Wikidata:Property proposal/form applies to sense

form applies to sense

Originally proposed at Wikidata:Property proposal/Lexemes

   Not done

Description	this form is only or mainly applicable this sense
Data type	Sense
Domain	form
Allowed values	sense of the same lexeme
Example 1	mouses (L1119-F4) → mouse (L1119-S2)
Example 2	hanged (L4116-F5) → hang (L4116-S3)
Example 3	flied (L277-F6) → fly (L277-S3)
Example 4	costed (L3953-F6) → cost (L3953-S2)

GZWDer (talk) 20:07, 14 January 2023 (UTC)

Discussion

•  Support--Tinker Bell ★ ♥ 19:16, 16 January 2023 (UTC)
•  Comment I can see this being useful, but I am unfamiliar with the usage you suggest in the first example - I've never heard "mouses" as a plural for the computer input device; "mice" works fine for this. ArthurPSmith (talk) 13:39, 17 January 2023 (UTC)
•  Comment This proposal is related to another proposal by GZWDer to (yet again) restructure all the Vietnamese lexemes, which is currently under discussion. – Minh Nguyễn ^💬 16:47, 17 January 2023 (UTC)
•  Oppose We already have requires grammatical feature (P5713) which allows us to associate a sense only with forms which have matching grammatical features. This way, if we add new forms, we do not have to worry about each new one being linked to an appropriate sense so long as the sense has a statement like this, and we do not have to check each form if retrieving information based on senses (to construct sentences for example). If linking in this direction is not viable for some reason, I would say that is reason to re-evaluate the way the lexeme is modeled - either by creating separate lexemes or by rearranging the grammatical features. This proposal does not have many examples, but if we take "hang"/"hung"/"hanged," this is somewhere that two lexemes would make sense. It is not entirely true to say "hanged" only applies to one sense; it is the older form of the past tense of this verb and in older texts it would apply to any of the senses of the verb. Rather than complicate things by further having to qualify that the form only applies to a sense after a certain point in time, we could create a new "hang"/"hung" lexeme which represents the verb senses for which "hung" has fully replaced "hanged" and treat it as a derivative of "hang"/"hanged" on which all senses except the one related to execution are archaic. --عُثمان (talk) 19:06, 17 January 2023 (UTC)

   Comment Judging by the examples given on the property requires grammatical feature (P5713) itself via Wikidata property example for lexemes (P5192) and Wikidata property example for senses (P5977), requires grammatical feature (P5713) would not allow us to achieve what this proposed property is intended to solve. (It seems requires grammatical feature (P5713) pertains to either the grammatical features of dependent lexemes in a syntax tree or to the addition of grammatical features (such as definiteness) to a use of the lexeme; if we wanted to be able to disambiguate between forms, we might have to invent artificial grammatical features just for this purpose, which would be a pollution of the knowledge base.) @عُثمان: But I concur with your other point (see below). ―BlaueBlüte (talk) 05:06, 1 February 2023 (UTC)

•  Oppose We should not need this. My feeling is that as a general rule, if a lexeme inflects differently depending on the sense, or if it isn’t self-consistent because it has, for example, form variants that aren’t consistent with all other elements of the lexeme including all senses, then it shouldn’t be one and the same lexeme. Split it even if the lemma might be the same or might look unusual to traditional-dictionary users (because we might add inflected forms for disambiguation, as in the case of a lemma like “hang/hung” proposed above by User:عُثمان; but people should already be familiar with such lemmas based on people stating their pronouns in this fashion: “xe/xem/xyr” and the like). In case we agree on this, we should of course state it somewhere as a guideline for modeling lexemes. (A lexeme that is self-consistent in this sense may still encompass variants, but those would be dependent only on external factors, such as regiolect.) ―BlaueBlüte (talk) 05:06, 1 February 2023 (UTC)
•  Oppose As a corollary to what BlaueBlüte wrote, if a representation of a form applies only to a certain sense, then the lexeme also needs to be broken up somehow. What this means in Vietnamese is that Nôm representations/forms should go in separate Nôm-specific lexemes that can fully express the nuanced relationship between the two writing systems. I would disagree with the use of the proposed property to shoehorn chữ Nôm into the same lexeme, pretend that it's akin to an unnamed grammatical feature, and twist the lexeme in knots to simulate a separate lexeme. – Minh Nguyễn ^💬 06:15, 5 February 2023 (UTC)
•  Oppose as per blaueblute--So9q (talk) 13:40, 17 February 2023 (UTC)
•  Not done Clear opposition to creating this and no attempt has been made to address the objections raised. - Nikki (talk) 19:03, 31 March 2023 (UTC)