How is "lightness" in the Black Speech? Is a translation question. As per Should questions saying "how do you say [thing] in [conlang]" allowed?

Yes, these questions should be allowed.

But there are already four close votes on it, one of them mine. Does this need more discussion? Do we want to avoid translation questions until out of private beta?

5 Answers 5


You may be making this too complicated — follow the lessons our successful language sites figured out long ago.

Acceptable questions contain a specific problem statement. But if someone asks how to say {x} in Klingtokiedwarfonto, get a dictionary. We're not here to do that work for you. This is not a translation service.

Where "Translation" Questions Do Work

If the author is having difficulty with an idiomatic expression or understanding the exact syntax of a situation, make sure the specific point of confusion is clearly stated in the question itself. Users often blithely express this as "what have you tried?" I'm not a big fan of dismissing users without further clarification, but you can politely ask for clarification and close a question if the question is clearly outside the purpose of asking experts to go out of their way to provide custom help for them.


I don't think we can avoid translation questions. To put it bluntly: we can't talk about constructed languages without talking about how things are said in them. An analogue might be if English Language Learners disallowed questions on how to say certain things in English, or if Stack Overflow only allowed high-level questions about programming, rather than code-based questions.

Pure translation questions might not be particularly interesting, but they're the bulk of what we do: translate. And it's undeniable that there are better and worse translations. Being uninteresting is no crime.

Constructed languages are dynamic and complicated, and it's not immediately obvious how to say certain things in certain languages. While we could orient the site towards purely high-level overview of how constructed languages tick, if we did that, we'd be missing something pretty significant.

Just as a few examples:

  • Certain languages may not have fully developed their vocabularies.
  • Certain languages may have minimalistic vocabularies that require grammatical construction, and are hard to parse if you're not familiar.
  • Certain languages may have complicated, non-obvious ways of forming words and others may have strange ways of forming sentences. How something is said is likely to be a major question one has about any conlang.
  • 2
    “Pure translation questions might not be particularly interesting, but they're the bulk of what we do: translate.“ I do not agree with this statement. The knowledge of what a particular word in a language by Tolkien is, and how to construct a language are entirely different skills. And seing a bunch of translation requests burying the interesting questions will only turn off experts in language construction, as they often cannot add to those questions. Feb 7, 2018 at 2:16
  • 2
    @Adarain Based on years of experience on Stack Exchange grappling with this exact problem, I think the "engineer to appeal to experts" line of reasoning is a pure logical trap. I've made that mistake in different forms on three sites now, and I'm starting to see why it's wrong. We should engineer the site to be what we, the users, would want to use it for. Nothing more, nothing less.
    – user23
    Feb 7, 2018 at 2:18
  • @Zyerah Just as long as you don't assume that users of this site want translation questions. That has to be proven.
    – curiousdannii Mod
    Feb 7, 2018 at 2:50
  • 1
    @curiousdannii It is proven by people asking them, and their quality is demonstrated through upvotes and downvotes.
    – user23
    Feb 7, 2018 at 3:34

How to say something in a conlang is important, but we need to beware lest the site become nothing more than a human-powered translation tool for obscure, minor or constructed languages.

Additionally, many "constructed languages" have a very small lexicon and grammar, and in many cases, requests for 'How do you say "x" in Y?' may not be able to be answered authoritatively by anyone but the language's author.

I believe that this site should be more about how to construct an artificial language than simple translation, so any requests for translation should be addressed as examples of language construction rather than discrete requests for translation, and where a question has already been asked about a particular aspect of a language, similar questions should be considered to be any that address aspects of the language that have been previously discussed, even if no one question addresses them all, and should be considered to be a duplicate if all aspects have been addressed previously, even if not all in the one question.

So, to make things clearer, I believe that pure translation should be off-topic, and allowable only if accompanied by a question of the particular characteristics of the language in question.

E.g. :

  • 'Please translate "x" into language Y' should be off-topic since there is little to be learned from it.


  • 'What is the antonym of "a" in "a b c" in the language D' should be on-topic because it can be accompanied by a discussion of how D handles antonyms and by extension inform readers of how other antonyms might be handled if they follow some regular pattern.


If the question shows no research, down vote it and move on. Of course, adding a comment that a simple Internet search would have given the answer, with a lot less typing, can be done as well. Otherwise, save your time, and attention, for questions you think deserve your attention.

Vote often, up or down as appropriate, and the simple questions will fade away, as they should. Some questions, after further reflection, seem like they do belong here. For example Right/Left side when it does not exist. Yes, I am one of the close votes, and now I'm thinking it is exactly the kind of question that does fit here. Not only is it exploring, and possibly developing an existing conlang. It also shows traps that can occur in the creation of a conlang for future creators. If/when the question is edited, it will be getting a reopen vote from me.

  • 2
    Does deciding by voting actually work in practice? I have no experience with StackExchange, but I can tell you that on reddit, it does absolutely not work and drives up all the low effort posts, burying the actually interesting ones, thus requiring a lot of active moderation. Feb 7, 2018 at 2:18
  • In the beginning voting will be semi-useful, mostly because of low user count. When a question, or an answer, receives enough down votes, or maybe a low enough overall score, it will fade (literally) and eventually disappear.
    – user4
    Feb 7, 2018 at 2:20
  • 2
    @Adarain Reddit has . . . other problems, as well, and we're pretty different. Sure, there are some side problems with voting - like the fastest gun in the west - but it does work, to a first approximation. We also have a much harder stance when it comes to moderation than Reddit.
    – HDE 226868
    Feb 7, 2018 at 3:38
  • @HDE226868 The moderation is the key, I believe, to the survival and health of every SE site. Witness the demise of Startups (twice).
    – user4
    Feb 7, 2018 at 3:47

I agree with Robert Cartaino that SE-Conlang is not a translation service. And while I agree with Adarain et al that, in general, translation questions ought to be off topic more than I disagree with the statement, I do think we need to retain some flexibility in regards to this kind of question.

The Toki Pona question referenced is a good example: it's not a pure "how do you say X" question, because the answer can not be "look it up in the dictionary". Rather this question, as I see it, delves more into how the language functions and the design parameters of the glossopoet.

The Black Speech question is also a good example because, well, there isn't a dictionary to look it up in. Also, online one can find many "reconstructions" and "cinematic additions" --- quasi-canonical, fan-generated languages and so forth for all the popular franchises. I think our expertise in the field as regards understanding & intuiting what might best be considered canon and what might best be called "fanciful additions to the canon" is well suited to this kind of question.

As for well settled, well known "standard" invented languages like Esperanto or LFN or Brithenig or Wenedyk (any language with an actual published dictionary or an accessible✳ online resource), sure we should be free to say "go look it up in the dictionary!" or "send an email to the author" and downvote or boycott or whatever we feel needful.

NOTE: webpages, especially older ones, have a tendency to disappear. If someone comes asking for help with a relatively well known invented language the online resource of which is now gone or hidden in some archive, I think it would be appropriate to a) give the immediate answer and b) point the seeker in the direction of said archive if there is one.

You must log in to answer this question.

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