Wikidata:Property proposal/Attribution Qualifier
Attribution Qualifier[edit]
Originally proposed at Wikidata:Property proposal/Creative work
Description | Relation of a creator to an art work when he's not a "straight creator" of the work. Use with sourcing circumstances (P1480) to qualify creator (P170) |
---|---|
Represents | attribution (Q50137645) |
Data type | Item |
Domain | qualifier of creator (P170) of artworks |
Example | Tobias and the Angel (Q1635039) creator (P170) Andrea del Verrocchio (Q183458), Attribution Qualifier: workshop of |
Planned use | Replace 9 specific properties (P1773 (P1773), workshop of (P1774), follower of (P1775), circle of (P1776), manner of (P1777), forgery after (P1778), possible creator (P1779), school of (P1780), after a work by (P1877)), with this generic mechanism, and accommodate an extensible set of qualifiers (21 are described by CCO & AAT). See Wikidata_talk:WikiProject_Visual_arts/Item_structure#Attribution_Qualifiers for more details. |
Motivation
There are currently 9 "work qualifier" properties used to describe situations when a creator is not a "straight creator" of the work. But these are not enough (CCO sec2 Creator Information p94 describes 19 and Getty AAT has 21) and this approach does not scale. I propose to replace them with a generic qualifier and a set of concept entities (eg "workshop of") that can be extensible, can easily be combined with each other, and together with sourcing circumstances (P1480) can capture all situations.
The current properties should be rewritten as follows:
- P1773 (P1773): sourcing circumstances: probably
- possible creator (P1779): sourcing circumstances: possibly
- workshop of (P1774): Attribution Qualifier: workshop of
- follower of (P1775): Attribution Qualifier: follower of
- circle of (P1776): Attribution Qualifier: circle of
- manner of (P1777): Attribution Qualifier: in the manner of
- forgery after (P1778): Attribution Qualifier: forgery after
- school of (P1780): Attribution Qualifier: school of
- after a work by (P1877): Attribution Qualifier: after a work by
- (the proposal will save us from the need to add 12 more ad-hoc properties like this)
See Wikidata_talk:WikiProject_Visual_arts/Item_structure#Attribution_Qualifiers for more details and examples. --Vladimir Alexiev (talk) 18:21, 25 February 2018 (UTC)
Notified participants of WikiProject Visual arts
@Multichill, Spinster, Oursana, Pigsonthewing, Marsupium, Shonagon:
Discussion
- @Jane023: Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:31, 25 February 2018 (UTC)
Oppose I don't understand why the missing qualifiers are not being proposed instead of changing the current method. Why? What doesn't scale? We are approaching 250,000 paintings so I think it scales pretty well. Jane023 (talk) 21:49, 25 February 2018 (UTC)
- These are not qualifiers but properties. Adding more and more props doesn't scale. Also, you cannot combine them (use them together).
- Attributed To is completely redundant because it says the same as Creator, Sourcing circumstances: probably
- Eg consider how the date at https://www.wikidata.org/wiki/Q32498#P571 says "inception 1300s: refine date: beginning of" where "refine date" is a qualifier.
- I'm proposing the same: to have one qualifier prop and an unlimited set of qualifier values, rather than an ever growing set of props.
- And would someone please comment how can I use the existing props in combination? @Jane023: maybe you? --Vladimir Alexiev (talk) 14:20, 1 March 2018 (UTC)
- The example you give above for Nine Worthies (Q32498) has nothing to do with the creator (P170) property. Maybe it would help if you could show a creator example that you feel is problematic and tell us how you need to combine properties and can't the way it is now. We have lots of paintings where the attribution is problematic and flips back and forth between various painters over decades. As far as I am concerned, each time a painting is sold as "by XXX" then you could possibly add a creator statement for "XXX" with that sale as a reference (and deprecate the statement if it's totally preposterous, etc). Currently I think most of the attributions we have are imported from Commons or collection databases and I see nothing wrong with refining that data with extra information, but I just can't follow your argument as you have explained it this far. Jane023 (talk) 13:02, 3 March 2018 (UTC)
Oppose ^^ (as per Jane) Multichill (talk) 00:18, 26 February 2018 (UTC)
- Comment I don't know much about this, but I clicked through to the "Visual Arts" project link and note this proposal was initiated almost a year ago there. It seems reasonable to me to do this more generically; on the other hand specific properties would also probably be fine. ArthurPSmith (talk) 15:13, 26 February 2018 (UTC)
- I'm of the same mind as Arthur. I don't wish to express a 'for or against' comment on this property itself, but I do see the merit of grouping all the different types of 'attribution qualifiers' under a common 'heading'. This would make it easier to search for all of the works which have any kind of unusual authorship model, and also would make it easier for people without expert knowledge of a specific work to insert the more generic property [rather than requiring them to know if it is a case of 'manner of' or 'circle of' or 'follower of'... It SEEMS on the face of it to be equivalent to the way we allow people to chose a more-general property and then drill-down to more specifics when/if they're known - similar to how in "located in the administrative territorial entity (P131)" you should input the most precise level that you know. Wittylama (talk) 13:54, 1 March 2018 (UTC)
I just stumbled upon subject has role (P2868), which I think fully fills the role of "Attribution Qualifier" (if you read Wikidata_talk:WikiProject_Visual_arts/Item_structure#Attribution_Qualifiers, you'll see I used a qualifier "role"). So there's no need for this property; but please still vote for the new representation method using a qualifier rather than a bunch of specific props that cannot be combined.
If the vote is negative, then I'll propose the creation of a bunch of specific props to cover CCO (as per @Jane023:) --Vladimir Alexiev (talk) 11:56, 3 March 2018 (UTC)
- Support this - however I believe you want object has role (P3831) - when qualifying a statement, the "subject" is the item the statement appears on, the "object" is the value of the item (the "creator" in this case). ArthurPSmith (talk) 18:08, 3 March 2018 (UTC)
Support, Neutral and Oppose
- Support for the disparition of P1773 (P1773) and possible creator (P1779). The main issue is here. The creators are not anonymous but the claim needs to be qualified. For so many artworks attribution is a debate, with variable and moving consensus. Having creator (P170) : Qsomebody and a qualification with sourcing circumstances (P1480) (or as proposed a new property) is a reasonnable approach for simple edition, adding references and easy reuse. And with this approach, we can enrich existing items with adding qualifications without changing existing and common claim.
- Another issue is when we have differents attribution. For exemple : https://www.wikidata.org/wiki/Q19904792#P170 Using P1773 (P1773) and possible creator (P1779) in qualification make difficult for such statments.
- More, edition with anonymous (Q4233718) with qualfication with P1773 (P1773) and possible creator (P1779) is really not a common approach and basiclly strange (the creator is not anonymous actually). Hard to understand for readers and new contributors ; more it makes it very complex for querying, artworks of a creator for exemple, only accessible for those who know very well this complexity and can manage it.
- Neutral for workshop of (P1774). There are variables approaches in art history. Example for artists in Velázquez's studio, art historians make now specific identification, like for del Mazo whose artworks were attributed to Velázquez before. Artworks from Da Vinci's studio are often refused for an attribution to Da Vinci. In the other hand, some artists are studios, like Rubens. Artworks of the studio of Rubens are considered as Ruben's artworks and claimed as it in Wikidata. A mixed approach based on art history could be a reasonable way : creator (P170):anonymous (Q4233718) with qualification workshop of (P1774) Leonardo da Vinci (Q762) ; creator (P170):Peter Paul Rubens (Q5599).
- Oppose for the disparition of follower of (P1775), circle of (P1776), manner of (P1777), school of (P1780), after a work by (P1877)
- Those propeties fills well for qualifications on unknow creator but linked to an another known (or sometimes to known creator linked to another known creator). Yes it could be considered an issue to have many properties but we keep the essential claim: creator (P170):anonymous (Q4233718) (or creator (P170):Han van Meegeren (Q436161) qualification manner of (P1777):Johannes Vermeer (Q41264) and it alllows simple way to enrich this claim.
- Here using the name of the linked creator first and to the real unknow by qualification could be a source of confusion for readers and contributors and lead to terrific challlenges for queries and reuse.
- Maybe forgery after (P1778) could be merged for manner of (P1777). The forgery is qualified with instance of (P31):forged artwork (Q29541662) (and subclasses) or could be with genre (P136)):art forgery (Q1400612).
- Best regards --Shonagon (talk) 00:24, 8 March 2018 (UTC)
- Comment I agree with Shonagon: if the essential part of the information is placed into the qualifier only, it's likely that people who use "creator=(unqualified)value"-only get misleading information.
--- Jura 00:48, 8 March 2018 (UTC)
- Comment I agree with Shonagon: if the essential part of the information is placed into the qualifier only, it's likely that people who use "creator=(unqualified)value"-only get misleading information.
Not done Lack of consensus. --Micru (talk) 15:08, 14 April 2018 (UTC)