Wiktionary talk:Translations

=Archives= Earlier page: Language considerations

Archive of discussion copied from Beer parlour - 22-May-2005=== then modified.

When a discussion becomes inactive, move it if necessary and link to it below.


 * Wiktionary talk:Translations/Whether non-English words should appear in English Wiktionary or not
 * Wiktionary talk:Translations/If present, what foreign word articles should contain
 * Wiktionary talk:Translations/Hiding translations
 * Wiktionary talk:Translations/Usage Notes
 * Wiktionary talk:Translations/Link to a PART of a foreign Wiktionary article
 * Wiktionary talk:Translations/Homographs
 * Wiktionary talk:Translations/Policy policy
 * Wiktionary talk:Translations/Multilingual Meta Policy
 * Wiktionary talk:Translations/"Translations to be checked" - a proposal
 * Wiktionary talk:Translations/Translation of meanings is misleading

Following are forked discussions. When one becomes inactive, remove it from here and move just its link up above.

= Wiktionary talk:Translations/Translation into lemma only =

= Wiktionary talk:Votes/2007-10/Translation into lemma only =

= Wiktionary talk:Translations/Noting lemma forms in WT:ELE =

= Wiktionary talk:Votes/2007-10/Lemma entries =

= Wiktionary talk:Votes/2007-11/Lemma entries 2 =

=Other discussion=

Including grammatical gender
Hello, everyone. I would like to include a link on this policy to the policy about including gender with language translations. (See Languages with more than one grammatical gender.) This policy about the inclusion of gender could use some discussion, as well. Anyone who can comment on this policy should please leave a message on its talk page. Thank you.--El aprendelenguas 21:58, 17 April 2006 (UTC)

Managing complex translations
Most good foreign dictionaries will contain explanations on certain words such as "if" or "the" which are difficult to tranlsate into foreign languages (especially those of non-Indo-European origin). However, there is no room to do that in the present Wiktionary style for translations.

I propose having complex translation entries look something like this:

if
 * Japanese: もし…なら、たら、ば、と (moshi... nara, tara, ba, to) (more)

where the if/jp page would provide a detailed description of how to translate English "if" into Japanese, in English. Thoughts?

Lots of points to consider
Here are some points that might need to be included in a policy like this:


 * As far as Wiktionary jargon goes, language and dialect are synonymous. The more common languages, such as the Chinese dialects, have standard names to appear in language headings and translations sections.
 * In full dictionary entries, Translingual and English are listed first, in that order, and all other languages are alphabetized.
 * In translations sections, all languages/dialects are listed alphabetically at the same level.
 * Inflections should be provided in the entry for the foreign word, not for the English equivalents. Short, irregular forms that do not share a root are exceptions.
 * Gender labels apply to the grammatical gender of the word, which varies between languages and is defined in the appendix.
 * Transliterations are parenthesized and unlinked, and for each language Wiktionary chooses only one standard for the Translations section. This is generally the most common transliteration in modern use. If there is some ambiguity, those that have the simplest diacritics are preferred.
 * They aren't pronunciations, for goodness sake! Only in rare cases should additional information be given, such as when no gender-neutral term exists. (See cousin.)
 * For foreign entries, each sense of the word or phrase should have a separate definition line, just as it would on the foreign-language Wiktionary.
 * Usually translations match fairly closely. Even so, we need to be sure to indicate which sense of the word is meant if there could be ambiguity. (More often than not, there is.)
 * It's necessary to be more thorough when translations do not match. In some cases there may not be a direct translation of the foreign word. In the opposite case, a single sense of the foreign word may apply to more than one sense of the English word.
 * Any standard transliterations may be entered as entries, not just the ones we prefer.

And for Christ's sake, could we please archive this page! I mean, that this is a multilanguage dictionary is pretty much set in stone right now, unless we were to just throw the better half of it away. Why do we have to draw attention to that aspect by making the page live again? DAVilla 04:01, 14 October 2006 (UTC)

Some lay-out questions from a beginner
Copied from the Tea Room 20 Oct 2006

Hi all,

I recently discovered Wiktionary and am enthousiastically adding Dutch translations wherever I know them. However, I come across big layout differences between pages. So here are some questions: Expect more questions of this sort. Are there more explicit guidelines? henne 15:54, 19 October 2006 (UTC)
 * In disambiguating between different meanings, the translations are listed in several, , groups.  Often, the meaning is mentioned above it in bold, sometimes it is given as a parameter to  .  Which is preferable?
 * Sometimes language names are themselves wikified. Is this desirable?
 * In Dutch, a lot of words are both masculin and feminin. Can I make my own  template?


 * The translation sections should be disambiguated by a bold heading at the top of each group. Older entries have numbers, which should be phased out.  The new system used in some entries uses a different, ,  format whereby the heading exists as a parameter to .  These templates will probably supersede existing layouts.  Language names should only be wikified if they are unusual or not generally known – that obviously doesn't apply to Dutch.  As for gender, I'm not aware of any m/f template, so I see no reason why you shouldn't create one.  There is some more info at WT:ELE and Translations.  Widsith 16:36, 19 October 2006 (UTC)


 * 
 * The use of a parameter in is still considered "experimental" at this point.  It would be better to follow the bolding convention until the new format has the various kinks worked out.
 * Language names that might be construed as "exotic" are wikified. Basically, if the country name doesn't match the language name, it might be considered "exotic" to an English speaker.
 * Please use for those.
 * Entry layout explained goes into pretty good detail regarding the current conventions. This is a fine place to ask similar questions.  --Connel MacKenzie 16:40, 19 October 2006 (UTC)


 * Thanks both.  is what I was looking for, as is the template index.  How can I know where the discussion about the  or  templates is leading? henne 10:38, 20 October 2006 (UTC)


 * There is also for "common gender", which I know is standard in Swedish, and I have heard used in reference to Dutch as well. --EncycloPetey 18:20, 20 October 2006 (UTC)


 * Well, Dutch dictionaries usually avoid the problem in that they only say ‘de’, i.e., the common article for m and f, and if it is explicitly m or f, then they add (m) of (f) after ‘de’. I am not sure what the right thing to do is.  You are probably right that c is more appropriate, as they are referred to by masculine pronouns.

Ok, one more style question that wasn’t answered by the ELE: When subdividing the translations into groups corresponding to different meanings, a description of the meaning is to be given. Is there a style guide for this? Should it be concise, telegram-style, should it contain articles, should it start with a capital letter?


 * Not yet, but there's active discussion beginning on Translations to work out basic policy and standards, which could then be used to update/modify the ELE. --EncycloPetey 18:20, 20 October 2006 (UTC)


 * Is that right? I was under the impression that we have always recommended entering a short gloss, containing enough common words from the definition above to link them unambiguously.  --Connel MacKenzie 18:25, 20 October 2006 (UTC)


 * That's right, but there's still no written style guide. I'm thinking that I may help DAVilla assemble information on the Translation section, including format, style, and content information. She's already assembled a great new page for handling the Translations to be Checked, but we still need more than we currently have on the standard section itself.  And what relevant info we have on the ELE is not well organized, in addition to missing some key points. --EncycloPetey 19:17, 20 October 2006 (UTC)

Templates and
I have taken liberties with these. They both do almost exactly the same thing, so trad is now a redirect to t.

Note that no policy decision on the use of these has been made. Some people like them, but we may end up subst'ing the lot.

Template t takes three parameters: the word in language, the language code and the gender. The gender must be one of f, m, mf, c, or n. (Note that indicators such as pl are not supposed to be in the tables, the pointer is to the singular lemma form. All other information is at that entry in the en.wikt, and in the other wikt!) Robert Ullmann 12:58, 14 November 2006 (UTC)


 * strike that, we have lots of plurals, and there isn't any present reason to make them go away; 4th parameter is s or p. Robert Ullmann 15:27, 14 November 2006 (UTC)

Community Portal
Shouldnt the community portal be linking to this page instead of the translation of the week?

Bearingbreaker92 03:43, 21 January 2007 (UTC)
 * No, it's the project of the TOW that should be linked. However, the link name should be changed to be less confusing. I'll take care of that. --EncycloPetey 03:47, 21 January 2007 (UTC)


 * Thanks, it looks better. Bearingbreaker92 04:21, 21 January 2007 (UTC)

Improvements to the policy page
I suggest add in Translations, 3:


 * This yields the pleasing result of:


 * French: cacher les traductions
 * Japanese: 訳語を隠す yakugo o kakusu


 * Korean: 역어를 잠추기 yŏgŏ rul jamch'ugi
 * Portuguese: esconder as traduções

Add a new 4 point :

4. If no sure about some translations, use:

=====Translations to be checked=====

Finally, include these links in the see also section:


 * How to check translations and Template:checktrans
 * Modular Wiktionary
 * Machine translation.
 * Policy - Transliteration
 * Translations of the week
 * Translations/Wikification
 * Languages with more than one grammatical gender
 * WikiSaurus: WT:ELE
 * Translation
 * Schemes.

--Mac 16:25, 4 March 2007 (UTC)


 * This looks good, and then we can finally do away with translations.--Williamsayers79 23:59, 16 May 2007 (UTC)

Translations for verb forms?
Hi, I don't remember how I got this information but, whatever, I've been believing that there should be verb translations only on the page of the infinitive form (for says, see say; for is, see be), so I have been removing them, informing "verb translations only on the page for the infinitive". So, what's the established policy here to be followed? Or is there any? -- Frous 11:38, 10 June 2007 (UTC)
 * Yes mostly, only the lemma form of the verb should have a translation to English. Other inflected forms should identify the inflection and point to the lemma.  However, the lemma is not always the infinitive.  In Latin and Ancient Greek, the lemma is the first-person singular present active indicative.  Part of the reason for this is that Classical language dictionaries use that form as the verb headword, another reason is that Latin has six different infinitive forms (and the infinitives tend to function grammatically as nouns). In any case, there shouldn't be translations given for non-lemma forms of any verb, noun, adjective, etc. --EncycloPetey 16:34, 10 June 2007 (UTC)
 * Sorry, I misundertood your question. You were talking about Translation sections on English pages, yes?  In that case you are correct.  A translations section should only appear on the page for the English infinitive, since that is the lemma form for English. --EncycloPetey 16:36, 10 June 2007 (UTC)


 * Yes, I meant the English verbs. So I haven't done that much disaster when deleting translations of e.g. third-person singular present indicative forms, thank God... ;) -- Frous 20:37, 10 June 2007 (UTC)

Rename
I would call this plage Wiktionary:foreign word definitions --77.210.55.105 06:57, 6 October 2007 (UTC)

Animate noun gender inflection

 * The following was moved here from Wiktionary talk:Votes/2007-10/Translation into lemma only. Rod (A. Smith) 18:10, 19 October 2007 (UTC)

The following references support the classification of words like as a non-lemma forms of lexemes with lemma forms like. If anyone can find some references that support the opposite position (i.e. that words like are distinct, gender-specific lexemes), please add them. Eventually, some of these may befome references list for Languages with more than one grammatical gender: Rod (A. Smith) 01:43, 18 October 2007 (UTC)
 * 2003, Luis D. Casillas Martínez, Gender Mismatches in Spanish and French N1/A de N2 Affective Constructions: Index agreement vs. Morphosyntactic Concord :
 * An important distinction that I must make before delving into these constructions at depth is that between inherent gender classification and gender inflection.
 * An inherently gendered lexeme comes from the lexicon with a fixed gender value. Most inanimate common nouns in Spanish and French are of this kind—but there are exceptions.
 * An inherently ungendered lexeme does not have lexical gender. It may be inflecting, with a form for each gender, or noninflecting, with a unique, gender-unselected form.  Many animate nouns are not inherently gendered, and show distinct inflectional forms; e.g. Sp. amigo ‘friend (.M, .F)’.  Some ungendered adjectives and nouns don’t inflect (e.g. Fr. imbécile and idiot).
 * 2004, Spanish Concise Dictionary, Harper Collins, Grammar reference page 192:
 * As in English, male and female are sometimes differentiated by the use of two quite separate words, e.g.
 * mi marido  mi mujer
 * my husband  my wife
 * un toro una vaca
 * a bull  a cow
 * There are, however, some words in Spanish which show this distinction by the form of their ending:
 * Nouns ending in -o change to -a to form the feminine → [1]
 * If the masculine singular form already ends in -a, no further -a is added to the feminine → [2]
 * If the last letter of the masculine singular form is a consonant, an -a is normally added to the feminine* → [3]
 * 2005, Greg Kobele, Agreement Bottlenecks in Italian :
 * :Nouns in Italian also vary their form depending on their number. Moreover, there are also roughly two classes of nouns, with respect to the forms they take in the singular and the plural. A class I noun is one which inflects like a class I adjective (holding gender constant), and a class II noun inflects like a class II adjective. (striken as unclear Rod (A. Smith) 23:16, 19 October 2007 (UTC))
 * 2007, Ricardo Bermúdez-Otero, Against nominal class features in Spanish :
 * §14 Theme vowels are involved in exponence of gender, but are not predictable from gender:
 * e.g. masculine nominals belong to the o-class by default, but the o-class also contains
 * feminine nominals man-o (F) ‘hand’
 * dual-gender nominals el testig-o (M), la testig-o (F) ‘the witness’
 * neuter demonstratives est-o (N) ‘this’, cf. est-e (M)
 * feminine nominals belong to the a-class by default, but the a-class also contains
 * masculine nominals dí-a (M) ‘hand’ [sic. Bermúdez-Otero gives “hand” instead of “day”]
 * dual-gender nominals el artist-a (M), la artist-a (F) ‘the artist’


 * The 2003 cite is quite clear, and the 2004 cite almost as clear.


 * However, I think you're misreading the 2005 cite; I think it's implying that a noun has exactly one gender. This is more clearly implied later on, where it says "In GAgr, each inflectional form of every adjective and noun is treated as a separate lexical item, which means that for every adjectival root there are four lexical items, and for every noun root there are two." (I think it's clear that at this point, the author considers adjectives to have four forms — ms/fs/mpl/fpl — and nouns to have only two, either ms/mpl or fs/fpl.) However, the article later goes on to say, "This grammar allows us to view class I pairs like zio/zia (uncle/aunt) as inflected forms of the same lexeme, zi." In other words, while the author by default considers these to be pairs of nouns, he says that his proposed approach makes it possible to treat them as single nouns. All told, I think this can be considered a reference for both points of view.


 * Similarly, I don't think the 2007 really goes either way; it uses the phrase "dual-gender nominal" to denote nouns that have a single form regardless of gender, and the phrases "masculine nominal" and "feminine nominal" to denote nouns that have a single form for a specific gender (regardless of whether there's an opposite-gender counterpart). Unless I'm missing something about the word "nominal", the implication seems to be that "artista" is one word regardless of gender, while "chico" and "chica" are separate words; but it's not a terribly strong implication, as, due to the topic of the paper, the author might simply have chosen the interpretation that lent itself best to his purposes (as should we do).


 * —Ruakh TALK 18:49, 19 October 2007 (UTC)


 * Hmm. I was surprised to ready your interpretation of the 2005 quotaton, but rereading the work, I see that the author isn't exactly clear about “whether grammar should be able to describe such ‘meta-paradigmatic’ relations”, saying such questions are “best resolved by the ability to account for psycholinguistic data”.  In the area I and you quoted, he is discussing the general case of nouns (giving  and  as examples, which he later describes as “inherently gendered”) in the context of why GAgr is “in a sense the worst possible case, and thus [only to] be adopted after all other avenues are explored.”  He doesn't even mention animate nouns until further on, where he says, “we can represent non-inherently gendered nouns like zi in the manner shown below.”  In the grammar he seems to prefer, he views  and  as two gender-inflections of *.  In any event, he is not terribly clear, so I have stricken the example.


 * In the 2007 example, though, I cannot credibly imagine that "dual-gender nominals" to mean "pairs of lexemes that vary in gender". Rather, it almost certainly means "individual lexemes, each of which has two genders".


 * Since I struck the 2005 example as ambiguous, I found another couple of references:
 * 1977, William J. Ashby, Clitic Inflection in French: An Historical Perspective ISBN 9062034691, page 14:
 * For inanimate nouns, which are arbitrarily either masculine or feminine, the use of the determiner to mark gender serves only a classificatory purpose. For some animate nouns, however, gender is more meaningful, in that it may reflect the sex distinction.  For a number of animate nouns there exist both masculine and feminine forms, which are distinguished by contrastive pairs of suffixes or by the addition of a suffix for one of the forms only. Examples are le fermier (masculine) - la ferière (feminine); le marquis (masculine) - la marquise (feminine).
 * 2002, Mark Harvey, A Grammar of Gaagudju ISBN 3110172488, page 149:
 * As is virtually universal in systems which mark gender, human nouns take concord according to the gender of their referent. Thus, a noun such as biibi ‘MF, MFZ’ will take Class I or Class II concord depending on the gender of the referent.
 * 1999, Francis Cornish, Anaphora, Discourse, and Understanding: Evidence from English and French ISBN 0198700288, page 129:
 * Gender, unlike number, is a lexical category for which most nouns in French are inherently specified. The a French NP's gender value derives from that of its head noun. Gender is therefore a category which, although ostensibly arbitrary—at least in the case of inanimate nouns—is related to the sense of the lexeme for which it is marked,15 and only derivatively to its reference; or, more accurately, to the reference of the entire NP of which the noun in question is the head, since nouns cannot on their own refer, but only denote.  The main exceptions to this generalization consist of human-denoting nouns whose gender is not lexically fixed, but is rather determined by the lexeme in question's occurring as head of an NP used to refer to a male or female person: examples are nouns like concierge ‘caretaker’ and secrétaire ‘secretary’.
 * I also found this reference, which supports the position that in German, the gender-differentiated animate nouns are distinct lexemes:
 * 1997, Marion Kremer, Person Reference and Gender in Translation: A Contrastive Investigation of English and German ISBN 3823349376, page 86:
 * Most human lexemes are gender-differentiable in German, whereby the male-denoting and grammatically masculine lexeme is typically the morphologically simpler member, and the feminine counterpart is a more complex derivative marked by the suffix -in (cf. (Id)).
 * Anyway, it's obviously not a cut and dry area of lexicography (or psycholinguistics for that matter). We should probably just consult existing dictionaries (both monolingual and bilingual) and follow the majority treatment for each given language.  Rod (A. Smith) 23:16, 19 October 2007 (UTC)


 * Re: 'In the 2007 example, though, I cannot credibly imagine that "dual-gender nominals" to mean "pairs of lexemes that vary in gender". Rather, it almost certainly means "individual lexemes, each of which has two genders".': I completely agree with this statement, but I'm afraid I don't see what point it makes? He uses "dual-gender nominals" in reference to things like "artista" (which can be either masculine or feminine), but not — so far as I can tell, anyway — in reference to things like "chico–chica" (where each has one gender). But regardless, to me this seems irrelevant: there's a lot of variation, as different papers use the interpretation that best serves their purpose. Wiktionary, too, should use the interpretation that best serves our purpose. —Ruakh TALK 00:19, 20 October 2007 (UTC)

Calques/word-for-word translations?
How should one indicate or notate calques (word for word translations)? For instance, staircase wit is a calque from French l'esprit d'escalier, hence I've listed the etymology as "Calque of French l'esprit d'escalier." Is this exemplary, or is a different form preferred?

If so, we should gloss & template calque, as it is technical.

Nbarth 02:19, 27 January 2008 (UTC)
 * Saying "calque of/from ... " is just fine. --EncycloPetey 23:24, 29 April 2008 (UTC)

Wikilinking to foreign phrases
In translation sections, sometimes the foreign translation of an English term is a phrase. Sometimes that phrase is a Sum of Parts where an entry shouldn't exist for the whole phrase, just for each word in the phrase. In this case, we should wikilink each word instead of the phrase (See comment WT:RFD). Should someone just use plain links, or ? I'd like to mention this general case in the main page, as whole phrases will get picked up by User:Tbot (or clicking newbies) and made into their own articles that shouldn't exist. --Bequw → ¢ • τ 23:15, 29 April 2008 (UTC)

Chinese translations
I created two templates recently:
 * produces: trad., simpl. (pinyin: Zhōngguó)
 * produces:

Please consider the suggested methods for Chinese Mandarin:

2 examples of translations with the suggested result:
 * Chinese: trad., simpl. (pinyin: Zhōngguó) (this entry has different traditional and simplified characters)
 * Chinese: (this entry has identical traditional and simplified characters)

I suggest to use "Chinese" as a default language for Chinese languages/dialects. Everything else, separate by *:, e.g. *: Cantonese, *: Min Nan, etc. Anatoli 04:08, 17 April 2009 (UTC)


 * We don't have a language header called ==Chinese== so why would we want to list translations under that word? If the language header is ==Mandarin== then the translations should likewise be listed under * Mandarin . All of the languages/dialects are listed alphabetically, not grouped topically. There are languages much more closely related than Min Nan is to Cantonese that are listed separately. Why does everyone who enters Chinese translations have to think that they're an exception to that? DAVilla 06:26, 17 April 2009 (UTC)


 * Wiktionary like any dictionary works with the written forms. If I need a translation into Chinese, then it is in Chinese characters, identical in 95-100% across all the Sinosphere. Once the characters are given, then subgroups may follow under the heading Chinese: as :Cantonese, :Min Nan. The characters will be the same in most cases, if pronunciation is known can be given. So, 中國 and 中国 is in Chinese, applies to all dialects. Mandarin is the default standard pronunciation in China and Taiwan, that's why I suggest to omit the word Mandarin. The large differences in pronunciation across the dialects are less of imprortance, if the written form is the same and they have identical entries.

Currently:


 * Chinese:
 * Cantonese: entry in Chinese characters (Cantonese pronunciation)
 * Mandarin: entry in Chinese characters (Mandarin pronunciation)
 * Min Nan: entry in Chinese characters (Min Nan pronunciation)
 * Wu: entry in Chinese characters (Wu pronunciation)

OR less often:

...
 * Cantonese: entry in Chinese characters (Cantonese pronunciation)
 * Mandarin: entry in Chinese characters (Mandarin pronunciation)
 * Min Nan: entry in Chinese characters (Min Nan pronunciation)
 * Wu: entry in Chinese characters (Wu pronunciation)

Suggested:

...
 * Chinese: entry in Chinese characters (pinyin this is based on Mandarin or Standard Chinese)
 * Cantonese: Cantonese pronunciation, add Chinese characters in Cantonese if different (rare)
 * Min Nan: Min Nan pronunciation
 * Wu: Wu pronunciation

This is not unsimilar to, even if the Arabic dialects usually differ more in writing, compared to written Chinese:


 * Arabic:
 * Egyptian:

Anatoli 06:47, 17 April 2009 (UTC)

Do not use these templates. Use template properly.

Do not enter anything on a line starting with "Chinese:", "Chinese" is not a language, it is only used to introduce the grouo (if used). Use "Mandarin"

Language lines within a group start with ** not *: (and use the full language name, e.g. would be "Egyptian Arabic" above)

If grouped, it is like this (the same standard as all other languages).

* Chinese:
 * Cantonese: (word)
 * Mandarin: (word)
 * Min Nan: (word)
 * Wu: (word)

Do use : and :  as qualifiers where needed.

Translation sections in non English entries (Translingual)
The WT:ELE states that "Variations for languages other than English... the translations section should be omitted."

The About Translingual says nothing about translation sections.

My parser have been found several words with the translation box within Translingual section, namely: @, Felidae, unununium, Iris, Hymenoptera, Ericaceae, =, Cetacea, Bluetooth, Corvidae, Lycopodiaceae, Equidae, E440, SAAB, titanium oxide, Russula adusta.

So I have the following question: should we remain it "as is", or these entries should be changed? -- Andrew Krizhanovsky 09:25, 30 March 2010 (UTC)
 * I guess the logic is that Translingual entries are by extension usually English words as well. But since they're translingual, in theory the translations should mostly be the same word! @ is blatantly wrong however, translations moved to at-sign. Mglovesfun (talk) 09:39, 30 March 2010 (UTC)


 * Thanks for the fast reply. -- Andrew Krizhanovsky 10:04, 30 March 2010 (UTC)

source-targetLanguages (or sclang)
We could use for translations in Wiktionary. --Diamondland 10:37, 19 November 2010 (UTC)

Options
I suggest an option in User's preferences, to show mainly the translation(s) to an specific language, the default elected by the user in the options.--Diamondland 10:39, 19 November 2010 (UTC)

Qotations
I suggest the user can also translate to a target language the quotations related with an specific entry.--155.54.178.240 07:43, 27 November 2010 (UTC)

Translation subdivisions
With the ongoing discussion at WT:BP, I think it would be good to see if we can add a section on this page about subdivisions of languages. It would be good to have common or desired practice in writing after all. Right now, the following languages are commonly split (I may have missed some):


 * Albanian
 * Gheg
 * Tosk
 * Arabic
 * varieties of Arabic
 * Chinese
 * Cantonese
 * Mandarin
 * Min Nan
 * etc.
 * Norwegian
 * Bokmål
 * Nynorsk

I propose to add the following (at least), but we'd eventually want to include all possible cases? Please suggest more!


 * Armenian
 * Old Armenian
 * Dutch
 * Middle Dutch
 * Old Dutch
 * English (older forms) - should this always be the first translation listed, in the same way that we list English first on pages?
 * Middle English
 * Old English
 * French
 * Middle French
 * Old French
 * Anglo-Norman? - I don't think this belongs here... but maybe someone else does?
 * Frisian - Like Chinese, not a language!
 * North Frisian
 * Saterland Frisian
 * West Frisian
 * Old Frisian
 * German
 * Middle High German
 * Old High German
 * Irish
 * Middle Irish
 * Old Irish
 * Primitive Irish
 * Low German
 * Middle Low German
 * Old Saxon
 * Occitan
 * Old Provençal
 * Persian
 * Middle Persian
 * Old Persian
 * Portuguese
 * Old Portuguese
 * Spanish
 * Old Spanish
 * Welsh
 * Middle Welsh
 * Old Welsh

How is this? 16:29, 11 November 2012 (UTC)


 * I think that Anglo-Norman certainly does belong under French, since it was always was and still is frequently referred to as "French". --WikiTiki89 17:08, 11 November 2012 (UTC)
 * I can understand that, but it might confuse users because it doesn't have French in the name so they won't expect to find it there. When looking for candidates I always wondered to myself "is it likely that someone will look under one of the other words in the name"? So Old French might be considered "French, Old", but with Anglo-Norman that isn't so clear. Then again, maybe Old Provençal shouldn't be listed there either, nor should Old Saxon... 17:21, 11 November 2012 (UTC)
 * But people who are looking up translations for Anglo-Norman, Old Provencal, and Old Saxon are very likely to know of their connection to French, Occitan, and Low German, respectively, and will likely attempt to look there. Also, if it is not their first time looking for it, they will already know where it is and finding a subsection is much easier than searching through a list of Old XYZs that are all alphabetized to the same place. Another thing is that most of these languages have a plethora of other names and the only constant is their modern-day equivalent. --WikiTiki89 17:38, 11 November 2012 (UTC)
 * I'm not so sure. Some editors who did a lot of work here a long time ago seemed to think that Dutch descends from Old Saxon... On the other hand, we could just call Old Provençal Old Occitan, couldn't we? That is the name Wikipedia uses after all. 18:06, 11 November 2012 (UTC)
 * Isn't Old Provençal just one dialect of Old Occitan, but the best attested one? Then it's like Old Icelandic, which is sometimes considered a synonym of Old Norse, but more precisely is just the best attested dialect of Old Norse. And speaking of Old Norse, where does it go in the list above? My preference would be to alphabetize it as "Norse, Old", but others may disagree. As for older forms of English, I say put them in alphabetical order under E, not at the top. —Angr 22:20, 11 November 2012 (UTC)
 * I don't exactly know what Old Occitan is, but the Wikipedia article treats them as synonyms. The alphabetization could be changed but I would prefer to discuss that separately as it's a different issue that doesn't directly affect this one (although they both affect where languages are found in the list). 22:30, 11 November 2012 (UTC)

Custom sorting order for certain languages
Following on from the discussion above and in BP, I wonder if it is desirable and/or technically feasible to implement sorting keys for certain languages, so that they are not sorted by the first letter of their name. It could place, for example, Old Norse under N. This idea is separate from the one above, so even if we decide not to make subdivisions we could still decide to sort Old English under E and Middle Dutch under D. 22:36, 11 November 2012 (UTC)


 * Personally, I would prefer to continue using the alphabetisation we use now, where "Old Norse" is after "Occitan" and before "Portuguese", etc. If we don't group languages, either generally or specifically, I oppose changing their sort orders: so for example, if "Middle High German" were not grouped as a *: -sublisting under "German", I would oppose sorting it as a * -listing as "High German, Middle" (though I also don't think this is being suggested), and I oppose sorting Old Norse as * Norse, Old . Given that sorting "Middle High German" next to German as * German, Middle High is just another way of grouping the Germans, I might not oppose it, but I would prefer *: -type grouping. But a language like Old Norse, which can't be grouped (because it is just as much a predecessor of Icelandic as it is a predecessor of Norwegian, etc), should continue to be sorted by its (full) name, IMO. (I'm aware that Old/Middle English is as much a predecessor of Scots as of English, but we also don't have "English" in translations tables, so OE and ME are special cases anyway.) - -sche (discuss) 03:05, 16 November 2012 (UTC)
 * I didn't suggest that we change the names of the languages, only their sorting order. So it might look like this: 03:35, 16 November 2012 (UTC)
 * * Galician
 * * Georgian
 * * German
 * * Middle High German
 * * Old High German
 * * Greek
 * * Ancient Greek
 * * Greenlandic
 * * Guaraní
 * * Guaraní


 * I object to that. My personal preference is to list "Middle High German" completely separately from "German", with other M-languages; but I'm O.K. with listing and sorting it as "German, Middle High", or even with listing it as "Middle High German" but nested under "German". (The reason I don't like the latter is that we use "German" to mean "Modern German", so nesting Middle High German under German would imply that Middle High German translations are Modern German, just as nesting Mandarin and Cantonese under Chinese implies that Mandarin and Cantonese translations are Chinese, and nesting Cyrillic under Serbo-Croatian implies that Serbo-Croatian Cyrillic translations are Serbo-Croatian. But despite my dislike, I could accept it.) What I'm not O.K. with is listing it as "Middle High German" but sorting it as "German, Middle High". That is very confusing. Someone looking for Greek should not have to recognize that it comes after Middle High German; they should be able to just scan the first letter of the language-names to find what they're looking for. —Ruakh TALK 03:46, 16 November 2012 (UTC)


 * "Literary Chinese" is already grouped under "Chinese" in many (perhaps all? IDK) entries, such as helium, although Literary Chinese is a predecessor of modern Chinese, not a form of it the way Mandarin (which is also grouped under it) is. I think users looking at any translation, whether nested or not, need to have at least the minimal knowledge of what the name of the lect/variety they're looking at means (or need to look it up upon encountering it). - -sche (discuss) 04:09, 16 November 2012 (UTC)


 * Re: first sentence: We don't use ==Chinese== as an L2 header. We only use it in translations tables, and only as a group-heading for languages nested inside it. So even though Literary Chinese is qualitatively different from the modern topolects, I don't see a problem with including it under the "Chinese" umbrella. ==German==, by contrast, has a specific meaning here, and we define it as not including Middle High German.  Re: second sentence: Absolutely. "Imply" may not have been the best word-choice; I wasn't expressing concern that someone would be misled, but simply that it would be wrong. No one would be misled by nesting Japanese inside German, but it would be wrong, because Japanese isn't German. Likewise, Middle High German isn't German, at least as we use the term "German". —Ruakh TALK 04:22, 16 November 2012 (UTC)


 * I agree with Ruakh that listing as "Middle High German" but sorting it as "German, Middle High" would be a needlessly confusing bad idea. - -sche (discuss) 04:09, 16 November 2012 (UTC)

Countries
Can we use after a translation? Like:,   for  ? If so, could this be added to the doc ? --Jerome Potts (talk) 22:09, 18 November 2016 (UTC)

Assited Cyrillic Uzbek
Assited Cyrillic Uzbek seems to get added under Serbo-Croatian. See synod, agrobiology, oncology. Matthias Buchmeier (talk) 20:31, 2 March 2019 (UTC)
 * I've tested the tool. It seems to work OK. These are probably typos of .Matthias Buchmeier (talk) 20:40, 2 March 2019 (UTC)
 * I've found that the tool works most of the time, but sometimes I do find this happens and will fix it if if I notice. Obviously, I didn’t notice these. Thanks, DPUH (talk) 20:53, 2 March 2019 (UTC)

Number of translations per language
Perhaps it escaped me, but: is there a limit to the number of translations per language? A problem of this nature has occurred at el.wiktionary (anon habitually adds more than 10 translations for English) and I was wondering what the general policy is. sarri.greek (talk) 05:19, 15 August 2019 (UTC)
 * There is no policy about this here. Because of the division in senses here there are rarely more than 4-5 terms per language. In other wiktionaries translating to english you might find that english has a whole lot of near-synonyms for many concepts leading to more than 4-5 terms per sense.--So9q (talk) 15:01, 13 September 2019 (UTC)
 * Thank you So9q. Translators would be delighted with 20 options, but a limit-5 seems sensible. sarri.greek (talk) 15:07, 13 September 2019 (UTC)