Wiktionary talk:Translations/If present, what foreign word articles should contain

from Wiktionary talk:Translations#If present, what foreign word articles should contain Rod (A. Smith) 21:48, 3 November 2007 (UTC)

--Richardb 05:36, 22 May 2005 (UTC) I was told it is policy to give translations only for non-English words, not proper definitions as English words get them, and to give English translations only for non-English words. I would like to know if that's true (Paul G and SemperBlotto claim so), and if it is, for what reasons this decision was made. I can't understand these rules, strongly oppose them, and want to see them discarded as soon as possible. Ncik 04 May 05

The truth is, that, until very recently, there were no writen or publicised policies/ standars / conventions etc at all, of any type, in Wiktionary. Only a few people who believed they had great insight and just knew what the policy should be! Please support the fledgling Policy Development process.--Richardb 05:50, 22 May 2005 (UTC)

Well, I think that requires a degree of discretion. If the word doesn't translate exactly to an English word, give it a rough definition and then link to similar English words (I've seen that in a Translations section a number of times). If there is more than one sense, but each sense corresponds to an English word, then link to those instead of a definition. I'm not one to give advice on this, really, but in some ways I see it as common sense. I think the rationale behind the decision is that all foreign words should be linked to at least one English word somewhere, and writing a definition for a word that is exactly the same as one for an English word is just unnecessary, and the English word should be favoured. Am I completely wrong here? --Wytukaze 19:55, 4 May 2005 (UTC)

I agree with User:Wytukaze, User:Paul G and User:SemperBlotto. In general foreign words should provide wikified translation(s) rather than a full definition. I also agree that there should be a measure of lattitude for words that do not translate easily or have no ready English equivalent. I would guess that this is pretty much the consensus opinion on the topic. Kevin Rector 20:04, 4 May 2005 (UTC)

<Jun-Dai 22:13, 4 May 2005 (UTC)> Wow. I don't know when this decision got made, but it's clearly in opposition to everything I've done since joining the Wiktionary. If the word is a clear cognate, it makes sense to link it to the English equivalent without adding much more, but for the bulk of most languages, we are providing an extremely limited service if we only link to translations of the word rather than providing full definitions. Or, to put it another way, it's extremely rare that words have a ready English equivalent, even for English's closest linguistic cousins (e.g., German), and the ones that seem similar to English words especially merit attention to the differences, since the possibility of confusion is so much greater.

I would like to see this discussion escalated to a policy decision asap, since if it falls against my favor, then we should probably remove a good portion of my contributions. </Jun-Dai>

The relevant history seems to be at Wiktionary:Multilingual coordination. --Connel MacKenzie 23:53, 4 May 2005 (UTC)

In the Ultimate Wikionay ALL words will be welcome, never mind if they have a translation to any other language or not. There will be ample room to provide an English/any other language definitions of the word, but it will not require a translation to ANY language. GerardM 00:35, 5 May 2005 (UTC)

<Jun-Dai 00:44, 5 May 2005 (UTC)> Connel, I'm not seeing this relevant history. All I see is a list of Wiktionaries, some discussion of interwiki coordination, and some discussion of language namespaces (which I wish we were using. I don't see the value in our current level-2 header arrangement). </Jun-Dai>

Jun-Dai, I was responding to Ncik's question as to whether it is true or not (that foreign word should be entered only as translations: answer is yes) and the history surrounding it being in that page's history, the discussion in it's discussion page, etc. On that page, is the sentence "Each language has its own Wiktionary project, which includes translations to and from that language, and descriptions in that language" followed by a grammatically incorrect explanation (that next sentence's grammar perhaps the victim of edit wars over this point of debate?) So, according to that sentence, we shouldn't have foreign word entries at all! A little further down however, it describes the interwiki links (which require identical spelling of article titles, to cross wikis) therefore the reason why we do need entries for the foreign spelling of words. --Connel MacKenzie 02:28, 5 May 2005 (UTC)

<Jun-Dai 07:14, 5 May 2005 (UTC)> Yeah, I did notice that pair of sentences. My interpretation of it, however, is greatly different from yours (despite my reading it several times). It seems clear to me that the two sentences mandate that we have foreign word entries: frei-tr->free and free-tr->frei both belong in the English Wiktionary it says (as well as the German one). This means that we need frei and free here, each pointing to the other. What the two sentences do not make clear is that we need provide simple translations. More importantly, the term translation is used somewhat ambiguously--even casually--and I get the strong impression that it does not stand in the way of providing a detailed explanation of the term frei in English, which could not happen without mention of the English word free by means of translation.

I would like to see whose sentences rewritten for clarity,. What I see the Wiktionary as, with regard to foreign words, is a source for seeing simple translations of English words into foreign languages (via translation tables, or better yet, translation subpages), and even more importantly a source for seeing explanations of foreign words in English--translations, yes, but also definitions, explanations, examples, and comments on usage. Whether this is the vision that everyone else here has or not, one thing is clear to me: if we mold the Wiktionary as a resource of this kind, we have the potential to provide a valuable resource that simply cannot be found elsewhere. If we limit ourselves to simple translations, then we are nothing more than a free, but fairly limited version of a poor-quality foreign-language dictionary (the good ones provide small definitions and example usages). What the hell are we here for? Let's provide as much usable information on each foreign term as we can, in concise, hierarchical, and digestible formats, so that someone looking up a foreign term will have as clear an idea of its usage as any non-human resource can provide. </Jun-Dai>

• As I wrote at User talk:Jérôme, Wiktionary isn't excused the need for definitions, usage notes, quotations, pronunciations, etymologies, synonyms, antonyms, homophones, inflections, references, derived and related words, and whatnot simply because a word isn't English. Uncle G 14:44, 7 May 2005 (UTC)

Here some reasons Paul G posted on my talk page about why the current policy should be kept to which I would like to respond as follows:

(1) It is stated on the main page that this is the English Wiktionary and definitions of foreign words are not given here.

Where???? You do mean the Main page, don't you?

(2) This is long-established policy.

I don't see a problem with abolishing long established polcies.

(3) Other Wiktionaries give meanings but in their own languages. The English Wiktionary tells you that rouge is French for "red" (among other things) but you can only determine this from the French dictionary if you understand French.

I don't understand what you are trying to tell me with these two sentences.

(4) Giving non-English translations of non-English words opens up the way to a slew of inconsistency and inaccuracy. For example, red might list translations for French, Italian and Spanish, while rouge lists translations for English and German, and perhaps gets Italian wrong. This Italian translation is then copied to the German page, which also lists Polish and Russian.

Who has been suggesting non-English translations of non-English words? —Muke Tever 01:03, 6 May 2005 (UTC)

Having wrong translations is not a problem that only affects non-English entries.

How am I to know that to get the Russian translation I have to go to the German page?

You don't know. But that way at least there is a chance you stumble over the Russian translation which otherwise wouldn't be anywhere at all.

Further, when a new translation is added, it must be added to (potentially) dozens of other pages.

Replace "must" with "should" and I agree.

When the Italian translation is discovered to be wrong, dozens of pages have to be checked. This is the single biggest reason for not giving translations other than English ones in non-English entries.

Copying and pasting translations (and other stuff) between different languages should be avoided if the person doing so doesn't know the languages he is copying to to a certain degree. This problem also exists within a language: E.g. when a user adds a wrong definition for a word and then, in analogy, gives similarly flawed definitions on the pages of morphologically derived words. So it's not that a new problem would be created if we changed the policy. We should simply discourage users to modify/create information involving a language they don't understand.

(5) By the same token, if definitions are included instead of translations, and a definition is added or amended on an English page, it will need to be added or updated on dozens of other pages too. More scope for inconsistency and inaccuracy, and the whole project becomes a nightmare to maintain.

I don't see the problem here (I have the suspicion I don't quite understand what you want to say). Since we don't link words to the meaning(s) (i.e. definitions) they have in that particular context but always to the whole page, changes there (i.e. on the page of the English word) wouldn't invalidate a link elsewhere.

One of Wiktionary's primary goals must be accuracy and consistency, and the current policy helps work towards this.

The current rules prevent accuracy since they only allow translations where proper definitions are needed.

Apart from that, I agree with Jun-Dai's, well articulated arguments. Ncik 05 May 2005

IMHO both a translation (if there is one) and a short definition should be given. (I'm doing this on la: when I can.) Especially since most words have more than one sense, and they don't always map equally: If all you do is say that Latin panis translates to 'bread', it doesn't tell you whether that can mean 'bread' in the sense of food generally (it can) or in the sense of 'money' (it can't). Look in any substantial bilingual dictionary: in many, many entries you are given the translation, and some disambiguating term, e.g. my en-ru dictionary has: under examination: "(inspection) осмотр; (exam) экзамен; (law) допрос" —and since Wiki Is Not Paper why can't we do better than that?—Muke Tever 01:03, 6 May 2005 (UTC)

• I wonder if this debate is an example of people talking past each other? I'm not sure. Anyway, I think we have to get past the idea of a tranlation being a single link to a single word. When the translation is ambiguous of course there needs to be the latitude to describe as accurately as possible what the word means in English (I don't think anyone is disputing this). So the entry for panis should note that it means bread in the sense of food, not in the sense of money. If the word bread is ambiguous to me then I can go look it up by following the link provided. However, there is no reason for panis to say "Baked dough made from cereals" when it can say "bread; as in food".
  • Well what I meant was that panis is both "bread" as in "baked dough made from cereals" and metaphorically "food" generally, just like "bread" (e.g. "our daily bread" = "panis noster cotidianus"). To take a harder example, English rod is basically Latin virgula -- but definining virgula as "rod" doesn't tell us anything about what kinds of rods "virgula" is used to refer to... (lightning rods? shepherd's rod? rods and cones in the eyes? Aaron's rod? membrum virile?) or whether 'virgula' has the same sense of a rod being used for disciplinary purposes (As in English "spare the rod, spoil the child").
  • Take a word with a metric buttload of English senses, like "set", then a word that translates as "set" in another language: a lot of the senses may transfer over, but likely a lot won't, and a lot will have different pragmatic uses, connotations, or whatnot... —Muke Tever 22:48, 6 May 2005 (UTC)
• There's conflation of translation tables (WT:ELE#Translations) and definitions (WT:ELE#Definitions) in this discussion. An entry for a non-English word should certainly fully explain the meaning(s) of the word, in its definitions section, and not be restricted to a one-word definition, for the reasons given by Muke. This is what paper dictionaries do. An (abbreviated!) example from Collins French-English Anglais-Français (1969):

  falloir v imp — to be necessary, must, need, take, require; vr s'en — to be lacking, be far from; il lui faut un voiture he needs a car; il m'a fallu une heure pour la faire [... and so on...]

  Wiktionary should certainly do no less than they. Whether articles for non-English words should have translation tables is a different thing, and the two issues should not be conflated. Uncle G 14:44, 7 May 2005 (UTC)

<Jun-Dai 05:58, 8 May 2005 (UTC)> There definitely agrees to be a degree of people talking past one another here. That said, I think the real debate is whether non-English entries should have as full an entry as normal English entries. I get the sense that some here (Wytukaze, Kevin Rector) feel they should be kept as short as possible, while others--including Ncik, myself, and Uncle G feel that foreign words should have as full a definition as an English word, excepting the existence of a translation table. I'm not sure where Connel and Muke stand in this debate. The main argument for not having a translation table on non-English entries (I don't think anyone's arguing this point, but I'll say it anyways) is that the English Wiktionary is not in the business of providing German->French translations--Everything we do should either be coming from English or coming into English.

One thing Connel has made quite clear is that our policy pages are extremely unclear and inconsistent on this topic. I suggest that those of us that are inclined to do so start proposing changes to the policy pages so that we can clean them up. They are, after all, left over from much older phases of the Wiktionary when it was much less clear what the Wiktionary was going to be, so much of it should be pretty easily cleaned up, as we have pretty much decided as a consensus to at least include non-English entries, translation tables on English entries to non-English entries, etc. While we are doing this, maybe we can come to some agreement on how non-English pages should word. I've defined some policy for Japanese entries, in specific, and it would be nice if we could do the same for other languages. More than this, however, we just need more entries added - the odds of any common Japanese word being in the Wiktionary are probably less that 1 in 100. For English it's more like 2 in 3 (I'm making these numbers up, by the way). </Jun-Dai>

If I made that clear, it was not my intent!  :-) I would like to thank Uncle G for clearing up my misconception on this topic - I was very much mired in that confusion. I was under the impression that foreign terms' definitions were to be kept bare minimum, just as the standing convention, without understanding why. With Uncle G's explanation (light bulb ==> on), I now get it. Even Ncik's inflamatory opening statement now no longer seems inflamatory, only a matter of misinformation (to him) and confusion. But still, I think a terribly verbose foreign definition is not a good idea - it should be only enough to link to an English word's particular meaning. (That said, I'm not likely to slash-and-burn a foreign language definition...knock yourself out if ya wanna.) Jun-Dai, did you just volunteer to rewrite that infamous paragraph (posted here for discussion first) in WT:ELE Wiktionary:Multilingual coordination?  :-) --Connel MacKenzie 07:22, 8 May 2005 (UTC)

My two cents:-

• things that should be included in a non-English word - Language, Part of speech, Pronunciation, Etymology (if different from an identically spelled English word of the same meaning), Translation (or brief description if no such English concept), Plurals and Conjugations etc, Related or Derived terms in the same language
• things that should not be included - Translation tables, Definitions copied from (or of similar length) to the English equivalent

SemperBlotto 07:37, 8 May 2005 (UTC)

<Jun-Dai 05:44, 9 May 2005 (UTC)>

• Translation (or brief description if no such English concept). . .

  This is the part that really confuses me. Very few foreign words "accurately" translate into English. There's rarely an equivalent English word to compare the meaning to. Even a fairly close word to English, like the Japanese デパート (pronounced "depāto"), which roughly means "department store" (the word from which it was borrowed), has enough of a different meaning in Japanese to merit an explanation of its own. If, for example, you link the German frei to English free and leave it at that, you are doing a disservice to the reader to ignore the fact that the German word can mean things that you cannot appropriately translate to free in English (frei nach Goethe, for instance, refers to a passage that was borrowed from Goethe). Additionally, there are bound to be numerous circumstances where you might use free in English, but frei would not be appropriate in German. Rather than being the exceptions, these cases are the rule; the exceptions--being those cases where the word pretty much maps to an English word, or a couple of English words--are going to be very few in number, if indeed they exist at all.

  Ultimately, I'm not sure what a "translation" entails, but if it is at all different from simply providing a full definition of the term in English, then I'm against it.

</Jun-Dai>

• • I often add a gloss to translations, so panis would read "bread (baked dough)" to make it clear that it applies to this meaning only. I think this is sufficient.

• • To reply to Ncik's reply to my comments on his page, which I have just seen here:

(1) It is stated on the main page that this is the English Wiktionary and definitions of foreign words are not given here.

Where???? You do mean the Main page, don't you?

Yes; I'm referring to this paragraph: "This is the English Wiktionary: it aims to describe all words of all languages, with definitions and descriptions in English only. For example, see frei (a German word). In order to find a German definition of that word, you would visit the equivalent page in the German Wiktionary (http://de.wiktionary.org/wiki/frei)." I agree that this does not reflect my point.

(2) This is long-established policy.

I don't see a problem with abolishing long established polcies.

True, "we've always done it this way" is not, in itself, a reason for keeping a policy.

(3) Other Wiktionaries give meanings but in their own languages. The English Wiktionary tells you that rouge is French for "red" (among other things) but you can only determine this from the French dictionary if you understand French.

I don't understand what you are trying to tell me with these two sentences.

If you want to know what "rouge" means and look it up in the French Wiktionary, you will get a definition written in French, which will make sense only if you understand French. The English Wiktionary tells you, in English, that it means "red".

(4) Giving non-English translations of non-English words opens up the way to a slew of inconsistency and inaccuracy. For example, red might list translations for French, Italian and Spanish, while rouge lists translations for English and German, and perhaps gets Italian wrong. This Italian translation is then copied to the German page, which also lists Polish and Russian.

Who has been suggesting non-English translations of non-English words? —Muke Tever 01:03, 6 May 2005 (UTC)

There are many pages here with non-English entries that translate the entry into multiple languages rather than just giving the English translation. The Esperanto page "vorto" was one such example (which has since been reduced to the English translations only - see the page history for the earlier version).

Having wrong translations is not a problem that only affects non-English entries.

How am I to know that to get the Russian translation I have to go to the German page?

You don't know. But that way at least there is a chance you stumble over the Russian translation which otherwise wouldn't be anywhere at all.

So it's better to have the translations all in the one place where people will think to look for them, namely the English entry.

Further, when a new translation is added, it must be added to (potentially) dozens of other pages.

Replace "must" with "should" and I agree.

When the Italian translation is discovered to be wrong, dozens of pages have to be checked. This is the single biggest reason for not giving translations other than English ones in non-English entries.

Copying and pasting translations (and other stuff) between different languages should be avoided if the person doing so doesn't know the languages he is copying to to a certain degree. This problem also exists within a language: E.g. when a user adds a wrong definition for a word and then, in analogy, gives similarly flawed definitions on the pages of morphologically derived words. So it's not that a new problem would be created if we changed the policy. We should simply discourage users to modify/create information involving a language they don't understand.

I disagree. I think people would go around copying and pasting if we were to change the policy, and then we would end up with an unworkable mess.

(5) By the same token, if definitions are included instead of translations, and a definition is added or amended on an English page, it will need to be added or updated on dozens of other pages too. More scope for inconsistency and inaccuracy, and the whole project becomes a nightmare to maintain.

I don't see the problem here (I have the suspicion I don't quite understand what you want to say). Since we don't link words to the meaning(s) (i.e. definitions) they have in that particular context but always to the whole page, changes there (i.e. on the page of the English word) wouldn't invalidate a link elsewhere.

The problem is twofold: unnecessary duplication of effort; and scope for inconsistency.

One of Wiktionary's primary goals must be accuracy and consistency, and the current policy helps work towards this.

The current rules prevent accuracy since they only allow translations where proper definitions are needed.

So include a gloss, as I mention above ("panis - bread (baked dough)"). This resolves this problem.

Apart from that, I agree with Jun-Dai's, well articulated arguments. Ncik 05 May 2005

I also support Ncik in agreeing with Jun-Dai's arguments :-) Sergio.ballestrero 09:12, 7 May 2006 (UTC)