Wikidata talk:WikiProject Religions

From Wikidata
Jump to navigation Jump to search


Nojhan Yair rand Runner1928 TomT0m Capankajsmilyo ArthurPSmith John Carter Nomen ad hoc Pictogram voting comment.svg Notified participants of WikiProject Religions,

Here is some queries and comments of the current states of Wikidata about religions :

I'm stopping here as clearly there is no strong coherent structure for the data on religion. Hopefully this project will sort that out.

There is a lot of work to do but plenty of low-hanging fruits too.

Cdlt, VIGNERON (talk) 23:33, 12 February 2016 (UTC)

PS: I rename the project in CamelCase to harmonise with other WikiProjects and so {{Ping project}} can work.

PPS: bonus, below is the {{Tree}} of subclass of (P279) of religion (Q9174) :

The subclass tree of religion (Q9174) needs some fixing. None of those classes apply to all religions (afaik). I'm also not sure about the ontology proposal given here. It doesn't seem correct that a religion would be both a subclass and instance of religion. It seems rather contradictory. --Yair rand (talk) 21:22, 16 February 2016 (UTC)
Regarding the subclass tree, I think that religion (Q9174) should inherit directly from social group (Q874405), the other seems correct for me.
I've proposed to use both instance of (P31) and subclass of (P279) mainly because of the historical perspective; For instance (haha), the first buddhism (the one existing during Gautama's life) is a religion with the name buddhism. But very rapidly comes different movement, that refine the idea of what is a buddhism, all in different ways. They all claim to be (the true) buddhism. The original one is not practice anymore, and one can consider that it is thus no more a religion, but as it was practiced as one during a given time, I failed to see why it would not be a religion. But then, how to tag the movement that refined the original idea? They are obviously subclasses of the original one.
Should we have two elements for each religion? One for the class and one for the instance? How to call them? Should we have buddhism (the class) and historical buddhism (the instance)? I feel this would be difficult to maintain two elements for each ones, thus my proposal to have both within each element. — nojhan () 09:15, 17 February 2016 (UTC)
For the subclass tree, I've got no idea. It seems pretty right to me.
For instance of (P31) & subclass of (P279), I'm confusing too by using the two on the same item. I'd rather prefer using only subclass of (P279) ; my understanding is that buddhism during Gautama's life or christianism during Jesus' life is not a religion, not fully, not yet, it's only afterward that people considered it as a religion (Gautama probably considered himself like a-sort of-kind of hindouist and Jesus as a jew). But this transition period are very akward and unclear.
Cdlt, VIGNERON (talk) 13:30, 17 February 2016 (UTC)
My point could be applied to any movement that have disappeared, not necessarily the "first" one. The difficulty is not to decide if it is a religion or not (albeit it is clearly one, IMHO), but rather to know: if Yungdrung Bön is only a sub-class of buddhism and bön (why not), then what is the (an?) instance of Yungdrung Bön? I failed to see how to rigorously decide what is an instance. — nojhan () 17:51, 17 February 2016 (UTC)
I realized that maybe I think of the wrong problem: what if an idea have no instance? Could we use only subclasses and never have any instance of religions? — nojhan () 18:41, 17 February 2016 (UTC)
After some research, this seems to be ã similar problem than the Wikidata:Requests for comment/Are colors instance-of or subclass-of color. — nojhan () 19:28, 17 February 2016 (UTC)
Another problem I'm thinking of is: how to differenciate influences from inheritance? Is Yezidism a sub-class of Islam influenced by Mazdeism? Of Mazdeism, influenced by Islam? A sub-class of both? Should Manicheism inherit from Buddhism, even if it's more a claimed syncretism, with very few philosophical influence? — nojhan () 10:01, 19 February 2016 (UTC)

Nojhan Yair rand Runner1928 TomT0m Capankajsmilyo ArthurPSmith John Carter Nomen ad hoc Pictogram voting comment.svg Notified participants of WikiProject Religions

I've finally opted for the inheritance only, used for general categories (i.e. the content of the ideology itself), and influence used for links at large (i.e the history of the religion, for example if a religon derivates from others). If somebody does not agree, please talk now :-) — nojhan () 17:40, 23 February 2016 (UTC)

Definition and social group[edit]

A definition :

: the belief in a god or in a group of gods

: an organized system of beliefs, ceremonies, and rules used to worship a god or a group of gods

: an interest, a belief, or an activity that is very important to a person or group

From this I infer that a religion is definitely NOT a subclass of social group. Instead there is a social group which defines the religion and claims its belief to the set of beliefs and practices the practices. author  TomT0m / talk page 18:12, 26 March 2016 (UTC)

@TomT0m: The existence of formal religious conversion processes in many religions complicates this. The attributes described in Webster's are not comprehensive. I agree, however, that religion is not a subclass of social group. I'm thinking we should consider religion/religious affiliation to be some form of sui generis attribute, with regards to human beings. (Less specific than human behaviour (Q3769299).) --Yair rand (talk) 17:52, 27 March 2016 (UTC)
@TomT0m: @Yair rand: I have added this link because of religions that are not clearly separated from a social group. This is a systematic case for ancient civilizations (for example "religion of Mittani"). In those cases, there is no practical separation between the two concepts. But I think I've understood your point and I now tend to agree, as no practical separation does not necessary have to be set in the theoretical description here. — nojhan () 08:53, 18 May 2016 (UTC)

I'd recommand to model stuffs by clearly separating the two class trees. Taking the example of christianism:

  • Christian (Q106039) View with Reasonator View with SQID for frwiki : Un chrétien est une personne adhérant aux doctrines et pratiques du christianisme, qui croit en la divinité et suit l'enseignement de Jésus de Nazareth rapporté par les Évangiles. translation : a christian is a person who adheres to the beliefs and practices in christianism, belives in god [...] So fat so good, consistent with the premices of this comment
  • Christianity (Q5043) View with Reasonator View with SQID

Other items what can be found digging with the "christian community" keyword:

Other relevant items :

  • cult (Q756820) View with Reasonator View with SQID - we should be able to include the practices associated to the religion as it is a part of a religion definition.

Ignoring The specific "The Christian Community", I wonder if we would need a new item "Christian community" that all christian would be member of "by default". Let's take this option.

Christian (Q106039) View with Reasonator View with SQID is the class of all christian. Any Christian is an instance of this class. We could define this class as "person who believes in christianism and/or practices a christian related cult and member of the crhristian community (as a whole)". We can note that practicing or not practicing can be a fracture inside the community. Anyway, ignoring this difficulty at first. A few things we can model consistently at that point:

I'd add, still consistently:

It's pretty consistent with definitions in articles, a property that shows imho that this is the right path to take. Wikidata should be able to reflect common definitions, and there is no reason we could not if we do things right. What's a little bit disturbing here is that the "christian" item shares a lot with the "christian community" one. Actually the instances of the former are(?) the members of the latter, which make me believe if we need two items for each community with a parralel structure beetween themselves. author  TomT0m / talk page 19:04, 26 March 2016 (UTC)

Some issues with this:
  • member of (P463) specifically excludes social groups.
  • While it may be useful to have hypothetical ontological clarity regarding these items, actual redundancy in statements is something we probably want to avoid. We probably don't generally need to actually link each person to the religion-associated social group. Further, I think there's a general consensus that direct linking of humans to non-Q5 items via instance of (P31), if not completely prohibited, should at least be avoided somewhat.
  • I suspect that Christianity and the Christian community may be simpler to deal with than most religious groups. What may work there may not work for others.
  • I don't see how there would be a use for a split between the class "Christian" and the group "Christian community".
--Yair rand (talk) 17:52, 27 March 2016 (UTC)
@Yair rand: there's a general consensus that direct linking of humans to non-Q5 items via instance of (P31), if not completely prohibited, should at least be avoided somewhat. => I know that but we don't have to explicitely do that. First, it should not make us avoid reasoning, and second, the link has not to be make in the database. Let Christian be the result set of the query{%0A%20%20%3Fperson%20wdt%3AP279*%2Fwdt%3AP31%20wd%3AQ5%20.%0A%20%20%3Fperson%20wdt%3AP140%20%3Fchristianism%20.%0A%20%20%3Fchristianism%20wdt%3AP279*%20wd%3AQ5043%20.%20%23%C2%A0christianity%20and%20subclasses%20%0A%20%20SERVICE%20wikibase%3Alabel%20{%0A%20bd%3AserviceParam%20wikibase%3Alanguage%20%22fr%2C%20en%22%20.%0A%20}%0A} then we get all "instances of" christian as defined in my definition. Please that "ontological purity", as you name this, I'd just say work with established definitions and a moderate bit or rigor, allows this query to work. This help us spot bad modelling. Some reasoning engines and languages, will automatically infer the class membership without the explicit triples, so it's actually not really important if we or not actually make the statement in Wikidata.
Lastly the community might just have an item because it's important for social sciences. author  TomT0m / talk page 09:46, 28 March 2016 (UTC)
@TomT0m: @Yair rand: I get the social group/religion separation, but I still don't get why we should use instance of (P31) instead of subclass of (P279) (or along with?) for religions. — nojhan () 08:53, 18 May 2016 (UTC)
@Nojhan: Imho it's a similar problem that this one : the classification of territorial entities. Some people in this project did not understand why we should not use statements like
< french commune > subclass of (P279) View with SQID < type of administrative division >
. But if you consider that Paris is an instance of french commune and that any instance of a subclass is an instance of the superclass, then it entails that Paris is a type of administrative division, which is wrong. Here I think our main objects are beliefs. Beliefs are somehow instanciated in the head the believers. Then if a religion is a set of belief, we can't say that Christianity is a subclass of religion because that would break the boundaries of class of ideas like Christian ideas and the set of religions which Christian ideas is an instance of. For example the frwiki article w:fr:Religion says Une religion se conçoit le plus souvent comme un système de pratiques et de croyances : "An instance of religion is a system of belief and practice" (commented by myself). A particular religion is a system of belief and practices, not the religion concept itself. author  TomT0m / talk page 09:58, 18 May 2016 (UTC)
@TomT0m: So you mean that « Buddhism » is an instance of « Religion » but that « Mahayana » is a subclass of « Buddhism »? — nojhan () 17:37, 19 May 2016 (UTC)
@Nojhan: Yep, après avoir lu en diagonale l'article frwiki sur Mahayana il apparaît que c'est une forme spécifique de boudhisme avec des spécificités, donc un cas particulier. Après le qualifier de religion ou de secte boudhiste, ça dépasse mes compétences :) Mais clairement secte boudhiste est au "même niveau" de classification que religion (probablement first-order metaclass (Q24017414) View with Reasonator View with SQID). author  TomT0m / talk page 17:43, 19 May 2016 (UTC)

A usecase : Evangelical Free Church of America (Q5415680)[edit]

Just ran other this item : Evangelical Free Church of America (Q5415680) used in a "religion" statement. Might be worth keeping as an example to discuss. author  TomT0m / talk page 11:11, 21 March 2016 (UTC)


Follow up to the discussion in the initial thread, another relevant item : religious community (Q2742167) View with Reasonator View with SQID. Not very expanded on frwiki and enwiki. Might deserve digging for members of this project. author  TomT0m / talk page 10:25, 28 March 2016 (UTC)

on gods and temples[edit]

I don't know if this WP is active at all but I'd like to raise worshipped by (P1049). There are several problems with this. What is a deity? The example given is "God (Q190) is the deity of Christianity (Q5043)"!! God (Q190) is a generic God, should it be the Christian God? Claiming that generic God is the Christian God, and presumably the Muslim God, and the Jewish God, seems to be POV. I would presume they are all different cultural entities, although that's my sociologist POV I guess.

Other religious properties are religious order (P611) and diocese (P708), which both seem pretty Christian.

I've been wondering how to handle the gods of Hindu temples, and how to link depictions of Hindu gods to the God themselves:

Nataraja (Q23641297) is the depiction of Shiva as Lord of the Dance. This obviously needs some property linking it to Shiva (Q11378). What should it be?

How do we connect Nataraja Temple, Chidambaram (Q1966341) to its resident deity Nataraja (Q23641297)? How do we connect it to Shaivism (Q234953)? Come to that, how do we connect Shiva (Q11378) to Shaivism (Q234953)? I'm not sure we have the properties to do this! Help! Secretlondon (talk) 22:04, 28 March 2016 (UTC)

This WP is active (there is a least one people working on it) and we're still discussing the ontology.
religious order (P611) and diocese (P708) are christian concepts, indeed, but it's not a problem, as far as the parent classes are consistent (I did'nt looked).
Your question "what is a deity" is related to the ontology of gods, which is a different question than "when and how to use worshipped by (P1049)". The second question is actually easier to answer: you muse use worshipped by (P1049) when a source says that this deity is venerated by this religion. My advice is to stay as close as possible to sources: you will have sources saying yes, other ones saying no. Keep both, be careful with ranks, use statement disputed by (P1310) (if necessary), remove unsourced items.
The question to know whether two gods with the same name are the same within each religion is a very difficult one. For religions, we propose to add a "religion" element each time a different "movement" is described (even if it's described with another name, like sect, school, church, etc.). THe problem is that it's difficult to keep the transitivity of subclassing, given that all those set of ideas are very inconsistent. Thus, the current proposal is to use a very "light" hierarchy, based on sources rather than on a complete and static description. Maybe this can be used for gods, too: Nataraja (Q23641297) being a subclass of Shiva (Q11378) (because "source"), avoiding "nature of".
For describing the religion(s) associated with a temple, use religion (P140) (see the discussion page).
For the temple/god relationships, I just don't know. But you're welcome to add the corresponding section(s) to the ontology (once you've found the best option), or to propose properties here. I think it's a good idea to indicate what properties not to use, in the ontology, when there is ambiguity. — nojhan () 13:35, 29 March 2016 (UTC)
I think we need another property, the opposite of deity of. I can say Kartikeya (Q380817) is the deity of Kaumaram (Q6378777), but I cannot make a statement on Kaumaram (Q6378777) linking Kartikeya (Q380817). Other religions seem to use named after (Christianity is named after Jesus Christ), or founded by (Islam is founded by Mohammed), neither work with this where we want a concept like 'worships' or 'has deity'. We could use this for temples as well as religious sects. What do people think? Secretlondon (talk) 14:41, 17 April 2016 (UTC)
We also need the property to link Islam (Q432) to Allah (Q234801) as the latter is a key component of the former, and yet we have no way of making the association. Secretlondon (talk) 14:47, 17 April 2016 (UTC)
See Wikidata:Property_proposal/Archive/18#P1049. The number of deities is presumably much larger than the number of religions, so "deity of" instead of "deity" makes sense. The items are already linked, an inverse is unnecessary. --Yair rand (talk) 16:54, 17 April 2016 (UTC)

Philosophico-religious concepts[edit]

Nojhan Yair rand Runner1928 TomT0m Capankajsmilyo ArthurPSmith John Carter Nomen ad hoc Pictogram voting comment.svg Notified participants of WikiProject Religions I feel there is something to discuss about Philosophico-religious concepts (those that the current ontology suggest to use with has quality (P1552)). Is there a common nature for those concepts? Should we restrict this huge set? Inverse the links? — nojhan () 09:36, 18 May 2016 (UTC)

If a religion is a set of ideas shared amongst the believers, then a philosophical theory is of the same nature. A philosophico-religious concept is then at the intersection of both, such as some philosophy entailed by a religion, or a philosophical theory interpreted in the context of some religion. author  TomT0m / talk page 09:46, 18 May 2016 (UTC)

Recent changes feed for religious groups[edit]

Here's a link to a recent changes feed of edits to items for religions/religious groups: Recent changes. Might be useful. --Yair rand (talk) 22:58, 7 June 2016 (UTC)

Encyclopedia articles lists[edit]

Over at wikipedia:Category:WikiProject prospectuses I've started a few pages on content in some reference sources which might be useful here. John Carter (talk) 22:19, 4 January 2018 (UTC)

Replace foundational text (P457) by main regulatory text (P92)?[edit]

Nojhan Yair rand Runner1928 TomT0m Capankajsmilyo ArthurPSmith John Carter Nomen ad hoc Pictogram voting comment.svg Notified participants of WikiProject Religions

I envisaged to replace the current use of foundational text (P457) in the ontology by the more generic main regulatory text (P92). The need is obvious if you think of religions re-using texts made by an older ancestor religion (like the Old Testament in christian religions). Is there any opposition or thoughts? — nojhan () 14:10, 10 February 2018 (UTC)

It’s not obvious to me that a foundation can not itself be derived from older text, nor that all religious texts can be assimilated to laws. author  TomT0m / talk page 14:17, 10 February 2018 (UTC)
That's why it bothers me, I can't find a good property to link a religion to its sacred/important texts. — nojhan () 13:39, 11 February 2018 (UTC)
A religion can have both a foundational text (P457) and a main regulatory text (P92), and they don't necessarily need to be the same text. I'm not sure how to link to sacred texts generally, but I don't think either of these properties would work well for that (as general properties). --Yair rand (talk) 18:20, 11 February 2018 (UTC)
Agree a faith can have both. I think the Book of Common Prayer probably qualifies as the main regulatory text in Anglicanism, but not necessarily as a sacred text. And, I suppose, particularly regarding the New Testament, it might make sense to indicate which version of the New Testament is being considered. Various forms of Buddhism may have similar issues. Christian Science might be seen as having similar distinction between Science and Health and it's organizational rules. As someone who has not edited anything other than the project page to list myself as a member, maybe creating a separate property for "sacred texts" or "inspired texts" or something along those lines might work best. John Carter (talk) 20:17, 12 February 2018 (UTC)
What about keeping both foundational text (P457) and a main regulatory text (P92), but removing the constraint that a main regulatory text (P92) should be a statute (Q820655)? It would avoid having to deal with the sacred notion, which could be a little bit tricky in some case.
Or should we want to embrace it and create additional properties in order to differentiate things like Quran (Q428), hadith (Q234343) and sharia (Q482752)?
I think I would prefer to keep only one property, because I know we will face weird corner cases. In that case, I would keep an unconstrained main regulatory text (P92), without even foundational text (P457)), because I realize that it does not really fit (what religion have a clearly identified text that have been created at some point it time and started it?). — nojhan () 08:49, 13 February 2018 (UTC)
I think we actually need a property specific to national/international law and not to church and faith jurisdiction. God can’t put you in jail, and this is a big difference (men law versus god’s law) author  TomT0m / talk page 11:26, 13 February 2018 (UTC)
Actually, that's a kind of fuzzy distinction. International law can't (usually) directly result in people being put in jail either, but in certain jurisdictions and areas, religious law can have immediate consequences implemented by courts, either government-affiliated or not. Centralized legislatures and justice systems acting as the agent of the governmental authority are not exactly universal ideas. I think we would need to look at a broader overview of legal systems and religious regulatory frameworks before we draw the line like that. --Yair rand (talk) 16:34, 13 February 2018 (UTC)
@Yair rand: That’s when the state, or the human jurisdiction when there is no state, choose religious laws. But it’s not always the case and if someone decides to eat pork when he’s a muslim, it’s mainly a matter of personal choice if you are in France, for example. Although it’s seriously prohibited in Islam. author  TomT0m / talk page 16:44, 13 February 2018 (UTC)
Although there have been prominent cases of individuals like priests monks and nuns who had to be tried by a religious court and (I think anyway) had to have that body turn them over to a secular court before they could be tried there. Gilles de Rais comes to mind. And I guess I should note that neither Christianity nor Judaism can be said to have foundational texts, because in at least both those cases the religion seems to have come before the books. John Carter (talk) 23:36, 13 February 2018 (UTC)
for the latter, I guess we can change the label and the definition of the property to include oral traditions, or accept « text » in a broader sense than « scripture ». Actually the frwiki fr:texte_article defines « texte » as « Un texte est une série orale ou écrite de mots perçus comme constituant un ensemble cohérent » (in english : « a text is an oral or written sequence of words perceived as a coherent body »), so not much of a problem.
For the former, I guess you may be right for the « organisation » part of the religion, any organization can have a regulation. And members can be excluded if they don’t follow the rules. This highlight the fact that there is a difference between the belief one can have about a religion and the « membership » of the religion, and so that there is an « idea » part and an « organisation » part in most religions. author  TomT0m / talk page 14:27, 14 February 2018 (UTC)
So, should we add a new property, like "sacred text", for the religious part and keep the other(s) for organisations? — nojhan () 12:34, 17 February 2018 (UTC)