Selecciona una palabra y presiona la tecla d para obtener su definición.


Cervantes: Bulletin of the Cervantes Society of America

Volume III, Number 1, Spring 1983







Vice President




Executive Council

Cervantes: Bulletin of the Cervantes Society of America
Assistant to the Editor: THOMAS A. LATHROP

Editor's Advisory Council

Associate Editors

Cervantes, official organ of the Cervantes Society of America, publishes scholarly articles in English and Spanish on Cervantes' life and works, reviews, and notes of interest to cervantistas. Twice yearly. Subscription to Cervantes is a part of membership in the Cervantes Society of America, which also publishes a Newsletter. $15.00 a year for individuals and institutions, $25.00 a year for husband and wife, and $8.00 for students. Membership is open to all persons interested in Cervantes. For membership and subscription, send check in dollars to Professor HOWARD MANCING, Secretary-Treasurer, The Cervantes Society of America, Department of Romance Languages, University of Missouri, Columbia, Missouri 65211. Manuscripts (submited in accordance with Cervantes, 2 [1982], 107 and books for review should be sent to Professor JOHN J. ALLEN, Editor, Cervantes, Department of Spanish and Italian, University of Kentucky, Lexington, Kentucky 40506.

Copyright © 1983 by the Cervantes Society of America



ArribaAbajo On Editing Don Quixote1

Daniel Eisenberg

Florida State University

Una sola edición de Don Quixote no puede satisfacer las diversas necesidades de los lectores norteamericanos del libro inmortal. Se distinguen tres clases de lectores, y se señalan las características de las respectivas ediciones apropiadas para cada una de ellas. Se destaca la importancia del juicio del editor al enmendar como elemento clave para la calidad de la edición, frente a lo que parece ser una sobrevalorización de la interferencia compositorial en cuanto factor decisivo para futuros criterios editoriales. Los compositores de Juan de la Cuesta no merecen el desprecio que últimamente han recibido. Cuesta mismo es el autor de una obra sobre la enseñanza de la ortografía.

Quisiera yo, que los tales censuradores fueran más misericordiosos, y menos escrupulosos, sin atenerse a los átomos del sol claríssimo de la obra de que murmuran [...] Es grandíssimo el riesgo, a que se pone, el que imprime un libro, siendo de toda impossibilidad impossible, componerle tal, que satisfaga, y contente a todos los que le leyeren.

Don Quixote, II, 3                

In this paper I propose to examine our current needs for editions of Don Quixote, comment briefly on the extent to existing editions satisfy these needs, and discuss what direction work in this area should take.

I begin with the following premises. First, textual science -to call it criticism is a bit misleading- is, like linguistics, international in its fundamentals, however national its application. While the myriad small decisions in editing -treatment of spelling details, accidentals, emendations based on knowledge of the language or of the author- are unique to each language's literature, the basic procedures and the major steps in the editorial process are universal. Since at present the leaders in textual scholarship are those working   —4→   on English and American literature, I am going to draw on them later for perspective and to make suggestions about editing.

My second premise is that a bad edition is worse than none, and our choices should be to edit rigorously, or not to edit. A bad edition is like a false indulgence: not only does it take the purchaser's money under a false pretense, it gives him a false sense of security, and he will not take the truly necessary steps to meet his spiritual or editorial needs. A bad edition, in short, makes a good edition less likely and harder to publish.2

Finally, my remarks are addressed to Americans, or at most to all those whose native language is English. I do not feel competent to assess the requirements for Quixote texts in another culture, and I think examining our own needs is a valid task for the Cervantes Society of America. This approach, I believe, clarifies some points.

Having set forth these premises, I would like to examine some myths about Quixote editions.

The first of these is that the establishment of the text of Don Quixote is exceptionally complicated and difficult, and perhaps there is another myth behind this one, namely, that the text of Don Quixote is quite corrupt. In fact it is not particularly corrupt, and is difficult to edit only because it is long, and because Cervantes' ideas and language are sophisticated. Compared with the problems of many other authors, Spanish as well as foreign, Don Quixote is straightforward. Shakespeare, for example, is a nightmare.3 Shakespeare scholars still do not know whether Hamlet should say «Oh, that this too, too, solid flesh should melt», or «too, too sullied flesh». We have little of this with Cervantes. To find the true texts of Calderón one must penetrate   —5→   a maze of fakes, forgeries, authorial revisions, sueltas, and last but not least, Mr. Vera Tassis.4 Alfonso el Sabio required an analysis of the variants of different manuscripts,5 and Lazarillo, whose first two editions are lost,6 has more recently required the same procedure of its editions,7 even to decide which text to take as copytext. In the case of Lorca's Poeta en Nueva York, which I have investigated in detail,8 an essential manuscript is missing, and without it we can only do a makeshift edition with much less authority.9 And of Celestina,10 which exists in dramatically different forms, the validity11 and authorship12 of which are currently intensely debated, más vale no hablar.

In the case of Don Quixote all the potentially relevant forms of the text have already been collected. Although new manuscripts, like that   —6→   containing works of Diego de San Pedro and Juan de Flores,13 do come to light, in the case of Cervantes this possibility is remote. There is already a well-founded consensus about the copytexts, the Cuesta principes, as we know, thanks to much work already done, that save for the well-known passages in the second Cuesta edition of Part I (which I will call «the second edition»), the first editions are closer than any others to what Cervantes wrote. The number of extant copies of these first editions -Part I and Part II- is blessedly limited.14 I had hoped to be able to give you an exact figure, but I have found that determining the number of extant copies, especially   —7→   of the confusing Part I, is a little project in itself. Robert Flores says, without offering any explanation, that «all» the copies of Part I share a certain characteristic which I will refer to later;15 I wish he had shared with us how many copies there are, and where they are located. Be this as it may, there are only around 15 of Part I, less of Part II, compared with about 100 of Shakespeare's first folio, every one of which has been examined by a textual scholar. We have at our disposal a great amount of help in the form of vocabularies, concordances, indexes,16 and scholarship about Cervantes' language, ideas, knowledge, and so on. With such a state of affairs, we may well wonder why a definitive edition of the most famous author of the Spanish language has never been published.

One obvious reason, which brings me to a second myth, is that such an edition would not necessarily be attractive to a publisher. It would be slow as well as expensive to publish -accuracy has its price- and would not necessarily sell as well as many might imagine, for it would be competing with many cheaper editions, some with distinguished names attached to them. In the long run, of course, such an edition would surely be profitable, but the investment in publishing it, to say nothing of the costs of producing the edition in the first place, would be substantial. Most Spanish publishers already have an edition of Don Quixote in their catalogues, and any new edition, especially a much better one, is a threat to the investment in those copies.

I do not mean to imply that publishers are petty, and that one or more would not jump at the chance to publish a Quixote edition which would eventually be widely accepted by scholars. I think some would publish it even at a loss, and subsidies for preparing and publishing such an edition should be available if the case for them is clearly and convincingly made. But this, plus the perception that existing editions are meeting needs adequately -they do sell copies- explains why the initiative has not come from the publishing industry.


I have been speaking about «an» edition, and deliberately so, but in fact this is the third and final myth I would like to comment on. Succinctly stated, this is the myth that one glorious edition of Don Quixote can meet all our needs. I doubt this is true for Spain; it is certainly not true for the United States. Our needs are too diverse, and I propose that future editors consider specialization.17

I have divided into three categories the potential American users and purchasers of a Quixote edition. First, there are STUDENTS and casual readers who want to read the work in Spanish, but whose linguistic skills and knowledge of Cervantes' culture are limited. These readers need a totally modernized edition of an accurate text, without textual notes. They do need abundant explanatory notes.18 They need to be told, looking at the first chapter for examples, what salpicón and duelos y quebrantos were, what tantum pellis et ossa fuit means, that Bucephalus and Babieca were the horses of Alexander the Great and the Cid respectively, even who Roland was, and so on. They do not need to know anything about the Caballero de la Ardiente Espada other than that he was the protagonist of a romance of chivalry by Feliciano de Silva, and no more about Silva than that he was a prolific author of such works.

Users of this edition would not need extensive references to scholarship. They do need an introductory essay placing the work in its literary context, and pointing out its major themes. This introduction might well include a section on «How to Read Don Quijote», calling attention, among other things, to the most important sections, an important piece of guidance for the many who are not going to read the work in its entirety. It should also include a discussion of Cervantes' language, and the edition could very usefully be graced with a vocabulary.


It would be a great convenience if the lines of this edition were numbered. And finally, two hidden sources of the popularity of Riquer's Juventud edition should be adopted: an index, and descriptive running heads.

No such edition, of course, exists. In fact, while Spanish-language editions of common text works have been published for American or British university use, there has never been such an edition of any but fragments of Don Quixote. Such an edition would surely sell very well.

A second category of readers, with different needs consists of SCHOLARS, whether they be professors or lay Hispanists. Here some of our needs are met by existing editions.

Such readers do not need to be told who Babieca, Bucephalus, or Roland were. They do need to be told that the opening sentence of the novel has been the subject of much scholarship, that in naming the University of Sigüenza Cervantes was being funny, and the overtones of the name Aldonça Lorenço.

How should the accidentals (spelling and punctuation) of such an edition be handled? Although each textual feature must be considered separately, as general principles I suggest the conservation of the significant accidental features of the text, and a bias against modernization, on which topic the English and American scholars are quite definite.19 I believe this especially important because Cervantes phonetics (and those of his characters) are all but unstudied,20 and we are not even sure of what modernization costs us.


It is, however, difficult to reach any consensus about what the «significant» accidental features of the text are, and such evaluation may be ultimately subjective. We can agree, however, that u / v can not possibly have any phonetic or etymological significance, and can be modernized for the reader's comfort; I see no purpose that is served by making a reader, even a scholar, read «tuuo» and «vuo». The same is true of i / j and i / y, and of imagen/ s. As Amado Alonso said of u / v, these are «dos dibujos de una sola letra».21

Similarly, I can not see how anything authorial is distorted by using modern accentuation, or through the use of the dieresis (¨) over «u». These are transparent elaborations on, rather than distortions of, the text's orthography. In Cervantes' autographs these signs are not found at all (Romera-Navarro, p. 22); the accents of Cuesta's texts are phonetic, even if not according to the current system. It is true that Golden Age word stress sometimes differed from that of the present, in which case, of course, Golden Age practice should be followed,22 and the accentuation of invented proper names (Alfefliquén, Trifaldí23 is sometimes no more than guesswork, but I believe that the editor is the person to deal with these questions, not every reader. The use of an accent on interrogatives is less intrusive than the inverted question mark, favored by James Crosby.24

There are two further modernizations which are less sound theoretically, but which I believe desirable: the division into paragraphs, and the silent resolution of abbreviations. An edition with no paragraphs will simply not be used; I have had the task of reading Golden Age novels without paragraphing, and feel that the lack of paragraph divisions was much an annoyance, and my comprehension increased insignificantly. The abbreviations were the textual feature most freely altered by compositors; their meaning is rarely ambiguous, and   —11→   this point is best addressed by the editor. The more acceptable of the other alternatives would be conservation of the abbreviations (tãbien), not their irritating resolution with italics (tanbien), which call attention to something of little importance.

Beyond this, however, given our present state of knowledge, I can see no justification for modernization. That the accidentals have been altered by the compositors is not, in my judgment, a justification for their modernization, nor is their irregularity grounds for regularization. They still reflect something of Cervantes' practice, even if it is not always clear just what, and the alteration was executed by Golden Age workers, with Cervantes' manuscript before them. This would make scholars read Aldonça and Pança, yet I do not view that as a great hardship; it also permits them to see that Sancho said vaziyelmo, not baciyelmo.

I would like to speak out, for the same reasons, against modernization of capitalization and punctuation. It was Golden Age usage to capitalize important words, such as Cavallero Andante, and I believe the distinction between that term and the uncapitalized cabrero worth conserving.

Golden Age punctuation has been mostly ignored. The punctuation of such authors as Milton25 and Shakespeare26 has, however, been shown to have important implications. «It is indeed a revelation to read a familiar play [of Shakespeare] for the first time in a Quarto or Folio text», summarizes G. B. Harrison. «The reader finds himself at once in the atmosphere of the Globe... The differences [in punctuation] are slight but subtle. It is just the difference between a piece of music as played by an amateur and by a master».27

James Crosby has examined the punctuation of Quevedo's Política de Dios, and found that it is based «en el ritmo retórico y en la intensidad o el término de las pausas,» rather than «en la lógica y en la sintaxis» (pp. 20-21). What Crosby has observed in Quevedo's text I have observed in Cervantes'. The punctuation is not haphazard or capricious, merely different, and in the case of a linguistically sophisticated   —12→   author, master of prose style, it is a considerable deformation to replace it with the system of a different age.

As this is the most innovative proposal I am making, I would like to look at some examples of the punctuation of the Cuesta texts. The punctuation of Don Quixote's challenge to the leonero reads in the original:

Leoncitos a mí, a mí leoncitos? y a tales horas?

(II, 17; fol. 60v).                

This becomes, in modern editions:

¿Leoncitos a mí? ¿A mí leoncitos, y a tales horas?

(II, 160 of Murillo's edition28).                

thus altering the value of the pauses. It is clear, incidentally, that it was permissible to use more than one question mark in a sentence, as in the following passage from the Curioso impertinente:

Y porque claro lo veas, dime Anselmo, tú no me has dicho que tengo de solicitar a una retirada? persuadir a una honesta? ofrecer a una desinteressada? servir a una prudente?

(I, 33; fol. 186v).                

and an example from Sancho:

Por ventura el que ayer mantearon, era otro que el hijo de mi padre? y las alforjas que oy me faltan con todas mis alhajas, son de otro, que del mismo?

(I, 18; fol. 78v).                

Here is an example with exclamation points, from a comment of Diego de Miranda:

Ta ta, dixo a esta sazón entre sí el Hidalgo,

(II, 17; fol. 60v)                

and the same thing from Pero Pérez:

Ta, ta, dixo el cura, Jayanes ay en la dança:

(I, 5; fol. 18 -misnumbered 15-r).                

All the recent editors supply exclamation points to these: «!Ta, ta!», and Rodríguez Marín makes the first of them «¡Ta! ¡Ta!», overemphasizing the strength of the words, which, in our first example at least,   —13→   are «entre sí». Adding exclamation points -and all the modern editions have more exclamation points than do Cuesta's- takes away from the force of those that are found, and they are found at points well worth emphasizing. An example is when Sancho says, «con voz admirativa y grande»:

Santa María, y valme, éste no es Tomé Cecial mi vezino, y mi compadre!

(II, 14; fol. 52r)                

and another is when he says:

A bodas de Camacho, y abundancia de la casa de don Diego, y quántas vezes os tengo de echar menos!

(II, 23; fol. 93r).                

There is no way that modern punctuation can conserve the effect of that of the following passage:

Cómo, y es possible, que ay oy Cavalleros Andantes en el mundo? y que ay historias impressas de verdaderas Cavallerías!

(II, 16; fol. 56r).                

or of this one:

Pero dezidme señores, si avéis mirado en ello? Quán menos son los premiados por la guerra, que los que han perecido en ella?

(I, 38; fol. 227v).                

Modernization of punctuation also eliminates the ambiguous structures of Golden Age prose, which can drift from statement to question or exclamation, or vice-versa, in the same sentence, and from narration to direct address in the same paragraph. I do not have a good example of the former from the Quixote, but here is one from the Espejo de príncipes, with the original punctuation:

O quánto era el contentamiento y gozo que el cavallero del Febo sintió con aquellas palabras assí por saber quién fuesse y que era Donzella humana en quien pudiesse poner sus pensamientos, como por ver el gran favor que le avía dado en dezir que su fama le uviesse traído allí.29

This sentence clearly begins as an exclamation, but by the end it isn't, and there is no logical break for a closing exclamation point. (I solved it, imperfectly, by adding an exclamation point, followed by comma, after «palabras».)


Though the passages are too long to read, Don Quixote drifts from narration to direct address in the speech of the canon which ends Chapter 47 of Part I, and the change marked in all recent editions, which use a sentence break, is artificial and, I believe, belated. There are similar switches in the first speech of Don Quixote in II, 22, and in the speech of the primo in the same chapter.

No modern edition conserves the effect of the very logically placed colons30 (which I have replaced with semicolons) in the following passages:

Paréceme señor hidalgo, que la plática de vuestra merced se ha encaminado a querer darme a entender, que no ha avido cavalleros andantes en el mundo, y que todos los libros de cavallerías son falsos, mentirosos, dañadores, e inútiles para la república; y que yo he hecho mal en leerlos, y peor en creerlos, y más mal en imitarlos, aviéndome puesto a seguir la duríssima professión de la cavallería andante, que ellos enseñan, negándome, que no ha avido en el mundo Amadises, ni de Gaula, ni de Grecia, ni todos los otros cavalleros de que las escrituras están llenas?

(I, 49; fol. 298r).                

Respondenles hía yo, que tanto la mentira es mejor, quanto más parece verdadera; y tanto más agrada, quanto tiene más de lo dudoso, y possible

(I, 47; fol. 289r).                

Pues yo te asseguro, dixo don Quixote, que ahechado por sus manos hizo pan candeal, sin duda alguna; pero passa adelante

(I, 31; fol. 172r).                

For an author with such control of the rhythms of his prose as Cervantes had, such modernization is not only unnecessary, but harmful.31 Anyone who can read the Cid in old Spanish can read Don Quixote in Golden Age dress. It may take a little bit longer, but this is not a major defect, and may be a small virtue; those who must read quickly can read a modernized version.


As already stated, these scholarly readers need annotations explaining the more difficult words and allusions, and references to the points which have been discussed by scholars and summaries of many of the discussions. They also need textual notes, so that they can determine where the editor has emended the text -and every modern editor emends the text somewhere, and rightly so. Editors, being human, make mistakes in emendation, and if other scholars are informed of their emendations, any curioso may verify for himself the wisdom of these changes. Textual notes are a fundamental feature of a modern scholarly edition.

I would like to give you an example of a well-known passage, the «pasaje más oscuro del Quixote», in which Cuesta's editions disagree and the modern editors do as well:

Es éste el mejor libro del mundo; aquí comen los cavalleros, y duermen, y mueren en sus camas, y hazen testamento antes de su muerte; con estas cosas, de que todos los demás libros deste género carecen. Con todo esso os digo, que merecía el que le compuso, pues no hizo tantas necedades de industria, que le echaran a galeras, por todos los días de su vida

(I, 6; fol. 21r).                

Cuesta's second edition reads «con otras cosas», and «el que lo compuso». Schevill and Bonilla read «otras» and «le», in Rodríguez Marín's text we find «otras» and «lo», those of Allen, Riquer, and Avalle-Arce have «estas» and «le», the readings of the princeps, and Murillo reads «estas» and «lo», so we have all possibilities represented. I will have more to say about this passage and what I believe the correct readings to be later; the point I would like to make now is that no edition since that of Schevill and Bonilla tells us that Cuesta's editions disagree, and that some scholars consider the readings of the second edition preferable; neither Rodríguez Marín nor Murillo tells us he has emended the text. And this is the most difficult passage of the work, about which 15 articles have been written.32

There has only been one edition of Don Quixote, that of Schevill and Bonilla, which had textual notes and which attempted to preserve the orthography of the Cuesta texts; this edition is now out of print. There has never been an edition which does not modernize   —16→   punctuation. We do have several well-annotated editions available for purchase. The annotations of Rodríguez Marín's «nueva edición crítica» retain their value, and they are accompanied by an index. Of the new, less expensive editions, the most extensively annotated is that of Murillo.

Finally, we most certainly need, though in my opinion less urgently than the preceding editions, a DEFINITIVE edition of Cervantes' text, as definitive as the state of the art can make it,33 an edition for cervantistas, to be consulted rather than read, an edition which no one will buy who does not already own at least one, and perhaps several, other editions. This edition would eventually be the source for the text of editions of the types outlined above, and would be used for linguistic study.

This edition should have no systematic modernization whatsoever, and not a letter would be changed without a record of that change being included in the edition. In fact, the most important feature of this edition would be its textual apparatus; it should include, among other items, the variants of all of Cuesta's editions. Since it would not be intended for reading, and explanatory annotations would be available elsewhere, explanatory notes should be dispensed with. This would make it a more lasting, as well as less expensive, edition.

The first step in the production of such an edition is to secure funding for its preparation. It can not, and should not, be prepared on a volunteer basis, and even at a very high price, contrary to the purpose of the edition, future royalties could not pay the costs. While some scholars would surely be willing to do some of the work without compensation, motivated by their interest in a definitive edition of Don Quixote, it would not be a wise use of the time of these highly-trained people; much of the work would be quite tedious, and could be done by persons without advanced training in Spanish literature.

The second step in producing such an edition is collation of the various copies of Cuesta's principes, in order to locate variants introduced during the printing process as the result of simultaneous   —17→   proofreading, and from broken type. The evidence that makes this step a requirement is overwhelming. There is no reason to suppose that Spanish printers worked differently than those of other countries -printing was an international trade- and the practice of English printers is well documented.34 The fact that different copies of Herrera's Anotaciones are not identical is stated in the book itself;35 Rivers collated six copies of Garcilaso's princeps, and found no two the same,36 nor are any two of twelve copies of Obras en verso del Homero español, studied by E. M. Wilson, the same;37 previous incomplete collation of Don Quixote has been most worthwhile.38

How many copies need be collated? That is a difficult question, and the authorities will not give one a definite answer.39 There is a   —18→   consensus, though, that five or six is a minimum,40 and I would suggest that if six copies needed to be collated for Garcilaso, and each of these was different, no less than six should be collated for Don Quixote. Some textual scholars from outside Hispanic studies that I have consulted, such as John Andrews of the Folger Shakespeare Library, say that if there are only fifteen, we ought to collate them all and have done with it.41

Once such collation has been done a modest but very attractive project is easily realized, one which I recommend to the Society: the publication of a more perfect facsimile. The problem with existing facsimiles is that they only reproduce one copy (most do not even tell us which), and since the copies differ, no single copy of Cuesta's editions represents even Cuesta's intentions. On the basis of the evidence collected from collation it is possible to choose, from different copies, the best state of each page, and by assembling these one can prepare a «reproduction» of Cuesta's editions as he himself would have produced them, were he not constrained by limitations of time and money. This has been done in the case of Shakespeare, producing a definitive facsimile of the first folio, the Norton facsimile, and the procedure used is described in a lecture by Charlton Hinman (cited in note 34). I suggest to the Society, then, as a first step in producing a definitive edition of Don Quixote, the preparation and publication of a definitive facsimile.42

Such a useful tool would not, of course, be a substitute for an edition, since Cuesta's editions, like all those of the period, have   —19→   numerous errors of different sorts. Here we have the heart of the editing process: the application of the editor's judgment and knowledge to the data assembled.

I will have some examples of this process shortly, but I would like, first, to call your attention to the Center for Scholarly Editions, sponsored by the Modern Language Association. This center, successor to the earlier Center for Editions of American Authors, has two main functions: the dispensing of advice, and the conferring of its emblem, «An Approved Text», to deserving editions. The Center has no prescribed rules to follow, although every edition must include «a statement or essay describing the history of the texts... and defending the rationale for any emendation of [the] copy-text», and a report of variant readings. The Center wishes editors to «explore the literature on editing, become acquainted with the issues, follow the arguments, and sort out what makes sense, as a scholar would do in approaching any other subject»; however, it does believe that «punctuation, spelling, and capitalization have an important bearing on the meaning and implications of written and printed texts; it believes that an author's practice in these matters is of interest in its own right and worth trying to establish». The Center also will insist on rigorous proofreading.43

The emblem «An Approved Text», whose commercial significance will be considerable, is a reasonable goal for any editor who wishes to produce a definitive edition; I recommend to the Cervantes Society editors that they seek it.

At this point I think I should make some remarks on the topic of compositorial analysis and its relation to Cervantes' accidentals. I am going to pause to make these remarks because some of us, including myself, have felt intimidated by the studies that have been published on this topic. Compositorial analysis is only one of the tools available to the editor. It is not, and was never intended to be, a substitute for good judgment. It was developed as a means of handling substantive -i.e., major- variants, not accidentals.


Flores' published goal is to identify the characteristics of spelling and punctuation of Cuesta's compositors, subtract these from his texts, and thereby to arrive at Cervantes' manuscript, or as Flores puts it, to recover Cervantes' orthography («The Need for a Scholarly, Modernized Edition of Cervantes' Works», Cervantes, 2 [1982]. 69-87, at p. 78; see also Compositors, pp. 87-89). To do such completely is impossible. It is theoretically impossible. Such has never been done for any author, and it certainly can not be done with Cervantes.

I am going to give you an example to illustrate why it is impossible. It is an example which, for simplicity, I am making up, but it is faithful in spirit to the problems involved. Let us assume that the number of commas found in the first edition of Part II varies, and let us further assume that it varies consistently by forms, so that we can identify two of the four compositors who worked on the edition44 as ones with higher proportions of punctuation than the others. The question then is, did two of them add commas, or did the other two subtract them?

This is a question which can be answered, at least in theory. It is answered by a basic procedure for this type of study: analyzing the work of the same compositors in setting another text, one for which there is a control, a text which they worked from and which can be examined. Although it would be worth checking, I doubt that any MS the Cuesta compositors used survives, since dead manuscripts were regularly recycled, but many of his publications were set from printed sources. In 1616, for example, Cuesta published an edition, of Josephus' Jewish War, a lengthy text on which at least some, and perhaps all, of his 1615 compositors worked. Josephus' text was certainly set from one of the two previous editions of that translation, and it would not be a difficult task to determine which. By doing a compositorial analysis of Cuesta's edition of Josephus one could, having identified the compositors and matched them with the compositors of Don Quixote, Part II, compare the work of the latter with a control text, and determine how each of them handled commas.

Let us further assume that this enormous piece of work has been done, and we have verified that the two compositors with many commas in Cervantes' text added commas to the text of Josephus; it   —21→   would then be a safe conclusion that they added commas to the text of Cervantes, and typesetters certainly did this sort of thing. So now we have confirmed that there are indeed too many commas in a portion of Don Quixote. To remove the excessive punctuation, however, would require us to identify which commas are the work of the compositors, and which are Cervantine. There is no way to do this. By analyzing the syntactic structures of the text -this would have to be a manual analysis, incidentally; there is no program to do it on a computer- one might be able to determine that with certain syntactical structures the one set of compositors consistently added punctuation. But Spanish syntactical structures, especially of this period, are too diverse for this to be done with every type of structure.

Let us make the same point another way. Compositorial analysis is a statistical method, and statistics depend on quantities of examples. It's one thing to take a word which appears fifty times in the text; it's quite a different thing with a word that occurs only five times. The ultimate case is the hapax legomenon, the word that an author uses only once, and with such a word compositorial analysis breaks down completely.

Flores' own published data demonstrate that one cannot generalize from how the compositors handled frequent words to how they handled less frequent words. The compositors themselves are not consistent. The same compositor (C) who spelled «barbero» with a «b» spelled «vozes» with a «v» and «vazia» with a «b», and compositor F spelled «vozes» with a «v» and «viuda» with a «b». The same one (E) who does not capitalize «general», «barbero», or «enano» does capitalize «gigante» and «ciudad». When we can not generalize about the treatment of the frequent words, we certainly can't about the infrequent words.

There is a further difficulty with the laudable goal of recovering Cervantes' orthography, and that is the assumption that Cervantes had a consistent orthography to be recovered, which assumption is basic to Flores' method as well as to his goals. Flores is very definite in speaking of «Cervantes' orthography». «Cervantes, who was writing at greater leisure [than the compositors]» -he tells us- «was not likely to fluctuate aimlessly back and forth from one spelling to the other» (Compositors, p. 88). «It is unlikely that Cervantes could equally well have written trahia, traía and traya, aora, agora and ahora, or barbero, Barbero and baruero» (p. 88). The «different spellings,   —22→   punctuation patterns, and setting habits [of Cuesta's compositors] destroyed the uniformity of the original manuscript» (Compositors, p. 63; italics mine). (Earlier, however, Flores had admitted that «it was improbable that Cervantes's orthography would have kept constant throughout his work» [p. 17].)

Now, everything that we know about Cervantes tells us that he was not a man who placed a high value on order, conformity, and consistency. His life and the diversity of his writings speak eloquently of his experimentation, his interest in novelty, his concern for the ideas and the «big picture» rather than the details. The compositors' inconsistency itself suggests struggles with an inconsistent authorial manuscript. But we have better evidence than these inferences about his orthography. I refer to the documents he wrote in his own hand, ten of which were reproduced and analyzed by Miguel Romera-Navarro.45 Here is what Romera-Navarro tells us about Cervantes' own spelling, undistorted by any compositors: «De los seis escritores clásicos con cuya escritura estamos algo familiarizados -Cervantes, Lope de Vega, Tirso de Molina, Quevedo, Gracián y Calderón-, el menos uniforme en la forma de la letra, en el empleo de mayúsculas y minúsculas, en el uso de ciertas vocales y consonantes, en ligar o no ciertos vocablos, en las contracciones y abreviaturas, en la puntuación y en el aspecto general de su escritura, el menos uniforme, repito, el más irregular, aun dentro de una misma página, aun firmando su nombre mismo, es Cervantes» (p. 2). Romera-Navarro also comments on Cervantes' punctuation: «Los documentos no traen un solo caso de coma, de punto y coma, de dos puntos..., ni el acento, las diéresis o el guión en la division de una palabra al fin del renglón... Jamás aparecen el paréntesis, el subrayado, ni otro signo ortográfico auxiliar, excepto el punto, y éste rarísimamente» (p. 22).

I would like to look at only one example of Cervantes' orthography, the most frequent example, his signature.46 Sometimes he signs documents «Miguel de cerbantes» and sometimes «Cerbantes». And here is a further point: he spelled «Cerbantes» with a «b», and this is one of the few consistencies to be found, a strong spelling preference, since he rarely used, in the autographs, «b» for   —23→   «v» (Romera-Navarro, p. 16). Yet in his books «Cerbantes» is only found inside his books (the Adjunta al Parnaso, the preliminaries of La Galatea, the description of La Galatea in Don Quixote, I, 6), and on none of the title pages, which spell «Cervantes» with a «v» or «u». Which is to say, the topic was not important enough for him to tell any of his publishers to be sure to spell his name as he signed it.47

So I believe that the theory that Cuesta wreaked havoc on a carefully-spelled MS of Cervantes will not stand. I do not doubt that some further progress in recovering Cervantes' accidentals is possible, and we should welcome it. But I must also point out that it is extremely slow and time-consuming research, and even were we to concede the possibility of a complete recovery of Cervantes' orthography, it is not even in the foreseeable future. As an illustration, I believe a fair one, of the utility of compositorial analysis in questions of accidentals, I would point out that this method has given us several very important facts, but what it has told us about Cervantes' spelling is that he wrote some forms of the auxiliary verb haber without the initial «h», and that he wrote «Dulzinea» with a «z» (Compositors, pp. 88-89; compare Romera-Navarro, pp. 16-18).

These are details, just as whether he wrote «cautivo» or «captivo», whether or not he capitalized «cura», even -a much more important case- whether he wrote «vuesa merced» or «vuestra merced» are all details. I do not mean that they are unimportant details, or that we have any reason not to accept emendations with open arms. But they are details all the same, and we should keep them in perspective. They were not particularly important to the author, either, as Flores himself tells us, when he states that Cervantes would be «hard pressed to restore his accidentals to Cuesta's texts» (Compositors, p. 89).

Having gone this far, I'm going to echar el resto and say something in defense of Cuesta and his employees. Flores is hard on modern editors, but he is poison on Cuesta's compositors. He spells out in words for us their «three thousand four hundred and sixty-nine important variants» (Compositors, p. 85). This is a «staggering number»   —24→   (p. 85). The compositors had an «unpardonable devil-may-care attitude towards their copy» (p. 87). They are guilty of «remorseless distortion of the authorial orthography» («Need», p. 85). These compositors «took on their shoulders all the rights and none of the responsibilities attached to proofreaders and editors» (Compositors, p. 85).

These are strong words. Were Cuesta and his employees fools?48 Of course, they weren't. They are seen as deficient because twentieth-century standards are being applied to their seventeenth-century labors. No recent scholar has claimed that Cuesta's shop was any worse than those of his contemporaries; Rudolph Schevill says: «comparado con otras primeras ediciones, v. gr., las de El Buscón, Guzmán de Alfarache y El Peregrino en su Patria, no parece [la edición príncipe de Cuesta] digna de tanto desprecio» (I, 10 of his and Bonilla's edition). There is evidence to support the view that Cuesta's work was better than average; Astrana Marín called his shop «una de las mejores de Madrid» (Vida, V, 610), and Francisco de Robles, who chose Cuesta to print Don Quixote, was the librero del rey, as well as the most famous and successful book dealer of his day. In setting the second edition of Part I, from which Flores takes his figures on Cuesta's inaccuracy, he was, as Flores has also shown (and Astrana stated, Vida, V. 629), working under great time pressure. It says something, I think, that Cuesta would subcontract work to the Imprenta Real, and the compositor of that shop who worked on the second edition was neither more nor less careful than Cuesta's men. And Flores himself, in another of his contradictions, says «Cuesta should not be commended or condemned on account of Don Quixote» (Compositors, p. 41).

There is some further information relevant to our assessment of Cuesta, in a book which I found while preparing this paper. It is the following:

Libro y Tratado para enseñar leer y escriuir breuemente y con gran facilidad cõ reta pronunciacion y verdadera ortographia todo Romance Castellano, y de la distincion y diferencia que ay en las letras consonãtes de vna a otras en su sonido y pronunciacion. Compuesto por Iuan de la Cuesta [...] En Alcala en casa de Iuan Gracian questa en Gloria. Año. 1589.


This book is a treatise on how to teach reading and writing, written by a teacher (maestro) for the use of other teachers. It first presents, like a cartilla, syllables for pronunciation exercises, and explains how to progress from letters to syllables to words; besides a discussion of the pronunciation of each letter, it includes an extensive list of Latin abbreviations the novice reader might encounter, with the resolution of each. In the section on writing there is a long discussion of how to choose and prepare plumas, how to hold the pen in the hand, how to place the pen and hand on the paper, and how to form the various letters with regard to an imaginary line; a number of woodcuts illustrate these points. Other topics discussed at some length include the correct and incorrect way to divide words at the end of a line, the use of the hyphen, and the use of capital letters.

This book is not an unknown one; it is found in the standard reference works on printing, such as Palau y Dulcet, Gallardo's Ensayo, and Juan Catalina García's Ensayo de una tipografía complutense (Madrid, 1889). It was used by Rufino José Cuervo49 and Amado Alonso;50 it is found in the Conde de la Viñaza's Biblioteca histórica de la filología castellana, and the section on writing is discussed extensively by Emilio Cotarelo y Mori in his Diccionario biográfico y bibliográfico de calígrafos españoles, I (Madrid, 1914), 237-43, in which the title page and some of the writing examples are reproduced. It has, I understand, also been used by writers on the history of education, as it suggests that advanced students be used to help with the teaching of the beginners (the relevant passages are reproduced by Gallardo and Cotareli). Yet I have been unable to find it in any Cervantine bibliography or, save a brief reference by Rodríguez Marín in his appendix on «La x de Quixote» (IX, 26-27 of his edition), any mention of it in a Cervantine context at all.

The question immediately arises: is this the same Juan de la Cuesta, or are we dealing with a homonym? Even on the surface it would seem that this is the same person; Juan de la Cuesta was not a common name, and I have run across no other men so named. If a Juan de la Cuesta had written on navigation, or veterinary medicine, or taking confession, or if he lived in Seville or Bilbao, we might well   —26→   suspect that there were two men of the same name; to suppose that the author of a treatise on reading and writing, head of a school located (according to Cotarelo) in Alcalá de Henares, should come to be a Madrid printer, through stages (Segovia and Alcalá) on which I will shortly elaborate, is not at all implausible.

Furthermore, this Juan de la Cuesta was not just any educator; he was a famous maestro, yet the documents concerning Cuesta the printer make no attempt to distinguish him from a famous homonym.51 The book in question was written «por comissión de los Señores del Consejo Real», the aprobación tells us, and the author declares: «como es muy notonio mi pupilage ha sido siempre tan grande que en esta mi ante ha sido el más copioso de el reino y de gentes muy principales no solamente de esta comarca sino de la corte y de hijos de criados y officiales de su magestad muy principales y de todos los reinos de España» (fol. 62r; the passage is also found in Gallardo).

The author is also someone who has looked at books closely, with an eye for their typography. There are frequent references to «impressores», to «moldes», to different types of letters found in books. When we add to this that the book was published by the widow of Juan Gracián, who dealt with the same Robles family of booksellers that Juan de la Cuesta the printer was to deal with,52 and that no one, ancient or modern, informs us that Cuesta the educator   —27→   and author and Cuesta the printer were different people, the necessary conclusion is that they were the same person.

This interest in written communication found in the book makes Cuesta's change of occupation plausible; a change from the teaching to the publishing profession is certainly one many more recent teachers have made. The identification of these two Cuestas also serves well to explain why we have so little information about the activities of Cuesta the printer before his appearance in Madrid in 1599, when he entered the printing establishment of María Rodríguez,53 widow of the printers Pedro Madrigal and Juan Íñiguez de Lequenica. That his entrance into the world of printing should be as employee rather than proprietor is not surprising; a teacher of reading and writing, even a famous one, could scarcely have accumulated the funds to enter the capital-intensive trade of printing any other way.

Cuesta's book, according to the dates of the preliminaries, was written by 1584, some five years before its publication. He had entered the printing trade no later than 1588, for in that year we find his name on a work of Horozco y Covarruvias, the Tratado de la verdadera y falsa profecía, published in Segovia. His name is also found on 1589 and 1591 editions of Horozco's Emblemas morales. He was presumably called to Segovia to print these works; Segovia did not have, and was not to have for a long time to come, a regular press.54 That a printer would pack a press and a box of type on a wagon and travel Castile's dusty roads is well known.

I find it inconceivable that María Rodríguez would employ, as regente of her well-known shop and replacement for the also well-known Íñiguez, someone who had not printed a book in eight years. It is almost as unlikely that the expensive equipment used to print   —28→   the Segovia books was left idle during the same period.55 Therefore Cuesta must have been in the printing trade during these intervening years in which his name is not found, and he could not have been living in Segovia, where no books were published; he was living somewhere else, and working for someone else. There is an obvious place where he would have lived, Alcalá, and there is an obvious employer, the publisher of his book, Juan Gracián's widow, María Rodríguez56 (Astrana, Vida, VII, 760). Her husband had died in 1587 (see the entries for 1587 in Catalina Garcia's Ensayo, cited on p. 25). It is reasonable to suppose that she would have sought out someone of linguistic skill and administrative ability to help her run the press after her husband's death,57 yet insisted that the name of her dead husband, whom she seems to have revered, appear on her books, as it consistently does. It is also possible if Cuesta was working in the shop of Juan Gracián's widow, where his book was published, that he set his own manuscript, in which case we have something quite rare -a book typeset by its author- and well worth reproducing in facsimile.

The full investigation of this book, which should be compared with Cuesta's Segovia books and the work of Flores' typesetter F, now identified as Cuesta himself,58 and an assessment of its significance for editors of Cervantes is not possible in this paper. What is clear, however, is that Cuesta was not an ignorant fool, negligent in his handling of Cervantes' manuscript, but a thoughtful man who had some stature as an expert on writing, spelling, and other accidentals. In fact, the printer-publishers, in Golden Age Spain as elsewhere,   —29→   were men of some culture,59 not exactly the passive recipients of authors' initiative and slaves of public taste that they are often presumed to be. It is scarcely surprising that the people who bought books, and most of the authors, expected a printer to act as a sort of editor,60 and especially so in the case of accidentals. (Modern punctuation, like standardized spelling, is a product of printers, not authors, as any reader of medieval manuscripts will attest.61)

Cuesta himself (typesetter F), presumably the most experienced and the one who was setting the standard for his shop, was the compositor who was freest in emending Cervantes' text. If he and his employees made some spectacular goofs with the text of Don Quixote, like «yelmo de Mambrino» for Sancho's deformation «yelmo de Malino», it must be said in his defense that Cervantes was very sophisticated and atypical in his use of language, using popular speech, foreign words, playing with words' meanings, inventing words -not at all the sort of text compositors were used to handling. The typesetters ignored things that were unimportant, and changed, in general, the ones they were expected to change. Modern editors do the same. It is only the definitions of «unimportant» and «expected» that have changed.

The text as published by Cuesta, with the spelling and punctuation added or altered, was not a bad text, then, by the standards of   —30→   the time. If we can improve on it, in substantives or accidentals, then by all means let us do so. But if we cannot recover Cervantes' accidentals, and I believe that in more than a minor way we cannot, it really does not much matter. If we somehow were to find the MS from which Don Quixote was set, or could reconstruct it using statistical methods, we might well find that it was as poorly punctuated as Cervantes' autographs, which have only periods, and those «rarísimamente». Would we want to respect the punctuation of such a manuscript? We would want to conserve and publish its readings, to be sure, but not to follow it in preparing a critical edition.62 Cervantes would disapprove of us if we did. Lacking evidence to the contrary, Cuesta's texts are punctuated and, with exceptions, spelled as Cervantes wanted them to be. Note that while Cervantes did criticize the «impressores» of Part I for the ruzio error (in II, 4 and 27), he makes no comment on their handling of accidentals.

The recent overemphasis on compositorial analysis of Cuesta's texts would seem to reflect a desire, shared by many editors, to reduce the establishment of a text to a series of rules that can be mechanically applied, and a reluctance to trust one's judgment.63 Editors fear receiving the criticism their predecessors have received as a result of their errors in judgment.64 However, an editor has the obligation to use his or her judgment and correct when he or she has a reasonable confidence that the original text is in error. To allow over-respect   —31→   for the first edition to inhibit emendation is not good editing.65 Let me give you some examples, all of which are from Chapter 6 of Part I.

The first of these is «el que le compuso», which the second edition, and those of some later editors, make «el que lo compuso». I believe that this change was mistaken. «Le» as a masculine direct object pronoun referring to an object, while no longer in use, is well documented in Golden Age Spanish, and there are other examples in the text of Cervantes.66 Why was this correction made at all? Because this phenomenon was in flux in Golden Age Spanish, and le was a minority usage; the compositor felt he was improving the text.67

My second example is whether we should read «con estas» or «con otras cosas que todos los demás libros deste género carecen»; «estas» is found in the first edition and «otras» in the second.

Now, no one in the last fifty years would say that because «otras» is found in the second edition, it is automatically right. But I would suggest to you that just because it is found in the second edition, it is   —32→   not automatically wrong, either.68 The passage clearly makes much more sense with «otras»; «estas» makes no sense at all, and that is the key criterion. It has nothing to refer to. «Estas» can not refer to the features of the Tirant the priest has just mentioned; he is adding to them with the preposition «con». But if he says that the book has the defects he has just pointed out, and other things, too, he is speaking in a normal way and adding strength to his criticism of Tirant. So even if the compositors made gratuitous emendations, judgment leads me to say that in this instance one of them was right, and we should be appreciative.

My final example is a few lines further on, when the examiners discover the pastoral novels in Don Quixote's library, and the priest wants to save them from the bonfire because they are «libros de entendimiento, sin perjuizio de tercero». Here we have a possible emendation which all the scholarly editors mention in a note, but no one since Rodríguez Marín adopts: that the priest should say «libros de entretenimiento». What the compositors did or didn't do is irrelevant to this point, since this emendation was not even suggested until the eighteenth century. Yet, of course, it is not wrong just because it is not found in any of the early editions. Examination of Cervantes' use of the two words, and of his ideas about the function of literature, leads me to the conclusion that this emendation is correct, just as Allen, in my opinion, was correct when he made two other emendations found in none of the early editions, the reading «llovía» for «vía» in I, 4 («toda aquella tempestad de palos que sobre él llovía»), and the relocation of the ruzio passage to a more appropriate home in I, 25.

«Entendimiento» was not a quality that books had. People can entender, inanimate objects can not, and therefore people, not books, can have entendimiento. People's entendimiento is repeatedly mentioned in Don Quixote:

El cura le estuvo escuchando con grande atención, y parecióle hombre de buen entendimiento

(I, 47; I, 566 of Murillo's edition).                


Y si es que la imitación es lo principal que ha de tener la comedia, cómo es possible que satisfaga a ningún mediano entendimiento?

(I, 48; I, 570 of Murillo's edition).                

Y cómo es possible que aya entendimiento humano, que se dé a entender que ha avido en el mundo aquella infinidad de Amadises

(I, 49; I, 577 of Murillo's edition).                

Ésta sí sera letura digna del buen entendimiento de vuestra merced, señor don Quixote mío

(I, 49; I, 579 of Murillo's edition).                

Ni es razón, que un hombre como vuestra merced, tan honrado, y de tan buenas partes, y dotado de tan buen entendimiento, se dé a entender, que son verdaderas tantas, y tan estrañas locuras, como las que están escritas en los disparatados libros de cavallerías

(I, 49; I, 583 of Murillo's edition).                

... ha se de advertir, que no se escrive con las canas, sino con el entendimiento

(II, Prologue; II, 34 of Murillo's edition).                

... y assí en esta segunda parte no quiso ingerir novelas sueltas, ni pegadizas, sino algunos episodios que lo pareciessen [...] y pues se contiene, y cierra en los estrechos límites de la narración, teniendo habilidad, suficiencia, y entendimiento para tratar del universo todo

(II, 44; II, 366 of Murillo's edition).                

... para componer historias, y libros de qualquier suerte que sean, es menester un gran juizio, y un maduro entendimiento

(II, 3; II, 64 of Murillo's edition).                

Literature, however, could offer entretenimiento:

Hanse de casar las fábulas mentirosas con el entendimiento de los que las leyeren, escriviéndose de suerte, que facilitando los impossibles, allanando las grandezas, suspendiendo los ánimos, admiren, suspendan, alborocen, y entretengan

(I, 47; I, 565 of Murillo's edition).                

... desta manera se harían buenas comedias, y se conseguiría felicíssimamente lo que en ellas se pretende, assí el entretenimiento del pueblo, como la opinión de los ingenios de España

(I, 48; I, 572 of Murillo's edition).                

Un Viriato tuvo Lusitania, [...] cuya leción de sus valerosos hechos, puede entretener, enseñar, deleitar, y admirar a los más altos ingenios que los leyeren

(I, 49; I, 578-79 of Murillo's edition).                

... rebien aya el curioso, que tuvo cuidado de hazerlas traduzir de Arábigo en nuestro vulgar Castellano para universal entretenimiento de las gentes

(II, 3; II, 59 of Murillo's edition).                


la tal historia es del más gustoso, y menos perjudicial entretenimiento, que hasta agora se aya visto 69

(II, 3; II, 64 of Murillo's edition).                

Allen proposes (I, 123, n. 23 of his edition) that the priest intended to distinguish the pastoral novels from the romances of chivalry, and in this he is surely correct. Yet we can explain this distinction in terms of entretenimiento. While the priest surely shares Cervantes' view that literature should contain some provecho to benefit the reader, it was normal to seek in books not lessons, but entertainment. This is what the priest is looking for, and he has just finished explaining how, with a few exceptions, he hasn't found it in the romances of chivalry, for which he has «ojeriza» (I, 48; I, 566 of Murillo's edition) and «rencor» (I, 48; I, 569 of Murillo's edition). After cutting the romances of chivalry to shreds, here, in the pastoral novels, are works which can offer something that fulfills the purpose of fiction, that is entretenimiento, harmless, presumably beneficial entretenimiento. The other candidate for mouthpiece for Cervantes' literary views, the canónigo, makes the same criticism of the romances of chivalry.

So for these reasons, I believe that the correct reading is libros de entretenimiento. Even though none of Cuesta's editions have it, I am confident that Cervantes would approve of this emendation.

Having cited these examples of judgment applied to the process of editing, I'm going to close, and I will close in the same way many self-help meetings do, with the serenity prayer, which is very appropriate to the textual scholar:

God, grant me the serenity
to accept the things I cannot change,
the courage to change the things I can,
and the wisdom to know the difference.