Wikidata:Property proposal/Narrative role

narrative role edit

Originally proposed at Wikidata:Property proposal/Creative work

Done: narrative role (P5800) (Talk and documentation)

Description	narrative role
Data type	Item
Domain	all types of narrative works
Allowed values	protagonist (Q215972), antagonist (Q245204), unreliable narrator (Q650118), ~~extra (Q658371)~~…
Example 1	A Study in Scarlet (Q223131) → characters (P674) → Dr. John Watson (Q187349) → [narrative role] → unreliable narrator (Q650118)
Example 2	A Study in Scarlet (Q223131) → characters (P674) → Sherlock Holmes (Q4653) → [narrative role] → protagonist (Q215972)
Example 3	Super Mario World (Q853143) → characters (P674) → Mario (Q12379) → [narrative role] → player character (Q1062345)
Example 4	Super Mario World (Q853143) → characters (P674) → Princess Peach (Q507001) → [narrative role] → damsel in distress (Q2293837)
Example 5	Princess Peach (Q507001) → [narrative role] → damsel in distress (Q2293837)

Motivation

the proposed property should be used only as a qualifier for characters (P674), similar to cast member (P161) and character role (P453). currently there is no way to express what the character means to the plot of the work. I have been using main subject (P921) to express which is the main character, but the main character and what the work is about, is not always the same thing. feedback is appreciated --Shisma (talk) 19:05, 2 July 2018 (UTC)[reply]

Notified participants of WikiProject Performing arts
--- Jura 13:56, 11 July 2018 (UTC)[reply]

Discussion

Comment On the one hand I like the idea of having an own property for this purpose, but on the other hand I'm not sure if it is really needed: I just use the more general object has role (P3831) (see *Buddenbrooks (Q326909), for instance) and it works quite well. Do you see any advantage of having this own property over using object has role (P3831)? - Valentina.Anitnelav (talk) 20:53, 2 July 2018 (UTC)[reply]

Comment @Valentina.Anitnelav: I did not know it exists and that is probably the best answer I can give you 😁. Why don't we use object has role (P3831) instead of cast member (P161)? I'd argue that specify makes it easier to find and more self-explanatory. But I don't know where I stand on the abstract vs. specific statements issue. --Shisma (talk) 16:01, 3 July 2018 (UTC)[reply]
Oppose There is already object has role (P3831) and subject has role (P2868) David (talk) 15:01, 3 July 2018 (UTC)[reply]
- @ديفيد عادل وهبة خليل 2: did we miss some points in the discussion below?
  --- Jura 05:20, 8 July 2018 (UTC)[reply]
Oppose use object has role (P3831) --Pasleim (talk) 16:29, 3 July 2018 (UTC)[reply]
Comment Thanks to Valentina.Anitnelav, I started experimenting with the value mentioned above using object has role (P3831). I think some people currently use other qualifiers (e.g. P161). In any case, some uses require object has role (P3831) and other subject has role (P2868) (I added a 5th sample above). It might be preferable to normalize this to one property, allowing an easy check on the values being used.
--- Jura 19:44, 3 July 2018 (UTC)[reply]
- Shisma, Jura: Good points. I actually prefer a way to indicate the narrative role as a statement on the character qualified with of (P642)+work (example 5) for the following reasons:

Especially in more complex narrations those statements can be contested (e.g. different critics with different interpretations). It seems much cleaner to me to indicate it at the character-item (with the correspondending sources, maybe qualified using nature of statement (P5102), statement supported by (P3680), statement disputed by (P1310)) than to add multiple characters (P674)-statements with different <narrative role>/object has role (P3831)-qualifiers (and confusing nature of statement (P5102)/statement supported by (P3680)/statement disputed by (P1310)-qualifiers) to the work.
There are statements about the narrative role that have to be qualified themselves (e.g. foil (Q5563727), used at Christian Buddenbrook (Q42324815) or The Little Man (Q5787753)).

I would support this new property if it can (also) be used in statements on the character-item (I would actually think about making it the main use). subject has role (P2868) works well, but there are opinions that it should only be used as a qualifier and "narrative role" as a label would be more speaking. It should be possible to use this new property also with literary archetypes (e.g. ingenue (Q117482), prince charming (Q1090843)). - Valentina.Anitnelav (talk) 09:22, 6 July 2018 (UTC)[reply]

Comment btw. the statement might also make sense as a qualifier for cast member (P161):
The 39 Steps (Q501105) → cast member (P161) → Alfred Hitchcock (Q7374) → [narrative role] → extra (Q658371)
If a role has no entity, as it is often the case with extras. --Shisma (talk) 07:11, 7 July 2018 (UTC)[reply]
- When checking current uses of "protagonist" with the "character role" qualifier, I found that some of these were conversions from the deleted qualifier "as". For this (incorrect) use, we do need a better solution. This proposed property would be perfect.
  --- Jura 07:33, 7 July 2018 (UTC)[reply]
Comment. 1. extra (Q658371) shouldn't be listed among the allowed values, as the new property would be dealing with characters, not actors. 2. You have to take into account the existence of narrator (P2438), which by the way should probably be redesigned, but should still cover some of the examples given here. Thierry Caro (talk) 01:27, 9 July 2018 (UTC)[reply]
- I'm not sure if the labels in all languages mean the same, but based on the statements on the item, it shouldn't: I deleted it.
  --- Jura 13:31, 11 July 2018 (UTC)[reply]
Support per discussion above. Seems preferable over the current solution with various qualifiers.
--- Jura 13:32, 11 July 2018 (UTC)[reply]

I tried to sort entities that I'd like to categorize with characters by their context in order to decide which entities make sense. some of them may require a different property.

Narrative role	Character stereotype	Casting choice	Video game mechanics
protagonist (Q215972), supporting character (Q4388180), sidekick (Q281525)…	Mary Sue (Q1254327), damsel in distress (Q2293837), villain (Q290691), companion (Q55497622) (see list of stock characters (Q21213748))…	ensemble cast (Q1154757), extra (Q658371), cameo appearance (Q191796), guest appearance (Q28228)	player character (Q1062345), non-player character (Q466466)

The column Narrative role should definitively only be used as a qualifier of characters (P674) because they only make sense in context of a story. The Character stereotype column only makes sense in Character entities. Casting choice should probably be a different Property and it should be used as a qualifier of cast member (P161) (though there are extra characters in non-action narratives). The column Video game mechanics makes only sense if you consider a game to have a narrative which in modern games is the norm rather then the exception. as soon as a Game entity has a characters (P674) property, it makes sense to assume the game has some sort of narrative. Objections? Ideas? --Shisma (talk) 10:20, 13 July 2018 (UTC)[reply]

Narrative roles my also apply to objects like deus ex machina (Q203698), MacGuffin (Q488733) or Chekhov's gun (Q4352783) but there is no property to describe The War of the Worlds (Q213019) → [is featuring] → bacteria (Q10876) → [narrative role] → deus ex machina (Q203698)--Shisma (talk) 10:34, 13 July 2018 (UTC)[reply]

I don't think that the separation into "narrative role" (as a feature of a character in the context of a story) and character stereotype (as a feature of a character itself) can be made that clearly, as almost every character stereotype does depend on the portrayal of a character in the context of a story: one approach can present a character as a villain, another one can challenge this presentation (one example from pop culture is Maleficent (Q1660555), which is portrayed as a stereotypical villain in Sleeping Beauty (Q215617), but a bit more nuanced in Maleficent (Q172271)). Especially to be a damsel in distress (Q2293837) does depend on the story. So one would need to qualify all those statements to a work.

As the boundary between narrative role and character stereotype is that vague (think damsel in distress (Q2293837), but also confidant (Q5160063)) both should be indicated at the same place. For the reasons I gave above I actually think that the character itself is the best place. To restrict the statement to a certain work one can just use of (P642). I would not mind redundancy by indicating narrative roles like protagonist, antagonist, supporting and minor character both at characters (P674) and at the character.

I agree that "Casting choice" should have a different property. For extra characters one can just use minor character (Q27623618). I would not make a difference between games and "conventional" narrative works. - Valentina.Anitnelav (talk) 10:51, 13 July 2018 (UTC)[reply]

@Valentina.Anitnelav: do you support the proposal?
--- Jura 12:29, 28 July 2018 (UTC)[reply]
Support - Valentina.Anitnelav (talk) 12:34, 28 July 2018 (UTC)[reply]
Weak support — Either this or something similar to this is needed. — Mr. Guye (talk) (contribs) 21:19, 31 August 2018 (UTC)[reply]

@Valentina.Anitnelav, Shisma, Jura1: Done: narrative role (P5800). − Pintoch (talk) 08:03, 7 September 2018 (UTC)[reply]