User talk:Msh210/Archive/Translation tables

Translation tables
I appreciate your position at Votes/2007-10/Lemma entries that non-lemma entries should direct readers to the lemma entry where they can find additional translations. Do you also think that translations into non-lemma target forms should go into that non-lemma English translation table? If so, which forms should go into that table? All of the forms into which the non-lemma English term can be translated or just the target lemma? Rod (A. Smith) 17:46, 2 November 2007 (UTC)


 * I do not understand your question: I think you'll have to reword it. The way I read (reed) your question, with an example, is: "I understand your position that the entry words should direct readers to the entry word for additional translations. Do you also think that translations of words should go into the translation table under the entry words? If so, which forms should be in that table: all translations of words or just translations of word?" But I highly doubt that that's what you meant, so please clarify.&mdash;msh210 &#x2120; 21:12, 6 November 2007 (UTC)


 * If you speak Russian or some other declension-heavy language, you example will work, although it's not worded as I intended. "all translations of words" is right if it means "each plural form of the translation of the lexeme word(s) into languages that have plurals and each form in languages that have no plurals" but "just translations of word" should be changed to mean "just the translations of the lexeme word(s) into the target lemma forms".  The ideal example would not have one-to-one translations from English to the target language, though.  What languages do you speak?  Rod (A. Smith) 23:12, 6 November 2007 (UTC)


 * English and a bit of Hebrew. (See the user-language categories on my talk page.) But, assuming I understand your most recent post here correctly, I understood your first one right too. The reason I doubted it is: Why would you include translations of word on the page words? No, include translations of words.


 * Just to make sure I understand you correctly: You were asking whether the page words should include verbum, verbi, and verbo or verba, verborum, and verbis, right? (See verbum.) Well, the latter. Hope this clarifies.&mdash;msh210 &#x2120; 21:16, 7 November 2007 (UTC)


 * That's close to what I meant to ask. For singular nouns, though, we list just one form of each translation.  People sometimes call that the "lemma" form or the "citation" form.  To simplify our vocabulary, I'm calling it the "main" form for the foreign lexeme.  In Latin, the main form of the lexeme verb(um|ī|ō|a|orum|īs|...) is verbum:.  So, word just lists verbum.  We intentionally do not list verbi: or verbo:.  I wondered whether you would want us to show verba, verborum, and verbīs or just verbum.  If you would prefer to give the multiple plural forms in the translation table, would you also change the singular entry translation table to give all three of the singular Latin forms?  Note that the number of foreign forms expands greatly for verbs, easily requiring dozens of forms for a single foreign lexeme.  Rod (A. Smith) 23:46, 7 November 2007 (UTC)


 * Oh, I see. Well, definitely any translation on the page words should be of words. As to whether we should include only verba or also verborum and verbis, I guess it would probably depend on language (and part of speech and perhaps the individual word's declension class or whatever it's called): perhaps in Latin it makes sense to include only the nominative, but perhaps in Estonian it makes more sense to include the nominative and illative or something. Likewise, maybe the page went should have only the past-tense, third-person, singular as its, oh, say, Hungarian translation, but all numbers and persons and both sexes as its Hebrew. Whatever makes most sense in the individual language. Guided of course (or not of course, but guided anyway) by considerations of space and neatness: if a language has forty-eight possible mood-person-tense-sex-number-etc. combinations for a particular word, you will not want to include them all unless you really need to. Really needing to, in this case, would follow from language-specific considerations, again.&mdash;msh210 &#x2120; 18:28, 8 November 2007 (UTC)


 * Hmm. so you would like each language project page to make its own independent decision about which forms to include in the translations tables, with contributors for some languages choosing to list multiple forms but others choosing just a single form.  Is it just me, or is that really confusing for both readers and contributors?  Rod (A. Smith) 18:59, 8 November 2007 (UTC)


 * It's not confusing for readers at all in my opinion. It is confusing for editors, but not much. It's a simple matter to check an About Language page, or the Translations page, to see what to do. And if an editor "messes up" and puts in too much translation, that's fine, too.&mdash;msh210 &#x2120; 15:33, 13 November 2007 (UTC)
 * Well, I'm not an editor of Hebrew entries, because I don't know enough about Hebrew. So, in my role as a reader, I wouldn't understand from the list described below whether each of those entries in the translation table is a different lexeme or a different form of one lexeme.  Rod (A. Smith) 16:07, 13 November 2007 (UTC)


 * Well, each would have a little gloss near it "third person" etc. That wouldn't tell you, though, which of them are forms of the same lexeme. I suppose that is important information, and we need a way to capture it.&mdash;msh210 &#x2120; 16:45, 13 November 2007 (UTC)
 * The way we handle it today is by just listing the “main” form of the foreign lexeme in English translation tables. Readers then know that each listed term is a distinct translation (as opposed to a distinct grammatical form that may or may not be used depending on the grammatical context).  Translators then click through to the foreign entry to determine which grammatical form to use in their particular situation.  Rod (A. Smith) 16:59, 13 November 2007 (UTC)


 * Yes, I know.  That's why you started the vote (that led to this discussion) in the first place. We were looking at alternatives to that.&mdash;msh210 &#x2120; 17:04, 13 November 2007 (UTC)
 * Of course. :-)  There's no rush, of course, but I look forward to your suggestion for improvement.  Rod (A. Smith) 17:12, 13 November 2007 (UTC)


 * No need to wait; my and opiaterein's respective prospective formats are at Wiktionary talk:Votes/2007-10/Lemma entries/words.&mdash;msh210 &#x2120; 20:14, 13 November 2007 (UTC)


 * Ah. There you gave the third person singular masculine past tense in the translation table.  So, if I understand you, you'd have contributors from each language choose one form to represent each of the usual English inflections.  For verbs, then, each language would have a form chosen to represent our simple past tense, one to represent our past participle, one to represent our present participle, one to represent our third person singular present tense, and one to represent our infinitive.  Right?  Rod (A. Smith) 20:26, 13 November 2007 (UTC)


 * No. Sorry for the ambiguity. I meant to have a number of such tables, corresponding to different persons, etc. (Although I'm not sure opiaterein's system isn't better.)&mdash;msh210 &#x2120; 20:28, 13 November 2007 (UTC)


 * Thanks for the clarification. Please let me know whether you would change anything about the Translations section below, which illustrates how I think you are suggesting to format translation tables.  Rod (A. Smith) 22:32, 13 November 2007 (UTC)


 * It's okay, but I think I like opiaterein's system better, actually.&mdash;msh210 &#x2120; 16:57, 14 November 2007 (UTC)
 * If you are refering to opiaterein's system of "drop downs within the drop down", I think that's how I've formatted it below. Do you have a different take on it?  Do you find it easy, as a reader, to determine whether there is a Hungarian translation anywhere in those nested tables?  Rod (A. Smith) 20:15, 14 November 2007 (UTC)


 * What I'm referring to is below, under the heading "Further translations". Maybe my calling it opiaterein's is a misattribution. (I'm not sure why you switched (or seemed to switch) to discussing a lemma in listing your translations under "Translations", below; I've switched back to a non-lemma form.) Note though that each language's editors will decide how many of its translations will go in the table, using readability, inter alia, as a guide.&mdash;msh210 &#x2120; 20:38, 14 November 2007 (UTC)


 * Ah. Thanks for the clarification.  As for your perception that I switched to discussing a lemma, that's not the case.  When I said, "the amount of work involved to add a new lexeme", I meant, "the amount of work involved to add each of the forms of a new lexeme".  In any event, you are no longer advocating that format, so it's irrelevant.  In the system you support, then, each language in the translation table shows a miniature conjugation table for each of the translated lexemes.  To show where each new lexeme begins and ends, readers just note the repetition of a tag, like "first person singular".  The second such tag would indicate a new lexeme.  If a language has no inflections that correspond to that of the English headword (e.g. like Chinese), the reader is directed to the English main (lemma) entry.  If a language has too many forms to fit well into the table (e.g. like Spanish), only a few forms are listed, omitting the others.  Presumably, you would also similarly expand the translation table in the main (lemma) English entries, because "speak" isn't just the infinitive, but also the second person singular present indicative, the first person plural present indicative, the future subjunctive, the imperative, etc.  Right? Rod (A. Smith) 21:10, 14 November 2007 (UTC)


 * I've been thinking about that last point, and I'm not sure, but my thoughts are as follows. Logically we should include all those translations s.v. speak if we do so (m.m.) s.v. spoke. Otoh, I suspect most people who look up speak will want to know what the lexeme means, not what the word (the second-person present tense, e.g.) means; they look up speak after coming across speaking or speaks (or, of course, the second-person present speak) and knowing a bit about how words are conjugated in English; or they look up speak after coming across speak (the bare infinitive) in an ESL book or the like. There will be of course some who look up speak (the second-person present) wanting to know what it means in their language, but I think that these will be fewer in number. Cluttering the page with multiple translations will merely confuse.


 * To put it in other words, s.v. spoke it's necessary, if we're to have translations at all, to have multiple ones (as how do we choose?); and it's necessary to have them so as to avoid making users go from entry to entry to look for their word. But s.v. speak it's not necessary to have so many, and they confuse.


 * The other user of the translations table is the native English speaker who wants to know how to say speak (second-person, masculine, singular present) in Hebrew. Thee ideal would be to have this translation s.v speak. But if we have the lemma form in the translation table s.v. speak, the user can then check that entry for the conjugation table; this is merely one extra step, and I think that this extra work for the user, while regrettable, is not as bad as having too many translations s.v. speak.


 * That logic ("one extra step is better than too much clutter") does not apply to the entry spoke (or other non-lemma entries), however. There, there are two extra steps. The foreign-language speaker looks up spoke, finds "past tense of 'speak'" and "see translations under 'speak'", looks up "speak", find translations into his own language, and then must extrapolate to his own language's past tense. (The two extra steps are looking up speak and extrapolating the translation.) The logic of "one extra step is better than too much clutter" also doesn't apply to spoke because, unlike s.v. speak, there is no lemma form of the translation to include, that the other forms are cluttering up. That is, s.v. speak, including non-lemma translations clutters up the lemma translations, obscuring them; s.v., spoke that's not true, as all the translations (if any) are non-lemma; so it's okay to have them.&mdash;msh210 &#x2120; 21:31, 14 November 2007 (UTC)


 * Wait, you'd seriously want went to list halakhti, halakhta, halakht, halakh, halkha, halakhnu, halakhtem, halakhten, and halkhu, as well as hafakhti, hafakhta, hafakht, hafakh, hafkhu, hafakhnu, hafakhtem, hafakhten, and hafkhu, as well as nine forms for every other translation? The sad thing is, even if we could keep on top of all this, we still wouldn't be giving every possible translation, as we'd be omitting the 12 "haya `ose" forms per verb, let alone all the vav-hahipukh'd future-tense forms, all the pausal-pronunciation forms, and all the direct-object including forms. (Actually, some of those we can get away with excluding on the grounds that they translate to more than just "went", but still.) —Ruakh TALK 20:14, 8 November 2007 (UTC)


 * In truth Hebrew was just an example in what I wrote above, and I didn't have it specifically in mind. But since we're discussing it: Yes, seriously. Nine forms is not that much. And as to the additional forms: "Haya holech" means "would have gone" or "used to go", not "went", no? Pausal pronunciations don't exist in past tense afaIk. The vav-hahipuch'ed future-becomes-past is not used with a pronoun (e.g., וָאֵשֵׁב but never *ואני אשב or *ואשב אני) and so are translations of "I sat" not "sat". (This argument doesn't apply to the vav-hahipuch'ed past-becomes-future words.) And the direct-object-including forms, as you mention, are certainly not translations of the bare past (or any) tense.


 * But more to the point: according to what I wrote above we'd leave it up to individual language users, and maybe Hebrew will wind up without many forms (or without many forms for specific parts of speech).&mdash;msh210 &#x2120; 15:33, 13 November 2007 (UTC)


 * Regarding your question: "He'd learn to read if he went to school" &rarr; "Haya lomed likro im haya holekh l'vet sefer.". Regarding pausality, hipukhativity, etc.: O.K., I'll take your word for it: my Biblical Hebrew is read-only. :-)   Regarding the part that's more to the point: O.K., fair enough. I just think that even if we do allow non-lemma translations (which I don't think we should), it would be nice for a language-neutral policy page to give some guidance, rather than encouraging large-scale (and potentially misleading)) variation from language to language. —Ruakh TALK 18:19, 13 November 2007 (UTC)

Translations

 * Chinese: 說話, 说话 (shuōhuà)


 * Korean: 말했다


 * Korean: 말했느냐


 * Korean: 말했어


 * Korean: 말했어


 * Korean: 말했어요


 * Korean: 말했어요


 * Korean: 말했습니다


 * Korean: 말했습니까


 * Korean: 말하셨다


 * Korean: 말하셨어


 * Korean: 말하셨어요


 * Korean: 말하셨습니다


 * Korean: 말하셨느냐


 * Korean: 말하셨어


 * Korean: 말하셨어요


 * Korean: 말했습니까
 * Korean: 말하셨습니까


 * Romanian: zisei, spusei
 * Spanish: dije


 * Polish:


 * Polish:


 * Romanian: ziseşi, spuseşi
 * Spanish: dijiste


 * Polish:


 * Polish:


 * Romanian: zise, spuse
 * Spanish: dijo


 * Polish:


 * Polish:


 * Polish:


 * Romanian: ziserăm, spuserăm
 * Spanish: dijimos


 * Polish:


 * Polish:


 * Romanian: ziserăţi, spuserăţi
 * Spanish: dijisteis


 * Polish:


 * Polish:


 * Romanian: ziseră, spuseră
 * Spanish: dijeron


 * Polish:


 * Polish:


 * Russian:


 * Russian:


 * Russian:


 * Russian:


 * Polish:


 * Latin: fabulābar


 * Latin:


 * Latin:

{{trans-bottom}
 * Latin:


 * Latin:


 * Latin:


 * Latin:


 * Latin:


 * Latin:


 * Latin:


 * Latin:


 * Latin:


 * Latin:


 * Latin:


 * Korean:


 * Korean:


 * Russian:


 * Russian:


 * Russian:


 * Spanish: dicho


 * Polish:


 * Polish:


 * Spanish: dicha


 * Polish:


 * German:


 * Polish:


 * Polish:


 * Spanish: dichos


 * Polish:


 * Polish:


 * Polish:


 * Polish:


 * Spanish: dichas


 * German:


 * Polish:


 * Polish:

Note: Many more forms belong in the table above, but should serve to illustrate the required structure. Note that many languages have peculiar forms, so they get their own collapsable section. Note also the amount of work involved to add a new lexeme in, say, Spanish. The structure is so complex, it's difficult in the editing window to determine which sense and which English verb form is being translated. Rod (A. Smith) 22:32, 13 November 2007 (UTC)

Further translations
(spoke, past tense verb) (notes on this table are above)


 * American Sign Language: see translations at speak [as Chinese, q.v. immediately below]
 * Chinese: see translations at speak [as Chinese, per you, above, doesn't inflect by tense)
 * Hebrew: דברתי ; דברת ; דבר ; דברה ; דברנו ; דברתם ; דברתן ; דברו
 * Korean: 말했다 ; 말했느냐 ; [about fifteen others for the formal, etc.]
 * Latin: fabulābar


 * Romanian: zisei, spusei ; ziseşi, spuseşi ; zise, spuse ; ziserăm, spuserăm ; ziserăţi, spuserăţi ; ziseră, spuseră
 * Spanish: dije ; dijiste ; dijo ; dijimos ; dijisteis ; dijeron

OK. Based on your "one extra step is better than too much clutter" principle, I modified on the dev server and created some entries for the forms of "to talk" to demonstrate the potential for sharing a translation table between the inflected forms of an entry. in that environment now accepts some arguments to show the translations in different grammatical forms, based on  . In http://wiktionarydev.leuksman.com/index.php/talked, the shared translations are shown with  . Using a system like the one there, editors can choose to share translation tables among the various inflected forms of an English lexeme. Each use of in the translation table corresponds to a distinct foreign lexeme, but editors can choose to show inflected-form-specific versions of the foreign lexeme to match the grammatical properties of the English headword. In the examples there, I wrote the inflected translations as decir, but that could easily also be decir if the editor feels there is enough space in the table. Each of the links, though, goes to the main (lemma) foreign entry, which hopefully has the most complete information about the lexeme. What do you think? Rod (A. Smith) 02:21, 16 November 2007 (UTC)

(Note: the colors there are crazy. Don't worry about that. Rod (A. Smith) 02:22, 16 November 2007 (UTC))

To address the MediaWiki transcluded section editing problems, I now have also changed on the dev server. As shown at talk and its various inflection entries, if a   argument is given,  shows an "[edit]" link in the main namespace that allows editors to edit the shared translation table. It also shows editors a notice in from the template namespace, explaining that the table is shared by inflections of the main entry. Hopefully that makes sense. If not, please send any comments/questions/suggestions my way. Rod (A. Smith) 23:09, 16 November 2007 (UTC)


 * Have you seen http://wiktionarydev.leuksman.com/index.php/talks et al.? Rod (A. Smith) 03:37, 21 November 2007 (UTC)


 * I glanced at it soon after you mentioned it here, saw it was complicated, and told myself I'll get back to it. I haven't done so yet. I still hope to. Sorry.&mdash;msh210 &#x2120; 17:19, 21 November 2007 (UTC)