7

The topic of what to do with language identification requests has come up several times already. However, based on how often regulars here have been closing such questions, (A.) I took the liberty of formally adding it to our list of off-topic questions. There was a time when I defended the inclusion of such questions. However, there are two reasons why I have changed my mind:

  1. linguistics.SE has been running for a few years now, and it seems that we don't have experts or enthusiasts of philological decipherment like the ones found at Language Log (see my previous link). All that seems to happen here is that someone already knows the language in question, and not only is it a native language of millions, it's the predominant language of a current geopolitical region.
  2. I think it's boring and rather unproductive to identify scraps of "mystical" languages such as Hebrew, Sanskrit, Tibetan, etc. again and again. I mean, these languages may not be mystical to its native speakers, but scraps of these languages tend to travel the world motivated by mysticism, and end up in places where they'd not be commonly recognized. What does this community gain by identifying Tibetan three times [1] [2] [3], Hebrew twice [4] [5], Sanskrit three times [6] [7] [8], and so on?

This is my wording for (A.):

Identification of a language or a script based on specific samples and/or its translation; questions about the scholarly aspects of translation and language identification are still welcome

If there are no objections, (B.) I'll also modify the following close reason from,

Questions seeking help with translation and localization are off-topic here unless specifically concerning the linguistic reasons for different translations.

to,

Questions seeking help with translation, localization and identification of languages are off-topic here unless concerning the broad field of scholarly study of language. <-- stricken through following the discussion between lemontree and curiousdanii.

Questions seeking help with identification of or translation of specific samples of text are off-topic.

Please let me know your thoughts on both (A.) and (B.).

14
  • 3
    My thoughts are: hurray!
    – curiousdannii Mod
    Jul 13, 2017 at 4:13
  • 1
    How about just "specific texts" instead of "specific scraps of text"?
    – curiousdannii Mod
    Jul 14, 2017 at 4:59
  • 1
    I'd go for "specific words or texts", because translation requests for single words are also common. In addition, I'd rename "identification" by "language identification", otherwise it's not clear what would be meant by identifcation of texts. And maybe put "translation" first because that's the more common off-topic reason. I.e. "questions seeking help with translation and/or language identification of specific texts or words"
    – Natalie Clarius
    Jul 14, 2017 at 10:43
  • 1
    @lemontree I can change it to "specific samples of text" if "scraps" has rude connotations. RE: "identification", I meant "[identification and translation] of [specific [samples of text]]. I prefer it in this order because that's usually the order in which people seek information. In my examples, [1,2,4,6,7,8] asks for identification, [3,5] asks for identification and translation in that order. After all, translation comes after the identification phase.
    – prash Mod
    Jul 17, 2017 at 23:32
  • I feel uncomfortable with the current wording "identification of or translation of specific samples of text", specifiacally with the "of or" here. I think it sounds more natural and better without the "of" before the "or".
    – Sir Cornflakes
    Jul 25, 2017 at 12:11
  • @jknappen That was done intentionally. With the double "of" I intended to make it clear that "identification" is meant to be read in the context of "specific samples of text".
    – prash Mod
    Jul 25, 2017 at 13:24
  • @prash In my English, the "of or" construction sound almost ungrammatical and I find it hard to parse. An alternative is repeating "specific samples of text" that would be a little more verbose but clear.
    – Sir Cornflakes
    Jul 25, 2017 at 14:02
  • @jknappen I prefer it without the repitiion :) In fact, search for "[n*] of or [n*] of" on googlebooks.byu.edu/x.asp and you'll see that the phrasing not unusual at all.
    – prash Mod
    Jul 25, 2017 at 15:15
  • @prash: i tried that search and here is my result "lack or loss of" 2.046 vs. "lack of or loss of" 85 (for American English). It is ungrammatical.
    – Sir Cornflakes
    Jul 25, 2017 at 16:51
  • @jknappen 85 matches is evidence of grammaticality. You can't judge the opposite, ungrammaticality from ngram searches. Contrast "naturalization of or resumption of" (54 matches) with "naturalization or resumption of" (0 matches). Would you say the latter is ungrammatical based on this evidence? Naturally, the former reads better to me.
    – prash Mod
    Jul 25, 2017 at 17:07
  • Just adding another aspect to it, should someone/ we open a 'Language Fans' / 'Translation/Identification' site (or an off-shoot of Linguistics SE if that's possible) ? so that all existing questions are not removed but migrated there.
    – WiccanKarnak
    Oct 1, 2017 at 1:32
  • @WiccanKarnak you can propose it at area51.stackexchange.com
    – prash Mod
    Oct 1, 2017 at 1:56
  • @prash i know, but I wouldn't make a case there as I am not a supporter of these types of questions and haven't really asked one myself. I was just concerned about the volume of questions that would be removed.
    – WiccanKarnak
    Oct 1, 2017 at 2:00
  • 1
    @WiccanKarnak It's hard to find supporters of such questions. These questions are usually asked by someone who sees a scrap of an unknown script, create a SE account to ask the question and disappear, never to be seen again, on any SE site. You'll get a general sense of this if you look at the karma scores of the people who asked the questions [1-9] that I linked above. I feel no regret at the volume of such questions that are removed.
    – prash Mod
    Oct 1, 2017 at 2:09

2 Answers 2

Reset to default
1

You make a good case for it, and I was finally convinced after reading user6726's answer. While I'd be fine with a properly scoped identification question, such as How to identify a foreign language from handwriting?, simply asking "what script is this" might not be a question of particular value.

Plus, I realized time ago that allowing various degrees of off topic questions just to improve our stats in the short term would have been detrimental to the growth of the site in the long term.

Therefore I personally agree with you and I'm fine with using your wording. Besides, we can always adopt a different policy if we deem it necessary.

I have a couple of suggestions:

My first one (which might have been your intention all along) is that we give some time to other users to post here to consider all perspectives in the matter. I'm sure everyone agrees that a wider consensus is always a good thing, and it's also better to find problems or flaws (if any) sooner than later.

Secondly, I think it'd be better if we adopted a case by case approach, so not just blindly closing questions that regretfully happen to have the tag, but rather evaluating - as we usually do - whether the question deserves to stay open, or whether it can be edited.

3
  • I agree with both your suggestions.
    – prash Mod
    Jul 13, 2017 at 10:43
  • I have changed the wording for both (A.) and (B.). Do you approve of the new wording too?
    – prash Mod
    Jul 13, 2017 at 20:04
  • Since it has been over a week, and the topic seems to have stabilized now, I have added the part about (B.) to the off-topic list. I need either you or @lemontree to approve it.
    – prash Mod
    Jul 21, 2017 at 8:13
1

I fully agree with user6726's opinion, in addition to the reasons that bytebuster and you mentioned.
I am therefore much in favor of both of the actions (A.) and (B.) - especially since this now finally formally catpures a policy that has implicitely been the common practice for a while, without it being clear whether this is actually the officially approved consensus to deal with such questions.

I have one improvement suggestion to make, namely that "the broad field of scholarly study of language" seems a bit vague to me. Would "the scientific aspects of these fields" be suitable?

BTW, what exactly is meant by "localization"? I had somehow thought this was used synonymously to what we here title "language identification", but I am now realizing it could also refer to adaption of a text to some (local variant of a) language as in "software localization"?

Not only for this on-/off-topic constraint, but also for other ones, we could also consider including examples of "good" and "bad" questions, so newbies get a better feeling of what kinds of questions are appropriate and what not. In order not to blow up the help site too much, one could do so by including this in some meta post discussing the issue that is being linked to from the rule (as is already done for some of the bullet points).

12
  • 1
    I like your suggestion of "the [X] aspects of these fields", because it would then be clear that we are talking about the three aspects previously specified. However, I'm reluctant to use the word "scientific" for [X], because I think it's too constraining. Instead, if we use "scholarly", it would mean we're taking about kinds of discourse that students, researchers, and practictioners of these fields deal with in their professional capacity, whether it be through publishing papers, collegial discussions, field work, classroom interactions, etc.
    – prash Mod
    Jul 13, 2017 at 9:58
  • @prash I'm okay with that.
    – Natalie Clarius
    Jul 13, 2017 at 10:02
  • You're right about "localization". The reason for clubbing all three aspects under one "close reason" is that linguistics.SE is still in beta stage, and we have already hit the maximum number of "close reason" items we could have.
    – prash Mod
    Jul 13, 2017 at 10:02
  • What am I right about - that it is ambiguous?
    – Natalie Clarius
    Jul 13, 2017 at 10:03
  • Sorry, I mean it's got to do with software localization, signboard localization, etc. That wording precedes my moderator status, so your understanding is probably as good as mine.
    – prash Mod
    Jul 13, 2017 at 10:06
  • 1
    Since "localization" is ambiguous but in any case falls either under translation or language identification, shouldn't we just remove this word?
    – Natalie Clarius
    Jul 13, 2017 at 11:26
  • 1
    Students study languages in formal scholarly contexts - are you thinking their questions should be on-topic? I'd rather not have any straightforward translation questions. I'm not sure we should really even have more involved ones, except as examples for questions about translation methodology and philosophy.
    – curiousdannii Mod
    Jul 13, 2017 at 12:39
  • @curiousdannii That's a reasonable point. Maybe the term "professional" (not exclusively scientific, might also be in professinal translation for example, but ruling out student homework requests) is a better fit? .
    – Natalie Clarius
    Jul 13, 2017 at 13:23
  • @lemontree Even then we don't want questions where the answer is a translation. I think we really need to get a list of actual past translation questions we do want to be on-topic because it's just too vague right now.
    – curiousdannii Mod
    Jul 13, 2017 at 13:29
  • @curoiusdanni Then what do you suggest as a close reason formulation? We would need this independently of a (certainly helpful) example list.
    – Natalie Clarius
    Jul 13, 2017 at 13:31
  • @lemontree Without good counter-examples I'd prefer a blanket statement like "Questions seeking help with translation and identification of languages are off-topic." I don't think we need to specify that translation theory is on-topic in the close reason.
    – curiousdannii Mod
    Jul 13, 2017 at 14:08
  • 2
    @curiousdannii I have changed the wording for both (A.) and (B.).
    – prash Mod
    Jul 13, 2017 at 20:04

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .