[This post is primarily meant for non-English speaking national EA groups who'd like to see translated EA-adjacent books (such as The Precipice) published in their countries. I felt especially motivated to write this after EAGxPrague, during which many people reached out to me and asked me about my experience as a translator of The Precipice into Czech.

Please note that if your group is doing something other than a book translation, such as translating an article or other type of EA-adjacent non-literary text, different considerations might apply. I don't want to intimidate people by how complicated the process of book publishing might seem. In fact, I'd be happy to advise people on their translation projects. I am also drafting an online workshop for EAs who'd like to translate or publish translated content in their countries (see below and please contact me if you're interested).]

Content summary:

In late May 2022, The Precipice was published in the Czech Republic. I am the professional literary translator who translated the book into Czech. In this post, I describe the publishing process and outline other ways the project could have been organized.^[1] I end with recommendations for publishing EA-adjacent books, which I think is the most useful part for my target readers.

These considerations and recommendations should be taken as my subjective opinions, albeit based on some experience studying literary translation and working as a literary translator.

In case you want the TLDR, my recommendations for those interested in publishing EA or EA-adjacent books are to:

1. Find the right publisher

2. Reach out to the publisher

3. Create a group of people who will help the translator as they translate

4. Get on one page with the translator (or find the right translator)

5. Create a glossary of terms

6. Proofread, proofread, proofread

7. Get in touch with other EAs who have done some thinking about translation

The process of publishing The Precipice

[Note about terminology: In this post, I will be talking a lot about "literary translation", "literary translators" etc. Since I've noticed some confusion about this term, I'd like to clarify that "literary translation" means translation of both fiction and non-fiction books. The case of The Precipice, as a non-fiction book, also falls under "literary translation", albeit not under "translation of fiction". The distinction I often make here is between "literary translation" and "translation of other texts", such as articles, documents, leaflets, web content and all things non-literary.]

Description of the process

The Precipice was published in the Czech Republic by Argo Publishers, in a series called Crossover. More information about the series and the publisher is provided in the sections Find the right publisher and Czech literary market.

For any translated book, fiction or nonfiction, in any decent publishing house (at least in the Czech Republic), here's how the process usually is:

1. The translator translates the book and submits it (usually using a word processing software such as Microsoft Word^[2]).

2. The language editor reads the translation simultaneously with the original. Not only do they correct errors, they should also actively offer alternative phrasing and comment on the translation.They send it back to the translator.
(Non-professionals are sometimes surprised by the quantity of the changes proposed by the language editor, so I thought it might be interesting to mention that even in a very good translation, it's absolutely normal to see ten or more proposed changes per page. I know this from my own experience as a translator and occasional language editor, as well as from many conversations with much more experienced language editors).

3. The translator reads through the comments, accepts the proposed changes they agree with, and discusses the proposed changes they don't agree with. The translator should always have the final say in this (as they are the owner of the copyright). The translator would, however, be silly if they decided not to accept any of the proposed changes. Also, if the publishing house is not happy with how the translator reacts to the language editor's feedback, they can decide to withdraw from the contract (in an extreme case).

4. The text is then sent to the publishing house to be typeset. (Typesetting is the process of arranging the digital text - the paragraphs, symbols, charts etc. - onto a page so that it's print-ready.)

5. After it is typeset, the language reviser (a different role from the language editor; different person, too) reads through the text. This time, they don't actively offer alternative solutions, but rather check for punctuation and typesetting errors.

6. After that, the text is sent to the translator again, and the translator again approves of the final version of the text.

7. For every book, there is usually also an in-house "responsible editor" who also reads the book before it is printed. For Argo's Crossover, this was Oldřich Vágner.

So for any translated book, you usually get 2-3 proofreaders and the translator usually reads the text a minimum of three or four times.

With projects like The Precipice, things get more complicated than that.

Argo decided from the start that apart from the usual proofreaders mentioned above, a Czech expert on x-risks would read the translation after it is finished, and add comments. For this task, they reached out to Jan Kulveit, Czech AI alignment researcher from Oxford's Future of Humanity Institute and Toby Ord's colleague.

Along the way, parts of the text were also consulted with Tomáš Gavenčiak (another AI safety researcher) and the whole text was proofread once more by Jiří Nádvorník, co-founder of Czech EA Association (CZEA). Jan Kulveit reached out to Tomáš Gavenčiak, while Jiří Nádvorník offered his help later in the process, after I became familiar with CZEA (see the Appendix My story if you're interested in details).

I also consulted some very small fragments with Toby Ord himself; see section Errors (or not?).

I am very grateful for all the help I received. Although the process was quite lengthy and I think it can be quicker and more efficient (as outlined below), I think the final product turned out well and all the people mentioned here contributed to it substantially.

Timeline

The whole process of publishing The Precipice in the Czech Republic took 20 months. The approximate timeline was:

October 2020 -- March 2021: I translated the book alone.

September 2021: Jan Kulveit and Tomáš Gavenčiak provided their comments on the translation, I implemented them and further discussed some problematic passages with Jan Kulveit.

September 2021 -- December 2021: Language editing of the translation took place. I then approved of the changes.

December 2021 -- February 2022: Typesetting took place.

February 2022: The book was typeset, language revision took place, I approved of the changes. Jiří Nádvorník, Jan Kulveit and Oldřich Vágner read the book again.

April 2022: The book was sent to the printing office.

May 2022: The book was printed, published.

Main challenges

While translating, these are the things I struggled with the most.

Unfamiliarity with the field of x-risks

I am a literary translator by education and trade. I first became familiar with EA precisely through the translation of The Precipice (you can read more details about my background in My story).

When I started working on the translation, not only was I unfamiliar with the field of x-risks; I also didn't know anyone who worked or did research in that area. (To be more precise: I did know Jan Kulveit existed, as I knew him from the media and it was pre-arranged by Argo Publishers that he'd help with the translation after it's finished, but back then I didn't feel like I could consult thIngs with him and reach out continuously, as I was translating). In practice, this meant that any topic I didn't understand, I had to research on my own, reading related papers online and consulting people from related areas I knew.

This was difficult work but it was also something I was prepared to do as we were trained to translate non-fiction texts like this at the Institute of Translation Studies. However, I can honestly say that never before had I translated something as complicated and highly scientific as The Precipice. The golden rule that "you can't write something in the translation unless you're 100% sure what it means and could explain it in your own words" failed me because despite all my research, there were still a couple of passages in the book I wasn't able to understand on my own (these were mostly technical details in the notes section of the book). But I bookmarked these to consult with Jan Kulveit, so I knew there'd be a safety net of some sort. Still, it would have been nice to have someone who would have helped on the spot (as proposed in Option B).

Note: At this point you might find yourself thinking: "Ok, but if it's such a struggle for the professional translator to even understand the original, why is a professional translator necessary at all, considering there are EAs who understand the original perfectly and speak the target language, too?" I talk about this in Option C.

I was alone

I only reached out to CZEA and became familiar with the community after I had translated the book, not while I was translating it. As mentioned above, things would have been easier for me if I had had a group of people to consult the translation with.

One other perspective from which to look at this is that had I been in touch with CZEA while I was translating, I could have made things easier for them, too. One member of the Czech EA community reached out to me after I'd submitted the translation, asking me how I translated one specific term. He wrote: "Could you look up for me how you translated this? Surely, you must have a glossary or a database of translated terms." Well, I didn't. I did help him - I remembered how I translated the thing he asked, and even if I hadn't remembered, I could have still looked it up in the Word document - but only then it occurred to me that I should have written a glossary of all the translated terms I had to look up or come up with. I didn't need it, sure, but it would have been useful for other members of the Czech EA community.^[3]

Errors (or not?)

Despite all my fascination by the source text, I occasionally found something that felt like an error. Who was I supposed to consult this with? Who even has this very specialized knowledge, such as the amount or type of magma ejected from a supervolcano? Eventually, I reached out to Toby Ord with some of my doubts. It turned out that three of the things I spotted were actually errors in the source text.

The style

"Glad you enjoyed it! I'm sure it was a challenging book to translate, with the mixture of clarity on technical points and poetic descriptions of humanity's broader potential and challenges," Toby Ord wrote to me, summing up what the main stylistic challenge of this translation was. While very poetic in some parts and very readable in general, it never stops being scientifically accurate (dozens of experts and scientists have been consulted to polish all the scientific details of this book).

Perhaps surprisingly to some, this didn't prove to be such a huge problem when it came to longer scientific explanations which I took the time to research and rephrase correctly in Czech (or I actively asked for an explanation). Rather, the first version of my translation proved to be wanting in small details in the more poetic passages, where I sometimes got carried away and used e. g. an adjective or verb that wasn't quite precise and revealed that my understanding of the topic was superficial (or that I don't usually read articles about that topic, for example). Jan Kulveit proved to be a huge help in this aspect, helping my translation to be as scientifically precise as the original is (while I still kept in mind its poetic and stylistical specifics).

How it could have been organized (other possible setups)

In this section, I'd like to lay out different options of how the translation of a complicated EA-adjacent book could be organized and its possible upsides and downsides.

Although I personally do have a favorite among these options (Option B) and my choice of this option is based on my personal experience with translating literature, I wouldn't like to force my opinions or intuitions on anyone. Precisely because of that, I also describe options I'm personally more skeptical about, such as Option D. You can try out any of those options if you're not convinced by my reasoning, and if you obtain a satisfactory result with any of them, I'd be happy to change my opinion as to which one of these is the most effective one.

Option A

In this scenario the initiative to publish the book comes from a publishing house, not the national EA group. The publishing house finds their own translator, possibly also their own consultants (Czech x-risks experts) to be consulted before the book is printed.

This is basically what happened in the case of The Precipice in the Czech Republic. Oldřich Vágner, editor-in-chief of the "Crossover" series in Argo Publishers, chose this book to be published and secured the rights. He then reached out to me, the translator, and to Jan Kulveit, the AI alignment expert. Jan Kulveit then reached out to Tomáš Gavenčiak. Meanwhile, I got to know the Czech EA community and eventually also got help from Jiří Nádvorník (see My story if you're interested in details).

One thing to understand here is that, in my experience, when a publishing house reaches out to an external expert consultant - which is quite common for a good publishing houses to do, especially when it comes to highly scientific non-fiction literature - the consultants usually only read the text after it is translated, sometimes even after it is revised. Publishers probably suppose this is the most effective way to go about it as it saves the expert's time; I argue it's probably not (at least not for EA-adjacent texts where you have a whole community of very knowledgeable people willing to help the translator).

Upsides of Option A

In this scenario, the publishing process is basically in expert hands. The editors and the typesetters are professionals who know what a quality book "should look like". If it is a good publishing house, there are people who'll buy the book only because it was published by that specific publishing house or even in that specific series (see section Czech literary market).

The translator is - presumably - skilled in writing in Czech (in Czech stylistics), which is extremely important to renowned publishing houses. As the editor at Argo Publishers said: "If there are factual imprecisions in the submitted translation, they can always be corrected; however, if the stylistics are bad, you can't possibly rewrite the whole book."

Downsides of Option A

The "literary experts'' can't be expected to know about EA, x-risks or any of the related fields. In this scenario, experts are consulted; however only after the translation is finished. Since experts in x-risks are usually busy people, they presumably won't jump on the translation and comment on it the moment it is ready. It then takes a lot of time to wait for the experts to read the translation and to add their comments; and often, after that happens, large parts of the text have to be reread and re-written, which is time-consuming for the translator and for all the concerned parties.

Option B

(And this is what I would recommend and what makes most sense to me, personally).

In this setup, the initiative to publish the book doesn't come from a publishing house, but rather from the national EA group. The EA group reaches out to a renowned publishing house in their country and lets them find a professional translator (an expert in stylistics and linguistics). However, they fill the translator in on the topics covered in the books and, ideally, create a group of experts which can, on a regular basis, consult the terminology and the complicated passages with the translator as they translate.

Upsides of Option B:

All of the upsides from Option A still apply, plus: this makes the whole process quicker, the translator doesn't feel lost and the experts and EAs have a certain control over how the translation is going (and if it's not going well, they can intercept sooner). I feel very optimistic about this scenario.

Downsides of Option B:

The only downside that occurs to me is that there is a chance that the translator won't get along with the group of experts, or that the experts won't understand their roles correctly and will try to modify the translation more than necessary, which in turn might generate conflicts that might slow the process down. This of course can happen, but it's all a question of effective communication and I don't think it is very likely.

Note:

I'd also feel optimistic about a mixture of Option A and Option B where the initiative for publishing the book comes from the publishing house, however the experts from the local national EA group check on the translation continuously.

Option C

Of course, the national EA group doesn't have to reach out to a professional translator. They can find the translator in their midst, or they can reach out to an expert on x-risks (not an expert on translation) to translate the book. They don't even have to contact any publisher in their country; they can publish the book themselves.

Upsides of Option C:

Since the translator would plausibly be somehow aligned and definitely hired by the national EA group, all of the concerned parties would also probably be "on the same page", and would have similar expectations from the project.

I also owe Laura González this note: "Self-publication (if the rights are secured) is a safe bet, while it's never guaranteed that a good publishing house will accept or be interested in publishing the book. I think publication with a publishing house is far better than this option (...) but sometimes the choice is not publishing house or self-publication, but rather self-publication or nothing." This is of course true and can be seen as an upside of this scenario. But I'd really encourage people to see self-publication as a last resort and rather try talking to local established publishers (see section Reach out to the publisher).

Downsides of Option C:

I'd be worried that in this setup, the translated book's literary value would be significantly diminished.

The danger here is that for the translator's role, a person would be chosen who is, say, well-versed in some of the topics the book covers and speaks the source language well, which in my eyes are not sufficient prerequisites for a literary translator. (A cliché saying among translators is that "Speaking a foreign language well doesn't make you a translator any more than having ten fingers makes you a pianist." You might have heard something similar in your language: the translators' frustration by people who think they know how to translate is universal.)

A literary translator is, above all, a person capable of spotting which aspects of the text are central to the author's style (as opposed to aspects which, for example, might be typical for the original language/culture but maybe do not constitute a central part of the author's style), and equipped with sufficient writing skills to find suitable ways to convey those stylistic aspects in their target language. This is a skill that, in my experience, takes years of training, and it doesn't come to you naturally just because you know the language and are familiar with a certain topic.^[4]

It is of course true that stylistics are even more important in works of fiction, and that for non-fiction books, it is factual correctness that is key. I'd argue, however, that stylistics play a very important role in non-fiction literature as well, and my hunch is that even readers who say they only care about a book's content (not its style or structure) always recognize, subconsciously at least, when something's off with the style. And then there is the author's point of view. Wouldn't it be a shame to just wipe off the style they created for the book?

I'm not saying that a non-professional translator must, by default, always produce a bad result.^[5] But I expect professional translators to have a better idea of how to render the author's style into the target language than non-professionals, and I think the probability that a professional translator (assisted by an expert in x-risks) will produce a quality literary translation is higher than the probability of a non-professional producing one, even if said person understands the topic and speaks the language very well.

It also comes to mind that while trained translators and literary experts, based on my experience, are quite likely to be aware of their own shortcomings in other fields (such as x-risks), scientists, on the other hand, might underestimate their own shortcomings in translation, writing, communication and stylistics (because those skills are sometimes seen as inferior to their more technical types of knowledge and skill). I find it likely that a professional translator reaches out to an x-risk expert for advice; I find this less likely to happen vice versa (and I think it's always important for the two sides to collaborate).^[6]

Also, if the national EA group decided to act without the publishing house, they might lose a significant number of readers who'd buy the book because it was published by an established publisher.

As for the question of finance, I suspect that in this scenario, all of the concerned parties would probably be better paid than the literary professionals, at least in the Czech Republic (see Czech literary market). This might seem like an upside at face value (and it definitely is for the translator). But as Laura González pointed out, if we accept that not everyone is fit for being a literary translator (despite their expertise in x-risks and knowledge of the language), then it's actually a downside for EA: you're paying a person more to do a worse job.

Option D

The national EA group might also disregard a human translator completely and opt for a machine translation, probably post-edited by relevant experts.^[7]

Upsides of Option D:

This would probably make the process significantly quicker and cheaper.

Downsides of Option D:

My intuition is that unless the local EA group hired a post-editor who is also trained in translation and stylistics, the book would lose a lot of its literary value, and if they did hire a post-editor trained in these fields, that person would have so much work post-editing the text that it would have been easier to have them translate it in the first place.

However, machine translators and language models are improving quickly and I wouldn't like to discourage anyone from trying this out. I'd be very interested in how the final product turns out.^[8]

Promotion of the Czech version of The Precipice

To promote the book, we are organizing a series of debates about it with x-risks experts. Two of these debates took place in June and more will happen in fall.

We are also planning to organize reading groups of The Precipice, record a podcast episode about the book with Argo Publishers and possibly reach out to some journalists with a press release, but this is still a work in progress.

Recommendations to publish the best possible product

To any national EA group that wants an EA-adjacent book published in their language, I'd personally recommend the following:

No. 1: Find the right publisher

Try talking to people who are well-versed in the literary market in your country (editors, literary translators, writers, literary critics, avid readers who happen to follow the situation at the literary market in your country closely) and ask them which publishers have a good reputation of publishing quality literature in good translations, without typesetting errors, with nice graphics etc.^[9]

You can also look for editions in which the EA-adjacent book of your choice might be published: in the Czech Republic, for example, both Doing Good Better and The Precipice were published by Argo in a series called "Crossover", a non-fiction series which "presents the most interesting and useful proposals to solve the problems around us." Apart from DGB and The Precipice, Rutger Bregman's Humankind or Alec Ross' The Industries of the Future were published in this series, and many others. Perhaps there is a similar series in your country as well?

No. 2: Reach out to the publisher

Obviously, it's hard to estimate how any given publisher will react to a proposal from the national EA group. But I'd be optimistic that they at least consider your proposal if your pitch includes the following:

• "This is a quality, well-researched book that sells well and has been published in other languages." (Show examples, present reviews of the book.)

• "This book is a very good fit for you as a publisher/ for one of your series." (Explain, present arguments.)

• "We, as the local EA chapter, formed a group of experts ready to help the translator whenever they need it." (See Option B; also explain why the translator will probably need it.)

• "There is a significant number of people who will probably read, buy and promote the book." (Present the local EA community.)^[10]

When reading my text, both [redacted] and Laura González pointed out that part of the work mentioned in No. 1 and No. 2 could also be done by literary agents. I honestly don't know much about how literary agents work nor how they choose the publisher for their book in a foreign country. It might be potentially high-impact to reach out to the literary agent (or maybe the author) of the EA-adjacent book you want published in your country. In that case, they'd be the ones pitching the book to a publisher in your country as insiders, and they could also include in the pitch all of the points mentioned above.

No. 3: Create a group of people who will help the translator as they translate (not after the translation is finished)

I'd recommend the national EA group to create a support group for the translator consisting of x-risks experts or local EAs who have a very good understanding of the topics the book covers. See Option B for details.

No. 4: Get on one page with the translator (or find the right translator)

If you're successful and the publisher reaches out to their own translator, be sure to meet with them as soon as possible and explain what you can do for them. Make clear to each other what your roles are and how often you'll consult the translation together. If you know a great professional translator who is somehow related to your national group, you can also recommend them to the publisher; perhaps they choose them for the translation and then the communication will be even easier.

In any case, check multiple times whether the translator is aware of everything the job entails: that it is a big project which a lot of people care about, that the translator will be the one responsible for its final form (although they will have a significant help from your national group) and that it's probably a commitment for many months (although this depends on the specific book you're translating).

Ondřej Hrách, a fellow translator, adds that it's important to take into account that literary translators are usually very poorly paid (you can read a little more about this in Czech literary market); and they're usually paid by word or page count, not by the hour. So it might be off-putting to them at first sight to see how much work this project entails. Explain how your group would help them and perhaps even that there are ways of providing additional funding (if you're happy with their work and if you're able to provide it, of course).

No. 5: Create a glossary of terms

Ask the translator to kindly create and update a glossary of terms as they translate. This will be very useful for your national EA group in the future. See this section for details.

No. 6: Proofread, proofread, proofread

You might think that since there was a group of experts checking the translation along the way, it has been proofread enough already. In my experience though, with a book so long and complex as The Precipice, errors always occur, even when it has been proofread by other people, too. Of course, it depends on the book you're publishing, but apart from the usual proofreaders with linguistic expertise, I'd recommend choosing at least two trustworthy EAs who will read the final version of the translation carefully and thoroughly before it goes to print.

No. 7: Get in touch with other EAs globally who have done some thinking about translation

Here are some people in the ecosystem I know of that are interested in the topic:

If you're interested in the impact you might generate by publishing an EA-adjacent book in your language, I strongly recommend reaching out to Open Phil's Eli Rose. His writeup on translation covers both literary translations and translations of other EA materials.

CEA's Catherine Low is generally excited about translation projects and will be able to connect you with more people.

Laura González is now working to get The Precipice and Doing Good Better published in Spanish. She's also coordinating a big project to get professional translations of many online resources (like the EA Handbook or the materials for the virtual courses).

[redacted] has translated EA content to German with a focus on longtermism and is happy to share his experiences.

The non-english-speaking-groups channel on the EA Groups Slack might be of interest too.

I would like to start an online translation workshop for EAs which covers how to translate EA content from scratch. I am actually now looking for a partner to do this with. I think I need a project manager, and I'd also be very grateful for someone who understands statistics, sociology or economy and/or someone who'd know where to look for answers to questions like: "How to find out which EA-adjacent book might be the best fit to publish in a specific country?", "Which translation projects have the biggest EV in specific countries?". If you're interested in more details, please contact me at [email protected]

Conclusion

I hope these recommendations and considerations will help EA-adjacent books to reach their relevant readers in non-English speaking countries.

While I think there are many possible ways of publishing an EA-adjacent book, I think some of them have a better shot at success, and I'd especially recommend what I outline in Option B.

I'd be happy to discuss this further and, as mentioned above, I'd also be happy to advise EAs globally on their translation projects.

I would also like to thank all the people who read parts of this text and commented on them: Laura González, Ondřej Hrách, Hana Kalivodová, Catherine Low, [redacted], Eli Rose, ES, and Jan Votava.

Appendix: How I got to translating The Precipice

Since people ask me this frequently, I decided to include a section on the Czech literary market and how I was chosen and trained for this task.

Czech literary market

Here are some pieces of what's considered common knowledge in the Czech literary and translating community. I hope this helps a person from a different cultural background get a better idea of the literary market we introduced the book into.

• There is a huge number of publishing houses in the Czech Republic: more than 7,600 publishers for only 10 million Czech-speaking people.^[11] To give you an idea, there are only around 900 publishers in the United Kingdom, which has almost seven times as many people as the Czech Republic.^[12] The competition is huge and it pushes the prices of the books and the salaries of people working in the literary market lower and lower.

• It is generally accepted that smaller, independent publishers publish fewer books of higher quality. Argo Publishers, where the Czech translation of The Precipice was published, definitely belongs to these smaller publishers with a better reputation. There are readers who actively only support publishing houses of this kind and buy books from them just because they were published by them.

• Most of the translators work freelance, as opposed to language revisers and editors, who very often work in-house.

• As far as translation goes, graduates of the Institute of Translation Studies of the Charles University have a good reputation on the literary market (based on what the publishers say and my colleagues' experience) and are frequently invited to collaborate by those smaller independent publishing houses.

My story

I had studied Translation and Interpreting at the Institute of Translation Studies at Charles University in Prague, Czech Republic. Both at Bachelor's and Master's, this is basically what you might call a double major, one being focused on translation (of mostly fiction), the other one on interpreting (simultaneous and consecutive). Even before finishing my Master's, I knew I wanted to become a literary translator and I had translated one non-fiction book before graduating. Just after my graduation in September 2020, Argo Publishers sent me the offer to translate The Precipice.

Argo chose me because I was recommended to them by a senior translator (also a graduate from the Institute of Translation Studies) and they also considered it a good sign that I had translated the previous book for Paseka Publishers (another smaller editorial with a good reputation). They wanted me to submit a sample of translation, and were happy with it when I did.

As for my original impressions, the book seemed interesting and well-written to me at first glance, however I was a little disappointed I wasn't assigned a fiction book instead, because I preferred reading those and considered them a more creative type of work. I really wanted to translate fiction. To be perfectly honest, my main incentive to accept the assignment was to establish cooperation with Argo Publishers, a renowned publishing house, more so than translating Toby Ord's book per se.

However, as I was working on the book, I became increasingly interested in the ideas it contained. After I finished the translation, I reached out to the Czech Association for Effective Altruism (CZEA) and after a year of being in touch with them, I started working part-time for CZEA. I still translate literature in my remaining working time.