Cross-posted to wiktionary and wikipedia; because I haven't seen a good example lately of how a merging might work; and because I've recently been through the cycle of slightly argumentative discussion that Ec seems to be falling into. Please reply to only one list, and feel free to add to or modify my example to illustrate your own ideas of what such a project might look like. SJ
On 6/18/05, Sabine Cretella sabine_cretella@yahoo.it wrote:
The differences between Gerard and me are based in a fundamental philosophical differences about the nature of Wiktionaries in particular and dictionaries in general. His position is a logical consequence from the premise that Wiktionary is just a translation dictionary; I disagree with his premise by considering Wiktionary to be much more than that.
It might be more constructive to think of this position (what I believe to be Gerard's) as a logical consequence from the premise that single-language Wiktionaries can be efficiently combined by taking advantage of translation-dictionary content (specifically their linkages between words)..
Each Wiktionary may be tasked with explaining all words from all languages, but it does so for the benefit of speakers of its own language. Gerard's Ultimate Wiktionary would work well if translatiions were simply questions of one on one relationships. As one example, the word "minister" exists in both Dutch, and you are probably safe to use the same word when going from Dutch to English. It doesn't work in the other direction. You can't translate the English "minister" to its Dutch equivalent when "predikant" is intended. Add in a third language and it can get very complicated.
Great, an example! Just what we need. Consider the following English-language content:
________________________________________________________________
=minister=
== [English] == minister ([n.]) 1. [A person trained to perform religious ceremonies at a Protestant church.] "The minister said a prayer on behalf of the entire congregation." || [Dutch]: predikant; ... || minister ([n.]) 2. [A person commissioned by the government for public service.] "He was newly appointed to be Minister of the Interior." ** || [Danish]: gesandt; [Dutch]: minister [m,f]; [French]: ministre [m,f]; [German]: Minister [m], Ministerin [f]; [Italian]: ministro; [Indonesian]: menteri; [Interlingua]: ministro; [Japanese]: 大臣 (だいじん, daijin); [Polish]: minister; [Spanish]: ministro [m] || minister ([n.]) 3. [A person who serves others.] minister ([vi.]) 1. [To tend to the needs of others] minister ([vt.]) 1. [To dispense, to administer]
== [Dutch] == minister ([n.]) 1. [A person commissioned by the government for public service.] "Zware voet jaagt minister Anciaux uit de bocht." ** || [Danish]: gesandt; [English]: minister; [French]: ministre [m,f]; [German]: Minister [m], Ministerin [f]; [Italian]: ministro; [Indonesian]: menteri; [Interlingua]: ministro; [Japanese]: 大臣 (だいじん, daijin); [Polish]: minister; [Spanish]: ministro [m] || ________________________________________________________________
In the above examplt: a) choosing your 'interface' language may change all text in [brackets], b) content between || double bars || is stored in the database, so that the two lists of translations for "minister (English, n., 2)" "minister (Dutch, n., 1)" are actually referencing the same list of database translations [marked above by a double asterisk **] c) that bugbear of having multiple definitions for the same word in some languages, but not in others, is finessed somewhat by relegating translations to the defintion level, not the word level. [1]
The Wiktionaries in individual languages are in a better position to explain this kind of problem in the target language for the translation.
As far as I can tell, the current independence of "[content in] Wiktionaries in individual languages" would remain, in the most detailed proposed concatenation of many Wiktionaries into one. Things that would change:
* All wiktionaries would share a single list of definition-linkages (definition 3 of word A is the same as definition 1 of word B); many of these linkages would be from one language to another, but others would be between synonyms in a single language.
* All synonymous definitions would share a single list of translations, so that this list need not be pasted 100 times (and updated 100 times for each update).
Sabine writes:
All this was explained more than once (in the discussionlists and on meta) and obviously you did not read it, but you only read and write what you like.
For reference, I have tried to follow this matter carefully. I have discussed metadata issues with linguists who are designing other Wiktionary extensions. I have asked similar questions myself, of Gerard on many occasions, and in conversations with both of you. :-) And STILL the most complete published plans for, or descriptions of, an "Ultimate Wiktionary" -- or a merging of many wiktionary projects into a single one under any other name -- are not clear. Please do not blame Ec for not seeing things exactly as you do.
Adding a third language is not complicated, as the relation and the
Adding a third, and then a tenth, language IS complicated. It is doable, but complicated, and necessarily slightly imperfect [1]. However, following the Wikimedia principle of doing useful things quickly and not worrying about theoretical perfection, this will be a useful project long before its nuances are completely satisfactory to all.
based on such an idea. International organisations, like Lisa (and they are THE language specialists for localisation) and Kennisnet (they are education specialist and work in many languages) are interested in it and believe it its value otherwise why would they have paid for the
Of course it has value to try to make such a project work. This does not mean it is not difficult; it is! These international organizations know how difficult it is; I imagine they are curious to see how we will proceed. Let us see if our efforts make something useful, despite the difficulty involved. I am confident that they will... whatever our project is called.
+SJ+
[1] There is a difficult question, which we are ignoring for now : just how precisely do all the translations of "minster (English, n., 2)" have synonymous definitions? When are two different words ever truly synonymous? But that is a discussion for another month.
Sj 2.718281828@gmail.com wrote:
[Interlingua]: ministro; [Japanese]: 大臣 (だいじん, daijin);
Will users of non-Roman scripts be able to see transliterations in their own scripts, or will they have to be stuck with Roman?
== [Dutch] == minister ([n.]) 1. [A person commissioned by the government for public service.] "Zware voet jaagt minister Anciaux uit de bocht."
Hopefully there's space for a translation of this example in the user's language.
b) content between || double bars || is stored in the database, so that the two lists of translations for "minister (English, n., 2)" "minister (Dutch, n., 1)" are actually referencing the same list of database translations [marked above by a double asterisk **]
What exactly is the translation table attaching to? The definition number? The auto-translating definition? or the individual language's translation of the definition? If the English definition is improved by someone, made more specific (and hopefully it would be) nothing stops a Dutch editor from adding translations (in several languages even) that relate only to the older definition that was faithfully done into his language. Also, the English editor mightn't even think to check an added translation--does he know that the Dutch editor isn't translating from the same definition?
- All synonymous definitions would share a single list of
translations, so that this list need not be pasted 100 times (and updated 100 times for each update).
Though all non-synonymous translations may have to be cut-and-pasted as many times as necessary for whatever near-matches apply. E.g. Span. "tu" has to be given as the translation for the definition "second person pronoun" (Eng "you"), for the definition "second person singular pronoun" (Lat. "tu"), for the definition "second person singular familiar pronoun" (OldEng "þú"), for the definition "second person singular masculine pronoun" (Heb. attah), for the definition "second person singular feminine pronoun" (Arab. anti), for the definition "second person familiar masculine pronoun" etc... Meanwhile "attah" can't go under the table for "second person singular feminine pronoun", and so forth.
And that's just a word with concrete semantics. When it comes to something a little more fluid, like "street", say, it might be more difficult.
[1] There is a difficult question, which we are ignoring for now : just how precisely do all the translations of "minster (English, n., 2)" have synonymous definitions? When are two different words ever truly synonymous? But that is a discussion for another month.
A minister is a special case. Nouns that refer to specific referents are likely to have synonymous definitions, but they're relatively rare; most words won't have ready one-word translations so easily. Try: walk, run, saunter, dash, sashay, hurry, skip, mosey, gallop (varying degrees of the same kind of action); die, expire, pass away, kick the bucket, buy the farm (varying degrees of the formality of an action);... sludge, slush, slime, mush, slurry... (various specific referents in the same close semantic field) -- different languages will handle them all differently (though possibly the common European languages will share more gracefully than others).
*Muke!
Muke Tever wrote:
Sj 2.718281828@gmail.com wrote:
[Interlingua]: ministro; [Japanese]: 大臣 (だいじん, daijin);Will users of non-Roman scripts be able to see transliterations in their own scripts, or will they have to be stuck with Roman?
When you look at the current practice in the nl:wiktionary, you will find that the different scripts are supported. This will also be the case for Ultimate Wiktionary. More interesting is how a right to left implementation will look like, like Arabic, Farsi, Hebrew ..
== [Dutch] == minister ([n.]) 1. [A person commissioned by the government for public service.] "Zware voet jaagt minister Anciaux uit de bocht."
Hopefully there's space for a translation of this example in the user's language.
There will be room for idiom, this line is problematic in that the meaning of "zware voet" is not clear in its own right without context. So as idiom it is not really great.
b) content between || double bars || is stored in the database, so that the two lists of translations for "minister (English, n., 2)" "minister (Dutch, n., 1)" are actually referencing the same list of database translations [marked above by a double asterisk **]
This practice will not be in the UW. Every meaning gets its own list of translations. The practices leads to many examples in the English wiktionary where the translation is absolutely wrong. This is particularly true when later meanings are added.
What exactly is the translation table attaching to? The definition number? The auto-translating definition? or the individual language's translation of the definition? If the English definition is improved by someone, made more specific (and hopefully it would be) nothing stops a Dutch editor from adding translations (in several languages even) that relate only to the older definition that was faithfully done into his language. Also, the English editor mightn't even think to check an added translation--does he know that the Dutch editor isn't translating from the same definition?
Meaning exist on a global level. When a meaning is added, it will be universally seen to be there. Meanings are linked to a word. Translations will be linked to meanings.
- All synonymous definitions would share a single list of
translations, so that this list need not be pasted 100 times (and updated 100 times for each update).
Though all non-synonymous translations may have to be cut-and-pasted as many times as necessary for whatever near-matches apply. E.g. Span. "tu" has to be given as the translation for the definition "second person pronoun" (Eng "you"), for the definition "second person singular pronoun" (Lat. "tu"), for the definition "second person singular familiar pronoun" (OldEng "þú"), for the definition "second person singular masculine pronoun" (Heb. attah), for the definition "second person singular feminine pronoun" (Arab. anti), for the definition "second person familiar masculine pronoun" etc... Meanwhile "attah" can't go under the table for "second person singular feminine pronoun", and so forth.
And that's just a word with concrete semantics. When it comes to something a little more fluid, like "street", say, it might be more difficult.
[1] There is a difficult question, which we are ignoring for now : just how precisely do all the translations of "minster (English, n., 2)" have synonymous definitions? When are two different words ever truly synonymous? But that is a discussion for another month.
Even when words are not "truly" synonymous for some, they may be "truly" synonymous for others. Some words have only a distinct meaning in certain subcultures. This can be adressed by giving a word multiple meanings. The truly synonymous meaning and the slightly differing meaning.
A minister is a special case. Nouns that refer to specific referents are likely to have synonymous definitions, but they're relatively rare; most words won't have ready one-word translations so easily. Try: walk, run, saunter, dash, sashay, hurry, skip, mosey, gallop (varying degrees of the same kind of action); die, expire, pass away, kick the bucket, buy the farm (varying degrees of the formality of an action);... sludge, slush, slime, mush, slurry... (various specific referents in the same close semantic field) -- different languages will handle them all differently (though possibly the common European languages will share more gracefully than others).
When it comes to explain what words like walk, run, saunter, dash, sashay, hurry, skip, mosey, gallop mean, it helps when you use methodology used in thesauri: some terms are included in others some may be specific to certain animals. This is something we hope to achieve with Ultimate Wiktionary. The GEMET thesaurus uses a basic set of relations. These will be hopefully a start for what will be an interesting experiment.
*Muke!
Thanks, GerardM
Gerard Meijssen gerard.meijssen@gmail.com wrote:
Muke Tever wrote:
Sj 2.718281828@gmail.com wrote:
[Interlingua]: ministro; [Japanese]: 大臣 (だいじん, daijin);Will users of non-Roman scripts be able to see transliterations in their own scripts, or will they have to be stuck with Roman?
When you look at the current practice in the nl:wiktionary, you will find that the different scripts are supported. This will also be the case for Ultimate Wiktionary. More interesting is how a right to left implementation will look like, like Arabic, Farsi, Hebrew ..
No, I know the scripts are supported. I meant, in the translation list, if the transliteration such as "daijin" would be transliterated (or transliteratable) into the user's script as well. I know that nl:wikt doesn't give a Hebrew transliteration of "大臣" anywhere, and that en:wikt has actually been deleting Cyrillic transliterations of Chinese (though those were added as entries, not translations).
== [Dutch] == minister ([n.]) 1. [A person commissioned by the government for public service.] "Zware voet jaagt minister Anciaux uit de bocht."
Hopefully there's space for a translation of this example in the user's language.
There will be room for idiom, this line is problematic in that the meaning of "zware voet" is not clear in its own right without context. So as idiom it is not really great.
I think that's meant to be a sample sentence using the word, not necessarily an idiom. But it should bear a translation into the user's language as well.
b) content between || double bars || is stored in the database, so that the two lists of translations for "minister (English, n., 2)" "minister (Dutch, n., 1)" are actually referencing the same list of database translations [marked above by a double asterisk **]
This practice will not be in the UW. Every meaning gets its own listof translations. The practices leads to many examples in the Englishwiktionary where the translation is absolutely wrong. This isparticularly true when later meanings are added.
So translation tables will not be shared across words then? That is good to hear.
What exactly is the translation table attaching to? The definition number? The auto-translating definition? or the individual language's translation of the definition? If the English definition is improved by someone, made more specific (and hopefully it would be) nothing stops a Dutch editor from adding translations (in several languages even) that relate only to the older definition that was faithfully done into his language. Also, the English editor mightn't even think to check an added translation--does he know that the Dutch editor isn't translating from the same definition?
Meaning exist on a global level. When a meaning is added, it will be universally seen to be there. Meanings are linked to a word. Translations will be linked to meanings.
So the problem of translations given to a language's out-of-date version of a meaning will still exist?
[1] There is a difficult question, which we are ignoring for now : just how precisely do all the translations of "minster (English, n., 2)" have synonymous definitions? When are two different words ever truly synonymous? But that is a discussion for another month.
Even when words are not "truly" synonymous for some, they may be "truly" synonymous for others. Some words have only a distinct meaning in certain subcultures. This can be adressed by giving a word multiple meanings. The truly synonymous meaning and the slightly differing meaning.
If a word really has different meanings, that's not a problem, and should be being done anyway.
But this doesn't work if the word doesn't actually *have* multiple meanings to native speakers. English "you" isn't divided into formal and familiar, or masculine and feminine, as it is in some languages. I hope this isn't what you're talking about.
A minister is a special case. Nouns that refer to specific referents are likely to have synonymous definitions, but they're relatively rare; most words won't have ready one-word translations so easily. Try: walk, run, saunter, dash, sashay, hurry, skip, mosey, gallop (varying degrees of the same kind of action); die, expire, pass away, kick the bucket, buy the farm (varying degrees of the formality of an action);... sludge, slush, slime, mush, slurry... (various specific referents in the same close semantic field) -- different languages will handle them all differently (though possibly the common European languages will share more gracefully than others).
When it comes to explain what words like walk, run, saunter, dash, sashay, hurry, skip, mosey, gallop mean, it helps when you use methodology used in thesauri: some terms are included in others some may be specific to certain animals.
Yes, they all have different meanings. They have different pragmatics too. For example it may be that the usual translation in one language actually has a different literal meaning, and that the exact translation is a kind of technical term nobody would use, or may not exist at all. (A well-documented example is that of color terms.) But if, as I understand from what you say above, translation tables are not being shared across words with "synonymous" definitions, then my previous comment on this point doesn't apply.
*Muke!
Muke Tever wrote:
Gerard Meijssen gerard.meijssen@gmail.com wrote:
Muke Tever wrote:
Sj 2.718281828@gmail.com wrote:
[Interlingua]: ministro; [Japanese]: 大臣 (だいじん,daijin);
Will users of non-Roman scripts be able to see transliterations in their own scripts, or will they have to be stuck with Roman?
When you look at the current practice in the nl:wiktionary, you will find that the different scripts are supported. This will also be the case for Ultimate Wiktionary. More interesting is how a right to left implementation will look like, like Arabic, Farsi, Hebrew ..
No, I know the scripts are supported. I meant, in the translation list, if the transliteration such as "daijin" would be transliterated (or transliteratable) into the user's script as well. I know that nl:wikt doesn't give a Hebrew transliteration of "大臣" anywhere, and that en:wikt has actually been deleting Cyrillic transliterations of Chinese (though those were added as entries, not translations).
Whatever my complaints about Gerard's UW, scripts are not one of them.
== [Dutch] == minister ([n.]) 1. [A person commissioned by the government for public service.] "Zware voet jaagt minister Anciaux uit de bocht."
Hopefully there's space for a translation of this example in the user's language.
There will be room for idiom, this line is problematic in that the meaning of "zware voet" is not clear in its own right without context. So as idiom it is not really great.
I think that's meant to be a sample sentence using the word, not necessarily an idiom. But it should bear a translation into the user's language as well.
We haven't really addressed this issue at en:wiktionary. My first impression would be that a quotation to illustrate the use of a word in a foreign language should be in its original, _and_ should be accompanied by its English translation.
What exactly is the translation table attaching to? The definition number? The auto-translating definition? or the individual language's translation of the definition? If the English definition is improved by someone, made more specific (and hopefully it would be) nothing stops a Dutch editor from adding translations (in several languages even) that relate only to the older definition that was faithfully done into his language. Also, the English editor mightn't even think to check an added translation--does he know that the Dutch editor isn't translating from the same definition?
Meaning exist on a global level. When a meaning is added, it will be universally seen to be there. Meanings are linked to a word. Translations will be linked to meanings.
So the problem of translations given to a language's out-of-date version of a meaning will still exist?
"Meaning on a global level" would need to be in some kind of abstraction. Without that you would need to give precedence to some language.
[1] There is a difficult question, which we are ignoring for now : just how precisely do all the translations of "minster (English, n., 2)" have synonymous definitions? When are two different words ever truly synonymous? But that is a discussion for another month.
Even when words are not "truly" synonymous for some, they may be "truly" synonymous for others. Some words have only a distinct meaning in certain subcultures. This can be adressed by giving a word multiple meanings. The truly synonymous meaning and the slightly differing meaning.
If a word really has different meanings, that's not a problem, and should be being done anyway.
How different is different? "Number" and "numeral" have closely related, but different meanings that are not recognized by everyone. The differential will not always be the same in all languages.
But this doesn't work if the word doesn't actually *have* multiple meanings to native speakers. English "you" isn't divided into formal and familiar, or masculine and feminine, as it is in some languages. I hope this isn't what you're talking about.
The 2nd person pronoun is full of problems. The English singular, "thou", is rarely used, so using that in translation would sound very strange to most hearers. In French politeness and social standing play a major role in the choice of singular and plural. Spanish takes politeness one step further by also using "usted", which takes a third person verb. New and old world versions of the language will treat these issues differently. You can't possibly sort this out through a multi-language translating dictionary.
A minister is a special case. Nouns that refer to specific referents are likely to have synonymous definitions, but they're relatively rare; most words won't have ready one-word translations so easily. Try: walk, run, saunter, dash, sashay, hurry, skip, mosey, gallop (varying degrees of the same kind of action); die, expire, pass away, kick the bucket, buy the farm (varying degrees of the formality of an action);... sludge, slush, slime, mush, slurry... (various specific referents in the same close semantic field) -- different languages will handle them all differently (though possibly the common European languages will share more gracefully than others).
When it comes to explain what words like walk, run, saunter, dash, sashay, hurry, skip, mosey, gallop mean, it helps when you use methodology used in thesauri: some terms are included in others some may be specific to certain animals.
Yes, they all have different meanings. They have different pragmatics too. For example it may be that the usual translation in one language actually has a different literal meaning, and that the exact translation is a kind of technical term nobody would use, or may not exist at all. (A well-documented example is that of color terms.) But if, as I understand from what you say above, translation tables are not being shared across words with "synonymous" definitions, then my previous comment on this point doesn't apply.
It's the subtleties that make the difference. Albert Camus' novel "L'Étranger" is variously tianslated as "The Stranger" or "The Outsider". It takes more than crude dictionary knowledge to sort out the distinction.
Before a UW project or any other kind of multilanguage translating dictionary can be an effective tool the material belonging to the separate component languages must first be functional reference tools for the native speakers of each language. English already has over 75,000 entries, but it still has a long way to go before it can be that kind of tool for English speakers. I can't imagine that the much smaller wiktionaries for other languages are any closer.
Ec
Hoi, This is the kind of reply that is helpfull. So I am really pleased to answer these questions. For the understanding of how I think about words, and to a large extend how UW may work, you have to appreciate that language word and meaning relate to each other. Highest in the order is the "language", the language has "word"s, a word has "meaning"s. Meanings have "translations". A translation is nothing more, nothing less than a relation to other words. In a list of translations, only the word will be shown. To get the full meaning of the translation, you have to go to the meaning of this word to appreciate all its subtelties.
Ray Saintonge wrote:
Muke Tever wrote:
Gerard Meijssen gerard.meijssen@gmail.com wrote:
Muke Tever wrote:
Sj 2.718281828@gmail.com wrote:
[Interlingua]: ministro; [Japanese]: 大臣 (だいじん,daijin);
Will users of non-Roman scripts be able to see transliterations in their own scripts, or will they have to be stuck with Roman?
When you look at the current practice in the nl:wiktionary, you will find that the different scripts are supported. This will also be the case for Ultimate Wiktionary. More interesting is how a right to left implementation will look like, like Arabic, Farsi, Hebrew ..
No, I know the scripts are supported. I meant, in the translation list, if the transliteration such as "daijin" would be transliterated (or transliteratable) into the user's script as well. I know that nl:wikt doesn't give a Hebrew transliteration of "大臣" anywhere, and that en:wikt has actually been deleting Cyrillic transliterations of Chinese (though those were added as entries, not translations).
Whatever my complaints about Gerard's UW, scripts are not one of them.
== [Dutch] == minister ([n.]) 1. [A person commissioned by the government for public service.] "Zware voet jaagt minister Anciaux uit de bocht."
Hopefully there's space for a translation of this example in the user's language.
There will be room for idiom, this line is problematic in that the meaning of "zware voet" is not clear in its own right without context. So as idiom it is not really great.
I think that's meant to be a sample sentence using the word, not necessarily an idiom. But it should bear a translation into the user's language as well.
We haven't really addressed this issue at en:wiktionary. My first impression would be that a quotation to illustrate the use of a word in a foreign language should be in its original, _and_ should be accompanied by its English translation.
Idiom will be welcome in UW, it will be treated as a word with the wordtype "idiom". Mind you the wordtype is defined with the word. As such it is possible to add translations as required. I do agree that idiom needs to be in the same language as the word itself.
What exactly is the translation table attaching to? The definition number? The auto-translating definition? or the individual language's translation of the definition? If the English definition is improved by someone, made more specific (and hopefully it would be) nothing stops a Dutch editor from adding translations (in several languages even) that relate only to the older definition that was faithfully done into his language. Also, the English editor mightn't even think to check an added translation--does he know that the Dutch editor isn't translating from the same definition?
Meaning exist on a global level. When a meaning is added, it will be universally seen to be there. Meanings are linked to a word. Translations will be linked to meanings.
So the problem of translations given to a language's out-of-date version of a meaning will still exist?
"Meaning on a global level" would need to be in some kind of abstraction. Without that you would need to give precedence to some language.
Please refer to what I wrote about relations at the top of my answer. When a meaning has been defined, it may have a list of words that are included because these words have the same meaning in the other language eg "you" can have many different words as a translation because of the way grammar works differently in different languages. When the translations of the words are seen as included in the meaning of the English word "you", you will have a list of words all translating to "you" but not being synonymous. From a translation point of view they DO translate to the English meaning of "you" even though they have a distinct meaning in the original language.
When meanings are added at this time, often the translations are not touched making it uncertain if these translations are correct. With UW translations are linked to meanings, these translations will not automagically transfer to a new meaning. One can assume that the orginal meaning stays within the boundaries of what was intended. If not, the translations have to be removed and reapplied by someone who understands both languages.
[1] There is a difficult question, which we are ignoring for now : just how precisely do all the translations of "minster (English, n., 2)" have synonymous definitions? When are two different words ever truly synonymous? But that is a discussion for another month.
Even when words are not "truly" synonymous for some, they may be "truly" synonymous for others. Some words have only a distinct meaning in certain subcultures. This can be adressed by giving a word multiple meanings. The truly synonymous meaning and the slightly differing meaning.
If a word really has different meanings, that's not a problem, and should be being done anyway.
How different is different? "Number" and "numeral" have closely related, but different meanings that are not recognized by everyone. The differential will not always be the same in all languages.
But this doesn't work if the word doesn't actually *have* multiple meanings to native speakers. English "you" isn't divided into formal and familiar, or masculine and feminine, as it is in some languages. I hope this isn't what you're talking about.
The 2nd person pronoun is full of problems. The English singular, "thou", is rarely used, so using that in translation would sound very strange to most hearers. In French politeness and social standing play a major role in the choice of singular and plural. Spanish takes politeness one step further by also using "usted", which takes a third person verb. New and old world versions of the language will treat these issues differently. You can't possibly sort this out through a multi-language translating dictionary.
A minister is a special case. Nouns that refer to specific referents are likely to have synonymous definitions, but they're relatively rare; most words won't have ready one-word translations so easily. Try: walk, run, saunter, dash, sashay, hurry, skip, mosey, gallop (varying degrees of the same kind of action); die, expire, pass away, kick the bucket, buy the farm (varying degrees of the formality of an action);... sludge, slush, slime, mush, slurry... (various specific referents in the same close semantic field) -- different languages will handle them all differently (though possibly the common European languages will share more gracefully than others).
When it comes to explain what words like walk, run, saunter, dash, sashay, hurry, skip, mosey, gallop mean, it helps when you use methodology used in thesauri: some terms are included in others some may be specific to certain animals.
Yes, they all have different meanings. They have different pragmatics too. For example it may be that the usual translation in one language actually has a different literal meaning, and that the exact translation is a kind of technical term nobody would use, or may not exist at all. (A well-documented example is that of color terms.) But if, as I understand from what you say above, translation tables are not being shared across words with "synonymous" definitions, then my previous comment on this point doesn't apply.
It's the subtleties that make the difference. Albert Camus' novel "L'Étranger" is variously tianslated as "The Stranger" or "The Outsider". It takes more than crude dictionary knowledge to sort out the distinction. Before a UW project or any other kind of multilanguage translating dictionary can be an effective tool the material belonging to the separate component languages must first be functional reference tools for the native speakers of each language. English already has over 75,000 entries, but it still has a long way to go before it can be that kind of tool for English speakers. I can't imagine that the much smaller wiktionaries for other languages are any closer.
The current wiktionaries are with respect, similar to a paper based dictionary. It does have little to no functionality that allows it to make a difference in this digital world. When we start to publish content in .dict or RFC 2229 format, people use in effect multiple resources to get their dictionary content. That will be new functionality. It means that new functionality can be bootstrapped onto an incomplete resource. When the OmegaT people are to use UW as a glossary resource in their translationtool, it will be new functionality. Functionality that will enhance the UW content. When UW can become a resource for Open / Free projects, it will gain many words that from a lexicological point of view may be incomplete but that will make it usefull as a resource to some. This will be new functionality.
In effect, there is no reason that I can see to improve the current wiktionaries when the same information can be had in the UW. There are reasons to use UW in stead of individual wiktionaries because the usefullness and the synergy factor is so much better.
Thanks, GerardM
Ec
On 6/23/05, Gerard Meijssen gerard.meijssen@gmail.com wrote:
Hoi, This is the kind of reply that is helpfull. So I am really pleased to answer these questions. For the understanding of how I think about words, and to a large extend how UW may work, you have to appreciate that language word and meaning relate to each other. Highest in the order is the "language", the language has "word"s, a word has "meaning"s. Meanings have "translations". A translation is nothing more, nothing less than a relation to other words. In a list of translations, only the word will be shown. To get the full meaning of the translation, you have to go to the meaning of this word to appreciate all its subtelties.
[snip]
Idiom will be welcome in UW, it will be treated as a word with the wordtype "idiom". Mind you the wordtype is defined with the word. As such it is possible to add translations as required. I do agree that idiom needs to be in the same language as the word itself.
I'm sure you've already considered this, but I'd like to mention it anyway. With idioms, and to some respect all of our possible entries, it could be quite important to include both a literal translation and a gloss; that is, with the idiom "black sheep" (as a very simple example), you would include (something like):
Dutch - Literal: zwart schaap Gloss: ongewone persoon
French - Literal: mouton noir Gloss: personne unordinaire
German - Literal: schwarzes Schaf Gloss: ungewöhnlich Person
Please forgive the basic formatting and my mistakes, in the highly likely chance that I made some. :)
Of course, with the idiom I have chosen, there just so happens to be a regular non-idiomatic usage, but it should give you the basic idea of what I'm on about, no?
[snip]
"Meaning on a global level" would need to be in some kind of abstraction. Without that you would need to give precedence to some language.
Please refer to what I wrote about relations at the top of my answer. When a meaning has been defined, it may have a list of words that are included because these words have the same meaning in the other language eg "you" can have many different words as a translation because of the way grammar works differently in different languages. When the translations of the words are seen as included in the meaning of the English word "you", you will have a list of words all translating to "you" but not being synonymous. From a translation point of view they DO translate to the English meaning of "you" even though they have a distinct meaning in the original language.
Again, I would hope that a note is made in the translations (and not just leave it up to users to figure it out for themselves) that the definitions do not correlate exactly, that "you" encompasses "tu" and "vous" and so on. This becomes important in situations where the word being listed as a translation does not yet exist, or where the languages are sufficiently exotic that one cannot expect the user to have any sort of background knowledge. (For example, how many of you would know exactly what the Ga word "nyɛ" meaning "you" specifically refers to?)
When meanings are added at this time, often the translations are not touched making it uncertain if these translations are correct. With UW translations are linked to meanings, these translations will not automagically transfer to a new meaning. One can assume that the orginal meaning stays within the boundaries of what was intended. If not, the translations have to be removed and reapplied by someone who understands both languages.
[1] There is a difficult question, which we are ignoring for now : just how precisely do all the translations of "minster (English, n., 2)" have synonymous definitions? When are two different words ever truly synonymous? But that is a discussion for another month.
Even when words are not "truly" synonymous for some, they may be "truly" synonymous for others. Some words have only a distinct meaning in certain subcultures. This can be adressed by giving a word multiple meanings. The truly synonymous meaning and the slightly differing meaning.
If a word really has different meanings, that's not a problem, and should be being done anyway.
How different is different? "Number" and "numeral" have closely related, but different meanings that are not recognized by everyone. The differential will not always be the same in all languages.
But this doesn't work if the word doesn't actually *have* multiple meanings to native speakers. English "you" isn't divided into formal and familiar, or masculine and feminine, as it is in some languages. I hope this isn't what you're talking about.
The 2nd person pronoun is full of problems. The English singular, "thou", is rarely used, so using that in translation would sound very strange to most hearers. In French politeness and social standing play a major role in the choice of singular and plural. Spanish takes politeness one step further by also using "usted", which takes a third person verb. New and old world versions of the language will treat these issues differently. You can't possibly sort this out through a multi-language translating dictionary.
A minister is a special case. Nouns that refer to specific referents are likely to have synonymous definitions, but they're relatively rare; most words won't have ready one-word translations so easily. Try: walk, run, saunter, dash, sashay, hurry, skip, mosey, gallop (varying degrees of the same kind of action); die, expire, pass away, kick the bucket, buy the farm (varying degrees of the formality of an action);... sludge, slush, slime, mush, slurry... (various specific referents in the same close semantic field) -- different languages will handle them all differently (though possibly the common European languages will share more gracefully than others).
When it comes to explain what words like walk, run, saunter, dash, sashay, hurry, skip, mosey, gallop mean, it helps when you use methodology used in thesauri: some terms are included in others some may be specific to certain animals.
Yes, they all have different meanings. They have different pragmatics too. For example it may be that the usual translation in one language actually has a different literal meaning, and that the exact translation is a kind of technical term nobody would use, or may not exist at all. (A well-documented example is that of color terms.) But if, as I understand from what you say above, translation tables are not being shared across words with "synonymous" definitions, then my previous comment on this point doesn't apply.
It's the subtleties that make the difference. Albert Camus' novel "L'Étranger" is variously tianslated as "The Stranger" or "The Outsider". It takes more than crude dictionary knowledge to sort out the distinction. Before a UW project or any other kind of multilanguage translating dictionary can be an effective tool the material belonging to the separate component languages must first be functional reference tools for the native speakers of each language. English already has over 75,000 entries, but it still has a long way to go before it can be that kind of tool for English speakers. I can't imagine that the much smaller wiktionaries for other languages are any closer.
The current wiktionaries are with respect, similar to a paper based dictionary. It does have little to no functionality that allows it to make a difference in this digital world. When we start to publish content in .dict or RFC 2229 format, people use in effect multiple resources to get their dictionary content. That will be new functionality. It means that new functionality can be bootstrapped onto an incomplete resource. When the OmegaT people are to use UW as a glossary resource in their translationtool, it will be new functionality. Functionality that will enhance the UW content. When UW can become a resource for Open / Free projects, it will gain many words that from a lexicological point of view may be incomplete but that will make it usefull as a resource to some. This will be new functionality.
In effect, there is no reason that I can see to improve the current wiktionaries when the same information can be had in the UW. There are reasons to use UW in stead of individual wiktionaries because the usefullness and the synergy factor is so much better.
Well, the main reason to improve current Wiktionaries would be so that they remain viable resources while we have no UW, of course.
Thanks, GerardM
Ec
Wiktionary-l mailing list Wiktionary-l@Wikipedia.org http://mail.wikipedia.org/mailman/listinfo/wiktionary-l
On a completely unrelated note, I gave our list a description, so those of you with clever email clients might notice a crazy name being inserted before the email address if you have the list as a contact.
Wytukaze
wiktionary-l@lists.wikimedia.org