Fantastoria

Welcome to Wikidata, Fantastoria!

Wikidata is a free knowledge base that you can edit! It can be read and edited by humans and machines alike and you can go to any item page now and add to this ever-growing database!

Need some help getting started? Here are some pages you can familiarize yourself with:

• Introduction – An introduction to the project.
• Wikidata tours – Interactive tutorials to show you how Wikidata works.
• Community portal – The portal for community members.
• User options – including the 'Babel' extension, to set your language preferences.
• Contents – The main help page for editing and using the site.
• Project chat – Discussions about the project.
• Tools – A collection of user-developed tools to allow for easier completion of some tasks.

Please remember to sign your messages on talk pages by typing four tildes (~~~~); this will automatically insert your username and the date.

If you have any questions, don't hesitate to ask on Project chat. If you want to try out editing, you can use the sandbox to try. Once again, welcome, and I hope you quickly feel comfortable here, and become an active editor for Wikidata.

Best regards! --Epìdosis 07:13, 20 July 2021 (UTC)

character types/stock characters

Hello Fantastoria,

I'm not sure about your changes to character type items (like villain (Q290691)). They are currently handled similar to occupation items (which also have instance of (P31) occupation (Q12737077), not subclass of (P279) occupation (Q12737077)). The idea behind is that villain (Q290691) is the type, not Palpatine (or any other character being a villain (Q290691)). Could you elaborate on the reasons for your changes? Thank you! - Valentina.Anitnelav (talk) 18:33, 27 October 2021 (UTC)

@Valentina.Anitnelav: villain (Q290691), antihero (Q110910)... are roles or archetypes that the characters represent in works of fiction, they are not characters themselves. In the same way that the protagonist or the supporting character are not characters in themselves but roles that the characters represent. They should not be subclasses of characters or instances of them. This does not imply that Palpatine, taking your example, is a character type because the type will remain the villain. That has not changed, only now the villain appears in the tree that corresponds to it as a role. Another thing is that the Wikidata paradigm has been constructed differently, in which case I apologize for the disruption it may have caused. --Fantastoria (talk) 10:01, 28 October 2021 (UTC)

The current modelling of character types is quite established (by practice). But it is still okay to put it up for discussion, once in a while, and the disruption has been a mild one, at most :). Just to respond to your points:

• "villain (Q290691), antihero (Q110910)... are roles or archetypes that the characters represent in works of fiction, they are not characters themselves. In the same way that the protagonist or the secondary character are not characters in themselves but roles that the characters represent."

instance of (P31)fictional character (Q95074) would be wrong here, you are right. But stock character (Q636497) is not a subclass of character (the English label might be a bit misleading) - it is a subclass of archetype, which is a subclass of class (Q16889133). So villain (Q290691)instance of (P31)stock character (Q636497) does not imply that villain is a character but that villain is a stock character [character type] < literary archetype < archetype < class. It may be thought of as the class of all characters that are of the character type villain (Q290691).

• "They should not be subclasses of characters or instances of them."

To take up my last sentence from the last paragraph: It can be thought of as the class of all characters that are of the character type villain (Q290691). So villain (Q290691) would be a class containing Gargamel (Q920224), Vladimir Harkonnen (Q1051419), Cruella de Vil (Q388605), Palpatine (Q51770) and White Witch (Q1997439). Some time ago character types were assigned using instance of (P31). Now we have character type (P9071) as a specialized property, to get things more structured.

• "This does not imply that Palpatine, taking your example, is a character type because the type will remain the villain."

Since we moved from instance of (P31) to character type (P9071) you are probably right - now it depends on interpretation if a character having character type (P9071)villain (Q290691) can be said to be an instance of villain. But if you interpret character type (P9071)villain (Q290691) this way, villain (Q290691)subclass of (P279)stock character (Q636497) would actually imply that Palpatine is a character type.

Which does not mean that it would not be possible to model certain character types as subclasses of character type. There as been some discussions about how to model profession (Q28640) (e.g. Talk:Q28640#Instance_or_subclass) with similar positions (certain jobs as a subclasses of profession (Q28640) vs. certain jobs as instances of profession (Q28640)). The second approach (instances of profession (Q28640)) prevailed (jobs are now instances of profession (Q28640) and direct or indirect subclasses of person (Q215627). To stay at least minimally consistent through domains and concepts I would still hold on to the same approach for stock characters/character types. - Valentina.Anitnelav (talk) 11:13, 28 October 2021 (UTC)

@Valentina.Anitnelav: On the last two points, as you describe it, character type (P9071) is a specialization of instance of (P31), so basically the first property is a way to declare an instance. Seen this way, Palpatine would not be a subclass of type character/stock character, but only an instance of the latter, since, while the classes are transitive, the instances are not (otherwise we would have a duplicate modeling).

On the other hand, the link that it provides on instance/classes is not clear to me that it can serve as a guide. After all, in the world of fiction there are also jobs: a character can be a lawyer, a villain and a protagonist. What I do see clearly is that they are three different "functions" that the same character can represent and none are characters (or subclasses of characters) in my opinion.

On the other hand, and as a brief offtopic that just occurred to me, in character type (P9071) the work of fiction in which that character plays that role should be suggestion constraint (Q62026391), in the same way that it is suggestion constraint (Q62026391) indicated in narrative role (P5800). In addition, perhaps we should have classes that distribute on the one hand the type characters (villain, antihero...) and on the other the narrative roles (protagonist, antagonist...). It seems to me that now they all "hang" from the same branch indistinctly.

--Fantastoria (talk) 09:43, 29 October 2021 (UTC)

character type (P9071) is not really a specialization of instance of (P31), unless you would call occupation (P106) a specialization of instance of (P31), too. But originally it was expressed via instance of (P31).

"Palpatine would not be a subclass of type character/stock character, but only an instance of the latter". It would just be wrong to call him an instance (instance of (P31)) of character type. An instance of a class is also an instance of every superclass - if you say that 1) Palpatine is a villain and that 2) villain is a sublass of character type you imply that 3) Palpatine is also a character type.

villain (Q290691)subclass of (P279)fictional character (Q95074) does not express that villain (Q290691) is a fictional character (Q95074). It does express something along the line of "villain (Q290691) is a type of character". The subclass relationship is the relationship used to express specialization. supervillain (Q6498903)subclass of (P279)villain (Q290691) does express that every super villain is also a villain. villain (Q290691)subclass of (P279)fictional character (Q95074) does express that every villain is a character.

You don't have to think of classes in Wikidata as sets or collections. You can also think of them as types or kinds of things.

I added some suggestions to character type (P9071) [1] and set the required qualifier constraint (Q21510856) statement to normal rank [2]. I also like the idea of splitting up stock characters in the literal sense from narrative functions. There are some borderline cases that may be tricky (some time ago we discussed them in Wikidata_talk:WikiProject_Narration#borderline_cases_for_narrative_Funktion_(P5800)/Charaktertypus_(P9071) - Valentina.Anitnelav (talk) 10:42, 29 October 2021 (UTC)

Hm, I understand what you're saying. According to the modeling, it is necessary that the villain, etc ... be subclasses of fictional characters even though they are really types of fictional characters so as not to produce absurdities in the items that "hang" from them. In order not to miss the point, we use instance of (P31) which is not transitive.

About splitting. There are currently almost 2,500 direct subclasses of fictional character (Q95074). We can create a class for archetypal/role character items such as "villain", "antihero" ... could it be called a "archetypal character"? and another for those of the narrative function character such as "protagonist", "antagonist" ... could it be called a "narrative character"? However, we could further subdivide the tree in view of the huge number of classes out there. Borderline cases can be kept where they are or discussed individually on their respective discussion pages.

--Fantastoria (talk) 13:36, 29 October 2021 (UTC)

I still have to think about it. Here my first thoughts: "Role type" or "narrative role" could be also a label for the class comprising "protagonist", "antagonist", etc. Maybe we could call the other class "character archetype". It could comprise all character types that are not exactly stock characters (e.g. villain, antihero, etc.). Both could be made sibling classes of stock character (Q636497). Then, stock character (Q636497) could be preserved for stock characters in the strict sense (damsel in distress, wise fool, etc.). - Valentina.Anitnelav (talk) 09:07, 3 November 2021 (UTC)

@Valentina.Anitnelav: I am a little confused. We have agreed that villains, antiheroes, protagonists ... are subclasses of fictional character (Q95074) to avoid absurdities. Seeing that there are 2500 classes of fictional character (Q95074), I have proposed to create intermediate classes (two initially: narrative role/narrative character and archetype character/character archetype). However, if we make them siblings of stock character (Q636497), then we would only be changing the value of «instance of» in those items, so the question of 2500 classes that hangs will still be present. For the current modeling to be maintained, narrative role/narrative character and archetype character/character archetype have to be subclasses of fictional character (Q95074) (reduction of the number of child classes) and that would not happen if we make them siblings of stock character (Q636497). Either I get very confused or I am more lost than an octopus in a toolbox :).

Regarding labels and aliases, I think that any of the proposed ones is valid. Perhaps it is the description that I find the most difficult, looking at the description of stock character (Q636497) and fictional character (Q95074). Starting a brainstorm. «archetype character/character archetype»: type of character according to the archetype played in the work. «narrative role/narrative character»: type of character according to the narrative role played in the work. Too tautological to begin with?

--Fantastoria (talk) 14:41, 4 November 2021 (UTC)

Hm... Now I'm a bit confused, too :) Maybe this graphical presentation helps (-> means "subclass of" the class above, * means "instance of" the class above):

I'm thinking about having something like

entity

-> collective entity

-> [...]

-> archetype

-> literary archetype

-> stock character

* crazy cat lady

-> role type

* protagonist

-> character archetype

* villain

* antihero

I'm not sure what you are thinking of. "Seeing that there are 2500 classes of fictional character (Q95074) [...]". Do you see a problem with having 2500 (probably direct and indirect) subclasses? Which problem do you see, exactly? - Valentina.Anitnelav (talk) 18:53, 7 November 2021 (UTC)

@Valentina.Anitnelav: It seems to me that I have made a mess with the links that I was consulting or wrongly copied the number (I put an extra zero, gulp) xD. Sorry. The good thing is that it has served to see that we talk about different things ... or, better, complementary. I follow your symbology. I proposed this:

Originally, now

fictional character

-> antagonist

-> main character

-> supporting character

-> villain

-> antihero

Proposal, after

fictional character

-> narrative character

-> antagonist

-> main character

-> supporting character

-> archetype character

-> villain

-> antihero

Complementary definitely.

Hundreds of instances of a class are not the same as hundreds of subclasses of a class. Some of these subclasses may be grouped into "interclasses" as soon as their descriptions are examined. That was my original idea when I started to check the descending tree of fictional character (Q95074). Of course, I started from the idea of thousands of direct subclasses, which I have found to be a mistake, although it does not affect my initial purpose. I had started by moving part of it to stock character (Q636497), but you have already shown me that it leads to absurdities. In the meantime, I came up with the above, which has led us to two complementary proposals. Therefore, I do not see a problem, but a huge structure with missing parts: an opportunity for improvement.

The only thing is that my "narrative character" and "archetype character" are different from your "role type" and "character archetype".

--Fantastoria (talk) 14:08, 8 November 2021 (UTC)

I see. I feel a bit uneasy with "narrative character" (character with a narrative role?) as this does not transport any information: every character has a narrative role. One could think about "archetypal character" as a superclass of villain, antihero, etc.

I definitely agree with you that a better structure is needed. My approach would actually have been to have a better classification of the facets in which characters are classified: One of the facets is the classification by character type (e.g. villain, hero...), another one by narrative role (e.g. protagonist, minor character...), a third one by species (e.g. human, anthropomorphic dog...), a fourth one by media (e.g. animated character, literary character...), etc. The reason for having so many subclasses is that fictional characters are classified by so many facets (and the abundant use of instance of (P31) for the description of characters).

Here are the results of a sparql query for all (direct) subclasses of fictional character with their "class type" as a graph: query (this is a good visualisation of "my perspective" on the matter).

You see that, here, the results are (partly) quite nicely structured: You have subclasses that are of the "character type" type, you have classes that are of the "fictional humanoid species" type, etc. But we have a lot of classes that do actually "hang" - all the subclasses that are classifications by media, for example. These could be grouped together.

So, to get back to the point: I'm sceptical with "narrative character", I'm okay with "archetypal character". I think I will definitely go for a better classification of "type classes" ("role type", "stock character", "character archetype"), especially with respect to the English label the current classification is wrong ("protagonist" is not a stock character). - Valentina.Anitnelav (talk) 08:58, 9 November 2021 (UTC)

conflicts-with constraint of publisher

Hello Fantastoria,

You set conflicts-with constraint (Q21502838) item of property constraint (P2305) written work (Q47461344) on publisher (P123), and it currently conflicts with subject type constraint (Q21503250) class (P2308) written work (Q47461344). I think I understand your rationale (publisher property is not of work level but of manifestation level in FRBR), but in many entities like Spanish Biographical Dictionary (Q41705771), some aspects of objects are bundled as one entity. In such a case, conflicts-with constraint (Q21502838) written work (Q47461344) seems to be a too strong constraint. What do you think about removing both subject type constraint (Q21503250) class (P2308) written work (Q47461344) and conflicts-with constraint (Q21502838) item of property constraint (P2305) written work (Q47461344)? This change will result in arousing an error when there is no appropriate class. In the case of Spanish Biographical Dictionary (Q41705771), there will be no error because online database (Q7094076) and book (Q571) are subclasses of allowed classes. We can use applies to part (P518) qualifier for designate the aspects of entity which corresponds to the publisher information for such a mixed entity if it is needed. --KAMEDA, Akihiro (talk) 03:14, 13 October 2022 (UTC)

Hello @KAMEDA, Akihiro:

If with this change it is maintained that the items that are strictly instances of written work (Q47461344) or of any of their subclasses will continue to maintain the restriction, I have no argument to oppose it. Implement the improvement.

I don't fully understand the internal relations of the conflicts between properties, items, restrictions or exceptions. I check as others have done before adding a restriction or an exception. Probably, I set off the occasional alarm from time to time. My apologies.

P.S. The same restriction in cover art by (P736) should also be corrected.

Un saludo. --Fantastoria (talk) 08:07, 13 October 2022 (UTC)

Thank you for your reply. I edited the constraint and checked that an error arouse in strict instances of written work (Q47461344) such as The Quincunx (Q246446). I'll wait feedbacks from others before editing cover art by (P736).

Un saludo. --KAMEDA, Akihiro (talk) 08:49, 13 October 2022 (UTC)