Wikidata:Property proposal/applies to name of object

applies to name of object edit

Originally proposed at Wikidata:Property proposal/Generic

Done: applies to name of object (P8338) (Talk and documentation)

Description	(qualifier) label or alias name (of the claim object) to which the claim applies
Data type	Monolingual text
Example	⟨ Uno (Q17267)    ⟩ named after (P138) ⟨ 1 (Q199)    ⟩ applies to name of object (P8338) ⟨ "uno" (Spanish) ⟩ ⟨ Vermont (Q16551)    ⟩ named after (P138) ⟨ Green Mountains (Q41507)    ⟩ applies to name of object (P8338) ⟨ "montagnes Vertes" (French) ⟩
See also	applies to name of subject (P5168), applies to part (P518)

Motivation edit

applies to name of subject (P5168) is intended to be used to indicate to which name a claim applies for the item the claim is on (the subject). However, it's naming (in English anyway) is ambiguous and could be interpreted as the same thing but with respect to a name of the claim's item value (the object).

This wouldn't be an issue for trivial cases where all of the following is true:

The subject and object do not have similar names
One of (or both) the subject and object is confirmed to have a single name

In such cases it can be implicitly determined which (the subject or object) the claim was intended to apply to. If they have similar names, then there's an obvious ambiguity. If both have multiple names then there's ambiguity in that (while it may be true the value matches a name listed only for one) it is unknown whether Wikidata could be missing name information which takes us back to the case of similar names.

I think, in general, it's better to explicitly specify data than rely on an implicit connection by matching the value to a name of either the subject or object and assuming this is always a trivial case. Additionally, by having these two separate properties it fixes any ambiguity in the use of P5168 which was originally intended to only apply to a name of the subject - which if we assume is still the case then this separate property is necessary for situations as described above anyway. I see this as analogous to subject has role (P2868) and object has role (P3831).

If this property is approved then I would also recommend updating the labels of P5168 to reflect that it is for the subject. Existing statements could be reviewed for misuse by filtering for cases where the value does not match any name listed for the subject. SilentSpike (talk) 15:33, 18 February 2020 (UTC)[reply]

Discussion edit

Support, but I think the datatype should be a lexeme:
⟨ Uno (Q17267)    ⟩ named after (P138) ⟨ 1 (Q199)    ⟩
applies to name of object (P8338) ⟨ uno (L44969) ⟩
--Tinker Bell ★ ♥ 19:43, 19 February 2020 (UTC)[reply]
- An interesting proposal I hadn't considered. However, would this not be too limiting? How would that work for the second example above? --SilentSpike (talk) 21:02, 19 February 2020 (UTC)[reply]
  - Can't we create a lexeme for the French word "montagnes Vertes" Germartin1 (talk) 07:56, 9 June 2020 (UTC)[reply]

Oppose This proposal is basically for a "tickbox" property with no actual target Wikidata item, from a data point of view, it just doesn't work (and I don't think we should want it to). Circeus (talk) 13:30, 28 February 2020 (UTC)[reply]

Huh? I'm afraid I don't understand your objection. It's not a boolean value, it's a monolingual string - a valid Wikidata datatype. It specifies which name of the claim object the claim applies to. Very similar to the existing property which specifies which name of the claim subject the claim applies to. Analogous to the existing role qualifiers for subject and object. There's precedent for properties of this type. --SilentSpike (talk) 13:49, 28 February 2020 (UTC)[reply]

Support I'd use "applies to name of value" or "applies to name of statement value" as label. --- Jura 10:11, 26 May 2020 (UTC)[reply]
- @Jura1: I figured it should be consistent with the "object/subject has role" properties, but am not opposed to such naming. --SilentSpike (talk) 13:00, 26 May 2020 (UTC)[reply]
  - I figured, but "object/subject has role" tends to be ill understood and in combination with "name of", it's likely to get worse. Maybe someone can come up with even a better label. --- Jura 13:05, 26 May 2020 (UTC)[reply]

@SilentSpike, Tinker Bell, Circeus, Jura1: applies to name of object (P8338) has been created. Pamputt (talk) 17:13, 16 June 2020 (UTC)[reply]