Example of Claude-designed translation workflow

[Tom Gally]

Some years ago, I translated a dozen or so essays by 寺田寅彦 for Iwanami Shoten’s 科学 magazine. They were later published in two books. Yesterday, an editor at another publishing company contacted me, asking me to write an essay about my translation process. I told him that I would prefer to write about the issues of AI and translation, and I probably will. The deadline is several months off, so I have some time to mull it over.

In the meantime, I had Claude start creating a knowledge wiki focused on the translation of such essays. After it worked on that wiki for a couple of hours, I had it design and run a workflow for translating one essay as a test. I haven't checked the translation, but I thought its detailed explanation of the workflow might be of interest:

https://gally.net/temp/20260514-torahiko-translation-process/index.html

Tom Gally

[Tom Gally]

Yesterday morning, in the shower, I wondered if it would be possible to create a system in which various elements of the translated text—register, literalness, etc.—could be varied along different axes. I tossed the idea at Claude and it spent the day extending and testing it, with a focus on translating essays by Terada Torahiko. Here are the results of the first test:

https://gally.net/temp/20260516-torahiko-translation-axes/index.html

Presumably similar systems could be built for other, more commercial purposes. That would make it easier to create translations tailored to the particular needs of a client.

Tom Gally

[Kevin Kirton]

Fascinating work, Tom, thanks for sharing.

The sudden quality improvements in machine translation brought about by LLMs was initially very dispiriting for me. I remember long threads on honyaku in the early 2000s about whether machine translation would ever reach human quality, but I'm assuming few translators would argue that it hasn't at least come close now. And when you factor in speed and cost, it's a harsh competitor for humans. But your experiments show how important and creative the human in the loop can be. Very interesting time to be alive.

Kevin Kirton 

Canberra, Australia

--
You received this message because you are subscribed to the Google Groups "Honyaku E<>J translation list" group.
To unsubscribe from this group and stop receiving emails from it, send an email to honyaku+u...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/honyaku/CACnS3Nfh2UuhJ85Tw8SxFN0a5NZu33v%3DftOhUOhpjKBbtQWpwg%40mail.gmail.com.

[Tom Gally]

Thanks for the kind words, Kevin.

I should point out, though, that my creative contribution to that project has been minimal. Yes, it was my idea to try varying translations across multiple axes, but it was Claude that actually came up with the specific axes for Terada Torahiko essays (other than two or three that I suggested). It also devised and implemented the idea of testing the validity of the parameters by having other LLMs evaluate sample translations at various settings. That would not have occurred to me. 

Regarding the issue of human vs. LLM-driven machine translation: At least for the sorts of translations I do and have been testing with LLMs, the quality differences had already entered the range of individual preference a year or more ago. Once it became possible to have LLMs check each other’s translations for omissions and mistakes, the differences from human translation came down to subjective judgment and taste. When I use a multi-LLM workflow to translate a speech now, I still make a lot of changes to their final draft before submitting it. But the changes I make are no longer corrections of the text as a translation. Instead, they reflect my own preferences for English writing and speaking style; other human translators and editors might very well prefer the machines’ versions.

And regarding that experiment I did with having the same text translated sixteen different ways with various combinations of parameter settings: Are we safe in saying that it would be impossible to get such a set of translations from human translators? At least, I myself would not be able to create a set of translations with the variations controlled to certain specified parameters like that, and it would seem difficult to assemble a team of human translators to do so, too.

Tom Gally

To view this discussion visit https://groups.google.com/d/msgid/honyaku/CANmm7qY5GnqWuUcVJjO-ktvFeyPqZE%3DX5iLBzeZxkcdhYrk-yQ%40mail.gmail.com.

[Brendan Craine]

Dear Tom,

As ever, your experiments both fascinate and terrify me.

To highlight a statement of yours that stood out to me:

Once it became possible to have LLMs check each other’s translations for omissions and mistakes,

the differences from human translation came down to subjective judgment and taste.

I think this is an extremely important point, and something that will define our profession in the coming years. 

LLMs can produce the trappings of rhetoric, theme, voice, and so on - but only as reflections of existing conventions.

As such, we as translators must be able to evaluate the effect of a word or sentence as much as its meaning.

Or, put another way - we will need to convince clients that our personal tastes, as professionals, are worth the extra cost. 

--Brendan C

2026年5月16日(土) 20:35 Tom Gally <tomg...@gmail.com>:

To view this discussion visit https://groups.google.com/d/msgid/honyaku/CACnS3NewgUW8ekW%2BHKpq5TibdSW%2BrjXmoQ9NXiJaAt%3D6uW%2B9DQ%40mail.gmail.com.

--

=============================================
『エクスポート・ジャパン株式会社』

ライティングチーム（リーダー）

クレイン・ブレンダン（Brendan Craine）

email: brendan@export-japan.co.jp

[Tom Gally]

Brendan,

we will need to convince clients that our personal tastes, as professionals, are worth the extra cost.

I agree completely.

A slightly related note: I have been impressed by the skills of Will Ireton, the interpreter working with Shohei Ohtani of the Los Angeles Dodgers. While he sometimes omits parts of reporters’ questions and Ohtani’s answers, he really gets the essence of what is being said on both sides. He also clearly knows baseball. I recently learned that he used to be a professional baseball player himself:

https://www.youtube.com/watch?v=pcOde_hZCao

Tom Gally

To view this discussion visit https://groups.google.com/d/msgid/honyaku/CAMhgjJFK7VgXgwPGc-mfuQcT2n9La0nwd5qTm4XM52w0H0qOww%40mail.gmail.com.

[John Stroman]

In response to Tom and Brendan, readers may find the following article entitled: "If AI can translate instantly, why learn another language?" useful for consideration. The subtitle is: "There’s a difference between using a tool to assist you, and using one to replace cognitive effort." This article takes about 5 minutes to read, but if you are in a hurry, skip to the final section entitled: "What AI translation can’t replicate."

https://theconversation.com/if-ai-can-translate-instantly-why-learn-another-language-280310

John Stroman

----------------

[Rene]

I read the argument, and find it convincing... at the current moment. But the author is talking about a moving target. AIs get smarter all the time, and already know how to manipulate human psychology better than psychologists. Human psychopaths and people with Asperger`s also do not naturally have empathy, but can absolutely learn and imitate it. And unlike human intelligence, AI is open ended. I feel the argument here sounds similar to the "God of the gaps" situation in religious discussions.

Rene von Rentzell, Tokyo

[Herman]

On 2026-05-22 05:56, Rene wrote:
> On Fri, May 22, 2026 at 7:27 PM John Stroman
> <stromana...@gmail.com <mailto:stromana...@gmail.com>> wrote:
>
> In response to Tom and Brendan, readers may find the following
> article entitled: "If AI can translate instantly, why learn another
> language?" useful for consideration. The subtitle is: "There’s a
> difference between using a tool to assist you, and using one to
> replace cognitive effort." This article takes about 5 minutes to
> read, but if you are in a hurry, skip to the final section entitled:
> "What AI translation can’t replicate."
>
> https://theconversation.com/if-ai-can-translate-instantly-why-learn-

> another-language-280310 <https://theconversation.com/if-ai-can-
> translate-instantly-why-learn-another-language-280310>

> John Stroman
>
>
> I read the argument, and find it convincing... at the current moment.
> But the author is talking about a moving target. AIs get smarter all the
> time, and already know how to manipulate human psychology better than
> psychologists. Human psychopaths and people with Asperger`s also do not
> naturally have empathy, but can absolutely learn and imitate it. And
> unlike human intelligence, AI is open ended. I feel the argument here
> sounds similar to the "God of the gaps" situation in religious discussions.
>
> Rene von Rentzell, Tokyo
>

There is that analogy, by what I glean from it is different. Those who
advance the God of gaps critique will often have constructed or
conceived something even more elaborate than God, such as the
Singularity, a geometric point which contains the totality of past,
present and future states of an infinite number of infinite universes
with infinite dimensions. Nice.

Similarly here. The AI, after all, is used by humans, and if you got a
powerful AI, what advantage does that give you over your enemy (or,
competitor, partner, etc.), who also has a powerful AI?

Herman Kahn

[Rene]

There is that analogy, by what I glean from it is different. Those who
advance the God of gaps critique will often have constructed or
conceived something even more elaborate than God, such as the
Singularity, a geometric point which contains the totality of past,
present and future states of an infinite number of infinite universes
with infinite dimensions. Nice.

Similarly here. The AI, after all, is used by humans, and if you got a
powerful AI, what advantage does that give you over your enemy (or,
competitor, partner, etc.), who also has a powerful AI?

Maybe. By the analogy of the "god of the gaps" argument, I simply meant the issue of addressing a moving target. The gaps in science where a God is needed as an explanation are constantly being filled, so God moves on to the next gap. Similarly, the gaps in AI where humans are needed (such as understanding human feelings in this example) are constantly filled by the next version of AI. So people find another gap where humans are needed.

Meanwhile, I wonder at what point the currently correct "AI being used by humans" morphs into "humans being used by AI"? AI is already massively successful in influencing humans right now, before the next monstrous data centers come online.

Rene von Rentzell, Tokyo

 

Herman Kahn

--
You received this message because you are subscribed to the Google Groups "Honyaku E<>J translation list" group.
To unsubscribe from this group and stop receiving emails from it, send an email to honyaku+u...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/honyaku/8b5c7a22-f4af-4a06-a67f-673ccbbb921f%40lmi.net.