 and 59, 1029).

     In Part II Cide Hamete lies and purposely misleads the reader. He also makes a series of observations about the characters, particularly Don Quijote, which are inconsistent with other facts in the narration.20 As stated above, Cide Hamete is a better historian than psychologist. Cervantes is inferentially aware of the increasing unreliability of his narrator and, although he continues to praise the veracity and accuracy of the narration, he offers a long series of comments on the Moor in such a way as to praise him and his work ironically, subtly reduce his stature, or criticize and ridicule him outright.
     The most frequent and obvious way in which Cervantes draws critical attention to Cide Hamete is by pointing out what the historian fails to say or what he omits: “fundándose no sé si en astrología judiciaria que él se sabía, puesto que la historia no lo declara” (II, 8, 631); “venían tres labradoras sobre tres pollinos, o pollinas, que el autor no lo declara” (II, 10, 646);21 “con cinco calderos,

     20 Allen also notes examples of the historian's “insensitivity” toward Don Quijote: Hero or Fool, II, 11. To Allen's observations I would add the statement in II, 72, when knight and squire are on their way home, that they travelled for a day and night “sin sucederles cosa digna de contarse, si no fue que en ella acabó Sancho su tarea” (pp. 1125-26), i.e., that Sancho finally completed his lashes and thus fulfilled the conditions for disenchanting Dulcinea. To pass off the accomplishment of the task that has most preoccupied Don Quijote since II, 10 as a mere afterthought is, to say the least, “insensitive.”
     21 The manner in which the identity of the animals ridden by the three girls is handled can not only cast doubt on the reliability of the historian but also impugn that of the editor.

1 (1981) The Metafictional Dialectic of Don Quijote 75

o seis, de agua, que en la cantidad de los calderos hay alguna diferencia” (II, 18, 709);22 “le tomó la noche entre unas espesas encinas o alcornoques; que en esto no guarda la puntualidad Cide Hamete que en otras cosas suele” (II, 60, 1037); “arrimado a un tronco de una haya o de un alcornoque —que Cide Hamete Benengeli no distingue el árbol que era—” (II, 68, 1100).
     Other lengthier and more noteworthy editorial commentary on the Moorish historian's work can also be found throughout Part II. In a long aside in II, 12, Cervantes appeals to extra-textual evidence to support a passing reference by Cide Hamete to the friendship between Rocinante and Sancho's donkey: “que hay fama, por tradición de padres a hijos, que el autor desta verdadera historia hizo particulares capítulos della . . . .  Digo que dicen que dejó el autor escrito que . . .” (p. 662). His defense of the historian's use of human analogies (“Y no le parezca a alguno que anduvo el autor algo fuera de camino” [p. 663]) only underscores the burlesque tone of the whole passage.
     Comparable in tone, but considerably reduced in scope, is the paragraph in II, 18 in which the morisco translator omits Cide Hamete's lengthy description of the home of Don Diego de Miranda “porque no venían bien con el propósito principal de la historia,” to which is appended the editorial judgment that the history “más tiene su fuerza en la verdad que en las frías digresiones” (p. 708). Thus translator and editor concur in criticizing the historian's methodology, implying somehow that there is a lack of truth in the excised descriptive passage.
     In the same spirit as the passages just discussed is a series of eight very important comic chapter introductions in which Cervantes engages in entertaining commentary on Cide Hamete Benengeli and his narration. Here, as much as in all other previously discussed passages combined, Cervantes brilliantly exploits the comic possibilities inherent in his historian / narrator. The structure of the passages is always identical: the opening paragraph of the chapter contains some sort of comic editorial commentary which reduces the stature of Cide

     22 In this passage it appears as though Cervantes is comparing Cide Hamete's text against others and notes the discrepancy in numbers. But whether the editor's comment is read so as to blame the inconsistency on Cide Hamete or to indicate that the latter's opinion is merely one of several possibilities, the effect is essentially the same as that produced by the other passages cited in this list.


Hamete Benengeli; the end of the critical passage is signaled by the use of the formulaic expression “cuenta la historia” or some variant of that phrase; narration is then resumed with no discernible carryover of the comic tone.23 This structure was used once in the first part (28, 298, cited above), but without any of the mild humor in the paragraph being directed at the Moorish historian.
     The eight chapters of Part I in which the technique of opposing Miguel de Cervantes and Cide Hamete Benengeli is used are the following:

     1) Chapter 5 begins with the translator's complaint that Sancho speaks “con otro estilo del que se podía prometer de su corto ingenio,” which is dutifully passed on by the editor in a literally parenthetical introductory statement which ends with the words “así, prosiguió diciendo: . . .” (p. 611). Thus the morisco directly and Cervantes inferentially (both here and in the subsequent parenthetical remarks on pp. 614 and 616) put into doubt the veracity of Cide Hamete's historical narrative. The irony is that Sancho's speech (and, therefore, also the author's narration) is perfectly appropriate; it is the translator's judgment which is questionable.

     2) The first words of Chapter 8 are Cide Hamete's enthusiastic “¡Bendito sea el poderoso Alá!” which, we are told, is repeated three times, along with an exhortation to enjoy the new exploits of Don Quijote (p. 630). The final words of the paragraph signal the editor's return to the narration: “y así prosigue diciendo . . . .”

     3) At the beginning of Chapter 10 it is stated that the author would prefer not to have to record the potentially unbelievable antics of Don Quijote, but fulfills his duty to

     23 Throughout the Espejo de príncipes y caballeros, 6 vols., ed. Daniel Eisenberg (Madrid: Espasa Calpe, 1975), the “translator/editor,” i.e., the author Diego Ortúñez de Calahorra, occasionally comments in first person on the “history” written by the wizard Artimidoro (assisted by the wizard Lirgandeo). This type of commentary is sometimes found in the opening section of a chapter and generally ends with some variant of “dice la historia que . . . ,” in a form very similar (except for the general absence of intentional humor) to that used by Cervantes. Particularly noteworthy is the long introductory passage to Chapter 38 of Book III (Espejo, VI, 88-92, to which Eisenberg calls attention as resembling Cervantes' practice in Don Quijote. While annotating this passage (p. 89, n. 8), Eisenberg notes that a scholarly study of the “traductor independiente” who comments on his text would be a valuable contribution to our understanding both of the romances of chivalry and of Don Quijote.

1 (1981) The Metafictional Dialectic of Don Quijote 77

record the facts in detail, “sin añadir ni quitar a la historia un átomo de la verdad” (pp. 642-43). Cervantes concurs (“y tuvo razón”) and then returns to his text in the first words of the second paragraph: “Y así, prosiguiendo su historia, dice que . . . .” In retrospect, after reading the chapter, the reader could object in two ways to the historian's concern. First, Don Quijote's actions are not nearly as unbelievable as was stated: he merely sees the reality of the three peasant women but eventually rejects this evidence in favor of the convincing interpretation of his faithful squire. Second, the “atoms of truth” alluded to obviously do not include such details as the type and gender of the animals which the three women ride.24

     4) Chapter 24 opens with the translator's faithful transcription of Cide Hamete's marginal note about what Don Quijote claimed to have seen in the Cave of Montesinos (pp. 761-62) before continuing with the narration: “Y luego prosigue, diciendo . . . .” As discussed above, the historian's suggestion that Don Quijote is either telling the truth about what literally happened or is consciously telling a lie, is unacceptable in the face of evidence that the experience was a dream. Furthermore, the statement about what Don Quijote might have said about the episode on his death-bed dissolves into ambiguous hearsay (“se tiene por cierto que . . . dicen que se retrató della”) and is absurdly meaningless.25

     5) Cide Hamete Benengeli enters into Chapter 27 with the words “Juro como católico cristiano” (p. 788). The translator speculates that the Moor does this in order to underscore the truth of what he has to say about Maese Pedro. But the thoughtful reader will realize that such a tacit admission of the superiority of Christianity over the historian's own Moslem beliefs is incongruous. An equally —or more— plausible interpretation of these words is that the Moor is consciously facetious in employing them; after all, for Moslems, Christians are lying dogs, and to swear like a Christian is to invalidate the oath. But no matter what the oath “means,” the passage is both comic and absurd. The first words of the second paragraph are “Dice, pues, que . . . ,” and the narration is resumed.

     6) Cervantes begins Chapter 40 with lavish, though (by this point in the novel) heavily ironic, praise for Cide Hamete: “Pinta

     24 See Howard Mancing, “Dulcinea's Ass: A Note on Don Quijote, Part II, Chapter 10,” Hispanic Review, 40 (1972), 73-77.
     25 John J. Allen, “Cide Hamete's English Translators,” Hispanic Review, 35 (1967), 367.


los pensamientos, descubre las imaginaciones, responde a las tácitas, aclara las dudas, resuelve los argumentos; finalmente, los átomos del más curioso deseo manifiesta” (p. 878). Then he addresses a series of ecstatic apostrophes to the four major characters (for by now Cide Hamete Benengeli has acquired the status and personality of a literary character): “Oh autor celibérrimo! ¡Oh don Quijote dichoso! ¡Oh Dulcinea famosa! ¡Oh Sancho Panza gracioso! Todos juntos y cada uno de por sí viváis siglos infinitos, para gusto y general pasatiempo de los vivientes.” After this highly comic, mock-emotional outburst, Cervantes signals the return to a translation of the manuscript with the words “Dice, pues, la historia que . . . .”

     7) The first words of Chapter 44 represent a brilliantly paradoxical reductio ad absurdum in the process of ridiculing the narrative structure of Don Quijote: “Dicen que en el propio original desta historia se lee que llegando Cide Hamete a escribir este capítulo, no le tradujo su intérprete como él le había escrito” (p. 906). That Cervantes the editor, who theoretically has the manuscript before him as he writes, should have to rely on hearsay (“Dicen que”) to report that while composing this chapter Cide Hamete wrote that his translator mistranslated it, that Cervantes would not consult with the translator, presumably working with him all the while, in order to correct any possible errors, or that the editor and/or translator would not either affirm or deny the validity of Cide Hamete's complaint, all combine to destroy any shred of truthfulness, reliability, or historicity remaining in the presentation of the story. The remainder of this long paragraph consists of the editor's sympathetic recounting of the author's sad lament concerning the frustrating restrictions imposed by his material, which is also (as was the case in I, 28, 298) a defense against the inevitable criticism of his not following a rectilinear narrative account. The next paragraph opens with a standard formulaic return to the story: “Y luego prosigue la historia diciendo que . . . .”26

     26 The reaction of Diego Clemencín to this chapter opening provides a classic example of how literal-minded readers can completely fail to understand, let alone appreciate, this sort of ironic narrative device: “Todo esto del principio del capítulo es una algarabía que no se entiende. Porque ¿cómo podía leerse en el propio original de la historia que no lo había traducido fielmente su intérprete? Ni ¿qué tiene que ver esto con la queja que tuvo el moro de sí mismo por haber tomado entre manos asunto tan seco y estéril? . . . Resulta de todo, que pudiera muy bien haberse excusado este largo y difuso preámbulo hasta donde vuleve a tomarse hilo de la narración.” See his edition of Don Quijote (Madrid: Castilla, 1966), pp. 1765-66, n. 1.

1 (1981) The Metafictional Dialectic of Don Quijote 79

     8) Chapter 53 begins with a sententious observation by Cide Hamete on the inevitable march of time, the seriousness of which is undercut when the historian reverses the order of the seasons of the year: “la primavera sigue al verano, el verano al estío, el estío al otoño, y el otoño al invierno, y el invierno a la primavera” (p. 984). Although editors of the novel often cite Covarrubias as proof that seguir can sometimes be used in the sense of perseguir ‘to search for,’ it is obvious from the beginning of the next sentence, “Esto dice Cide Hamete, filósofo mahomético,” that seguir is to be used in its standard meaning of ‘to follow.’ That Cide Hamete Benengeli is now a Mohammedan philosopher, and that he has everything backwards, can be considered a synthesis of his role in Part II of Don Quijote. After his commentary on the author's statement, Cervantes, for the first and only time, simply returns to the narration of events without using “dice la historia.”27

     Very similar in effect to these chapter openings are the titles, presumably written by Cide Hamete Benengeli, of many chapters. Whereas in Part I chapter titles are generally straightforward and accurately descriptive,28 in Part II there are several which are untrue or, at least, misleading. In the title of II, 6, it is stated that it is “uno de los importantes capítulos de toda la historia” (p. 617), while the chapter, though interesting, is not important by any standards. In the title of II, 23, Cide Hamete anticipates his subsequent comment on Don Quijote's version of what happened in the Cave of Montesinos by stating that the “imposibilidad y grandeza” of the knight's story necessitates that it be considered apocryphal (p. 750). (The word apocryphal here is ambiguous at best, suggesting perhaps the existence of extra-textual conflicting material.) In II, 34, the Moor's

     27 I have not included in this list the beginning of II, 50: “Dice Cide Hamete, puntualísimo escudriñador de los átomos desta verdadera historia, que . . .” (p. 958). The praise in the editor's brief remark is humorous enough, for any reference to the “atoms” of the story is by this time bound to evoke at least an indulgent smile on the reader's lips, and the formulaic reference to the text is present, but the extreme brevity of the passage sets it apart from the others discussed together.
     28 The only real problem with chapter titles in Part I concerns those which seem to be misplaced. The title of I, 10 refers to the adventure with the Biscayan, which took place in the previous chapter, and to the encounter with the Yanguesans, which occupies I, 15. The titles of I, 29 and I, 30 are reversed. The battle Don Quijote has with the wineskins in I, 35 is announced in the title of the following chapter.


description in the title of Merlin's prophesy as “una de las aventuras más famosas deste libro” (p. 843) is at least a gross exaggeration. His statement in the title of II, 64 that Don Quijote's defeat in Barcelona is the adventure “que más pesadumbre dio a don Quijote” (p. 1076) is, as observed above, not consistent with the text. The historian reveals himself as not very perceptive of the differences between the two parts of the work when he uses the word “aventura” (a problematic term in Part II) almost twice as often as in chapter titles in the second part (25 times) as in the titles of the first part (13 times). It is also worth noting that it is only in the second part that Cide Hamete includes several chapter titles that are completely frivolous:

     II, 9, 638 Donde se cuenta lo que en él se verá.
II, 28, 795 De las cosas que dice Benengeli que las sabrá quien le leyere, si las lee con atención.29
II, 31, 812 Que trata de muchas y grandes cosas.
II, 40, 878 De las cosas que atañen y tocan a esta aventura y a esta memorable historia.
II, 54, 990 Que trata de cosas tocantes a esta historia, y no a otra alguna.
II, 66, 1086 Que trata de lo que verá el que lo leyere, o lo oirá el que lo escuchare.
II, 70, 1108 Que sigue al de sesenta y nueve, y trata de cosas no escusadas para la claridad desta historia.

     Allen has clearly perceived that it is Cide Hamete Benengeli's “perspective on Don Quixote that is unreliable, because it does not change as the character changes, and so Cervantes contrives to alienate the reader from him.”30 I would go further and suggest that in the second part of the novel Cervantes, in his role as editor, establishes a clear opposition between himself (and the reader) on the one hand, and the author, Cide Hamete Benengeli, on the other. Cervantes, who most directly influences the reader's perceptions of all the characters (including Cide Hamete), implicitly stands relatively close to his protagonist and draws attention to the insensitive, unperceptive, lying author. Directly and indirectly, subtly and openly, Cervantes mocks Cide Hamete and makes him an object of the

     29 The phrase “dice Benengeli que” is obviously the editor's addition to the historian's chapter title.
     30 Hero or Fool, II, 5.

1 (1981) The Metafictional Dialectic of Don Quijote 81

reader's laughter; in fact, Cide Hamete Benengeli emerges as the most consistent comic character in Part II of the novel.
     The reader laughed heartily at Don Quijote throughout the first half of Part I when the knight's chivalric madness was at its height. As Don Quijote reached his sad accommodation with reality later in Part I, this laughter decreased. A reluctant and ineffective Don Quijote inspires more pity than laughter. In order to maintain a humorous tone and avoid his work's establishing a dominant note of pathos, Cervantes makes Cide Hamete turn on the protagonist and reveal his own comic shortcomings.
     The reader of Part II of Don Quijote continues to laugh at a supremely comic work of art. What is generally not realized is that the butt of this laughter is now less the comic antics of the mad knight-errant, and more the absurdities of the historian / author and the narrative structure of the work. The main role of Cide Hamete Benengeli, “filósofo mahomético,” in Part II is to replace —or at least rival— Don Quijote as an object of laughter.31


     31 This essay, in slightly different form, makes up a section of my forthcoming book, The Chivalric World of Don Quijote, to be published early in 1982 by the University of Missouri Press. I would like to acknowledge the Research Council of the University of Missouri-Columbia for financial support in the preparation of the manuscript and the editors of the press for permission to publish this excerpt.

Fred Jehle Publications of the CSA HCervantes