Say No to term "full post-editing"

[John Stroman]

Fellow 'Yakkers,

Below is a link to a 42 min YouTube interview by Slator entitled "Jakub Absolon, CEO of ASAP-translation.com, on the role of post-edited machine translation and why he thinks full post-editing is a misnomer."

https://www.youtube.com/watch?v=DjQiw-e_PEw

Here is the explanation from the weekly Slator newsletter.

Post-Editing Politics

Heard this week from a manager of a large in-house corporate language services unit: “I don’t like the term light post-editing and we don’t offer it.” While much of that unit’s volumes are going through a post-editing workflow, it’s not light but full post-editing.

Now full post-editing is a term that doesn’t sit well with Jakub Absolon, CEO of ASAP-translation.com, a boutique LSP and guest on this week’s SlatorPod. Absolon knows a thing or two about post-edited machine translation (PEMT), managing a business that has transitioned to PEMT over the past few years and having done research in the field since 2016.

Absolon argues that there are three types of translation: raw machine translation; PEMT, where the translator (or language expert) removes critical (but probably not all) errors and meets time and cost constraints; and human translation done leveraging the latest TB, TM, MT, LLM language tech. In short, Absolon says that full post-editing is simply human translation and should be priced and timed as such. 

Caveat: I have not had time to watch this yet, but intend to do so this afternoon. If the last sentence of Slator's explanation is accurate, it may provide human translators with a good argument when that will enable us to maintain or obtain sufficient rates for our services negotiating with customers.

John Stroman

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

[Carl Sullivan]

Team,

I am doing these--I would agree with this statement. It all benefits the company, and the expectations are for the translator to carry all the wood. 

Carl Sullivan

Former Administrator, Japanese Language Division, American Translators Association

--
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 on the web visit https://groups.google.com/d/msgid/honyaku/CAJVTx9781TqO1J-ErDzU7jzzvJL%2BQG47X5pygn2%3DX%2B1Eg%3D93%3DQ%40mail.gmail.com.

--

Carl T. and Masae Yonaiyama Sullivan

MasaCa Translation Services LLC

masacatransl...@gmail.com

MasaCaTranslations.squarespace.com

801-369-1501

CDR Carl Thomas Sullivan Sr., US Navy Retired

Disabled Combat Veteran

[John Stroman]

To follow up on my earlier post, although the interview is 42 min, the discussion about full post-editing per se begins at 15.32 and lasts until about 5 min before the end. I found the discussion interesting because ISO standards have a definition for full post-editing, and Jakub Absolon explains that in consideration of the time and expertise required, the final product requires the same functions as those performed by an experienced and productive human translator. Therefore full post-editing should fall under the category of human translation with no discount in terms of pricing. His main point is the assignment of liability for quality, especially in cases such as a machine instruction manual where an error in translation might lead to injury or death of the user. In that respect, only a human can ultimately assume responsibility for the quality of the final product.  When Absolon uses the term "domain" in the video, he means a specialized technical field that requires expertise beyond bilingual ability.

John Stroman 

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

[Jon Johanning]

I'm glad you watched it, John, because I stopped watching during the long initial run-down of the interviewee's life history and location, his company's details, etc., etc. I didn't make it through the whole 15 minutes. What a terrible way to start an interview!

I agree that, not only for documents that might lead to injury or death if mistranslated, but for all translations, frankly, responsibility needs to be assigned to humans, because only humans can understand language. But these days we seem to be living in a world in which most consumers of translations want quick and cheap, and accuracy be damned.

I'm doing a series of jobs for one agency which involve "post-editing" translations of medical case reports which were done by what appear to be native Japanese speakers who mostly get the technical terms right but make the standard errors in punctuation, number, etc. In many cases, they seem to be using DeepL and not correcting its errors. So in effect I'm doing full post-editing of DeepL. I am getting paid my rate for editing, but I'd rather be doing the translations from the start; apparently they are working for cheaper rates than my translation rate. You can't fault a business for wanting to pay its workers as little as possible.

Jon Johanning

[John Stroman]

Jon,Thanks for your input. The first 15 minutes tried my patience as well.

Since you mentioned technical terms in medical reports, I can relate the following: My experience with DeepL is that it tends to take medical terms (especially long strings of kanji), break them down, and then recombine them. This often leads to literally translated output or layman's terms rather than the kind of specialized terminology that a medical translator would use. For example, DeepL came up with "patients with bile duct cancer" for 胆管癌の患者, but not "patients with cholangiocarcinoma." The former is literally correct, but the latter is what an experienced medical translator would use based on the English abstract of the report that was supplied by the authors.

On a separate note, I have a sneaking suspicion that DeepL is monitoring my keystrokes when I copy/paste the DeepL output into Word and then rewrite it.  (I use the free desktop version). My only evidence is that the first time DeepL encountered the above Japanese, it spat out  "patients with bile duct cancer," but in subsequent encounters in the same file DeepL came up with "patients with cholangiocarcinoma." This kind of thing has also happened in the past. I'm not paranoid about it because that actually makes my job easier, but I'm curious if others have had a similar experience. Hints of Big Brother in Nineteen Eighty-Four.

John Stroman

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

[Jon Johanning]

Well, yes, you can fault them, and workers do it all the time. What I meant was: you have to expect them to economize by paying what they can get away with. Workers need to organize to fight that, but organizing is very difficult in this trade.

Jon Johanning

[John Stroman]

Fellow 'Yakkers,

As followup, below is the link to the results of a poll taken by Slator last week on use of the term "full post-editing."

https://www.datawrapper.de/_/5RttL/?utm_source=Slator+Newsletter&utm_campaign=6766afb622-Slator+Weekly+-+Dec+11%2C+2020_COPY_01&utm_medium=email&utm_term=0_28e0ad4ca5-6766afb622-161362853&mc_cid=6766afb622&mc_eid=fb7469eae2

John Stroman

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

--

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 on the web visit https://groups.google.com/d/msgid/honyaku/cc25a2ff-a21d-4b58-9bee-835edeb1ea18n%40googlegroups.com.

[Matthew Schlecht]

On Fri, Oct 27, 2023 at 9:36 AM John Stroman <stromana...@gmail.com> wrote:

Fellow 'Yakkers,

As followup, below is the link to the results of a poll taken by Slator last week on use of the term "full post-editing."

https://www.datawrapper.de/_/5RttL/?utm_source=Slator+Newsletter&utm_campaign=6766afb622-Slator+Weekly+-+Dec+11%2C+2020_COPY_01&utm_medium=email&utm_term=0_28e0ad4ca5-6766afb622-161362853&mc_cid=6766afb622&mc_eid=fb7469eae2

I listened to most of the video, and I'm not sure what all the fuss is about.

I think the market for "light post-editing" is dwindling or gone at this point. And, rightly so.

Just out of curiosity, I experimented with that a few times some years ago, and I found the experience distinctly unfulfilling.

So, the objection is that the "full" part of the phrase is superfluous?

The point is made during the video that possibly the label implies the final product quality to be somewhat less than that of a "human translation", but this is obviously not the case. Or, shouldn't be.

The translation product turned in by a (full) post-editor should not differ in quality from what an "ab initio" human translation would produce. The mechanics, sequence, procedures, and timing will differ, but the quality must not.

Matthew Schlecht, PhD
Word Alchemy Translation, Inc.
Newark, DE, USA
wordalchemytranslation.com

[John Stroman]

Matthew,

Well put. I agree. Since I leverage my work with MT and AI, but still rewrite nearly 100% of the output for internal consistency, etc., I still consider myself a translator in the traditional sense of the word.

John Stroman

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

--

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 on the web visit https://groups.google.com/d/msgid/honyaku/CALnYxcwsKjYwLL%2B0kzE0wAD%2BqGZ8HZ3BAjK0%2BHHZ%3D8gxLMQ4zw%40mail.gmail.com.

[Jon Johanning]

John Stroman: "Since I leverage my work with MT and AI, but still rewrite nearly 100% of the output for internal consistency, etc., I still consider myself a translator in the traditional sense of the word."

Applies to me, too. Anything someone puts out under (or over) their name has to be carefully reviewed and corrected by that person, so nothing  in the translation world can be produced purely mechanically. If this flat statement is untrue, I'd like to see a counterargument.

Jon Johanning