Wikidata:Property proposal/usage example

Jdh

usage example

usage example

Originally proposed at Wikidata:Property proposal/Lexemes

   Done: usage example (P5831) (Talk and documentation)

Description	example of how a lexeme might be used
Represents	example (Q14944328)
Data type	Monolingual text
Domain	lexeme
Example 1	kindly (L16097) → He was kindly 'Uncle Walt', the amiable old man all American children loved like a member of their own family. (English)
Example 2	mieszkaniec (L13356) → Ten człowiek jest mieszkańcem Warszawy. (Polish)
Example 3	Wiktionary (L3402) → Two automatic systems provide links for the content pages of Wiktionaries. (English)

demonstrates form

Originally proposed at Wikidata:Property proposal/Lexemes

   Done: subject form (P5830) (Talk and documentation)

Description	(qualifier) the form of a lexeme which is demonstrated by a usage example
Represents	inflected form (Q4423888)
Data type	Form
Domain	lexeme
Example 1	kindly (L16097) → L16097-F1
Example 2	mieszkaniec (L13356) → L13356-F5
Example 3	Wiktionary (L3402) → L3402-F2

demonstrates sense

Originally proposed at Wikidata:Property proposal/Lexemes

   Done: subject sense (P6072) (Talk and documentation)

Description	(qualifier) the sense of a lexeme which is demonstrated by a usage example
Data type	Sense
Domain	lexeme
Example 1	kindly (L16097) → L16097-S1
Example 2	mieszkaniec (L13356) → L13356-S1
Example 3	Wiktionary (L3402) → L3402-S1

Motivation

I found this set of properties created on test server by @Lucas Werkmeister (WMDE), Lucas Werkmeister: see 1, 2 and 3. There is example for this property here and here. I think this system of citation nicely connects forms and senses in examples. KaMan (talk) 10:45, 29 August 2018 (UTC)

Discussion

 Support I've been wondering how we should handle supporting quotations - this approach makes sense to me, though I'd be open to other options too. I assume the idea is this would be a statement on the lexeme, with the mentioned qualifiers. The other approach would be to make it a statement on the sense (or the form?) - in which case I think the third property here would not be needed. ArthurPSmith (talk) 17:55, 29 August 2018 (UTC)

Yes, IMHO this makes most sense on the lexeme (though I wasn’t sure of this at first). Each example features both a particular form and a particular sense, so putting the example on the sense feels asymmetrical to me. --Lucas Werkmeister (talk) 15:54, 30 August 2018 (UTC)

 Comment to be honest, I’m not sure if this is data we actually want to collect or if usage examples belong on the individual Wiktionaries…? --Lucas Werkmeister (talk) 15:54, 30 August 2018 (UTC)

Polish words at pl.wiktionary usually have examples, Polish words at en.wiktionary and others usually do not have examples. If examples of usage would be centralised on Wikidata then more wiktionaries could benefit. KaMan (talk) 16:06, 30 August 2018 (UTC)

• Sometimes the sample could refer to a specific spelling of a form (e.g. "en-gb" on visualize/visualise (L11), not "en"), so maybe a language qualifier is needed to. If it just relates to the capitalized form, maybe we should specify that too.
  --- Jura 07:21, 1 September 2018 (UTC)

  I agree, if you select form then all spellings of form are displayed see https://www.wikidata.org/w/index.php?title=Lexeme%3AL123&type=revision&diff=737745674&oldid=728703222 but on the other hand I think we have language qualifiers already available like language of work or name (P407) or original language of film or TV show (P364). Perhaps we need additional property (qualifier) of general usage to point to language code applicable to all properties pointing to forms like object form (P5548). KaMan (talk) 08:03, 1 September 2018 (UTC)

•  Support Wostr (talk) 21:04, 6 September 2018 (UTC)
•  Support Peter Bowman (talk) 12:23, 8 September 2018 (UTC)
•  Support lets get it on! --Shisma (talk) 15:30, 10 September 2018 (UTC)
•  Question what do we need demonstrates form for? isn't it possible to add a usage example as a statement to a form rather then the lexeme? --Loominade (talk) 09:59, 11 September 2018 (UTC)

  No, because there can be always sentence which demonstrates more than one form at once, like Wiktionary and Wiktionaries in sentence "The purpose of this work is to discuss a new proposal based on two previous ones that have all been made regarding a possible support of Wiktionary through Wikidata via a provision of structured language data that can then be represented in the Wiktionaries." Of course we can choose sentences to avoid such situations, but there are well and widely known citations from books, moovies worth to quote with such cases. The same goes to senses, because one can argue that usage example should be in statement of sense. KaMan (talk) 10:20, 11 September 2018 (UTC)

  is it still valid to add the new statement to a form? @KaMan:--Loominade (talk) 08:13, 14 September 2018 (UTC)

  @Loominade: I suppose it is, but you have to consider that in the future it will be hard to prepare mechanism to share examples with Wiktionaries if they would be partially in lexeme level and partially in form level. For the same reason it will be also difficult to write queries once lexemes will be enabled in queries. KaMan (talk) 08:22, 14 September 2018 (UTC)

  @KaMan: now I understand that subject form (P5830) is a direct reference to a form. Now it makes sense to me to use it only on the lexeme. --Loominade (talk) 10:27, 14 September 2018 (UTC)

@Peter Bowman, Wostr, KaMan, ArthurPSmith, Lucas Werkmeister, Jura1: @Shisma, Loominade:  Done: usage example (P5831). − Pintoch (talk) 07:30, 14 September 2018 (UTC)

@Pintoch, Jura1: I suppose the last one is ready to go? — Finn Årup Nielsen (fnielsen) (talk) 11:41, 19 October 2018 (UTC)

@ArthurPSmith: can you please create this third missing property? Similar to subject form (P5830) but for senses. KaMan (talk) 11:19, 21 October 2018 (UTC)

•  Oppose 2nd qualifier as proposed.  Comment I think it would be sufficient to create a more generic property, e.g. "applies to sense" instead. --- Jura 17:34, 21 October 2018 (UTC)

  Right, @Jura1:, I have changed property name according to your proposition. I don't mind name, I need functionality of pointing sense. More general property can be better. But I think form related property should be more generic to so I proposed such change: Property_talk:P5830#More_generic_usage. KaMan (talk) 06:18, 22 October 2018 (UTC)

  Looks good. I edited the desc. Thanks. --- Jura 17:45, 23 October 2018 (UTC)

•  Support If we have a "demonstrates form" qualifier for usage examples, then we should absolutely have a "demonstrates sense" or "applies to sense" qualifier. Edit:  Neutral Given the added sample, it seems as though this "applies to sense" is fulfilling a very different function than what was originally proposed for qualifying usage examples with the senses they demonstrate. Perhaps once "demonstrates sense" has been approved, a separate proposal could be made for this "applies to sense" qualifier. Liamjamesperritt (talk) 23:29, 22 October 2018 (UTC)
•  Oppose the generalized version, since Jura1 has not provided any examples for it and it’s not clear to me at all where else it’s supposed to be useful. Right now, to me this feels like a generic property you could use any time you need a sense-type property, which does not result in useful data (similar to the “as” property we used to have). --Lucas Werkmeister (talk) 21:10, 24 October 2018 (UTC)

  @Jura1: could you please answer to vote of Lucas Werkmeister and provide some examples where else it could be useful. KaMan (talk) 07:20, 29 October 2018 (UTC)

  I added a sample, but I don't agree with the comparison. Not only we still have "applies to part", but the split of another property seems to have created more problems that may have solved and is ill understood even by Wikidata admins. We also had to create an additional property to "unsplit" it for some fields. Obviously, if eventually we find that for some uses we should create another property, I don't see a problem with that. --- Jura 16:14, 29 October 2018 (UTC)

  @Lucas Werkmeister: What do you think about sample provided by Jura1? Sorry for speeding you up guys but I'm in need of using this sense related property and I would like to find some consensus between you two. I use a lot usage example property and there is increasing lack of pointing to the senses. KaMan (talk) 09:05, 31 October 2018

Trungdongnai (UTC)

• @KaMan: well, there was plenty of consensus for the original “demonstrates sense” property before the proposal was changed, so I don’t think the delay is my fault ;)

  @Jura1: I’m not sure if I understand the example correctly, but if you want to have statements that L:L33839#F1 (“Euroland”, uppercase) applies to L:L33839#S2 (“Dutch retail store chain”) whereas L:L33839#F2 (“euroland”, lowercase) applies to L:L33839#S1 (“currency area within the EU”), then I strongly disagree: if the uppercase spelling and the lowercase spelling mean two different things, then IMHO they should be different lexemes (though linked through some property like derived from lexeme (P5191), of course), instead of merging them into one lexeme and then trying to fiddle them apart again through “applies to sense” statements.

  At this point, I would suggest to go ahead and create “demonstrates sense”, and then we can discuss a) whether we want to have “applies to form” and “applies to sense” at all, and b) whether we want to have them as new properties or widen the scope of the existing “demonstrates form” and “demonstrates sense” properties. Both of those questions affect forms and senses equally, so I don’t think it makes sense to discuss the forms half over on Property talk:P5830#More generic usage and the senses half here in the property proposal. --Lucas Werkmeister (talk) 23:13, 31 October 2018 (UTC)

  I don't see where you get that interpretation from. I don't really see the problem with the current name, but there isn't really a need to hurry this, we have older proposals that still need to be created. --- Jura 07:25, 1 November 2018 (UTC)

  I completly agree with Lucas Werkmeister. The example provided by Jura1 is wrong. It should be in separate lexeme. There was already consensus about original wording of property so I restored it. KaMan (talk) 16:20, 2 November 2018 (UTC)

now  Support the more specific version again. (But also, since I’m a WMDE employee, I’m not sure if my vote should count anyways, even when I’m using my private account…) --Lucas Werkmeister (talk) 23:19, 2 November 2018 (UTC)

•  Support If there is a "demonstrate form" qualifier, then we need a "demonstrate sense" or "applies to sense" qualifier. I don't mind the name. Mfilot (talk) 10:33, 28 October 2018 (UTC)
•  Support Karol Szapsza (talk) 17:27, 2 November 2018 (UTC)

@Peter Bowman, Wostr, KaMan, ArthurPSmith, Lucas Werkmeister, Jura1: @Shisma, Loominade, Mfilot:  Done: subject sense (P6072). Feel free to adapt the label if needed. − Pintoch (talk) 17:38, 3 November 2018 (UTC)