Wikidata:Property proposal/is mentioned in

mentions edit

Originally proposed at Wikidata:Property proposal/Person

Not done

Description	Wikidata Eigenschaft eines (nicht-fiktionalen?) Werkes, das Objekte erwähnt (und nicht beschreibt), die nicht Zentralthema sind (de) – (Please translate this into English.)
Data type	Item
Template parameter	n/a
Domain	book (Q571), article (Q191067), etc.
Allowed values	Named Entities with Q-numbers
Example 1	Lebenslauf Friedrich Adam Scholler (Q113679331) mentions Benk (Q792973)
Example 2	Lebenslauf Friedrich Adam Scholler (Q113679331) mentions Nicolaus Zinzendorf (Q76336)
Example 3	Lebenslauf Friedrich Adam Scholler (Q113679331) mentions seminary (Q233324)
Planned use	We have marked up Named Entities in texts, which we would like to add as mentions (without them being descriptions nor main subjects)
Expected completeness	always incomplete (Q21873886)
Robot and gadget jobs	the inverse relationship ("is mentioned in") should perhaps
See also	main subject (P921), described by source (P1343), present in work (P1441)
Single-value constraint	no
Distinct-values constraint	no
Wikidata project	Moravian Knowledge Network Research (Q113678653)

Motivation edit

We, the Moravian Knowledge Network (Q113620869), are working with (and have digitised and OCRed/digitally transcribed) 1000s of texts - reports, letters, talks, biographies - written by members of the Moravian Brethrens. These texts often mention Named Entities (esp. people, organisations, places) which are not the main subject of the text. Using main subject (P921), described by source (P1343), or present in work (P1441) would be semantic misuse, since these mentions are neither, as already stated, the main subject nor do the particularly describe an object nor are the works fictional. It is simply the case that there is a noteworthy mention, which might be important due to the mere connection. The property "is mentioned in" would be the inverse relationship. Is there a need for a separate proposal? MKNetwork (talk) 12:07, 23 September 2022 (UTC)[reply]

Discussion edit

Comment Is using described by source (P1343) really bad, though? tree (Q10884) has one claim for Armenian Soviet Encyclopedia (Q2657718), but the whole encyclopedia is not about trees. This is how I've seen it used. AntisocialRyan (Talk) 14:29, 24 September 2022 (UTC)[reply]

Ah, I see that this has changed to "mentions", which is a better property in my opinion. AntisocialRyan (Talk) 14:29, 24 September 2022 (UTC)[reply]

It's not clear right now whether this is the proposal for 'mentions', or for 'is mentioned in'. The page name is one of these, the subheader another. I think you do need distinct proposals for each.

I'm a little concerned about the cardinality of both of the presumed proposals. If you have 1000s of texts, each having, what, tens of mentioned concepts, how quickly are you going to fill items for the text object and or items for things like seminary (Q233324). (By fill, I mean, there is a maximum amount of data that will fit in a single item, a finite limit.) Do you have any estimates of how many 'is mentioned in' a WD item might get as a result of this proposal and your scanned texts?

This is an interesting exercise, of course. I'm not convinced that wikidata is the right vehicle for it. --Tagishsimon (talk) 16:26, 26 September 2022 (UTC)[reply]

Indeed mentions/is mentioned in could lead to uploading the index of a text to WD, which is overkill. The cross-link between 2 items needs to be substantial. Vicarage (talk) 16:52, 26 September 2022 (UTC)[reply]

Support The existence of a property shouldn't be rejected because there is some theoretical potential fear that it may be overused or misused. If that was the basis of rejection, then we would have to reject many other properties that could also be misused. Even if we were to agree that the cross-link needs to be "substantial", then that is no reason to reject the existence of the property, which is clearly useful in many contexts. The meaning of "substantial" would be quite subjective and depend on the context and who is making the judgement.

PhotomediaTN (talk) 19:05, 15 March 2023 (UTC)[reply]

Oppose there are too many valid claims to list them all in an item. ChristianKl ❪✉❫ 17:37, 28 September 2022 (UTC)[reply]

Support Perhaps for some items, there may very well be be too many valid claims to list them all, but that's no reason to reject the existence of the property. If too many claims are made using it, then these can be rejected. For many, most items, there will not be too many valid claims to list them all.

PhotomediaTN (talk) 19:08, 15 March 2023 (UTC)[reply]

Not done no support for creation of this property --DannyS712 (talk) 01:11, 3 December 2022 (UTC)[reply]

Strong support This is a very useful property that would allow WD to include linked open data of "aboutness" of sources. The objection that the property might prove to be too popular, and therefore so useful that it may result in too many uses (i.e., "cardinality", how many "mentions" per item) is a very weak argument against allowing for the growth and development of WD as knowledge base.

PhotomediaTN (talk) 18:58, 15 March 2023 (UTC)[reply]

I think is a very useful property. There's no way right now to further describe what a work is about other that main topic. I fail to see how further interlinking based on mentions could be detrimental. Fberrizbeitia (talk) 19:56, 15 March 2023 (UTC)[reply]

@Fberrizbeitia "I don't understand why people opposed this" is not a good argument if you want to speak in favor of creating a property. ChristianKl ❪✉❫ 12:49, 16 March 2023 (UTC)[reply]

@ChristianKI "we might end up with a property that is so useful that it results in too many uses" is not a good argument if you want to reject a property. PhotomediaTN (talk) 20:23, 24 March 2023 (UTC)[reply]

On hold PhotomediaTN (talk) 20:06, 15 March 2023 (UTC)[reply]