THE PROBLEM OF TRANSLATIONS (AND A COUPLE OF PROBLEMATIC SOLUTIONS). PART 1: THE ROLE OF MACHINE TRANSLATION

This is the English translation of the article in two parts originally in Catalan, which I published in the web El Biblionauta (https://elbiblionauta.com/ca/, November 2021). Here is the second part.

In Douglas Adams’ humorous novel The Hitchhikers’ Guide to the Galaxy (1978), a small, yellow animal known as a Babel fish is used, inserted into one’s ear, as the solution to the problems of universal interlinguistic communication. We’ve all dreamed of a device that fulfils a similar function without the inconvenience of having a living creature near the brain, and a lot of effort is being put into that. Google, it seems, is working to build a universal interpreter inspired by Star Trek, an app (I suppose) easily installable on cellphones. While finding a way to translate live speech, as we all know Google Translate, other services like DeepL and Word itself (in an often overlooked menu option), help us translate written texts from one language to another with an increasing degree of efficiency.

Five years ago, Google Translate, launched in 2006, was in fact transformed into Google Neural Machine Translation (GNMT), a service that uses an AI-managed neural machine algorithm capable of processing contextual meaning, a feature that partly explains the drastic improvement of machine translation. I guess DeepL works the same way. In recent weeks, as I have already written here, there has been a heated debate about the use of this type of translation in the Spanish subtitles of the Netflix mega-success, the South Korean series Squid Game. As ATRAE (Asociación de Traducción y Adaptación Audiovisual de España) has complained, the multinational Iyuno, of which Netflix is a client, has used automatically generated subtitles for the first time, later revised by a human translator, charging a third of the usual fee (that is, between 60 to 100 euros for a 100-minute film). Post-edition (as this practice is called), ATRAE protests, threatens to destroy many jobs and lower the quality of subtitling. AVTE (Audiovisual Translators Europe) already published last September the Manifesto on Machine Translation (https://avteurope.eu/2021/09/13/press-release-avte-manifesto-on-machine-translation/) where it warned of the deep damage that machine translation will do in the short and long term in the audiovisual field and where it defends the need to achieve better collaboration between human translators and companies that offer powerful machine-based machine translation services.

This debate has not reached the literary world, but I want to start it by taking science fiction and the Catalan language as a case in point. I do not have a clear idea of the fees charged by translators but I understand that a novel of 300 pages in English can cost a few thousand euros (between 2000 and 4000?) to translate. If we think of a sales figure between 100 and 500 copies, we already see that business is limited, even impossible. I don’t quite understand, if I think about it, why there is a certain secrecy around the money it costs to publish a science-fiction book in Catalan, but when I asked several publishers what volume of business they hoped to generate I only got elusive answers. I was going to write that this is a topic for another debate but it is quite the opposite: it is a matter for this one. If we do not know clearly what a translation costs, it is difficult to solve the problem of how to fill the gaps in the publishing market for science fiction in Catalan. I leave aside the delicate issue of subsidies, which is perhaps the real focus of the debate.

The proposal I make below will not please anyone and could even shock many. I take as a case study the English author Richard K. Morgan, whom I just interviewed at Festival 42 and whose work I know entirely in the original English. Morgan has published nine novels (https://www.richardkmorgan.com/books-more/), six of which have been translated into Spanish (one of them, Altered Carbon, twice due to the author’s disagreement with the first translation and the subsequent breach of contract with the publisher). His trilogy about super-soldier Takeshi Kovacs, recently adapted by Netflix, is about to be completed in Spanish, the language in which you can read the first novel (Carbono modificado), the second, Ángeles rotos, and soon the third, Furias desatadas. Also translated are his first novel, Leyes de mercado, and the fantasy trilogy Sólo el acero, El gélido mando, and La impía oscuridad. In contrast, the author’s favorite novel among all the ones he has written and, for me, the best, Black Man (known in the United States as Th3rteen) will probably never be translated into Spanish (unless, of course, Netflix also adapts it). When I insisted to his publisher in Spanish that this was a good work, he replied that he had no doubt that it was, but that it is a novel too long and too little known to be translated. I understand that. It could always be the case that a larger publisher takes over Black Man but assuming that this does not happen I make here a controversial proposal: I would recommend Morgan, and all authors in a similar situation, to subscribe to a machine translation service, pay a translator to review the generated text, and self-publish, either on their own website or on platforms like Amazon (or Lektu … or El Biblionauta).

If no publisher is interested in paying for a translation into Catalan and publishing it, or has no resources, I think Morgan (or any other author in a similar situation) could follow the same method and self-publish in our language. I anticipate the furious protests of publishers and translators, but in all honesty, what should an author do who wants to find a new market in a new language but cannot find a publisher? Is it fair for a work to go unpublished in another language because it’s too expensive to translate or publish? The authors have so far accepted the rules of the game according to which a foreign publisher is the one who chooses to buy the rights and commission the translations, and surely they already have enough work to write for them to embark on new and strange adventures in the world of self-publication. As far as I know, authors never commission translations but expect foreign publishers to do so because it is logically cheaper for them. It’s all a matter, however, of working out expenses. If authors conclude that it is worthwhile to self-publish a translation managed by himself (or his agent), whether using human translators or revised machine translation, there is no obstacle for them to move forward. It all depends, as I say, on what expenses they care to assume.

I have no intention of antagonizing translators, a professional guild that deserves all my respect, nor publishers, but, perhaps because of the imagination of science fiction authors, many things are changing in the field of translation. I had the impression that the use of machine translation was much less widespread than it is in institutions, business and professional fields, but friends who are professional translators have frankly acknowledged to me that they are now basically engaged in revising texts translated by AIs. It could be argued that machine translation is too little advanced and requires deep revisions as expensive as a translation from scratch but this is a diminishing hurdle, as those of us who use machine translation know (I mean in non-literary tasks).

The vision of a world where only AIs are translators and there are no human beings trained in the profession should frighten us all, and it frightens me very deeply, yet I must make the problem of where translation is heading visible. It would be somewhat ironic that science fiction becomes the genre in which revised automatic translations into Catalan could proliferate, but it would be an irony consistent with the very nature of this genre. Perhaps rules can be set, so that only works that no publisher wants to publish, or for which there is no human translator into Catalan, are translated by combining the work of AIs with human work, but it is truly a pity that we cannot access works in other languages because the laws of the publishing market hinder it. If there is no market for some works in some languages (I’m not just talking about English) it would be logical to look for other strategies. These, by the way, should always be legal, no translation can ever be made disrespecting the rights of authors on their work. Thinking of the authors, I think, as I say, that revised machine translation and self-publishing are the most appropriate paths.

If you find this proposal unacceptable, we can focus for the time being on the first proposal (see the first part of this article) and turn El Biblionauta into the seat of a polyglot council of wise readers that can help publishers make beneficial choices for everyone, relating to what science fiction could be translated into Catalan, using human translators and beyond the English language.

I publish a post once a week (follow @SaraMartinUAB). Comments are very welcome! Download the yearly volumes from https://ddd.uab.cat/record/116328. Visit my website https://gent.uab.cat/saramartinalegre/. The Spanish version of the posts is available from https://blogs.uab.cat/saramartinalegre/es/

THE PROBLEM OF TRANSLATIONS (AND A COUPLE OF PROBLEMATIC SOLUTIONS). PART 1: THE SELECTION OF ORIGINALS

This is the English translation of the article in two parts originally in Catalan, which I published in the web El Biblionauta (https://elbiblionauta.com/ca/, November 2021).

It is common to celebrate from time to time the novelty of the publication in Catalan of foreign works of science fiction or fantasy, but it is not so common to reflect on the dynamics that make it possible for these works to reach our language. And on the contrary: although not so often, we are happy when we receive news of the translation of a work of the fantastic in Catalan into a foreign language, despite not even knowing how these little miracles happen. I therefore open a reflection on this topic that will lead, as will be seen, to two bold proposals described in two different articles, one of which is sure to create controversy (see part 2).

So far, things work as follows: publishers decide independently which authors and books they want to translate into Catalan, buy the rights, commission a translation, have it corrected, publish it, and sell it to the reading public with more or less success. However, there is no committee that carries a list of works which would be interesting to translate into Catalan (or from Catalan to other languages), so that in the set of translated works there are always important shortcomings of both classics and novelties. Some works were translated a long time ago but are out of print, others were not translated at the time of their highest popularity and it seems that they will never be translated, and current authors do not find anyone to publish them in Catalan even when they are known. in their language and, why not say it, in Spanish.

The first proposal I make, then, is to make El Biblionauta the headquarters of a committee of science fiction, fantasy and horror readers in Catalan that can advise local publishers and turn the market for books translated into Catalan into an environment. much more consistent than it is now. I’m well aware that readers are volatile and that we don’t always buy the books we want to read (which is why there are libraries, friends, and various illegal resources). I would say, however, that if between 100 and 300 people express the opinion that it would be desirable to translate certain foreign titles, Catalan publishers would so do more confidently than simply relying on their own intuition, or sales in the original language.

The committee’s idea is also applicable to the translation of Catalan into other languages. When I translated Mecanoscrit del segon origen into English, a novel that had already been translated into fourteen other languages but incredibly not into English, I realized that neither the publishers nor the institutions (whether the Institut Ramon Llull or directly the Conselleria de Cultura) monitor which Catalan books are translated into other languages. To be fair, the IRL does offer a database of books in Catalan that could be of interest to translate (https://www.llull.cat/catala/literatura/books_catalan.cfm ) but this is not specific enough in relation in science fiction, fantasy and horror. I don’t see why the readers of El Biblionauta shouldn’t be in charge of managing a list of Catalan works in these genres that would be desirable to publish in other languages. Obviously, it would be easier for Catalan publishers to look at the list of foreign works recommended by readers than for foreign publishers to look at a list of Catalan works, but it’s all about getting started.

In the middle of writing this article I had the pleasure of being a spectator at the new Festival 42 (https://www.barcelona.cat/festival42/ca) of the round table ‘New genre classics in Catalan: A boom with Adams, Dick, Le Guin, Butler, Matheson, King, Poe, Bradbury, Lovecraft and those who will soon follow…’, moderated by Miquel Codony and with the participation of Jordi Casals, Jordi Llavoré, Antoni Munné-Jordà, Martí Sales and Isabel del Río. The table was a celebration of the work that publishers such as Males Herbes, Mai Més Llibres, Chronos, Laertes, Raig Verd, L’Altra, Periscopi, Pagès, Kalandraka and Edicions SECC, among others, have been doing for about ten years in two ways: expanding the list of Catalan translations of foreign science fiction, fantasy and horror classics and recovering out-of-print editions, updating them. This is a very laudable job, without a doubt, but I myself was in charge of questioning a very important point in a brief intervention, when I protested, as an English philologist, that English is too important in this boom. The word ‘classic’ can’t be limited to English-language science fiction, I insisted, but that’s what’s happening right now.

This is not a new opinion in my thinking but it’s true that a conversation during the festival with Italian publisher and novelist Francesco Verso opened my eyes a little bit more. Verso commented to me that, as the University of Rochester’s Three Percent website warns (https://www.rochester.edu/College/translation/threepercent/about/), only 3% of all books published in the United States is a translation, including books of all genres. Rachel Cordasco, a friend of Verso, has an impressive database of speculative fiction works translated into English on her website SF in Translation (https://www.sfintranslation.com/) and has just published Out of This World: Speculative Fiction in Translation from the Cold War to the New Millennium (2021), described as a guide. Verso himself is pursuing a truly international language policy as a publisher, looking for translators of all possible languages, as he told me, and remunerating them in the same way as English translators to encourage them to do more work. The website of his publishing project (https://www.futurefiction.org/) includes a world map where many authors can be found outside the Anglo-American sphere.

A very important problem, then, is that neither readers nor publishers of genre fiction in Catalan know enough about other languages. To be better informed you can use resources such as Francesco Verso’s map, Rachel Cordasco’s website and guide, or academic books such as Dale Knickerbocker’s, Lingua Cosmica: Science Fiction from around the World (2018). This book is part of the growing wave of interest in the Anglo-American academic world for speculative fiction in other languages, of which the new book Science Fiction in Translation: Perspectives on the Global Theory and Practice of Translation, edited by Ian Campbell and in which I myself participate, is also part. In my review of Knickerbocker’s volume (https://www.revistahelice.com/revista_textos/n_27/Helice 2019 Fall-Winter MARTIN ALEGRE BABEL FISH URGENTLY NEEDED.pdf) I complained about how frustrating it is to read a book of this kind full of very attractive reading suggestions lacking translations. The editor, on the other hand, complained about the lack of academic specialists in speculative fiction written in languages other than English or in non-Anglo-American territories (there is, for example, African science fiction in English).

It is easy to understand why the current translation boom is basically linked to the Anglo-American classics since they are the ones we all know, but I think there is an important contradiction between the status of Catalan as a small language among those spoken in world, and the little attention we pay to science fiction in languages similar to ours. This leads me to think that the committee of wise readers I was talking about should be polyglot, if not individually at least as a whole. Both Francesco Verso and my co-editor at Hélice magazine, Mariano Martín, are admirable polyglots, and their mastery of diverse languages gives them a comparative knowledge of the space of international science fiction that is simply incomparable. Hearing them engage in a conversation about Bulgarian science fiction a few days ago was a pleasure but, again, a frustration because no text is translated into Catalan.

So I get to the point where I have to express a very strange feeling: I miss the Catalan translation of genre books (science fiction, fantasy, horror) from other languages whose existence I am unaware of. As Francesco Verso told me, we have reached a situation in which not only first-class classics but also second- and third-rate works in English are being translated because they reach us through the powerful Anglo-American distribution machinery. Meanwhile, first-rank works in other languages –both classics and novelties, in large or small languages– go unnoticed, just as Catalan works go unnoticed among international readers. I understand that it is too much to ask that Catalan readers and publishers suddenly become polyglots aware of the current state of the science fiction published abroad, beyond the English language, but that is what we need. Either we do this or we look for bilingual or polyglot people who can inform us and, above all, who can translate into Catalan other traditions yet to be discovered.

In the second part of this article, I explain the role that artificial intelligence could play in this process. Keep reading …

I publish a post once a week (follow @SaraMartinUAB). Comments are very welcome! Download the yearly volumes from https://ddd.uab.cat/record/116328. Visit my website https://gent.uab.cat/saramartinalegre/. The Spanish version of the posts is available from https://blogs.uab.cat/saramartinalegre/es/