Wikidata:Property proposal/grammatical number

grammatical number

Originally proposed at Wikidata:Property proposal/Lexemes

   Not done

Description	grammatical number of the word
Represents	grammatical number (Q104083)
Data type	Item
Domain	Lexemes and Forms
Allowed values	singular (Q110786), singulative (Q1450795), dual (Q110022), trial (Q2142560), paucal (Q489410), plural (Q146786), collective (Q694268), partitive case (Q857325), indefinite number (Q53998049)
Example	dog (en) → singular (Q110786) planètes (fr) → plural (Q146786)

Tubezlob (🙋) 19:32, 17 April 2018 (UTC)

Discussion

Statement directly in the lexeme for the main form, and in forms for specific forms. Tubezlob (🙋) 19:32, 17 April 2018 (UTC)

• How does this make sense on the Lexeme? --Denny (talk) 19:38, 17 April 2018 (UTC)

  @Denny: For example, the word "mathématiques" (in French) is just plural (a plurale tantum (Q138246)). So it has to be possible to indicate that the lexeme is a plural (Q146786). Tubezlob (🙋) 20:13, 17 April 2018 (UTC)

Wouldn't it be sufficient to simply not have a singular form? --Denny (talk) 20:15, 17 April 2018 (UTC)

@Denny: OK, but where do you indicate that the lexeme is plural? Tubezlob (🙋) 20:17, 17 April 2018 (UTC)

Forms are plural, not lexemes, I'd think. Yes, some lexemes like "mathématiques" only have plural forms, but to me, it seems, this has no import for the lexeme itself. (I might be entirely mistaken and I'm happy to learn, this is not my area of expertise). So yes, on Forms there definitively should be a number property - I am just not sure about Lexemes. --Denny (talk) 00:38, 18 April 2018 (UTC)

•  Support at least for forms,  Comment it could be useful for lexemes if we adopted the several lexemes for a lexeme solution (like we talked a bit about, on Wikidata_talk:Lexicographical_data#Senses_vs_forms or Wikidata_talk:Lexicographical_data#Form:_text_or_lexeme for instance). Cdlt, VIGNERON (talk) 06:33, 18 April 2018 (UTC)
•  Oppose - already covered by the "grammatical features" functionality of Lexemes. Deryck Chan (talk) 09:19, 26 April 2018 (UTC)
•  Comment - but it must have also "indefinite" form, per Basque using it. -Theklan (talk) 13:17, 23 May 2018 (UTC)
• @Tubezlob, Denny, VIGNERON, Theklan: Now that lexemes are out, I believe Deryck Chan above is correct that this is covered by "grammatical features" for forms, which specifically indicate that the form is plural or singular (or I suppose other if needed). Do you disagree? If so can you provide a clear example where "grammatical features" does not cover what you are trying to do? ArthurPSmith (talk) 19:46, 23 May 2018 (UTC)

  @ArthurPSmith: OK it's covered by "grammatical features" for forms. But I cited the example of "mathématiques" in French that is a plurale tantum (Q138246). How can we indicate this if we can't add "grammatical number: plural" in a statement? Tubezlob (🙋) 19:52, 23 May 2018 (UTC)

  Hmm, are you allowed to indicate plurale tantum (Q138246) as the Lexical Category, instead of noun? ArthurPSmith (talk) 19:58, 23 May 2018 (UTC)

  I tried this on Lexeme:L123, it seems to be accepted. Not sure what other effect it may have... ArthurPSmith (talk) 20:01, 23 May 2018 (UTC)

  @ArthurPSmith: not entirely sure but agreed to wait before creation, there is not rush, we should take our time. Tubezlob: "mathématiques" is an interesting example, some sources says it's plurale tantum (Q138246) but meanwhile some mathematicians use the singular "mathématique" (mostly in old text but some modern texts too, it regains some interest, for example Cédric Villani (Q334065) use almost only the singular), Lexeme:L471 will be a nightmare to model too. I still think a property may be useful and do more than the "grammatical features" as properties can have qualifiers and references. Cdlt, VIGNERON (talk) 20:11, 23 May 2018 (UTC)

•  Oppose this is not a property of a Lexeme, but a grammatical feature of a Form. Since grammatical features are modeled directly, no Property is needed. For things like "avionics", use instance of (P31) plurale tantum (Q138246). I'd recommend against using plurale tantum (Q138246) as a lexical category; it's more useful to stick to established parts-of-speech like noun (Q1084) there. -- Duesentrieb (talk) 12:10, 24 May 2018 (UTC)
  • I noted that Lexeme:L362 has the lexical category given name (Q202444). I suppose that is not an "established parts-of-speech"? — Finn Årup Nielsen (fnielsen) (talk) 16:51, 28 May 2018 (UTC)
    • In my opinion, it is not. "Lydia" is a given name (Q202444), but its lexical category should be proper noun (Q147276). -- Duesentrieb (talk) 11:19, 29 May 2018 (UTC)
•  Comment As the suggestion appears now it seems to mix up with the grammatical features. However, there does seem to be a necessity for some form of property that indicates what kind of grammatical numbers are possible for the feature. It relates (also) to mass noun (Q489168) and count noun (Q1520033). I suppose that instance of (P31) could be used as mention above by Duesentrieb? Example. One Danish word that I am unsure of is "vand" (water). Usually that is mass noun (Q489168), so you do not say "one water" nor "two waters". However, in one sense of "vand" you can say "one vand" (short for "one sodavand", soft drink (Q147538)). — Finn Årup Nielsen (fnielsen) (talk) 16:47, 28 May 2018 (UTC)
  • The same ambiguity exists in German: "Wasser" in general (the plural, "Wässer", exists but is so rare it's rather obscure and strange sounding) vs. "ein Wasser" as "a drink of water", the the plural being the same: "zwei Wasser bitte!". One could argue that these should be two separate lexemes, since they have different morphology (different plural forms). Strangely, we do not do the same with food... -- Duesentrieb (talk) 11:31, 29 May 2018 (UTC)
    • @Duesentrieb Counterexample: “zwei Eis am Stiel”. I previously did not consider this as a real plural but now tend to think otherwise, cf. my remarks in German (“Ich frage mich an der Stelle, was wir alles als Plural ansehen sollen.”). -- IvanP (talk) 17:50, 31 May 2018 (UTC)
• @Tubezlob, VIGNERON: is this still needed?
  --- Jura 05:36, 26 July 2018 (UTC)
  • @Jura1: I still think this could be useful, at least to duplicate the grammatical features if we need to add a references. That said, we're still at the beggining so a suggest to  Wait until there is more discussion. Cdlt, VIGNERON (talk) 07:59, 26 July 2018 (UTC)

 Not done Not enough support.--Micru (talk) 17:24, 7 November 2018 (UTC)

---

• @Tubezlob, VIGNERON:: As this was closed as not done, maybe we should formulate an alternative to address the need. --- Jura 17:33, 7 November 2018 (UTC)
  • @Jura1, Tubezlob: do we really need to? I would prefer to find some references to see how to structures this before doing a new proposal (especially as most plurale tantum (Q138246) do have a singular form, I just added it for mathématiques (L19807) for instance and tried to put the data into the "Grammatical features", not sure if it's a best way though). Cdlt, VIGNERON (talk) 17:57, 7 November 2018 (UTC)
    • I think it generally works. I'm not really convinced by your sample. In any case, the problem is that if you want to add references, you'd need some property. Maybe a generic "grammatical features" property to add references should work. --- Jura 18:01, 7 November 2018 (UTC)