Wikidata talk:WikiProject Visual arts/Item structure

Artist

creator (P170) – artist parameter (normally filled with commons:Template:Creator)

It would be nice if the rkd number could be assigned as a property of the creator. The rkdimages number (already created) is a property of the artwork. Jane023 (talk) 20:01, 3 June 2013 (UTC)[reply]

added to Wikidata:Property proposal/Authority control – The preceding unsigned comment was added by Zolo (talk • contribs) at 06:36, 5 June 2013‎ (UTC).[reply]

Creator option problem

It has to be considered how to deal with the Option-parameter of commons:Template:Creator. I cite the doc: Available options are: "workshop of", "circle of", "school of", "or follower", "or workshop", "and workshop", "attributed to", "after", "formerly attributed to", "follower of", "manner of", and "near". These options will add those phrases to the author's name in the user's language. It would certainly be a bad idea to state <artwork> creator <creator> (like this), even with a qualifier creatoroption <workshop of> since it says that the workshop-owner is the creator if I'm not mistaken. A solution could be a construction like <The Temptation of Saint Anthony (Q2397537)> creator <an workshop>, qualifier: workshop of <Hieronymus Bosch (Q130531)>. We could even create an item for every workshop we need. But that won't work for "after", "follower of", "manner of" and so on. Any ideas? --Marsupium (talk) 14:01, 19 October 2013 (UTC)[reply]

Addendum: Cf. the solution described by the CDWA Lite, see the section "4.6. Sub-element: Attribution Qualifier Creator" of http://www.getty.edu/research/publications/electronic_publications/cdwa/cdwalite.pdf. --Marsupium (talk) 00:56, 3 December 2013 (UTC)[reply]

Interesting, thanks. I also notice that they have two different parameters for role played (say painter / engraver) and part (background / furniture). Currently, we only have part of (P361). --Zolo (talk) 08:37, 3 December 2013 (UTC)[reply]

Title

P357 (P357)

There are traditional and modern titles such as Portrait of Pietje Puk, versus the newer Portrait of a man, formerly thought to be Pietje Puk, or worse, "Portrait of Joe Shmoe", with all loss of the historical data on Pietje Puk. – The preceding unsigned comment was added by Jane023 (talk • contribs) at 09:08, 30 May 2013‎ (UTC).[reply]

Yes, "original" title is very bad here, but it seems that the "original" will be dropped. As you mention, there are many types of title. The standard Wikidata solution for this kind of cases is to use qualfiers rather that separate properties (that makes it easier to cope with disputable case). -Zolo (talk) 08:24, 1 June 2013 (UTC) milder Zolo (talk) 21:34, 17 October 2013 (UTC)[reply]

P357 (P357) seems to use the string datatype? In future, we should state the title(s) of an artwork with language information (perhaps unless the title does not use a natural language at all), which means we should use properties with monolingual text datatype and multilingual text datatype once they have become available. If we already use P357 (P357) we should keep in mind that most statements with that property should later be replaced by statements with other datatypes … --Marsupium (talk) 13:11, 17 October 2013 (UTC)[reply]

There has not been much news about monolingual/multilingual strings. It seems that is not really a priority. Actually, we have the same issue for book title anyway, so I think we can use P357 (P357) for now. Another possibility would be to use aliases instead of properties, at least, when there is no official (say artist-given) title, making a title property relatively unimportant. --Zolo (talk) 21:34, 17 October 2013 (UTC)[reply]

Think so too. However, aliases don't seem to be a real solution to me. I had yet thought about them. But they have a different meaning, they have do be given for every language in any case and "using alias would not allow qualifiers" as you declared. ;) In most cases language qualifiers (such as in Diary of Anne Frank (Q6911)) would be appropriate for statements with P357 (P357). However, I think that we should be very careful with the use of this property. At the moment it really does not seem to be well-defined. And there is original language of film or TV show (P364), too, which state is just as deplorable. You know, of course – just to keep hold of it here. --Marsupium (talk) 00:54, 18 October 2013 (UTC), expanded 13:25, 18 October 2013 (UTC)[reply]

Medium

made from material (P186) (can use qualifiers)

Need to think of an external tool to recreate the natural-language feel of commons:template:Technique. Separate properties for material and technique ? – The preceding unsigned comment was added by Zolo (talk • contribs).

Yes, we would like to differentiate between watercolour paintings and oil paintings, and then clarify by panel paintings vs canvas, etc. --Jane023

I was thinking about things like "intaglio" or "collage". both "oil" and "panel" are materials so I think they should use the same property (for panels with applies to support like in Mona Lisa (Q12418). Raw Wikidata does not look good very nice, so we will need external reproscessign anyway. --Zolo (talk) 10:27, 30 May 2013 (UTC)[reply]

OK, but won't that decrease participation? "Oil-canvas" is so popular on Commons it is often not even listed in the file. This should be the default somehow. – The preceding unsigned comment was added by Jane023 (talk • contribs) at 07:54, 1 June 2013‎ (UTC).[reply]

Dimensions

Length
Width
Height
Depth

Is this [the length, comment by Marsupium] enough for architecture, sculpture, and oil paintings with or without a heavy frame? – The preceding unsigned comment was added by Jane023 (talk • contribs) at 07:54, 1 June 2013‎ (UTC).[reply]

Here also, I think qualifiers will be the only practical solution. --Zolo (talk) 08:24, 1 June 2013 (UTC)[reply]

Qualifier to which statements? A height qualifier for a width statement for a painting e.g.? Why shouldn't we use a single statement for each dimension? --Marsupium (talk) 13:11, 17 October 2013 (UTC)[reply]

I meant, we can proboably have three properties (width, height, depth), but we will need qualifiers to state "with frame" or "without the pedestal". --Zolo (talk) 21:34, 17 October 2013 (UTC)[reply]

OK, I'm stupid. --Marsupium (talk) 00:54, 18 October 2013 (UTC)[reply]

Object type

instance of (P31)

Hmm what does this mean exactly? (instance of architecture, sculpture, or oil paintings?) – The preceding unsigned comment was added by Jane023 (talk • contribs) at 07:54, 1 June 2013‎ (UTC).[reply]

As artworks are material objects, they should also be instances of meterial objects, ie statue, mansion, or painting (Q3305213), see Help:Basic membership properties. Usually, we should use the most precise item available, but that may sometimes be tricky. Note that the property can take several values. --Zolo (talk) 08:24, 1 June 2013 (UTC)[reply]

As it is indeed rather tricky we should document the possible values somewhere I think. There might even be artworks which are not really material objects. I am not sure whether to call an artwork of digital art (Q860372) a material object. And there will be difficulties with conceptual art (Q203209), too. Cf. w:Sol LeWitt#Wall drawings: "The wall drawings, executed on-site, generally exist for the duration of an exhibition; they are then destroyed, giving the work in its physical form an ephemeral quality.[21] They can be installed, removed, and then reinstalled in another location, as many times as required for exhibition purposes."

So I propose all objects of our concern should have a statement instance of (P31) <work of art (Q838948)> or instance of (P31) <[any subclass of work of art (Q838948)]> ~~rather than necessarily~~ or of a material object class (cf. this, too). I think it is one of our tasks to set up an class tree (or how that is called). The objects of commons:Template:I18n/objects should appear in this tree. --Marsupium (talk) 14:41, 17 October 2013 (UTC), expanded 13:25, 18 October 2013 (UTC), corrected 17:49, 15 November 2013 (UTC)[reply]

The basic idea of P31 is that the value should be precise item available. This is the right place to indicate that the object is a statue, a painting etc. Using another property for that creates various sorts of problems. Given that P31 is transitive, that should not create real problem, other than technical. If it is too complex to find that the item is an artwork based on data stored in another item, we can also add "instance of artwork" but that should be in addition to more precise values.

Actually we have artworks that are not material objects, and objects that are not artworks but should have similar item structures (for instance, furniture and archeological artefects use Template:Artwork on Commons. There is no need for a closed list of possible values, though we might want to list values actually used. The difficulty imho is that the distinction between object type and other properties like genre may not always be completely clear, but actually it seems manageable. --Zolo (talk) 05:52, 19 October 2013 (UTC)[reply]

Just what I meant … The non-artwork material objects are a good point (may need mention on the task force's home page, too). So each item of this task force's concern should be at least one of:

an instance of <artwork> or of one of its subclasses or
an instance of <material object> or of one of its subclasses.

A closed list would be nonsense, however, one should be able to find out the values actually used (as subclasses of work of art (Q838948) and of <material-object-item> (physical object (Q223557)?)). Yet anything on wmflabs that facilitates it? --Marsupium (talk) 14:01, 19 October 2013 (UTC), corrected 17:49, 15 November 2013 (UTC)[reply]

We have http://208.80.153.172/wdq/?q=tree\[838948][279] (i thought there was a way to show it as a tree, but can't find it). There is some cleanup to do ! --Zolo (talk) 08:51, 20 October 2013 (UTC)[reply]

You obviously meant the Wikidata generic tree: http://tools.wmflabs.org/wikidata-todo/tree.html?q=838948&rp=279 which "is based on Wikidata query". You are completely right with the cleanup. Yesterday, I have begun to conceptualise a system for the class tree by mapping the values currently used by commons:Template:I18n/objects to the Art & Architecture Thesaurus (Q611299) at User:Marsupium/Artworks task force item trees. Together with commons:Template:I18n/objects/wikidata you made up it may be the beginning of a useful class system for works of art or physical objects in general at Wikidata. --Marsupium (talk) 15:15, 2 November 2013 (UTC)[reply]

That sounds great. We do not have an "AAT ID" property yet, do we ? We should create one. --Zolo (talk) 16:06, 2 November 2013 (UTC)[reply]

I have yet thought about it. I was just to shy to propose it. --Marsupium (talk) 17:40, 2 November 2013 (UTC)[reply]

proposed here :).

Working on Wikidata:WikiProject Visual arts/AAT Objects Facet Mapping, I noticed that the transclusions of commons:Template:I18n/objects collected by the category commons:Category:Unsupported object are an unpleasant problem in addition to the messy parts of commons:Template:I18n/objects. That's why I wonder if you have yet an idea about the way, order or steps to import the commons:Template:Artwork data? Thanks in advance, --Marsupium (talk) 01:18, 31 December 2013 (UTC)[reply]

Object history, credit line

commission: commissioned by (P88)
use qualifiers for statements with owned by (P127)?
Support. --Marsupium (talk) 13:11, 17 October 2013 (UTC)[reply]

Obviously the overleaf standing significant event (P793) is also meant to be used for the description of the object history. I am not decided whether it's a good idea or not. With the use of qualifiers for statements with owned by (P127) we could claim the ownership or a span of time whereas significant event (P793) could claim a change in the ownership. significant event (P793) should have change of ownership (Q14903979) and similar items as value with point in time (P585) as a qualifier. The use of one of the two possibilities may depend on what we know about an artwork: We may have sources about a sell only and use significant event (P793) then or we may have sources for the presence in a collection and use owned by (P127) with qualifiers. We have to consider the special case that a unique event has itself an item. At first I thought that significant event (P793) was meant to be used with unique events only. So I stated: <Light and Colour (Goethe's Theory) – The Morning after the Deluge – Moses Writing the Book of Genesis> significant event <Frankfurt art theft>. However, now my solution does not look good to me. Anycase Light and Colour (Goethe's Theory) – The Morning after the Deluge – Moses Writing the Book of Genesis (Q219271) should be link to Frankfurt art theft (Q1792594) or the other way round. --Marsupium (talk) 14:41, 17 October 2013 (UTC)[reply]

I think in this case, we should do things exactly the same way, except that we should add statement is subject of (P805) as a qualifier. It should probably also be shown in Q1792594 as well (for easier queries, and to make sure the list is complete). Would that be ok ? --Zolo (talk) 21:34, 17 October 2013 (UTC)[reply]

Support --Marsupium (talk) 00:54, 18 October 2013 (UTC)[reply]

we need a qualifier for getting back the stolen artwork --Oursana (talk) 23:38, 2 February 2014 (UTC)[reply]

Institution, current location

owner owned by (P127)
collection collection (P195)
current location location (P276)
Accession number inventory number (P217)

~~make it a qualifier of the collection ?~~ no, not really convenient for French museums) – The preceding unsigned comment was added by Zolo‎ (talk • contribs).

However, we should make it the other way round (in some or most cases): take the collection as a qualifier for the accession number as I've yet stated elsewhere. --Marsupium (talk) 13:11, 17 October 2013 (UTC)[reply]

Good idea, as long as the collecion is also shown in a separate property (otherwise, that may get very unclear). --Zolo (talk) 21:34, 17 October 2013 (UTC)[reply]

Of course. --Marsupium (talk) 00:54, 18 October 2013 (UTC)[reply]

Exhibition history

Notable enough for Wikidata ? – The preceding unsigned comment was added by Zolo (talk • contribs) at 09:15, 4 April 2013 (UTC).[reply]

If it is there, it's always notable, because the artwork is notable, so for example the artwork probably influenced another younger artist somewhere who saw it in an exhibition, etc. Though the Mona Lisa was "seen" in print by many artists before travel became easier to do, recording its whereabouts is definitely notable for other reasons (security issues, influences on modern artists or public opinion, art history publications, etc.) – The preceding unsigned comment was added by Jane023 (talk • contribs) at 09:08, 30 May 2013‎ (UTC).[reply]

Added it to proposals. Zolo (talk) 08:52, 1 June 2013 (UTC)[reply]

Thanks, --Marsupium (talk) 13:11, 17 October 2013 (UTC)[reply]

Inscriptions

P438 (P438)

There is a limit of 400 characters. What do we do if an inscription contains more than 400 characters ? Pyb (talk) 18:34, 15 November 2013 (UTC)[reply]

Oh, in fact, that might get a problem. And ATM no clean solution comes to my mind. Two messy ones could be 1) to use several statements and 2) to give the full text at the talk page and state this in the 400-character-text. I hope that we will find a better way to solve it. Do you have any examples, Pyb? --Marsupium (talk) 18:58, 15 November 2013 (UTC)[reply]

When the inscription is really important for the object (say, the Rosette Stone), it may be worth creating a Wikisource page and linking to it using full work available at URL (P953) ? In the Rosetta Stone case, I don't know which linguistic version of Wikisource we could use though... --Zolo (talk) 08:46, 16 November 2013 (UTC)[reply]

Good idea! The best solution for such cases could be to combine two ways, give as much of the inscription as possible in the P438 (P438)-statement and add a full work available at URL (P953)-statement in addition. For now I have only added the

<Rosetta Stone (Q48584)> full work available at URL (P953) <https://el.wikisource.org/wiki/Στήλη_της_Ροζέττας>

. The three available Wikisource language versions do not seem to be very reliable to me, so I did not copy anything. The Greek version might be the most original one, so I have chosen it. A future sitelink integration of Wikisource may simplify to the choosing problem. --Marsupium (talk) 12:38, 16 November 2013 (UTC)[reply]

References

References for individual statements directly in the "sources" field of the statement
Databases can use special properties:

Joconde Joconde work ID (P347)
RKDimages RKDimages ID (P350)

Some museums (Louvre, Metropolitan Museum, British Museum) do not use any explicit identifier for objects in their internal database. We can retrieve a number identifying the artwork in the URL but that may not be very stable. – The preceding unsigned comment was added by Zolo (talk • contribs) at 09:15, 4 April 2013 (UTC).[reply]

Catalog number (I am thinking for example of the numbers by famous collectors and cataloguers, like the "Lugt number" which is used by the RKD and the British museum) – The preceding unsigned comment was added by Jane023 (talk • contribs) at 07:54, 1 June 2013‎ (UTC).[reply]

The is a proposal at Wikidata:Property_proposal/Creative_work#Catalog_number_.2F_Numero_di_catalogo, though I am not sure about how best to render it in Wikidata. --Zolo (talk) 08:24, 1 June 2013 (UTC)[reply]

a "reference" field for external resources that are not about a particular statements

That sounds like something like foaf:isPrimaryTopicOf as used by http://dbpedia.org/page/Mona_Lisa to express the relation to http://en.wikipedia.org/wiki/Mona_Lisa. Do we have a such a property? --Marsupium (talk) 17:49, 15 November 2013 (UTC)[reply]

Permission

copyright license (P275)

Should we have a way to indicate publc-domainness of the object ? – The preceding unsigned comment was added by Zolo (talk • contribs) at 09:15, 4 April 2013 (UTC).[reply]

Isn't it enough by linking out to the image on Commons? You can always link out to something like "This is a non-free image", or "Submit a photo of this artwork!" – The preceding unsigned comment was added by Jane023 (talk • contribs) at 07:54, 1 June 2013‎ (UTC).[reply]

Sometimes, Commons files need two licences: one for the photo and one for the photographed object. If Commons can fetch its data from Wikidata, it could be useful to store the object-license info here, but I agree that it is probably too early for this. --Zolo (talk) 08:24, 1 June 2013 (UTC)[reply]

The object-license is a good point. I support the purpose. If we do not be in need of several statements for the permission with qualifiers for the countries they are valid for, we might state that an object belongs to a certain group for which there are certain rules for the permission in a certain country? I'm quite inexperienced in permission questions, sorry … --Marsupium (talk) 13:11, 17 October 2013 (UTC)[reply]

Various Commons templates to map

commons:Template:Technique
commons:Template:ProvenanceEvent
commons:Template:Depicted person
commons:Template:Title
commons:Template:Inscription
commons:Template:I18n/objects: the object catalogue can be shown here on a documentation page with links to the items, the translations can perhaps become labels of these items?
templates in commons:Category:Multilingual tags: Techniques and in commons:Category:Multilingual tags: Provenance: similar
…

I think that – apart from commons:Template:Artwork – we have to map at least these Commons templates, please add other missing ones! --Marsupium (talk) 13:11, 17 October 2013 (UTC)[reply]

Depicted person is included depicts (P180). ProvenanceEvent is P793, but we still need to define more precise guidelines for qualifiers.

I have started commons:Template:Technique/wikidata, but this is a long and rather boring task, help would be welcome ! --Zolo (talk) 21:34, 17 October 2013 (UTC)[reply]

Oh, I'll have to take a look at it later. --Marsupium (talk) 13:25, 18 October 2013 (UTC)[reply]

Other things

Parts of the descriptions, like those using Template:Depicted person or Template:Depicted place can use the "depicts" Property (P180) – The preceding unsigned comment was added by Zolo (talk • contribs) at 09:15, 4 April 2013 (UTC).[reply]
Much of the category content is caught by other properties – The preceding unsigned comment was added by Zolo (talk • contribs) at 09:15, 4 April 2013 (UTC).[reply]
Forensic evidence for provenance and research, such as dating of the tree rings on wood panels, x-ray photos of paint layers, analysis of linen canvas, and other restoration actions that are becoming more prevalent today. ----Jane023
I think the determination method or standard (P459) qualifier property should be used for that. --Zolo (talk) 10:27, 30 May 2013 (UTC)[reply]

Open questions

distinction genre / instance of / depicts

distinction genre / instance of / depicts. Where should things like "family portrait" or "Madonna and Child" go ? --Zolo (talk) 09:59, 5 April 2013 (UTC)[reply]

See my comment on "original title" above. This gets tricky really fast and I don't know what to do about it. See for example the portrait of Jan Asselyn by Frans Hals - I still can't figure out who is "right" about that. Jane023 (talk) 07:54, 1 June 2013 (UTC)[reply]

We need a property for the Iconclass (See iconclass code 73D241 for example for the commonscat Commons:Category:Last Supper) of an artwork: Q1502787. Jane023 (talk) 08:24, 4 June 2013 (UTC)[reply]

That's an interesting database, but using it in Wikidata may be tricky as well. We could

use it in item about individual artworks. I

I dont think it is the best solution. We probably also need to have "depicts: the Last Supper", because if we want to be able to find artworks depicting the last supper, we should not have to use external ontologies. That would create much redundancy and additional work. Also, controlled vocabularies are discouraged by designed on Wikidata.

Link from the artwork item to an item about a theme, using "depicts" or another property, and use the iconclass in the item about the theme as an "authority control"

I think that is better, provided that the item really corresponds to the RKD icon class (would that be Last Supper (Q51633) or Last Supper in Christian art (Q6494691) ?) --Zolo (talk) 06:50, 5 June 2013 (UTC)[reply]

iconclass code 73D24 Last Supper--Oursana (talk) 12:16, 21 March 2014 (UTC)[reply]

We can do this with depicts (P180)--Oursana (talk) 23:43, 2 February 2014 (UTC)[reply]

Use of genre (P136)

Coyau, your use of genre (P136) like here does not seem to be useful to me. Compared to instance of (P31) genre (P136) is not well-defined (cf. Property talk:P136). I do not see any advantage of genre (P136) to instance of (P31) ATM. But you might claim any? Currently, genre (P136) is not listed overleaf. However, we should agree on the modelling of this information, maybe at #Object type. If we agree on not using we should revise all these items: http://208.80.153.172/wdq/?q=claim[31:(claim[279:838948])]_AND_claim[136] (222 now) (though work of art (Q838948) is not the best highest class for this task force (as yet pointed out elsewhere)). --Marsupium (talk) 14:03, 8 November 2013 (UTC)[reply]

It is just common sense and common knowledge about art. A portrait is a representation of a person by any means, painting, sculpture, literature, etc. And it is a genre and it is well defined, see w:Hierarchy of genres. I'm sorry if it not usefull to you, but to some others, it might. --Coyau (talk) 15:18, 9 November 2013 (UTC)[reply]

I know what you mean, you may have misunderstood me, Coyau: It is of course commons sense that artworks have a genre. And the genre described by w:Hierarchy of genres is also probably well-defined. And the information is of course useful to me and to many others. The point is the property used to model this information. And the property genre (P136) is not well-defined in my eyes. instance of (P31) is. Cf. also #Object type. So let us use the last to model the genre information and revert this. ;) Best, --Marsupium (talk) 22:53, 9 November 2013 (UTC)[reply]

Thinking about it. If "portrait" is a subclass of "artistic genre" then "Mona Lisa: instance of portrait" would mean that Mona Lisa is also an instance of artististic genre, which does not seem to make sense, so I guess a separate property for the genre is the way to go. --Zolo (talk) 13:49, 16 November 2013 (UTC)[reply]

To me <portrait (Q134307)> subclass of (P279) <artistic genre> is nonsense, such as the currently stated instance of (P31)-statements for portrait (Q134307): It should be a class, not an instance, but not a subclass of <artistic genre>, the item should not describe the term, but the class of all artworks (if we talk about "portrait" in art history) that are portraits, cf. portraits in the AAT. --Marsupium (talk) 15:30, 16 November 2013 (UTC)[reply]

Yes, actually, "portrait" is defined as an artwork depicting someone, so p31 seems to make sense, but that may not be the case for everything. Can we say p31: Landscape art ? I see that the AAT has an "art genre" [1] but I cannot find what it contains. Somewhat relatedly, how do we deal with stylistic indications, or how do we rattach an art movement (how do we say that Les Demoiselles d'Avignon (Q910199) is cubist ?). --Zolo (talk) 08:53, 17 November 2013 (UTC)[reply]

Though I changed my mind a little bit recognising a sense in Coyau's statemtent this modelling does not seem to be useful or practical to me, we should agree that "portrait" is defined as an artwork depicting someone. So I think we should revise the items concerned.
You can find the "hierarchy" of <art genres> at [2]. But that is not very helpful.
We should redefine landscape painting (Q191163) as the class of all landscape artworks or so. That is actually covered by the content of the linked Wikipedia articles though their lemmata do not refer to a class for some reasons. We can state then:

<landscape painting (Q191163)> Art & Architecture Thesaurus ID (P1014) <300015636>

<The Abbey in the Oakwood (Q334360)> instance of (P31) <landscape painting (Q191163)>

We could perhaps state:

<Les Demoiselles d'Avignon (Q910199)> has period/style <cubist>

<cubist> Art & Architecture Thesaurus ID (P1014) <300021495>

We do not seem to have a has period/style property yet. Though this modelling question is perhaps advanced and not the most urgent. I think an adaption (eventually with deviations) of the AAT's classification of objects and styles and periods might be the best way. I have done some research and I suppose that we will not find better classifications than those.

--Marsupium (talk) 17:04, 30 November 2013 (UTC)[reply]

Use of AAT

We might take in consideration to adopt parts of Art & Architecture Thesaurus (Q611299) (which shall be published as linked open data in 2014 afaik) or similar controlled vocabularies. Is somebody experienced in those things here? Perhaps we should seek advice? --Marsupium (talk) 13:11, 17 October 2013 (UTC)[reply]

Now we have Art & Architecture Thesaurus ID (P1014). --Marsupium (talk) 14:03, 8 November 2013 (UTC)[reply]

Additional items for images?

I would suggest to create additional items for the pictures in cases where one object carries more than one picture like The Allegory of Good and Bad Government (Q2401235) or Woman Carrying a Pail of Water (Q14856931) (this item refers – being fussy – just to one picture and not to an object at the moment). That would allow a meaningful use of properties which refer to the picture like depicts (P180) and P438 (P438) (in some cases). Properties like those for the dimensions or image (P18) could be used for the object and another time for each picture. --Marsupium (talk) 14:01, 19 October 2013 (UTC)[reply]

A far future project would be to link from items to image sections in the form of file shapes created by commons:Help:Gadget-ImageAnnotator, cf. the HyperImage project. That would be more precise than the applies to part, aspect, or form (P518)-statements. --Marsupium (talk) 19:04, 15 November 2013 (UTC)[reply]

Where is the Item structure for Materials? (was: Which property to use to link color varieties of a pigment?)

It would be nice to begin to see "some" structure of Material modeling on the Item structure page. Medium is already there, so we should have Visual Art Materials also or just Materials?
For example, someone in the history on ochre (Q12353401) mentioned, as I already know, that pigments have color variations. But which property can be used to link a parent pigment, ochre (Q12353401) --> to it's children, red ocher (Q54806372) ?
Or is it generally known that pigments will always subclass from parent pigments, and thus no further property is available or needed?

--Thadguidry (talk) 20:46, 16 September 2020 (UTC)[reply]

Root item

Texts, musical works etc. are categorized as work of art (Q838948) but they have very different requirements from material artworks. So I think we need another root item. Actually so that all visual artworks are together. Which one should be used ? Actually, does it really need to have something to do with art ? I do not think so, as it is a sometwhat complex and controversial concept (it seems somewhat pointless to enter into discussion about whether jewels or amulets are works of art). I do not see anything wrong with classifying paintings at work of art, but only if at the same time they are also categorized in a class that makes it clear that they are unique material objects. What should this general concept be ? The AAT seems to use "object facet [3]) Should we have something like that ? Can we simply the AAT concept even though the AAT is copyrighted ? In a way "artefact" seems technically correct too, as it refers to any kind of object, but I am wary that in real usage, it is actually restricted to archaeological artefacts. Of course, we should also find solution for conceptual art of for artworks that exist in several copies. --Zolo (talk) 11:17, 13 November 2013 (UTC)[reply]

Strong support, more detailed comment later. BTW: I like the your idea of Talk:Q838948/Instance structure. --Marsupium (talk) 15:53, 13 November 2013 (UTC)[reply]

By now I think that we will need at least two root classes (in the form of root items): <material object> (physical object (Q223557)?) and something like <anything not necessarily material with any form of designing(in German: Gestaltung)> (including the products of conceptual art (Q203209)) (i. e. some form of visual image in the broad use of the term in contemporary art history literature including designed objects such as a knife and so on, cf. the messy collection of WP-articles under visual culture (Q861716)). For a lot of things/items these classes will overlap, for some they will not. I think <artefact> could then be a subclass of both root classes and will surely have far more instance than the rest of the two root classes. I think the term includes paintings, statutes and so on, but also knifes and chairs. That is good. Reflecting these two root classes this project should probably be called WD:Objects and visual images task force or WD:WikiProject objects and visual images (cf. WD:Bot requests#Rename "task force" to project"). --Marsupium (talk) 17:49, 15 November 2013 (UTC)[reply]

Can "Wikiproject:Visual Arts" do ? That may not be really correct but that is a bit less cumbersome :). I guess that some projects overlapping with this one will have to be created (material culture, archaeology, whatever). ~

So design (Q1519454) should be a superclass of work (Q386724) ? If I understand correctly, in German: "Werk" = work that is sufficiently original to be eligible to copyright and "Gestaltung" = any type of man-produced idea or object, but we have some mess in the interwikis. --Zolo (talk) 08:57, 16 November 2013 (UTC)[reply]

Visual arts is probably a better term, at least for the bulk of our users and editors. It fits well with the pages under Wikipedia:WikiProject Visual arts (Q13627814), too. The capitalisation should perhaps be WD:WikiProject:Visual arts, but I am not sure about that.

Malheureusement, <work (Q386724)> subclass of (P279) <design (Q1519454)> would not be correct, since "Gestaltung" describes the process which produces <anything designed> / <etwas Gestaltetem> / <quelque chose dessinée/modelée/formée/façonnée>. "Werk" can almost be everything similar to the Latin "opus". I think work (Q386724) describes not necessarily a physical object. So I propose the following statements:

Unfortunately, Artefact (Q231515) collects disambiguation pages and I have not found an "artefact" item that does not emphasise the archeological meaning as archaeological artefact (Q220659). I'm quite indecisive how to link all that. BTW: In my eyes virtual artifact (Q10855008) is sort of a Contradictio in adjecto (Q1129113), I'd assume that the term "artefact" only refers to physical objects, but that may be false. --Marsupium (talk) 12:41, 16 November 2013 (UTC)[reply]

physical object (Q223557) does not work very well as spme interesting properties are related to the fact that it is a human-processed object. I did not find any suitable item, so I created artificial physical object (Q15222213). I suppose part of the structure should simply be inherited from physical object (Q223557), but there is no way to indicate that in {{Class documentation}}, so I mostly copied what we had for "work of art" in Talk:15222213/class. Note that I have added a "required" parameter to the template. That should simply mean "this property should be fillable for all instances of this class, but it will surely need a lot of fine tuning. --Zolo (talk) 13:49, 23 November 2013 (UTC)[reply]

That seems to be a very good solution for now. Thanks, --Marsupium (talk) 17:04, 30 November 2013 (UTC)[reply]

Deletion of property:moveable object location

Now that we have ranks for statements, I would like propose location (P276) for deletion. I think it is more sensible to switch to P766 (P766) + rank = 'preferred' to indicate that it is the most current value (actually the English label is already out of synch with other languages). --Zolo (talk) 17:06, 15 February 2014 (UTC)[reply]

~~{{o}}~~ to the solution with 'preferred' since that should be used to state that we think an information in a source is more likely than those of the other sources, shouldn't it? We cannot use it for both things. And Wikidata collects statements, not truth about the world (cf. somewhere in m:Wikidata/Data_model). Regards, --Marsupium (talk) 09:35, 26 February 2014 (UTC), edited 17:55, 27 February 2014 (UTC)[reply]

What the development team has said in the most recent posts I have said is that "preferred" mostly means the value that should be retrieved in priority in queries and infoboxes. The example that was given (I dont remember exactly where) was that for demographic data, we should mark as preferred the most recent reliable data (it is much simpler, and more efficient, that letting the software compute it). "Deprecated" on the other hand really means data that seem of low value or have been proved wrong. That indeed means that two concepts are conflated in ranks, maybe it should be changed, or maybe it more manageable than having two different parameters, but that is the way it seems to be currently conceived. --Zolo (talk) 21:01, 26 February 2014 (UTC)[reply]

OK, I see! Anyway you should probably be very careful, especially with time handling I assume, and think about all (sorts of) existing property-value-couples and to which new ones they will be transferred! And P766 (P766) would need a qualifier that says that the event is "being", right? Regards, --Marsupium (talk) 17:55, 27 February 2014 (UTC)[reply]

Oppose if I consider the german translation of both properties, they are different as P766 (P766) is not a permanent property, so it cannot replace location (P276).--Oursana (talk) 08:18, 27 February 2014 (UTC)[reply]

That's imho wrong. --Marsupium (talk) 17:55, 27 February 2014 (UTC)[reply]

Usage (at time)

How to express that Palmesel (Q318914) is used on Palm Sunday (Q42236)? I post it here since it appears to be a general question. Thanks, --Marsupium (talk) 03:31, 17 April 2014 (UTC)[reply]

https://www.wikidata.org/w/index.php?title=Q318914&diff=next&oldid=121253266: significant event (P793) Palm Sunday (Q42236). Regards --Oursana (talk) 12:52, 18 April 2014 (UTC)[reply]

Thank you, good idea! This does not reveal the kind of relation between the Palmesel (Q318914) and the annual event Palm Sunday (Q42236), but better than nothing! --Marsupium (talk) 15:10, 21 April 2014 (UTC)[reply]

"key event" is used for one-time events experienced by the item and can be used to create a timeline. I am not sure it makes sense for this sort of use. I would rather use has use (P366) for now, even though it is not very precise either. --Zolo (talk) 19:20, 21 April 2014 (UTC)[reply]

Currently has use (P366) has a constraint to instance of (P31) <building (Q41176)> items. --Marsupium (talk) 15:45, 25 April 2014 (UTC)[reply]

If you see de:Schlüsselereignis it really can't be used here. So I suggest we change the constraint of has use (P366) also to objects.--Oursana (talk) 22:42, 2 December 2014 (UTC)[reply]

Properties vs Qualifiers

Hi, I noticed that though the English Wikipedia has 762 painting stubs alone, there are only about 500 paintings showing up in Reasonator. On the other hand, there are lots more items that link back to property "painting" (Q3305213). I started to check some paintings that I had added and saw that some lacked items while others lacked the painting property. My question is, whether to put the date, genre, creator of the painting as qualifiers to the painting property, or just list them on the item page? Jane023 (talk) 13:19, 31 May 2014 (UTC)[reply]

Hi Jane,

Wikidata Query shows 7952 results to me at the moment searching for items with instance of (P31) <painting (Q3305213)> or subclass of (P279)s thereof. Do you have a link for the paintings displayed to you in Reasonator? I do not really get where the 500 come from.

What do you mean with "painting property"? A instance of (P31) <painting (Q3305213)> statement? If so the other properties you mentioned, date, genre and creator, should be added to a painting item in seperate statements like in The Milkmaid (Q167605). --Marsupium (talk) 19:38, 1 June 2014 (UTC)[reply]

Thanks! I realize my question was double (why only 500 and how to use the qualifiers). I will look at the milkmaid, thx. Here is the link in reasonator: only 496 items. Jane023 (talk) 11:19, 2 June 2014 (UTC)[reply]

Reasonator only displays 500 items in the "From related items" section. --Coyau (talk) 16:42, 4 June 2014 (UTC)[reply]

Theme

How do I say that The Virgin and Child with Four Holy Virgins (Q17320997) depicts Virgo inter Virgines (Q17331271). I have used depicts (P180). I think that is technically correct. And if we assume that p180 is transitive, it should imply: depicts: Mary, a garden, etc. But that transitivity may not be very robust. What if one of the features of Virgo inter Virgines (Q17331271), because some of the usual features of a Virgo inter virgines may be missing. Also it is not very easy to grasp at first read. We may also decide that we should not rely on the transitivity, and state explicitly "depicts a) Virgo inter Virgines b) Mary. But that may be even more confusing (does it depict one Mary + another Mary in a depiction of Virgo inter Virgines ?). Should we use a different property ? main subject (P921) ? I do not think it is exactly that ? A new "artistic theme" property ? --Zolo (talk) 16:20, 10 July 2014 (UTC)[reply]

@Coyau (since you edited The Virgin and Child with Four Holy Virgins (Q17320997)): I think there are three possibilities:

	definition/"description" of Virgo inter Virgines (Q17331271)	classification; Property:P1256/Property:P1257	propertyX for `<The Virgin and Child with Four Holy Virgins (Q17320997)> propertyX <Virgo inter Virgines (Q17331271)>`	relation of Virgo inter Virgines (Q17331271) at an item with it to (other) depicts (P180)-statements at the item
1	about "group of Mary, Jesus Child and at least one other Saint Virgin"	`instance of (P31) <group of humans (Q16334295)> or the like`; Iconclass notation (P1256)	depicts (P180)	transitivity, no other independent depicts (P180)-statements describing the group → Virgo inter Virgines (Q17331271) has to be specified with qualifiers, e.g. has part(s) (P527)
2	[class of all instances of an] "image with a Virgo inter Virgines"	`subclass of (P279) <image (Q478798)>`; depicts Iconclass notation (P1257)	instance of (P31)	independent description with depicts (P180)-statements → more redundancy than solution 1
3	"artistic theme with Mary, Jesus Child and other Saint Virgins, usually situated in a hortus conclusus"	`instance of (P31) <artistic theme (Q2160811)> or <motif (Q1229071)> or the like`; Iconclass notation (P1256)	depicts (P180) or based on (P144)	extra description with depicts (P180)-statements still necessary for precision → more redundancy than solution 1

Solution 1 would reduce redundancy, however, it would make querying for a part of a theme complex – in our case Virgo inter Virgines (Q17331271) – quite complicated, this would be even worse with theme complexes not only grouping physical things like Massacre of the Innocents (Q15676570) and adoration of the Magi (Q488841). Thus I strongly discourage this solution.
Solution 2 and 3 both produce redundancy in some way. Solution 3 with depicts (P180) would require caution when querying depicts (P180)-statements to distinguish possibly redundant (physical) objects and themes in the range, using based on (P144) does not have this disadvantage. (But, Coyau, why did you decide to prefer based on (P144) to depicts (P180)? At least the property's description – "the work(s) used as basis for subject" – does seem to fit. We could even create a new property to connect an image work to its theme, what do you think?) I have pointed out the conditions under which I would prefer either solution 2 or solution 3 here.

The current state follows solution 3. The Iconclass tagging of the Rijksmuseum follows a similar way. That can be OK. But the problem is complicated. I think we should have a look in CDWA, CCO and CIDOC CRM prior to a final decision. I hope I get to do that soon, the horror of changing the system grows each day … So much on my two cents. What do you think about the solutions, are there yet other things to consider, (dis)advantages of any solution? --Marsupium (talk) 22:09, 31 July 2014 (UTC)[reply]

Discussions somehow related:

--Marsupium (talk) 10:49, 2 August 2017 (UTC)[reply]

Theme: expressing the identity of things with complex descriptions

@Coyau (since it concerns many of the claims you create): We probably have to talk about the use of depicts (P180) anyway!

I think that it would be great to use depicts (P180) to describe the things an image depicts as precise as possible. That includes unambiguousness. When one depicted thing is described by more than one depicts (P180)-statement the description is no more unambiguous. Regarding the current group of statements <The Virgin and Child with Four Holy Virgins (Q17320997)> depicts (P180) <Mary (Q345)> , <mother (Q7560)>, we can not know if Mary (Q345) is depicted once or if there is a representation of the Virgo inter Virgines (Q17331271) and one of Mary (Q345) separately and one more of an instance of mother (Q7560). So I discourage from this use. I propose
1. not to repeat redundantly statements that other items state: Because of the statement <Mary (Q345)> child (P40) <Jesus (Q302)> the group of statements <The Virgin and Child with Four Holy Virgins (Q17320997)> depicts (P180) <Mary (Q345)> , <mother (Q7560)> is redundant and should be discouraged since the depiction of Mary (Q345) is the meant depiction of mother (Q7560). (Other examples: <Henri Cordier (Q15526285)> depicts (P180) <Henri Cordier (Q722268)> , <man (Q8441)>, <Henri Cordier (Q722268)> instance of (P31) <human (Q5)> ; sex or gender (P21) <male (Q6581097)>; Perseus and Andromeda (Q3900153).) The information about the yet described instances on the images can simply be inferred. That is the idea of the semantic web or structured data.
2. to use qualifiers to specify the description of one instance on an image: <Shepherdess with her Flock (Q15876256)> depicts (P180) <woman (Q467)> , <herder (Q12059906)> should be replaced by <Shepherdess with her Flock (Q15876256)> depicts (P180) unknown, qualifiers: instance of (P31) <herder (Q12059906)> ; sex or gender (P21) <female (Q6581072)>.

If we agree on this, creating depicts (P180)-statements will be a bit less trivial than it is now, but the information will not be as trivial at the same time.

To remind you and me: The use of shown with features (P1354) and or perhaps more specific properties has to be figured out! For example occupation (P106) might be usable for qualifiers of depicts (P180)-statements. I wonder if we should only allow subclasses of physical object (Q223557)s for the range of depicts (P180) ({{Constraint:Type|class=Q223557|relation=subclass}}) and shift all other things to other properties.
To make the data more structured and thus more rich for queries, Christ Child (Q942467) in the range of depicts (P180) should probably be replaced by Jesus (Q302) and a (yet to create!?) qualifier stating the age.

I hope we can make the description of images more powerful than the current handling is. Thanks for any input, --Marsupium (talk) 22:09, 31 July 2014 (UTC)[reply]

@Coyau: Since you add a lot of statements of the concerned form (that I propose to delete) I would very much appreciate a response to the issue from you to prevent unnecessary extra effort later! Thanks in advance, --Marsupium (talk) 14:16, 20 September 2014 (UTC)[reply]

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘ Jumping in with my two cents. For my particular interest (costume research), I want to be able to indicate that in this portrait the sitter is wearing a black silk brocade doublet trimmed with gold and silver bobbin lace and ouches set with red gems, with thread-wrapped buttons and pockets on the sleeves, lined in pink fabric (probably linen), over a shirt with blackwork embroidery and a narrow ruffle edged in black-and-gold twisted cord with four bandstrings at the neck. I don't expect anyone will have an answer about how we structure this, but it's what I'd really like, something to think about. (And I'd like to be able to find this picture by searching for 'sleeves with pockets'.)- PKM (talk) 17:31, 9 December 2014 (UTC)[reply]

Some time ago I thought about the problem and thought that we should perhaps create a new property for structured descriptions as opposed to depicts (P180) which is de facto mostly used for tag (Q658349) almost as messy as those created by the ARTigo game. --Marsupium (talk) 02:24, 25 May 2015 (UTC)[reply]

Field of work

For painters, should the field of work (P101) be a genre of painting (Q16743958) or just art of painting (Q11629)? - PKM (talk) 19:35, 1 December 2014 (UTC)[reply]

Renaissance painting (Q15885248)--Oursana (talk) 02:08, 2 December 2014 (UTC)[reply]

I suppose painting is fine for the real generalists who dabbled in everything, but there were lots of specialists, such as people who painted horses for a living. So it really depends on the painter. I can also see specifying frescoes or water colors. Jane023 (talk) 21:41, 2 December 2014 (UTC)[reply]

Series or Cycle?

I just stumbled across painting series (Q15727816) which seems to mean the same thing as cycle of paintings (Q16905563). There are more than 100 items tagged as painting series (Q15727816). How shall we handle this? - PKM (talk) 01:40, 9 December 2014 (UTC)[reply]

I'd think "cycle" is a series that has some coherent order like in "narrative cycle". So that would be a subclass of series. Marie de' Medici cycle (Q2558141) would a cycle because it tells a story with a beginning and an end. Rouen Cathedral Series (Q637414) does not have such a clear order, and even though it is a series, it is not a cycle. --Zolo (talk) 21:32, 9 December 2014 (UTC)[reply]

Can we move this discussion to site Questions as at least similar questions are discussed there--Oursana (talk) 11:10, 10 December 2014 (UTC)[reply]

@Oursana, Zolo: Yes, copying now.- PKM (talk) 19:56, 10 December 2014 (UTC)[reply]

Sale of Collection

Do we have a standard way to record the sale and dispersal of an art collection?

In the case of Barbara Piasecka Johnson Collection (Q18793192), which was sold at Christie's, I have recorded significant event (P793) => change of ownership (Q14903979) with point in time (P585) of the sale date and the Christie's catalogue link as the reference. But "change of ownership" doesn't indicate that the collection was broken up rather than sold as a whole. Is there a better way to record this? - PKM (talk) 02:22, 17 January 2015 (UTC)[reply]

Full list of properties per item type for GLAMs

I was wondering whether it would be a good idea to update this page with all the properties currently used for the instance painting. I think it would be really interesting for GLAMs to have access to to representative listing of all the properties used for a given type of objects. It would help them to understand Wikidata and to compare it with their own data. I think it would also help mappings activities. Similar page could be maintained for different types of cultural heritage objects.--Valentinec89 (talk) 13:24, 21 April 2015 (UTC)[reply]

@Valentinec89: I agree, can you help update the list? Having it for other types of objects makes sense too. Multichill (talk) 16:53, 21 April 2015 (UTC)[reply]

I did a rough query:

This should help to make the list more complete. Some of these properties are qualifiers or references btw. Multichill (talk) 17:13, 21 April 2015 (UTC)[reply]

I have previously created a list of properties for maps. Do you think I could copy it here? --Susannaanas (talk) 11:15, 25 July 2015 (UTC)[reply]

Yes please! Jane023 (talk) 17:26, 25 July 2015 (UTC)[reply]

I will make it a subpage first, but I think a table allowing to spot the differences between artwork types will be required. --Susannaanas (talk) 06:34, 21 October 2015 (UTC)[reply]

Maps subpage

I poured the maps metadata stuff on a subpage Maps. Please have a look and suggest how to invite the relevant community! Cheers, Susannaanas (talk) 08:40, 21 October 2015 (UTC)[reply]

Issue with the table in the #Qualifiers section on the other side

The examples and descriptions in the source code of this table are not rendered. Perhaps someone can investigate and solve the problem. At least, it is pointed out by now. Cheers, --Marsupium (talk) 23:31, 4 January 2016 (UTC)[reply]

Which statements shall groups of works carry?

At the moment some items for groups of works (series of creative works (Q7725310) and subclasses) have statements that refer only to a part of them (which might or might not have an item for itself), some leave out those statements. The question how to handle those cases exists for example for statements with collection (P195) and location (P276), but also for those with image (P18).

An example for an item for which some statements refer only to parts of it is Sunflowers (Q157541). One might indicate this with applies to part, aspect, or form (P518) qualifiers or encourage the creation of items for the parts with the statements and the deletion of the statements on the group item?
An example for an item for which all parts have the same collection is basilica cycle (Q17379582). Here it might be more useful to state this at the group item?

I'd probably prefer to move statements that do not refer to all parts to separate items. What do you think? --Marsupium (talk) 01:15, 20 January 2016 (UTC)[reply]

Thanks for bringing this up. I ran into a lot of these series (including the Sunflowers (Q157541)) and I prefer to have only statements on the series item that are true for all members, so no messing around with applies to part, aspect, or form (P518). For a lot of these series not all members have been created yet so I left the statements on the series item alone for now. Multichill (talk) 20:05, 20 January 2016 (UTC)[reply]

I suppose the issue is that one can only use applies to part, aspect, or form (P518) if there is an item one can give as its object; but if there is an item, then why not put the statement on the item directly.

Worst case I have found for this is Hours of Étienne Chevalier (Q1671114), a manuscript that has been broken up, for the property collection (P195) and also the different values for inventory number (P217). Most of the surviving leaves are in a museum in France, but the rest are spread all across the world. (See c:Category:Hours_of_Étienne_Chevalier).

I don't know if it makes sense to have items for each of the individual leaves (but perhaps each accession number?) On the other hand it's rather misleading to apply P195 and P217 to the whole collection, on the basis of a single leaf. It's not "the Hours_of_Étienne_Chevalier" that's in the British Library, but just one image from it. Jheald (talk) 23:04, 20 January 2016 (UTC)[reply]

Well I would leaved the large group alone (don't split it into leaves), but create an umbrella group for the book (which existed presumably at some point in time as one unit, though I know of examples that were split during the reformation and then added to later) and then each separate collection can be created and become part of it. --Jane023 (talk) 13:08, 21 January 2016 (UTC)[reply]

@Multichill: That is exactly what I'd also prefer. I tried to point this out.

@Jheald: applies to part, aspect, or form (P518) is also used with generic part descriptions, cf. Wikidata:WikiProject Visual arts/Item structure#Individual objects and parts, an example is Basilica Santa Maria Maggiore (Q15117526).

@Jheald, Jane023: Surely, one item for the group is anyway necessary. However, I think it should be useful to have an extra item for each splitted part since each of them gets own provenance with accession numbers and exhibition history and through structures here to document that is probably messy, especially as there are no qualifiers on qualifiers.

Would you agree to add to Wikidata:WikiProject Visual arts/Item structure#Individual objects and parts something like this (language and other correction appreciated!):

The group item should preferably only have statements that are true for all its members. Each part that was separated from any other part at some point in time should have its own item. As long as those items for the parts do not exist, the statements are accepted in the group items. If possible those statements should at least be specified through qualifiers with applies to part, aspect, or form (P518) indicating the part by a generic part item in the way explained below and/or if known a point in time (P585) (or if known better a start time (P580) and a end time (P582)) qualifier so that if those qualifiers are the same on different statements for the same property it is clear that they refer to different parts. [Here comes the paragraph to applies to part, aspect, or form (P518).]

? --Marsupium (talk) 17:42, 22 January 2016 (UTC)[reply]

Yes, that sounds reasonable. --Jane023 (talk) 17:48, 22 January 2016 (UTC)[reply]

Linking work/discipline/profession

On Wikidata:WikiProject Visual arts/Item structure#Types of visual artworks we have a (incomplete) table with work, the discipline (activity) and the profession. How should these types be linked together?

The missing items should be created and everything should be linked together. That should also prevent accidental merges. Besides that we still have:

art of painting (Q11629): P??? -> painter (Q1028181)
painting (Q3305213): P??? -> painter (Q1028181)
painter (Q1028181): P?? -> painting (Q3305213)

Any idea if we want to link these together and how? Multichill (talk) 18:14, 5 February 2016 (UTC)[reply]

The creation of the missing items is a good idea. A new property of painting (Q3305213): P??? -> painter (Q1028181) could perhaps be a superproperty of manufacturer (P176). I don't see a need to create the properties for the three missing links though there might be one. --Marsupium (talk) 15:35, 7 February 2016 (UTC)[reply]

user:Spinster started the work on the missing items. Awesome! Multichill (talk) 21:28, 9 February 2016 (UTC)[reply]

Provenance

I'm preparing several datasets to be imported to Wikidata (regarding this project, though the page is a bit outdated) and I'm wondering which P should I use to express provenance. donated by (P1028) is quite straightforward and I don't have any problem with donations, but I don't know how to proceed with all the other ways an artwork can enter a collection (acquisition, deposit, bequest…). Shall I use a combination of significant event (P793) + acquisition (Q22340494), deposit (Q5260774) (though it's not exacty the same), bequest (Q211557) + qualifier for the date? Thanks! --ESM (talk) 12:20, 1 March 2016 (UTC)[reply]

Wonderful to read this - welcome! I think we will probably need to propose some properties for this. I would just look at what data fields you want to match with Wikidata properties and add those to our current item structure. What is missing should be proposed at Wikidata:Property proposal/Creative work. --Jane023 (talk) 12:36, 1 March 2016 (UTC)[reply]

@Jane023: Hello, sorry for my late reply, I completely forgot to check if there was an answer. So, I've made a list of the possible ways an artwork can enter a collection (at least IME there aren't many, but still):

acquisition (Q22340494)
unknown (which Q should we use here? We are leaving it blank ATM)
deposit (Q5260774)
gift (Q707482)
under investigation (we are leaving it blank ATM)
bequest (Q211557)
award (Q618779)
ownership (Q6498684)
(some sort of retrieval. I'm unable to find a suitable translation to English)
(some sort of transfer between administrations. I guess this happens when several museums join their collections in a big, single one)
usucaption (Q313326)

What we've been doing (with @KRLS:) is expressing provenance by creating statements with significant event (P793) and the qualifier point in time (P585). In the case of donations (and the donor is someone notable) we add the statement donated by (P1028), with the qualifier point in time (P585). Would that be fair enough? Thanks for your help! --ESM (talk) 14:17, 30 March 2016 (UTC)[reply]

This looks good! I was just discussing recovery vs discovery for paintings that are re-acquired after theft. Is that what you mean? In the Netherlands we also have restitution guidelines for works restored to heirs after being stolen by the Nazis in WWII. We should probably propose a property for restitution since I don't think it is the same as recovery after theft. --Jane023 (talk) 18:42, 30 March 2016 (UTC)[reply]

@Jane023: Hello, I got an answer from Museu Nacional d'Art de Catalunya (Q861252) regardinf the two items of my list that I didn't know how to translate:

The first one is used when an artwork enters the collection after having been lost or unused. In this case would it be OK to consider it as a recovery?
The second one is used when an administration (which actually doesn't know the provenance of the artwork) transfers said artwork to another administration. I still have no idea about how to translate this one.

Step by step, I guess! --ESM (talk) 20:26, 4 April 2016 (UTC)[reply]

Yes step by step. In these cases I would just use discovery for an artwork that was previously unidentified as such (for example the Rijksmuseum had photos made when it was built in the 19th century, but didn't consider them artworks until now when early photos are appreciated for other reasons than their original purpose). For switching administrations then the collection changes, so that is not a change in the identification number, but the collection name. Jane023 (talk) 12:43, 5 April 2016 (UTC)[reply]

Still Life with Gingerpot 2 (Q18710646) is an interesting item. I updated it a bit based on the info here. Multichill (talk) 17:52, 10 April 2016 (UTC)[reply]

How to indicate the relative position of a part of an object within the whole object?

@Nurni: I am not sure if it is the best idea to indicate the relative position of a part of an object within the whole object like this:

⟨ Adoration of the Madonna (Q23797845)  

 ⟩ part of (P361) ⟨ Q23784077  

 ⟩
applies to part, aspect, or form (P518) ⟨ central part (Q17525443)  

 ⟩

Though I haven't seen a better solution on another item. Do you have ideas how to describe it? --Marsupium (talk) 02:14, 14 April 2016 (UTC)[reply]

@Marsupium: I thought it over and agree with you that it isn't the best solution. Though there are 155 similar cases where an item has part of (P361) with qualifier applies to part, aspect, or form (P518):

SELECT ?itemLabel ?item ?sLabel
WHERE 
{
  #?item wdt:P31 wd:Q3305213  .
  ?item p:P361 ?stat .
  ?stat pq:P518 ?s

  
   SERVICE wikibase:label {
    bd:serviceParam wikibase:language "en,pl,de,fr" .
   }
}

Try it!

I have no good idea how to improve it. Nurni (talk) 07:43, 14 April 2016 (UTC)[reply]

Oh! Thanks for the query! We probably need a new property like "position of part". This could perhaps be used as a qualifier for part of (P361) and as well for has part(s) (P527) statements. That might solve some of the 155 cases including those concerning objects of visual arts. The playing card use cases seem nonsensical to me. @Poulpy: But I probably miss the point? --Marsupium (talk) 08:06, 14 April 2016 (UTC)[reply]

Object history or provenance

I've just become aware of the list at Wikidata:WikiProject_Visual_arts/Item_structure#Object_history_or_provenance and notice that theft (Q2727213) is not one of the entries, which rather surprises me given that many notable artworks have been stolen at some point. Thryduulf (talk) 14:42, 18 August 2016 (UTC)[reply]

Sketches and final artworks

I've been searching for a property to state that an element is a sketch/project/mould… of other elements. Example: Q26756861 is the model of Q26756064 and Q26756461, but can't find a way to express this relationship between these elements, and to change their P31 to Q17489659 doesn't seem the proper way to do so. Any help will be appreciated, thanks! --ESM (talk) 14:17, 2 September 2016 (UTC)[reply]

Artworks associated with a culture, rather than a person

There are of course many thousands (millions?) of artworks in major museum collections that have no designated creator, and will likely never have any, since they come from societies where it was not common practice for artists to put a signature on their work. The current guideline has two suggested ways of dealing with such a situation. The first, Use of creator (P170) in uncertain cases, suggests that we attribute an ethnic group as the "creator" of the artwork, as if we actually know the individual artist's ethnic identity (rather than merely the cultural tradition they are working in), which is unsatisfactory. For example, it would not be at all uncommon that a creator of Roman art would be Greek or Middle Eastern or something else.

The second, Ethnic group / culture / ... (The context of the visual arts) suggests that we attribute it directly to the culture, but doesn't provide a particular property - I suggest that we specify culture (P2596), which although it is currently used mostly for archaeological sites, would seem to be very well-suited to importing metadata on artworks associated with particular cultures.--Pharos (talk) 15:30, 3 January 2017 (UTC)[reply]

Historical period as context

I would think that, additionally, we'd want to specify historical period as a suggested property in The context of the visual arts, as in time period (P2348) = Eighteenth Dynasty of Egypt (Q146055). Though it rather is broad, this is often the scope of dating information available for many more ancient artworks.--Pharos (talk) 21:37, 3 January 2017 (UTC)[reply]

Attribution Qualifiers

Notified participants of WikiProject Visual arts

@Multichill, Spinster, Oursana, Pigsonthewing:

Wikidata:WikiProject Visual arts/Item structure#Use of creator (P170) in uncertain cases describes 9 props to be used in cases where someone is not a "straight author" of a work.

These were proposed and approved en-masse at Anonymous works qualifiers.
But I think they are also applicable to non-anonymous works, eg creator John Smith, forgery after Rembrandt.

Multichill wrote: "The Rijksmuseum was so kind to share their collection manual. It's a 100+ document touching onto a lot of problems we face too, including anonymous works." But the Bible is CCO and sec2 Creator Information p94 lists 19 Attribution Qualifiers (20 are found in Getty AAT):

ATTRIBUTIONS TO A KNOWN CREATOR: attributed to, formerly attributed to, probably by, possibly by
UNKNOWN CREATORS LINKED TO KNOWN CREATOR: studio of, workshop of, office of, atelier of, assistant to, pupil of, associate of, manufactory of
Not Working Directly with Known Creator: follower of, circle of, school of
Influenced by Known Creator: style of, after, copyist of, manner of

If we merge the two sets, we'll probably get 23. But who's to say that's the end of it? Adding more and more properties is not a sustainable approach.

Furthermore, I see these problems:

You can't combine qualifier props, eg "formerly attributed to the workshop of" or "possibly by the workshop of".
You can't easily query for works related to a master, no matter what the nature of relation. There are no sub-props on WD, so you'd have to enumerate all the props.
More specific contribution roles (eg Sculpture (Cast made by), Casted by, Engraved by, Gilded by) are captured using Creator with qualifier has use (P366) "role", as I see at Mona Lisa: Creator
- We can combine specific roles with Anonymous Works Qualifiers eg like this
  - possibly by: Rodin; role: sculpture; point in time: 1890

WD promotes a relatively small number of props but a huge number of entities. So my proposal is to:

Get rid of these properties
Introduce a new prop Attribution Qualifier, or just reuse has use (P366) "role"
Reimplement the 9 WD props and the 20 AAT concepts as WD entities

As an added benefit, we can organize them in a hierarchy as in CCO & AAT.

The examples mentioned above could map to:

creator: John Smith; Attribution Qualifier: painter
- creator: Rembrandt; Attribution Qualifier: forgery after
creator: Rembrandt; Attribution Qualifier: workshop of
- sourcing circumstances: formerly
creator: Rembrandt; Attribution Qualifier: workshop of
- sourcing circumstances: possibly (presumably)
creator: Rodin; Attribution Qualifier: sculpture; point in time: 1890; sourcing circumstances: possibly
- creator: John Smith; Attribution Qualifier: casting (manufacturing process); point in time: 1905

The original 9 properties would be replaced by the following representations:

P1773 (P1773): sourcing circumstances: probably
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
possible creator (P1779): sourcing circumstances: possibly
school of (P1780): Attribution Qualifier: school of
after a work by (P1877): Attribution Qualifier: after a work by

If you disagree to remove the props, then let's expand the list to include all the ones from CCO & AAT. --Vladimir Alexiev (talk) 17:52, 7 March 2017 (UTC)[reply]

Pinging again:

Notified participants of WikiProject Visual arts

@Multichill, Spinster, Oursana, Pigsonthewing: --Vladimir Alexiev (talk) 07:44, 27 March 2017 (UTC)[reply]

Attribution Qualifier Discussion

For subproperties on Wikidata cf. subproperty of (P1647).

But you need to add another branch to your query, and hope all the relevant props have subproperty of (P1647) --Vladimir Alexiev (talk) 08:50, 8 April 2017 (UTC)[reply]

I see there might be advantages in using a "role" property. On the other hand I prefer to be able to state The Night Watch (Q219831)creator (P170)Rembrandt (Q5598) without having to set it with a qualifier The Night Watch (Q219831)creator (P170)Rembrandt (Q5598)rolepainter and especially as a data consumer I do not want to query a "role" qualifier if I (just) want to have the creator. How do CCO, Categories for the Description of Works of Art (Q5051819) and especially CIDOC Conceptual Reference Model (Q624005) handle this matter? --Marsupium (talk) 10:58, 27 March 2017 (UTC)[reply]

The case of "straight creator" is simple and will stay as is

CCO says you need to express Qualifications but doesn't specify how

CDWA has indexingCreatorSet/attributionQualifierCreator, so it's similar to what I propose (see https://github.com/VladimirAlexiev/rnc/blob/master/CDWAlite-nocomment.rnc#L113)

CIDOC CRM doesn't say how to handle it, but for the British Museum we handled it by adding appropriate attribute inside a Production event, see 3 sections here https://confluence.ontotext.com/display/ResearchSpace/BM+Association+Mapping+v2#BMAssociationMappingv2-ProducedByCloselyRelatedGroup. So it's again similar to my proposal: the qualification is an attribute somewhere, it's not a different kind of link to the person --Vladimir Alexiev (talk) 08:50, 8 April 2017 (UTC)[reply]

Hi Vladimir, I don't really have the mental bandwidth to think this through at the moment... I agree with you that the general museum and cultural sector vocabularies are better to base our model on than the temporary Rijksmuseum list we now use, and I've also encountered cases where I thought we need a better system than the current one. But now I don't have the time to dive into this more deeply. On the other hand I don't think we should make the system more cumbersome 'because adhering to standards is the thing to do', but because we are actually needing and will be using all options, and because the system on Wikidata will make querying and research more accessible.

One thing we don't solve well, is the workshops. That shouldn't be in a qualifier or property IMO - it would be good if we can, for instance, easily query all works created by 'the workshop of Rembrandt', and this workshop is actually also an agent which is a bit similar to an organisation.

I beg to disagree because the membership of such orgs often changed in time and is unclear. Furthermore, would you also handle "Circle of" and related groups (See British Museum link above) as such unclear vague Organizations? --Vladimir Alexiev (talk) 08:50, 8 April 2017 (UTC)[reply]

Personally, I would prefer to discuss this in a longer process with input from specialists, after having gathered real world examples that occur often and that we can use as examples. I think in the meanwhile it's really OK to work imperfectly, we can always improve things we input now where we find better solutions later on. Spinster 💬 11:19, 27 March 2017 (UTC)[reply]

Same with me as Marsupium and Spinster. I think on commons we have also and workshop which is quite important and on commons I miss or workshop. We have to keep things easy. Looking forward to solve it with examples--Oursana (talk) 12:20, 27 March 2017 (UTC)[reply]

It seems there is a request for some real world edge cases and an explanation of how they would work in the current system vs. how they would work in an alternative system. I think <applies to part> needs to figure in this: sometimes the face or main figure is by the master and the background is thought to be workshop.

In the meantime, I would support keeping the current process and adding the "missing" attribution qualifiers from AAT. - PKM (talk) 20:18, 27 March 2017 (UTC)[reply]

Maybe that's a good way for the time being, until someone is ready to write a bot to migrate the statements. The risk is that adding the missing ones will institutionalize the current (wrong) approach and "cast it in stone". --Vladimir Alexiev (talk) 08:50, 8 April 2017 (UTC)[reply]

Hello Vladimir Alexiev. Yes, creator (P170) without qualifier is enough in many cases (just one creator and no debate on attribution). Sometimes, as you point, it is not so easy and I agree that to qualify the relation between an item and a creator item is the best approach. It is the common approch in cataloging (for examples in AAT model or in French database Joconde (Q809825) http://www.culture.gouv.fr/documentation/joconde/fr/partenaires/AIDEMUSEES/methode.htm#AUTR ). Even more, it is a good approach for simple edition and easy reuse. And with this approach, we can enrich existing items with adding qualifications without changing existing and common claim (for example: if it is not the painter himself but his workshop, we have only to add the information as form of qualifier and we keep the relation item<->P170<->creator).

An example where specialist don't agree on an attribution : https://www.wikidata.org/wiki/Q19904792#P170 . Here we need and have 2 statements (one affirmative, one presumed) with reference for each one. In that case, it is difficult to imagine using the property P1773 (P1773) in qualifier.

A very good example. "sourcing circumstances: presumably" matches the qualifications I propose, but we need to use the standard vocabulary used in CCO --Vladimir Alexiev (talk) 08:50, 8 April 2017 (UTC).[reply]

Another issue with the several proporties in qualifiers is the complexity for querying on a creator and to display the creator (P170) of an item. It is complex and only accessible for those who know very well this complexity and can manage it.

I agree with your proposal for a specific property for those kind of qualification instead of several properties. In practice it should works and could be a great way to keep easy edit, with more complex information and specific references, and easy reuse (the goal of Wikidata). It could be very useful. Thanks. Best regard --Shonagon (talk) 01:42, 28 March 2017 (UTC)[reply]

copyright information

I started to add copyright information for artworks.

the property public domain date (P3893) with a date and as qualifier applies to jurisdiction (P1001) and if needed start time (P580) and end time (P582) for begin and end of validity of this date (we can have multiple dates as copyright rules might change over years)
as a source: laws applied (P3014) with a link to national legislation (law or specific article) and determination method or standard (P459) to include a copyright term (Q5169289) that has been used for the calculation, for example 70 years or more after author(s) death (Q29870196)
see as an example Truncated View of the Broekzijder Mill on the Gein Wings Facing West (Q19883472)
there is also a proposal for a property for a copyright owner Wikidata:Property proposal/copyright owner --Hannolans (talk) 15:00, 9 May 2017

(UTC)

two-dimensional work of visual art

How about Two-dimensional work of visual art (set, superset is Q41806013)? --Fractaler (talk) 08:10, 1 November 2017 (UTC)[reply]

horizontal_depth

See Wikidata:Property_proposal/Creative_work#horizontal_depth discussion --Jarekt (talk) 01:32, 15 July 2018 (UTC)[reply]

Engraving

Any idea how The Congress of Vienna (Q66120851) should mention the engraver and the person who drew it?

I went for creator (P170)=engraver and after a work by (P1877)=drawer. The other items I looked at seem to support at leas the first part.

Commons seems to have only prints of the engraving (Wellington is said not to be in the drawing).

Ideally we would have additional items for some of the reproductions, quality varies largely. --- Jura 14:17, 5 August 2019 (UTC)[reply]

Distinguishing an artwork and its editions

Trying to clean up the items for Non-Violence (Q3142180) and Tuffsen (Q15220016). Both are cases where a work of art was created and then expressed in multiple physical editions. In neither case is there an obvious original. In the the first case the various editions do not necessarily share dimensions, in the second it exist in different materials.

My spontaneous thought is that there should be an item per physical manifestation where that item indicates it is an edition of some main the item containing the shared propritrs (creator, depicts etc). My struggle is figuring out what to tag the "main" item so that its ibvious that it shouldn't carry statements about position etc which are not shared by all editions. Any suggestion are welcome. /Lokal _Profil 11:16, 15 November 2019 (UTC)[reply]

I used “group of paintings” with “has part” for Armada Portrait (Q4792503) ~~(the individual painting item has been tagged “part of series”, so~~ you might want to use “painting series” instead). - PKM (talk) 19:34, 16 November 2019 (UTC)[reply]

And details for one of the portraits are still attached to the group... will fix. - PKM (talk) 19:38, 16 November 2019 (UTC)[reply]

I’ve cleaned these up and removed “part of the series” since it was causing a constraint violation. It’s a better example now. - PKM (talk) 22:44, 16 November 2019 (UTC)[reply]

Thanks for the clear example! I saw the option of using group of works (Q17489659) (or a subclass therof) but what stopped me is the sense that it should be made up of distinct but related works. I would argue that in my example there is only one work, but distinct representations therof. I.e. there is no creative dofference between them. Similar to multiple inabridged editions of the same book-work. /Lokal _Profil 10:47, 23 November 2019 (UTC)[reply]

Instrument used

Descriptions of art objects sometimes include instrument used, like ball pen or brush. What property can be suggested ? (beginning of discution at WD:PC#Instrument used) - Kareyac (talk) 06:47, 20 February 2020 (UTC)[reply]

Collections modelling

Moved to Wikidata talk:WikiProject Visual arts#Collections_modelling to get broader input. Multichill (talk) 12:03, 27 November 2020 (UTC)[reply]

Copies of bronze sculptures

I am wondering if we have a consensus for the modelling of the relationship of individual instance copies of bronze sculptures to the parent item, and to each other? For example, The Thinker (Q18003128) (Rodin) is a famous bronze sculpture which has many copies in museums around the world. The notable feature of this relationship is that there [potentially] many copies that are all equally "real" and legitimate - they are not derivatives or fakes like the word "copy" might imply in other artforms.

There are several ways that this is currently being done in different items, or could be done. I'd like if we could do it consistently...

based on (P144) is one option (also know as “copy of”, “derived from” and “modelled after"). It is used in this item The Thinker (Q154571).
instance of (P31) is another option, used in this item The Thinker (Q77440071).
Another possibility is part of the series (P179) for the individual works and has part(s) (P527) for the parent item. This has precedent for use with items about paintings and tapestries.

I assume different from (P1889) (or perhaps series ordinal (P1545) would be the only ways we'd wish to connect the different individual items to each other.

What do people think? Wittylama (talk) 16:41, 24 May 2021 (UTC)[reply]

As the person who is asking - my instinctive preferences is based on (P144), due to the aliases covering the concept that this issue represents quite well, IMO. My instinctive reaction to the third option (has part(s) (P527)/part of the series (P179)) is that this is't appropriate as that is for when the individual works are intended to be different from each other. Wittylama (talk) 17:01, 24 May 2021 (UTC)[reply]

Similar question asked here one year ago, with no response. Good luck. -Animalparty (talk) 17:41, 24 May 2021 (UTC)[reply]

@Wittylama: - I don't have easy answers but thought I'd give some encouragement and a possible way for improvement - at one of our GLAM edit-a-thons a while back, even one of the world's experts on sculpture said that the art history community gets this wrong, and there are no easy answers. The quote she had was, "Welcome to my misery." Her video critique of the Wikipedia article about "The Greek Slave" is instructive and the recording made the advice useful for future editors as we've gone back to it a number of times since to improve content. The video that starts at 10 minutes and 55 seconds gives you 5-minutes of insight into how we might improve our writing and modeling, at least of marble sculpture - en:Talk:The_Greek_Slave#Scholar_reflects_on_this_article. With my work with the Smithsonian, we have access to Dr. Lemmey in case we come up with a more comprehensive plan for Wikidata modeling that I'm confident she would be happy to give feedback on. -- Fuzheado (talk) 11:16, 25 May 2021 (UTC)[reply]

ivory carving (Q351853)

I wonder if ivory carving (Q351853): carving of animal tooth or tusk by using sharp cutting tools shouldn't be a subclass of sculpture (Q860861): three-dimensional work of art instead of carving (Q18448934): act of using tools to shape something from a material by scraping away portions of that material ? Weft (talk) 07:59, 21 December 2022 (UTC)[reply]

I think this could validly be both (and not instance of scuplture). PKM (talk) 00:46, 2 January 2023 (UTC)[reply]

Photographs - subclasses vs. genres

I've been making items for concepts from the Thesaurus for Graphic Materials (Q52441805). I have made expedition photography (Q115761211) and postmortem photography (Q115923787) and a few others as subclasses of photograph (Q125191) (following the existing railway photography (Q113190018)), but I wonder if these should all be ~~subclasses~~ instances of photography genre (Q3100808) instead?

The discussion of subclasses/genres upstream is several years old now - how do you all feel these days? - PKM (talk) 00:44, 2 January 2023 (UTC) [edited][reply]

Never mind, resolved by further comparisons. Should be both instances of photography genre (Q3100808) and subclass of photography (Q11633). Off to make changes ...

Provenance2

@Jane023:, @Jarekt:, @Multichill: I remember we discussed it meanwhile, but can't find it. Can we move or at least link to the discussion here.

Please look at Talk:Q48976445 Oursana (talk) 14:42, 20 January 2023 (UTC)[reply]

@LAP959:

Copying my response from there to here: It's fine to have two owners at the same time in owned by (P127) - just use qualifiers to explain or clarify, together with significant event (P793) for art theft (Q1756454). See Wikidata:WikiProject Provenance to discuss in future. Jane023 (talk) 15:37, 20 January 2023 (UTC)[reply]

Oursana, there was a lot of discussions at Wikidata:Property_proposal/Sister_projects#"beforehand_owned_by"_&_"afterwards_owned_by". --Jarekt (talk) 01:51, 21 January 2023 (UTC)[reply]

Thanks for reminding me, Jarekt! I have discussed this before/after issue with Multichill from time to time and my head explodes whenever I think about it. Rather than "previous owner" or "next owner" we need something more specific for the agent of the transaction who may or may not use their own money in the transaction (a bequest may involve taxation but no money). Note this is off the topic of theft or auction and is purely a modelling question. I think we need some sort of "handoff" property (maybe like a baton in a relay race). If two things like "handoff releaser" versus "handoff holder" existed, then we could use an extra agent property for large transactions such as complex purchases involving consortia. I think in such cases the consortium probably needs its own item. In general I have not started modelling art deals separately from their art dealers, but maybe we need to think about that too. Jane023 (talk) 08:42, 21 January 2023 (UTC)[reply]

Thanks for your answers, I cannot follow Jane023 completely, but we need more a property for possession without ownership. As the illegitim possession is no ownership. I cannot model my afore mentioned example Q48976445 with the available properties.--Oursana (talk) 10:50, 21 January 2023 (UTC)[reply]

Illumination styles

I have posted the following at WikiProject Manuscripts.

Our items for national and regional illumination styles are not consistently modelled. Here are examples from before I started playing with them:

compare:

and

art of Romanesque illumination (Q1276812)subclass of (P279)art of illumination (Q22669085)
art of Romanesque illumination (Q1276812)Commons category (P373)Romanesque manuscripts

I like using the metaclass art style (Q1792644) for these, which is a subclass of "genre". I would like to propose a standard for illumination styles like so:

I am a little uncertain about using a Commons class of manuscripts for an art style, but not totally opposed. Does anyone have comments on this point, or any other suggestions for improvement on the model? PKM (talk) 23:37, 9 December 2023 (UTC)[reply]

Update: Per discussion on Wikidata Telegram and at WikiProject:Manuscripts, I am going to follow the model suggested above plus I will make and link separate items for art styles and manuscripts, with the commons links going to the subclasses of "illuminated manuscript". There will be at least two phases. Phase 1 is harmonizing all of the art style items (complete). Phase 2 is separating manuscripts (product) from art styles (started). A possible phase 3 will be creating art style items for Persian and Indian manuscript styles, which we don't have today (or I haven't found them). - PKM (talk) 00:11, 23 December 2023 (UTC)[reply]

Collecting info about the curatorial departments in art Items

Please see Wikidata talk:WikiProject Visual arts#Collecting info about the curatorial departments in art Items DaxServer (talk) 14:59, 24 May 2024 (UTC)[reply]