Published in Manuscripts, 45 (1993), 13-21.

The Story of a Cervantine Discovery

Daniel Eisenberg

     From 1983 dates my first contact with a text with which I have subsequently spent much time. It is evidently a chapter—on this all agree—from an unsigned work of prose fiction. The fragment is the oldest item in a volume of unrelated manuscripts assembled and bound in the eighteenth century. Paper, spelling, and letter shapes indicate a manuscript from within Cervantes' lifetime, as all who have seen the manuscript agree. I believe the author of this text was Cervantes. The manuscript bears creases that indicate it had been folded as a booklet before being opened and bound into the present collection. On what would be the back cover of the booklet is a descriptive line, commonly taken as the title, "Dialogue between Selanio and Cilenia, on Country Life." Because of its location and because these words only describe a portion of the text, I have rejected it as a title and identified the chapter as being from Cervantes' lost Weeks in the Garden. Finally, I believe that the manuscript is in Cervantes' hand, thus his only fictional autograph ever discovered and his only literary autograph whose location is known. If authentic, it is the biggest Cervantine find since the discovery of the lost Numancia in the eighteenth century.
     This may seem like a daring or even reckless series of claims. However, each part of the above—that the fragment is by Cervantes, that it is a fragment of the Weeks in the Garden, and that it is an autograph—has been proposed previously, though by three different scholars. One reaction I expected, and which has occasionally been said, is that I didn't do anything new. So let me clarify just what my claim to fame is. First, I called attention to the text, which was completely forgotten by Cervantine scholars. It was so forgotten that in none of the modern discussions of Cervantes' lost works is this text even mentioned. Second, I have edited the text, and published with it a reproduction of the manuscript, so that anyone interested may study the hand, spelling, and punctuation, as well as check my editing. Most important, I have not just suggested that it might be authentic, but made the case for its attribution, putting my reputation at stake. Save a disreputable scholar who first published it in 1874, no one else has been willing to do this. One can see where previous editors, such as Schevill and Bonilla, simply set it aside. For if one argues that it is Cervantine, then it is a piece of a lost work, and Cervantes' only literary autograph to boot. Rather than argue this package, previous Cervantine scholars, all now deceased, have let it disappear from view.
     The beautiful text, which enchanted me from the first, consists of a dialogue between a man, Selanio, and a woman, Cilenia, who talk outdoors one summer afternoon. Selanio told Cilenia of the descent of Truth to this world in female form. A figure parallel to Christ, she was to lead people to God. Unable to tolerate the corruption of the city, where she was not wanted, Truth took refuge in the country. Cilenia set out to find her, and one who seeks Truth can always find her. After a mystical encounter and union with Truth, Cilenia now carries Truth, like Christ, within her. Thus she is a Truthful woman, and Selanio can worship her without apostasy. The text offers a fictional resolution of the conflict between human and divine love. Revealing the text to be a fragment are references to past and future conversations, and other characters not present. It has considerable parallels with Cervantes' Dialogue of the Dogs, one of his most accessible works after Don Quixote and a model for the psychoanalysis of Freud (who read Cervantes in Spanish).
     My first contact with this manuscript and text was not romantic in the least, unless one finds bibliographies romantic. Before writing my A Study of "Don Quixote" (Newark, Delaware: Juan de la Cuesta, 1987), I reread Cervantes' complete works, and then felt I should look at the attributed works. A list of works whose attribution to Cervantes had been proposed is found in a standard reference book, the Suma cervantina of E. C. Riley and Juan Bautista Avalle-Arce. In that list, I saw that many attributed items came from a single book, with the sensational title Various Unpublished Works of Cervantes (Madrid, 1874). Looking at this today-forgotten book, the authenticity of this text, and only this one, seized me. My task was to test—to undermine, if possible—that first impression of authenticity.
     The editor of Various Unpublished Works, Adolfo de Castro, was a scholar of no reputation because he had previously forged a Cervantine text and published it as authentic. This was The Squib (El buscapié), famous in its day and translated into several languages before the forgery was exposed. However, in contrast to The Squib whose manuscript he would never produce, in this case Castro named the library that owns it, and supplied the call number. The library, located in the Cathedral of Seville, is commonly called the Columbine for its founder, the bibliophile Ferdinand Columbus, illegitimate son of Christopher. It is a very logical library for a Cervantine manuscript to appear in, since Cervantes spent much time in Seville. The only other surviving prose manuscript of Cervantes, a non-autograph copy of the attributed Pretended Aunt, is found in that library. A now-lost Sevillian manuscript known as the Porras manuscript also contained Cervantine texts.
     I wrote the library, sent them the money they requested, and a microfilm of the 15 pages arrived in some weeks. I did not tell the library that I suspected they might have a fragment of a lost Cervantine work, assuming that such a suggestion might well delay the microfilm, perhaps indefinitely. Looking at a printout from the microfilm increased my anxiety, because the hand looked authentic. I had already worked somewhat with Cervantine autographs, including another forgery, in studying his phonetics. It dawned on me that I was the expert on Cervantes' hand. This was by default: no one else living had studied it.
     The reader may reflect on what he or she would do upon stumbling across a possibly sensational find, which, however, had been published by a famous forger. To announce the discovery would be to invite careful scrutiny at best, derision and hooting if one erred. The first step was of course to be sure myself. I worked two years with the text without telling anyone other than my fiancee (now wife), and my former professor David Kossoff. My wife said later that I seemed "possessed."
     My task was of course to convince myself. A study of Cervantes' hand would not have proved authorship, only that Cervantes copied the text. Even this was not possible; there was no suitable control. We possess only a single letter written as much as thirty years earlier, and some financial documents. Also, Cervantes' hand is quite unstable, varying considerably on a single page, according to the deceased author of the only study of it. My argument for the fragment being an autograph is based on the small insertions and stylistic revisions it contains, in the same hand: the fragment's author must have written out this copy. If Cervantes is the author, it is an Cervantine autograph. But was Cervantes the author?
     The lack of a concordance or electronic text of Cervantes' works made authentification using word frequencies impractical. My approach took two avenues: to find parallels between the fragment and the known works of Cervantes, and to rule out other potential authors. The fragment was very well written, and it seemed anything but the work of a beginner or an unknown. The sentences of the text are carefully structured, and it is rich in both ideology and vocabulary. Many lesser lights could not have written it.
     Sixteenth-century Spanish fiction, because it set the stage for Don Quixote, is a well-surveyed area, even if much work remains to be done. Fortunately, it was my first area of specialization, and I had published a book and a checklist on sixteenth-century chivalric fiction. I owned microfilms of many of the texts not available in modern editions, and others had been purchased by my university library at my request. I therefore set out to skim through fiction of the period, everything I could get my hands on, looking for another potential author. My fear was that I would announce the text as Cervantine, and two weeks later someone would reveal its "true" author.
     I did find three other authors with whose writings this text shared some features. However, we know Cervantes read and admired two of them. The third said in a prologue that he was imitating Cervantes. So this reinforced authenticity rather than undermining it.
     Another part of the job was to edit the text. My original intention was to do a paleographic edition. One of the hardest decisions was to abandon this plan: I could not distinguish between small and capital letters, between words divided and words linked. From the microfilm I could not confidently identify punctuation, dots on i's, and accents, all of which seemed chaotic. Also, it was beyond my word processor of the time to record the abbreviations used. I therefore decided to prepare a modernized and regularized edition to accompany the facsimile. I included the line numbers of the manuscript in my edition, and recorded in footnotes all emendations and modernizations other than the most common ones, of which I gave examples. To date no one has criticized this decision, nor any reading in the edition, nor the modernization of the fictional characters' names, which was linguistically the trickiest step. Years later I returned to the text to complete a study of Cervantes' consonants (phonetics) which I had set aside when it first surfaced. I found that the modernized, regularized, but annotated edition was more helpful than a paleographic edition would have been. Furthermore, the phonetics of the fragment are completely compatible with the phonetics of Cervantes' known works.
     What I would do differently if I had it to do again, and had more money than I had at the time, would have been to go to Seville and edit from the manuscript rather than from microfilm. An application for funding to make that trip was not successful, and I was not able to go to Seville at my own expense until 1987, when my edition was already set in type. My troubles in finding the person in the Cathedral of Seville who could locate the manuscript for me, and reading it in a room the size of a closet, with two stations and a tiny bulb, would itself be a colorful story. Omitting this, I expected the manuscript to be more legible than the microfilm, but found the difference even larger than I expected. Many letters at the edge of the page that I had thought lost were hidden by paper curls, and a difficult spot revealed itself to be a hole in the paper, which the photographer had simply ignored. Fortunately none of the restorations in brackets in my edition proved to be incorrect, but a number were not necessary. Working from the original would not only have been less risky than working from the microfilm, it would also have been much quicker.
     As two years went by, I became increasingly confident that no critic of my thesis was going to be able to suggest any other potential author. I had also found such extensive parallels in ideology and wording between this text and Cervantes' known works that they filled 75 pages when later set in type. Convinced, then, that the text was authentic and that my position was, by my own standards, unassailable, I announced the discovery in 1985.
     In retrospect I was naive to expect otherwise, but initial reaction to my announcement was disappointing. Scarcely anyone paid any attention to a new text of Cervantes, a great contrast with the attention given to the much briefer and much less important putative Shakespeare discovery of the moment, "Shall I Die?" A friend helped to arrange a small panel on it at the annual meeting of the Association of Teachers of Spanish and Portuguese, in New York. About ten people attended; none of the invited news media, including the Spanish language news media, showed up. One would think that a report on the recovery of a fragment of a lost work would interest the Cervantes Society of America, on whose Executive Committee I serve, but its program chairman rejected a proposed paper for two years in a row, after which I stopped proposing it. There was not much interest in Cervantine discoveries. The whole area, thanks to the forger previously mentioned, had a bad reputation. A new, authentic text would also mean that many people would have to do a lot of thinking about Cervantes, and thinking is sometimes work.
     On another front, I ran into difficulties in getting my study published. In retrospect, it is not a very desirable project for a publisher, who had much to lose and not much to gain from bringing it out. If I was right I would get the credit, not the publisher, but if I was wrong the publisher would share in the blame.
     My study was originally accepted, enthusiastically, by one of the coeditors of a scholarly journal called El crotalón, which published in huge volumes articles that sometimes went well over a hundred pages. He knew my previous work and was ready to take a chance. However, the article never appeared. The 1985 volume, in which it was to appear, was the final volume of the journal. Dissension among the editorial board caused by my study seems to have been a significant factor in its demise.
     The article was moved to a book series sponsored by the journal, which seemed more acceptable to its editorial board. My study, now a book, appears in a catalogue published at the end of the final volume of El crotalón. This volume never appeared either, although the volume that followed it in the series did. As more years passed—my study never appeared until 1989—my concern grew, and I started protesting more and more loudly. I suspected that the delays and pressures on the editors were another sign that I was right. Finally the other coeditor of El crotalón, who wrote me a formal letter of apology for his misconduct, got the volume published in a literary series of the Salamanca state government. The publications director of the Salamanca government promptly lost his job. His successor never answered any of my letters, nor sent out any review copies. I did, however, receive the number of copies my contract specified, and sent out review copies myself.
     At this point the tide began to turn. There have now been five reviews and one review article. No reviewer has said that the fragment cannot be by Cervantes, nor suggested any other author. Again to my surprise, the four reviews by Spanish speakers have been more favorable than the two by English speakers, who said they were unconvinced, though not that I was wrong. The book got a surprisingly favorable review in Ínsula, the Spanish monthly that is something like the TLS in importance. I was interviewed on Spanish radio and television, and there were a number of newspaper articles. I was also named a corresponding member of the Royal Academy of Letters, an appointment which in an indirect way is a confirmation of the validity of my discovery.
     This episode has been very educational. It has revealed to me something of how the world of scholarship works, at least in my small and contentious field. Cervantes is a powerful cultural symbol, and fierce, competitive pressures strive to control him and to use him to promote careers. I have come to see that "who you know" and "how much pressure you can bring to bear," although meaningless with the perspective of decades, does have some bearing on what scholarship gets published and what books get reviewed. Still, one convinced that he or she is right can generate a certain amount of favorable pressure simply on the strength of one's conviction.
     The adventure has also given me the benefits of danger faced and vanquished. Having taken a big risk and survived it, I am much more confident. My name will forever be associated with this beautiful text.
     I have become much more secure in my understanding of who Cervantes was and what he believed in. In fact, I have subsequently come to attribute other texts to Cervantes, and indeed to see him as a ghost writer during some of the obscure periods in his life. I have matched another lost work of Cervantes—a report on the festivities celebrating the birth of prince Philip IV—with an anonymous published report on that topic. A lengthy report on the Sevillian jail, whose date corresponds precisely with Cervantes' imprisonment there and whose ostensible author is an unknown, I believe also was by Cervantes, and have said so in an unpublished paper. I am about to suggest that he may well be the author of two historical works on Algiers, where he was held prisoner for five years. After all, it is Cervantes himself who tells us that he was the author of "works that circulate without his name on them."

     For further information and for the text of the fragment of the Weeks in the Garden, see my Las "Semanas del jardín" de Miguel de Cervantes (Salamanca: Diputación de Salamanca, 1988 [1989]), "Repaso crítico de las atribuciones cervantinas," Nueva revista de filología hispánica, 38 (1990), 477-92 (a revised version in a collection of my articles, Estudios cervantinos [Barcelona: Sirmio, 1982]), and "Cervantes' Consonants," Cervantes 10.2 (1990 [1991]), 3-14. The fragment has not yet been translated into English, nor has its significance for Cervantes' thought or for the interpretation of his other works been studied.

Daniel Eisenberg <>