Talk:Q31629
Autodescription — type of sport (Q31629)
- Useful links:
- View it! – Images depicting the item on Commons
- Report on constraint conformation of “type of sport” claims and statements. Constraints report for items data
- Parent classes (classes of items which contain this one item)
- Subclasses (classes which contain special kinds of items of this class)
- ⟨
type of sport⟩ on wikidata tree visualisation (external tool)(depth=1) - Generic queries for classes
- See also
- This documentation is generated using
{{Item documentation}}.
superset
edit"type of sport" - subset, "type" - set --Fractaler (talk) 14:19, 26 October 2016 (UTC)
What is the meaning of this item?
editUsually, when we say that something "X is a type of Y", we mean "X is a class that is a subclass of Y", and so that "every instance of X is also instance of Y". For example, we have that church (Q16970)instance of (P31)building type (Q811102), because every (instance of) church (Q16970) is also a (instance of) building (Q41176).
Now, I see that many sports (maybe all of them?) are instances of type of sport (Q31629), and I don't get the meaning of this. For example, in which sense we say that association football (Q2736)instance of (P31)type of sport (Q31629), and why don't we just say that association football (Q2736)instance of (P31)sport (Q349)?
Unless I'm missing something, from a semantical viewpoint the current situation is a mess, and I would like to propose an alternative arrangement.
I would distinguish three variants of items:
| Type of item | Description | Example | What's new? |
|---|---|---|---|
| instances of type of sport (Q31629) & subclass of sport (Q349)[1] | it is a family of sports, but not a sport (thus it is not instance of sport (Q349) or any of its subclasses) | football codes (Q1081491) | Nothing. |
| instances of sport (Q349)[1] | it is a specific sport, whith its own specific rules. | association football (Q2736) |
|
| instances of competition class (Q22936940)[1] | a "variant" of a sport mainly used for classifying participants in a given competition. | women's association football (Q606060) | It depends on the specific item. In the example of women's association football (Q606060), it should be instance of women's sports (Q920057), instead of a subclass of it. |
In simple words, association football (Q2736), basketball (Q5372), etc. are sports, while football codes (Q1081491), team sport (Q216048), mind sport (Q1188693), etc. are types of sports. Note that we usually say "I play association football (Q2736)" or "I play basketball (Q5372)", but we don't say "I play football codes (Q1081491)" (intended as a family of sports) or "I play team sport (Q216048)".
Notified participants of WikiProject Sports Horcrux (talk) 16:37, 9 December 2023 (UTC)
- Yeah it's a mess, you are 100% correct. Nevertheless, I can't support your proposal. Some rather unsorted remarks regarding the status quo:
- Initially we used to build a sports ontology that relates instances of type of sport in a subclass-of hierarchy. This is an ill-fated approach as there is not really a subclass-of relation applicable in most of these situations; much rather, some instances of type of sport commonly share a set of similarities or have a related history, but there is rarely relationship that justifies a subclass-of relationship. The result of this attempt can be seen here for instance. Since at some point all types of sport already had a subclass-of relation which connects them to the sport item, it was considered inappropriate to use this value for P31 as well. We use the metaclass item "type of sport" instead.
- There is a language aspect as well. As a native German speaker, I can confirm that the "<association football> <is a/instance of> <type of sport>" is indeed the way how this is expressed here. Sport as in Q349 is a super abstract concept, IMO one cannot "instantiate" it and thus it is not a useful value for P31. In German it would sound super odd to say "<association football> <is a/instance of> <sport>" as you propose. I am sure this applies to other languages as well.
- We also need to consider that there are many data items which only exist because there is some Wikipedia article somewhere, and users (correctly) try to integrate everything into the Wikidata graph. This is often difficult since many of these items are kinda unique specimens that do not fit anywhere into the current data model. This naturally leads to inconsistencies.
- And there is incompetence sometimes as well. Users try to "fix" something along their personal preferences, but in contrast to the established model. That creates some mess as well.
- It is a great idea to consider alternatives to the subclass-of hierarchy that we currently have, but I do not think that this involves the "instance of: type of sport" claims here. —MisterSynergy (talk) 20:25, 9 December 2023 (UTC)
- I oppose too. I consider sport (Q349) as an abstraction of some (rather broad) activities while association football (Q2736) as a class of more specific activities. Thus I can imagine that instance of (P31)association football (Q2736) would be the game which my son has played last week (so it has the place and the time - specific spatio-temporal entity (Q58415929)), and instance of (P31)sport (Q349) can be my cross-ski run last Advent. So usually we have subclasses of these classes in Wikidata. In opposite it makes sense to have some instance of (P31)type of sport (Q31629) and probably even P279 of it. --Infovarius (talk) 11:58, 12 December 2023 (UTC)