There seems to be a new Japanese-to-English reviewer who insists that translations mirror the original word-for-word. I always put in the effort to make my translations sound natural and fluent in English, so it's disappointing when I'm penalized for that extra effort.

For Job #91128XXX, there were two sentences that explained how a product could be customized. To improve the flow in English, I moved one of the noun phrases from the second sentence to the first sentence. The resulting meaning was identical. This was marked as a "syntactic error" because "the corresponding source segment was included in the sentence that follows."

Rearranging words is not a valid reason to mark an error, so I filed a re-review request with the following comment:

>I intentionally rearranged the wording to improve the flow. This is not an error under Gengo's guidelines.
>"Being accurate in a translation means accurately conveying the message of the source, not the order of the written words or sentences."
>"Always aim to re-order sentences and don’t stick to the word order of the original."

Since the reviewer did not allege an error in meaning, there is not much more I can say. My request was rejected as "not processed:" "Insufficient explanation. There is a lack of congruence in reference/resource you have provided. If you wish to have your request processed, please provide legitimate resources to back up your dispute. "

Here is another example. For Job #905XXXXX, I used the sentence [omitted]. The reviewer marked this as an omission with the following reason: "This seems to be the match for [omitted]; although this broadly works in this sort of context where "this" can directly point to the relevant text, it would be safest to handle this with some kind of term match just in case (the client might edit in images between the line and subsequent text, or any number of other context changes could occur)." In other words, the reviewer wants me to translate it as something like [omitted]. But the colon already indicates that "this" refers to what follows. I did not bother filing a re-review request.

I would appreciate it if reviewers could be reminded to only mark clear errors under Gengo's guidelines, and to allow plenty of room for non-literal translations. I do admire the reviewer's thoroughness.


  • 9

    Hi Stefan,

    I work in the same language pair, and I believe I know which LS you're referring to.

    I think it's inevitable that reviewers have rather different approaches to reviewing translations, and I'm fine with that, but I definitely feel that said reviewer tends to be a little overzealous in going through every passage with a fine-tooth comb and marking every single spot that can possibly be improved on as an error.

    For example, I've been penalized by this reviewer in the past for:

    - proper nouns that in their opinion should not only have been transliterated but for which I should also have included an explanatory note

    - translating 水溶性 as "soluble" instead of "water-soluble" (even though that's implied when we say something is "soluble")

    - leaving out 約 (approximately) when translating a facility's distance from the nearest train station (even though no one would read "5-minute walk" and actually think that the facility is exactly a 5-minute, 0-second walk from the station)

    ... you get the idea ;)

    As I've said in the past, I firmly believe that there should be a hard distinction between an error (which I think of as something that is objectively unacceptable and would appear "off" to everyone conversant in both languages) and a plausible rendition of the source text that may not necessarily mirror what the reviewer prefers or has in mind as their ideal translation. Marking every instance of the latter can be helpful for translators from a pedagogical perspective, especially if they are marked as suggestions instead of errors, but the final score of the translation should be reflective of the overall quality of the translation.

  • 9

    The first example is truly baffling, especially since I've had my translations marked down multiple times for being "too literal". I'd like to hear some explanation from Gengo's Quality Control Team regarding their official stance on this.

    Perhaps @Rica Tero could look into it and provide us with some insight on this matter?

    The second example seems to be a clear-cut case where the "error" should have been marked as a "suggestion", since he/she admits that the translation "broadly works" in this context. In fact, I've had similar cases marked as a suggestion by other reviewers in the past, and it makes me curious what this reviewer thinks the "suggestion" category is for, if not for exactly this kind of situation. It makes you wonder what kind of onboarding training Gengo provides to these reviewers to ensure that their scoring is consistent with each other.

  • 13

    I’ve noticed that the reviews have been quite chaotic over the last year or so, and it’s hard to tell what exactly is going on behind the scenes since the reviews themselves are so inconsistent. I don’t know if it’s miscommunication about policy or the personal discretion of an LS.

    To start on a positive note, I’ve seen remarkable improvements in a large number of the reviews recently. The LS will give fair, well-reasoned criticism with insightful comments, while also leaving positive feedback regardless of the score. I’ve been on this platform since around 2014 and I hardly remember ever getting positive remarks till recently, even for 10/10 reviews. It’s helpful to know what you’re doing right just as much as what you’re doing wrong.

    Unfortunately, the issues you all have pointed out have been ongoing as well. A perfectly fine translation will come back riddled with “minor errors” after a review. Then once I take a closer look, I’ll notice that the LS has essentially rooted through the translation to find words with varying plausible translations, and then pick whichever term I didn’t use to mark it down as a minor error. There is typically a preference for literalness in whichever alternative term I’m given. I’ve even had alternatives presented to me that are so overly literal that they’re arguably worse than what I chose. 

    I’ve also had a few odd situations where the reviewer will simply say “This doesn’t sound right to me” about a sentence and leave it at that without any further explanation. This makes it virtually impossible to make an evidence-based counter argument against their criticism, which we all know is a requirement on our end when requesting a re-review.

    Obviously there are many cases where word choice matters, and not all synonyms should be seen as valid for every situation, but these particular “errors” often feel arbitrary and pointless given the context of the translation, or even the needs of the customers (who often put requests for 意訳 “non-literal” translations in the comments these days). To add to the confusion, I’ve had at least one somewhat recent review stating that I shouldn’t be too literal. I totally agree with this of course, but it makes for mixed messaging to say the least.

    Addition/omission errors play a similar role here. It’s standard industry practice to make additions and omissions when necessary, particularly in this language pair, so it’s bizarre to see reasonable alterations for readability marked down as errors due to some sort of extreme adherence to the source. Strict faithfulness to the source material may take precedence in certain cases, but a translation of a casually written business memo shouldn’t be treated as if it were a philosophical text for review purposes. Most of the work we do on here has to involve some degree of localization and modifying to achieve a natural-sounding final product, which means that the target text may need to be rearranged when appropriate, as does some amount of rephrasing or omitting to prevent awkward redundancies in the target text. It should also go without saying that additions are a fundamental aspect of Japanese to English translations due to the differences in syntax, writing styles, etc.

    These unfortunate cases do tend to be the minority of reviews, at least in my experience, but it’s still worrying considering the severe impact they can leave on your overall score. I should also say that one of my most recent reviews had positive comments about my choices and it was a non-literal translation, so I don’t want my remarks above to be seen as criticism for every single LS doing these reviews. It’s a noticeable problem, but it doesn’t seem to be universal.

    Lastly, I don’t want to make it sound like pointing out errors should be forbidden. Thoroughness in the reviews is great for quality control, and I’m all for that. I just don’t think the aforementioned tactics are the way forward in facilitating better translations on the platform.

    (Edit: Apologies, I didn't think my comment would end up being so long.)

  • 3
    Rica Tero

    Hi everyone! 

    Thanks for sharing this concern and some insights in the forum.

    @Stefan - We edited your post to delete some information about the jobs as it violates our forum guidelines, particularly the Translator Agreement. I hope everyone will be careful when sharing concerns related to jobs in the forum.

    I already talked with the Quality Team, and they are investigating this case. I will update here once I get their response to this concern.

    Thank you.

  • 4

    Ok, so it's not just me, then.

    I had similar reviews recently, and I bet it's the same reviewer. It makes me reluctant to take up some jobs, because this reviewer will nearly always reject the re-review request, and even if they grant it they refuse to admit that they're being ridiculous. They make (incorrect) interpretations of the source text (which I tried to explain that I actually know, since I was a regular translator for the customer), and then they reject a review on the basis that "we don't accept re-reviews for interpretation differences" or some nonsense to that effect.

    Please, do something about this reviewer.

  • 4

    Well, with curious timing I believe this same reviewer has just struck me yet again. They marked as an error one section and in their reasoning they *explicitly stated* that it was because it wasn't a literal translation.

    I'm not wasting my time asking for a re-review, I'm just going to go straight to support. This is unacceptable.