Earlier this year I joined a book club reading Don Quijote and got deep enough into it to seek out the “false second volume of Quijote” of Alonso Fernández de Avellaneda (1614), the book that Don Quixote catches Don Jerónimo and Don Juan reading in Chapter 59 of Cervantes’ actual Part II.

“Why do you want us to read all that nonsense, Don Juan? Nobody who has read the first part of the history of Don Quixote de la Mancha can possibly derive any pleasure from reading this second part.”

“All the same,” said Don Juan, “it’ll be as well to read it, because there’s no book so bad that there isn’t something good in it.”

(Don Quijote II.59, trans. Rutherford.)

Avellaneda’s “false Quijote” was translated into English by Alberta Wilson Server and John Esten Keller in 1980 (Amazon), and eventually I bought that translation and read it. It’s entertaining and accurate, if somewhat riddled with typos in the 2009 edition. But the first English version of “Avellaneda” I’d found was Yardley’s, which isn’t a translation of Avellaneda’s Spanish original at all: it’s a translation of a 1704 adaptation by Alain-René Lesage, in which Lesage deviates significantly from what we usually expect of a “translator” — cutting some subplots, expanding others, dragging in whole episodes from elsewhere (including from Cervantes’ own Part II), and completely changing the ending!

Realizing the extent of this metafictional morass, I came across an interesting essay on exactly this subject — “Avellaneda’s Don Quijote, its relationship to Cervantes, and its adaptation by Lesage” — the doctoral dissertation (1906) of one Martin Wolf, later published in the Journal of Comparative Literature Studies (1909). Only one problem: it was in German! So I translated Wolf’s essay from German to English. (I promise I didn’t knowingly commit any Lesage-level trickery in this translation.)

• Read Martin Wolf’s essay in the original German.

• Read my English translation.

I leaned heavily on machine translation, pasting a paragraph at a time into Google Translate and then tweaking until the sense seemed right. Google Translate, like most “AI” (large-language-model) tools, is actually really helpful for this kind of human-supervised task. Unsupervised, of course, it would have made a dog’s breakfast of the text in multiple ways. Wolf’s academic language is sometimes “not un-difficult” to parse, or even ambiguous; and the essay incorporates text in German, Spanish, and French as well as small amounts of English, Italian, Latin, and Swedish.

Last year I used Google Chrome’s built-in “Translate this page” feature to read Maurice Renard’s science-fiction tale Le péril bleu (1911); on that kind of single-language straightforward pulp fiction, machine translation actually works quite well, as long as you’re willing to treat the text a little impressionistically, and to mentally retranslate a few tricky words and idioms. For example, in The Blue Peril Google’s machine translation often confused French voler “fly” with voler “steal.”

For the HTML and CSS styling of my transcriptions, I am indebted to Bill Thayer’s Lacus Curtius website, which I find amazingly readable; and for the idea of using dotted lines for hyperlinks (much less obtrusive than the default blue-text-solid-underline style) I’m indebted to gwern.net.

In the German transcription I’ve tried to be completely faithful to what Wolf wrote, except for fixing simple typographical errors; if he got a page number or a name wrong, I’ve just marked it up with a little CSS tooltip for the correction. In the English translation, I erred in the other direction, quietly fixing errata without any markup, and using CSS tooltips only for places where I think I myself have technical issues with the translation (e.g. phrases I guessed at).

I also used CSS to color-code each quotation from Cervantes’ Quijote; from Avellaneda’s; and from Lesage’s; and I hyperlinked Wolf’s sources to archive.org and/or Google Books whenever I could track them down.

If you find an error, or (being fluent in German) can improve the translation of any line, or confirm any translation I’ve marked up with the “unclear” CSS class, please send me an email and let me know!