Wikidata:Property proposal/mentions named entity

From Wikidata
Jump to navigation Jump to search

mentions named entity[edit]

Return to Wikidata:Property proposal/Creative work

   Under discussion
Descriptionnamed entity mentioned only by name in a work (use P1441 for entities present in the events of a work; use P840 for locations that are part of the setting of a work)
RepresentsName-dropping (Q1424112)
Data typeItem
Domaincreative work (Q17537576)
Allowed valuesAny entity with a property that's instance of (P31)Wikidata property to indicate a name (Q19643892)
Example 1Nobody Speak (Q26885626)Donald Trump (Q22686) (named as (P1810) → "Trump")
Example 2That's Not Me (Q18164235)Gucci (Q178516)
Example 3We Didn't Start the Fire (Q1448949)Budapest (Q1781)
Example 4Paper Towns (Q1138063)The Mountain Goats (Q145515) ("Ben and Radar showed up at eight on the dot. I got in the backseat. They were shouting along to a song by the Mountain Goats.")
See also

Motivation[edit]

I think this is interesting cultural data to capture, for instance, it provides data on things like which fashion brand is name-dropped the most in rap. However, there is also application to literary works - with the distinction from present in work (P1441) being that these entities are not actually present in the events of a work, only mentioned by name. See discussion at project chat here: Wikidata:Project_chat#How_can_we_state_that_a_person/entity_was_named_in_a_creative_work

The difference from other proposals such as Wikidata:Property_proposal/mentioned_in_work and Wikidata:Property_proposal/Mentioned_at is the better defined scope/application of this property not currently covered by existing related properties (and also this property is a statement on the work item, since an entity could be mentioned in thousands of works - leading to excessive statements on the entity item). --SilentSpike (talk) 16:44, 8 November 2019 (UTC)

Discussion[edit]

  • Symbol support vote.svg Support Seems useful. A lot of editors have been using a lot of different properties to model this relationship, this makes it a lot clearer. If this goes through, I propose a "mini task force" to identify and migrate previous edits using "main subject", "present in work", etc. to map mentions. Moebeus (talk) 17:00, 8 November 2019 (UTC)
  • I would like to see this property be able to be extended (with a qualifier) to cover allusions as well as specific mentions. E.g. Everything Zen (Q5418126) referencing Life on Mars? (Q2077201) ("Mickey Mouse has grown up a cow") or "Yes, I's Finished/Abie Baby" from Hair (Q1165499) paraphrasing Gettysburg Address (Q214524). (Alternatively, rather than a qualifier there could be two separate properties.) - Jmabel (talk) 18:06, 8 November 2019 (UTC)
  • Symbol support vote.svg Support I like SilentSpike's example of name-dropping in rap and I think that Moebeus (above and in the discussion at Project chat) and Jura (in the discussion at project chat) have a good point: There are already attempts to model this kind of relationship (with reasonable motivations), just that other properties are exploited for that. - Valentina.Anitnelav (talk) 11:13, 9 November 2019 (UTC)
  • Pictogram voting comment.svg Comment Responding to some of Jura's questions (from project chat) here for archival purposes:
What type of work would this apply to? How many such statements should we have? Most works have a person or place index should these all be included?
—Jura
  1. I think this can apply to any sort of work that has written or spoken content. My main motivation is the application to lyrical works as that's where I see the most value in how this represents cultural influence and reach (i.e. if many songs name the same entity, then it is shown to be culturally significant). However, the same line of thinking can be extended to written works (see example 4 of the proposal) where the existing property for something's presence in the work doesn't quite cover the case of an entity only mentioned by name.
  2. As many as are valid statements? I don't see an abundance of statements existing for most works since the use case and allowed values of this property limit what it applies to (i.e. persons/locations actually present in the work are already captured via other properties and this is also only applicable entities with names - aka "grass" does not mean there should be a statement for grass (Q643352)). As for works that are very referential, that seems to me like a valid use case where there may exist many values and the use of this property helps to establish that the work is heavily referential in a machine readable way.
  3. Good question, I'm not sure to be honest. My gut reaction is no since those seem like metadata about the work rather than actually being part of the work. However, perhaps they should with a qualifier to indicate that they're part of the index (also keep in mind, persons/places that are present in the work are already captured by other properties and don't belong in this property about entities only named).
--SilentSpike (talk) 11:40, 10 November 2019 (UTC)
  • Pictogram voting comment.svg Comment Is this proposal equivalent to Wikidata:Property proposal/references ? Presumably one wouldn't need both? Should the proposals be merged? Jheald (talk) 17:46, 11 November 2019 (UTC)
    • I don't believe so since there's a distinctly different set of allowed values. That property would cover works paying tribute to other works by borrowing ideas, concepts or events specific to them (which cannot really apply to any given entity). Whereas this property can apply to any given entity explicitly named (only) in a work. I feel like that's not the best explination possible, perhaps Valentina.Anitnelav can shed some light on the difference more succinctly. --SilentSpike (talk) 19:55, 11 November 2019 (UTC)
    • Currently they are not exactly equivalent (as explained by SilentSpike) and I admit that I'm also a bit partial to having a distinct property for references between works. On the other hand it might be actually possible to subsume this ("mentions") proposal under Wikidata:Property proposal/references by broadening the latter's scope to include all kinds of entities. If such a property should be accepted (in spite of concerns that it might be too broad) one should probably follow Jmabels proposal to indicate the kind of reference via a qualifier, maybe a new one along the lines of <kind of reference> that may take values like "mention", "allusion", "parody", "copy of style", etc. - Valentina.Anitnelav (talk) 20:08, 11 November 2019 (UTC)
  • Symbol neutral vote.svg Neutral Generally too many in long books; don't you think so? Nomen ad hoc (talk) 20:15, 12 November 2019 (UTC).
    • Do you have an example? I can't really respond since I can't think of any books that include many mentions of entities which aren't characters or locations in the setting. --SilentSpike (talk) 20:34, 12 November 2019 (UTC)
      • But what about essays? Nomen ad hoc (talk) 20:47, 12 November 2019 (UTC).
        • Good question, I suppose it's worth considering how (or if) this should apply to non-fiction works which may mention many entities due to their nature (e.g. Dictionary of Greek and Roman Biography and Mythology (Q1210336)). I almost feel like it shouldn't since in those cases the entities are part of the subject of the work and not merely mentioned by name. --SilentSpike (talk) 21:39, 12 November 2019 (UTC)
          • For non-fiction books that aren't dictionaries nor encyclopedias, perhaps could we consider that the property should be used chapter by chapter (each one having an item)? Nomen ad hoc (talk) 21:53, 12 November 2019 (UTC).
  • Symbol oppose vote.svg Oppose For now, If papers with 2,000 authors are causing problems, then how are we going to deal with works that mention 2,000 - or 20,000 - people or things? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:04, 17 November 2019 (UTC)
    • This seems like a vote to Symbol wait.svg Wait. Presumably until there is a solution/rule/agreement in place to handle items which could have many statements? I'm still personally sceptical of the existence of works with so many name-dropped entities (outside of non-fiction works which this could be refined to exclude), but admit that it's a possibility that one could exist. --SilentSpike (talk) 00:24, 19 November 2019 (UTC)
  • Symbol support vote.svg Support i need a property like this.--Shisma (talk) 17:09, 18 November 2019 (UTC)