Localities in works of art

[Jegelewicz]

Would "object" be appropriate for several busts of celebrated scientists such as Linnaeus, Humboldt, Buffon, and St. Hilaire? These are also part of our collection of casts. We were thinking of identifying them as Homo sapiens, but is there a way to perhaps identify them as an agent, or both? For location, we thought we would put down their burial site. I'm curious as to your thoughts on this.

@AJLinn can you give them some good advice? Thanks!

[AJLinn]

Object would be the part name. I would build it in the data entry page like this: Scientific name: [build -->] A {string} / taxon name = unidentifiable / ID string = Bust, Carl Linnaeus [Fine Art would probably describe it first as "Sculpture, Bust, Carl Linnaeus"]

Attribute: Description - describe who the bust depicts, what the bust looks like and is made from, with measurements In the agents section you'd list the artist, not the subject Locations would be associated with the manufacture / use / collection of the bust, not the subject matter. If you want to list where they're buried, make them agents and list it there. Someone can cross reference to the agents table.

However, I wonder if it would be useful for our Art collections to have an agent role "subject" for portraits? I know they do something with locality regarding what place is depicted in landscapes. Maybe @marecaguthrie or @krgomez could weigh in on the usefulness of that as a new agent role? It's sort of related to that author role I brought up in #1863 .

[dustymc]

Yea all agreed.

"object" is the generic part for collections in which the 'what it is' data are stored in identifications.

You CAN use the "A {string}" formula to link the object to Homo sapiens, but I'd assume that means it's carved out of a person.... If the object is carved out of ivory then "Odobenus rosmarus {Sculpture, Bust, Carl Linnaeus}" would be a useful identification (for the specimen bulkloader - you can add as many taxa as you want, and don't have to deal with the formulae, after the record is created).

It's always worthwhile to let the relational nature of Arctos work for you. If you record Linnaeus' burial site in a specimen record then it applies to that specimen record, and if you need to correct it you need to correct all records using the data. If you record it with the agent then it applies to everything that references the agent, and if you need to correct it you do so once and that propagates out to all bajillion child "records." ("Records" are entirely arbitrary things in Arctos - an Agent's address is "part of the specimen record" if you want it to be, that's only a matter of perspective.)

agent role "subject"

Do you mean in http://arctos.database.museum/info/ctDocumentation.cfm?table=CTCOLLECTOR_ROLE? If so I don't think I have any objections....

depicted

http://arctos.database.museum/info/ctDocumentation.cfm?table=CTSPECIMEN_EVENT_TYPE

Maybe an observation? "Specimen was detected and not killed or removed from context; No biological samples were taken. Human sightings, camera traps, and GPS telemetry data are appropriate here." doesn't seem entirely wrong, although it could be tuned up for this use.

http://arctos.database.museum/info/ctDocumentation.cfm?table=CTMEDIA_RELATIONSHIP (if you have media) works as well (although https://github.com/ArctosDB/arctos/issues/1608 is still hanging around) - 'created from collecting_event' is the place the camera was (and it's probably small), 'shows locality' (which should be 'shows collecting_event') is what's depicted in the image (and is probably large).

[krgomez]

However, I wonder if it would be useful for our Art collections to have an agent role "subject" for portraits? I know they do something with locality regarding what place is depicted in landscapes. Maybe @marecaguthrie or @krgomez could weigh in on the usefulness of that as a new agent role? It's sort of related to that author role I brought up in #1863 .

What we have done for images depicting agents is to use the "shows agent" media relationship for the media that shows the collection object.

This is also the strategy that we were planning to use for artworks depicting places, that is, to use the "shows locality" media relationship. However, we never did this as I think there was uncertaintly that this was the best way to go about documenting localities depicted in artworks. See #1608. We could use some advice on this. Right now this information is only documented in description attributes.

As for a new agent role "subject", this seems like it might be a better way of documenting this information than through media. These relationships feel more buried than if the information was somehow more connected to the object information.

[marecaguthrie]

Yes- I very much like the agent role- subject

On Fri, Apr 12, 2019 at 11:33 AM Karinna Gomez notifications@github.com wrote:

However, I wonder if it would be useful for our Art collections to have an agent role "subject" for portraits? I know they do something with locality regarding what place is depicted in landscapes. Maybe @marecaguthrie https://github.com/marecaguthrie or @krgomez https://github.com/krgomez could weigh in on the usefulness of that as a new agent role? It's sort of related to that author role I brought up in

1863 https://github.com/ArctosDB/arctos/issues/1863 .

What we have done for images depicting agents is to use the "shows agent" media relationship for the media that shows the collection object. [image: Screen Shot 2019-04-12 at 10 58 19 AM] https://user-images.githubusercontent.com/40579515/56059966-f489a780-5d11-11e9-9a67-877a2cfba94e.png

This is also the strategy that we were planning to use for artworks depicting places, that is, to use the "shows locality" media relationship. However, we never did this as I think there was uncertaintly that this was the best way to go about documenting localities depicted in artworks. See

1608 https://github.com/ArctosDB/arctos/issues/1608. We could use some

advice on this. Right now this information is only documented in description attributes.

As for a new agent role "subject", this seems like it might be a better way of documenting this information than through media. These relationships feel more buried than if the information was somehow more connected to the object information.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/ArctosDB/arctos/issues/2044#issuecomment-482696452, or mute the thread https://github.com/notifications/unsubscribe-auth/AlUox-p7aCLZC4_eyOaE8kvRvAmQ3u6bks5vgN-LgaJpZM4csNM5 .

-- Mareca Guthrie Curator of Fine Arts & Associate Professor of Art University of Alaska Museum of the North 1962 Yukon Drive P.O. Box 756960 Fairbanks, AK 99775-6960 mrguthrie@alaska.edu

University of Alaska Museum of the North: www.uaf.edu/museum UAF Art Department: https://www.uaf.edu/art/ https://www.uaf.edu/art/ Colors of Nature: http://www.colorsofnature.org/

[dustymc]

documenting localities

Localities are places. Events are places at times. You always have temporal information - you know your photograph wasn't taken tomorrow, was taken some time after cameras were invented, etc. The idea is stable, https://github.com/ArctosDB/arctos/issues/1608 is just moving the relationships to a more appropriate data object.

[krgomez]

Maybe an observation? "Specimen was detected and not killed or removed from context; No biological samples were taken. Human sightings, camera traps, and GPS telemetry data are appropriate here." doesn't seem entirely wrong, although it could be tuned up for this use.

http://arctos.database.museum/info/ctDocumentation.cfm?table=CTMEDIA_RELATIONSHIP (if you have media) works as well (although #1608 is still hanging around) - 'created from collecting_event' is the place the camera was (and it's probably small), 'shows locality' (which should be 'shows collecting_event') is what's depicted in the image (and is probably large).

Do you think making an observation event might be the best way to document a depicted place that is different from the locality assigned to the manufacture event? Since it sounds like shows locality is going to change to shows collecting event, I suppose one option for us is then to create an observation event that describes a locality depicted in the artwork? Basically we're trying to find a way to deal with, for example, a woodcut depicting a glacier at Mt. Chamberlain, where the block was cut and printed in Fairbanks. Or a painting of Denali that was made in an Anchorage studio. Is there another idea of how we could deal with this kind of information? Our manufacture events relate to the act of making for which the locality is sometimes/maybe often not the same as that which is depicted. All we have now is text describing whatever is depicted in the artwork in our description attribute field, but we see value in recording depicted places as localities somehow.

[dustymc]

Given the current choices, yes but maybe only sorta.

MAYBE we need a new event type, but I don't think I see a clear line between 'artistic picture of Denali' and 'camera trap picture of Denali (that happens to have a wolf standing in it).' I'd generally rather over-use one value (users might find some stuff they don't care about) than use multiples arbitrarily (users won't find all of what they care about).

I'm less comfortable with a painting than a photograph. My 4 year old isn't a very realistic artist. He swears this scribbling on the electric bill is Australia. Nobody wanting to see how things have changed over time has any reason to care. They probably DO want your photos, even if they were produced by an "artist" and cataloged by a fine art collection instead of a naturalist and cataloged by a natural history collection.

I suppose that's just a judgement call regarding how accurate the depiction is. Paintings which can be used to get at natural history questions certainly exist!

Our manufacture events relate to the act of making for which the locality is sometimes/maybe often not the same as that which is depicted.

That is the distinction media relationships "created from collecting_event" and "shows collecting_event" are trying to get at - given those two things you should be able to more or less re-create the photo. http://mvz.berkeley.edu/Grinnell/

@ccicero @mkoo thoughts on this?

[marecaguthrie]

Paintings which can be used to get at natural history questions certainly exist!

Absolutely! Even highly distorted imagery can tell us what aspect an artist was emphasizing or focused on. And completely non-representational works of art can be of value for scientists too such as the type of mark, type of tool or type of pigment that was used. I absolutley hope that scientists will make use of our collection in Arctos but I'm also thinking of the value that the artwork would have for our collection's primary users: artists, art historians, freelance art curators, other art museums (both academic and non-academic), documentary film producers, poets, historians, and humanities scholars in general.

The imagery in an artwork that is inspired by or a reaction to a particular location doesn't have to be photorealistic to have value to these users. A painting of a particular location may or may not have value for a climate change scientist but it can say a lot about individual experiences a location and/or speak to the larger complex emotional meaning that a location holds for a particular culture at a particular time even if this information is more subjective and the type of research is more qualitative than quantitative.

Subject matter location could be /should be listed in the description of the object but mapping it would have considerable value because someone who is researching or organizing an exhibition of artwork related to a particular region might not know all the names of those localities within that region. I can also see a fantastic project mapping amount and type of color used across various localities or mapping different artistic mediums used for various locations.

I'm not sure that this type of location information fits best under event types. A lot of artwork is rather temporally nebulous- it can be a depiction of a memory that is mixed with the present or a representation of an imagined future. We have examples in our collection that are photographs that are deliberately contradictory as to what time period they are occurring in. I'm really seeing this tracking of location as a form of subject matter rather than a particular event. Give me a few days to think on this and I'll post specific examples.

[dustymc]

Thanks @marecaguthrie, very useful.

I hadn't considered the "natural history" aspects of the materials-n-such. I suspect that, like a "cultural" object that happens to be made of something with interesting (to geneticists, the stable isotope crowd, etc.) molecules, your user base is in part an artifact of who can find your material. Maybe the dendrochronologists would like to have a close look at your frames if they could find them on GBIF by searching for "Betula+Alaska+before {date}"....

It still seems to involve a fair bit of judgement - I'd probably use a text field to get at "distorted" place-time data - but I think I'm pretty comfortable with the idea that there's value in 'object depicts place+time.' If we're all agreed on that, then we need to decide if that's an 'observation' (where I'm currently leaning) or some new type of specimen-event.

[Jegelewicz]

It still seems to involve a fair bit of judgement - I'd probably use a text field to get at "distorted" place-time data - but I think I'm pretty comfortable with the idea that there's value in 'object depicts place+time.' If we're all agreed on that, then we need to decide if that's an 'observation' (where I'm currently leaning) or some new type of specimen-event.

I like "observation" and when the time is distorted it can be represented with a wide range in the began-end date fields. Place distortion might be a bit tough, although these probably wouldn't be georeferenced, so maybe that doesn't matter as much?

[marecaguthrie]

I need to think about this because it isn't about observing the object- it is about what is visually or conceptually observed in or by the object itself. I'm worried that if observation events are used in multiple ways then when someone wants to search for results the data will be too garbled and not helpful. Perhaps a new type of event called "depiction"? "artistic depiction"? I would need to think this over a bit more...In the documentation, we could describe it as not just visual depiction of a location but also conceptual depiction. I'm trying to think about how a non-art collection might use this. I keep getting hung up on the fact that it is an event- I'm just not sure this is the right fit- I don't really see it as an event. Is there any other way of documenting this outside of events? Does anyone have any better ideas? Karinna and I have both looked at the Getty Vocabulary and can't find a method that they use for recording locations as a subject matter in the artwork.

[dustymc]

depiction

I could be convinced.

artistic depiction

Too subjective, I think, although there might be some utility in somehow segregating "photo of {place}" and "some sort of less-realistic depiction of {place}" or something.

event

"Events" are fundamentally just links between specimens ("things to which someone assigned a catalog number") and place-time data.

[marecaguthrie]

Yes, that's what I was thinking- some way to distinguish between the artist/maker's intention- where the choices (framing, color, lighting) were more considered and deliberate for the purpose of creating something that gets at larger ideas, emotions, concepts vs. something that was used primarily as a form of documentation.

On Wed, Apr 24, 2019 at 1:15 PM dustymc notifications@github.com wrote:

depiction

I could be convinced.

artistic depiction

Too subjective, I think, although there might be some utility in somehow segregating "photo of {place}" and "some sort of less-realistic depiction of {place}" or something.

event

"Events" are fundamentally just links between specimens ("things to which someone assigned a catalog number") and place-time data.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/ArctosDB/arctos/issues/2044#issuecomment-486429538, or mute the thread https://github.com/notifications/unsubscribe-auth/AJKSRR6OQP3QKSOXT64NPLTPSDEXTANCNFSM4HFQ2M4Q .

-- Mareca Guthrie Curator of Fine Arts & Associate Professor of Art University of Alaska Museum of the North 1962 Yukon Drive P.O. Box 756960 Fairbanks, AK 99775-6960 mrguthrie@alaska.edu

University of Alaska Museum of the North: www.uaf.edu/museum UAF Art Department: https://www.uaf.edu/art/ https://www.uaf.edu/art/ Colors of Nature: http://www.colorsofnature.org/

[Jegelewicz]

depiction

I could be convinced.

I like it.

[marecaguthrie]

I'm okay with depiction if you don't think it would be too vague. Karinna do you have any thoughts? Can you help me come up with a description for depiction?

"Visual or conceptual depiction of place, such as the location that is the subject matter represented in a work of visual art, music or poetry." (?)

Does it need to be the subject matter? Can it be depicted without being the subject matter and if so do we want to track it? Should we word it to also include historic photographs and field photos or are those tracked in a different way?

On Wed, Apr 24, 2019 at 1:31 PM Teresa Mayfield-Meyer < notifications@github.com> wrote:

depiction

I could be convinced.

I like it.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/ArctosDB/arctos/issues/2044#issuecomment-486434884, or mute the thread https://github.com/notifications/unsubscribe-auth/AJKSRR2AVZEKIGRXTO5OUBLPSDGSTANCNFSM4HFQ2M4Q .

-- Mareca Guthrie Curator of Fine Arts & Associate Professor of Art University of Alaska Museum of the North 1962 Yukon Drive P.O. Box 756960 Fairbanks, AK 99775-6960 mrguthrie@alaska.edu

University of Alaska Museum of the North: www.uaf.edu/museum UAF Art Department: https://www.uaf.edu/art/ https://www.uaf.edu/art/ Colors of Nature: http://www.colorsofnature.org/

[dustymc]

historic photographs and field photos

I think not, but I'm pretty waffly on that. Photos (videos, audio, whatever, and wherever they're cataloged) still seem more like 'observation' events.

I also think nobody else is dealing with this sort of thing as cataloged_items (it's media supporting cataloged items) so maybe answering that isn't critical just yet.

"Subject matter" seems overly specific - eg, I suspect most users would want both a painting of Denali and a painting with Denali in the background; they may not much care what you think the subject is.

[Jegelewicz]

historic photographs and field photos

I think that depicts still works in this case. Dusty is right that from the natural history perspective, we generally aren't cataloging field photos (but we should be - in a separate collection - then we would have an actual DAMS see #1738 ). Instead, we just upload them as media and use the "shows event" relationship.

BUT WAIT! that relationship does not exist, just "shows locality"! I thought we were doing away with that!!!! #1608

http://arctos.database.museum/info/ctDocumentation.cfm?table=CTMEDIA_RELATIONSHIP

[dustymc]

(but we should be - in a separate collection

We used to. Collections are arbitrary - cataloging them separately doesn't inherently do anything special, other than take a tiny bit more work/access to relate them to other stuff. AFAIK we can magic a DAMS out of anything that might be called a document (so anything) on demand, and of course Arctos contains much more information than any other DAMS. Find me a use case....

BUT WAIT!

PLEASE, get the AWG to tell me what to do, or help do, or whatever, with that.

[krgomez]

"Visual or conceptual depiction of place, such as the location that is the subject matter represented in a work of visual art, music or poetry." (?)

Does it need to be the subject matter? Can it be depicted without being the subject matter and if so do we want to track it? Should we word it to also include historic photographs and field photos or are those tracked in a different way?

I like the idea of "depiction" events. Maybe leave out the subject matter language? Visual or conceptual depiction of place as seen in works of art? Visual, photographic or conceptual depiction of place as seen in works of art?

I wonder about all the 1400+ photographs in the Art collection. For many of these we have a verbatim location of where the photo was taken recorded by the artist. When that is the case, we have included those localities in the manufacture events. Would we also want to create depiction events for these photographs? Or if the place depicted is the same as the place of manufacture, would that not be necessary?

[dustymc]

as seen in works of art?

Too specific - this might work for my not-so-artistic hand-drawn map of my trapline or something.

I'm also like 51% sure that "photographic" is an observation and not a depiction.

Would we also want to create depiction events for these photographs?

I'd vote yes, assuming you have the data and resources.

Or if the place depicted is the same as the place of manufacture, would that not be necessary?

I suspect having both would always be useful for searching. I think "same place" would almost always be a symptom of low precision - if there's enough 'place' depicted for you to consider this approach, the creator probably isn't occupying all of what's depicted.

[krgomez]

How about just: Visual or conceptual depiction of place.

[krgomez]

Is it okay to create the new event type "depiction" defined as visual or conceptual depiction of place ? We would like to start using this if you are all agreed it's good to go.

[campmlc]

I agree.

On Tue, May 14, 2019 at 12:13 PM Karinna Gomez notifications@github.com wrote:

Is it okay to create the new event type "depiction" defined as visual or conceptual depiction of place ? We would like to start using this if you are all agreed it's good to go.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/ArctosDB/arctos/issues/2044?email_source=notifications&email_token=ADQ7JBFBHRM64GMXLN7N7A3PVL6MRA5CNFSM4HFQ2M42YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODVMKNSI#issuecomment-492349129, or mute the thread https://github.com/notifications/unsubscribe-auth/ADQ7JBCEBPWKVFR7GV3INVTPVL6MRANCNFSM4HFQ2M4Q .

[dustymc]

@campmlc I think you have superpowers - feel free to express your agreement by updating the code tables....

[Jegelewicz]

OK, definition police - using the word in the definition is a no-no. How about:

depiction - visual representation or conceptual portrayal of place

[Jegelewicz]

Added

[krgomez]

Thank you!!

[krgomez]

I have a question regarding searching localities + event type.

I have started creating depiction events and in doing so realized that the searches aren't working the way I want them to. That is, it doesn't seem like one can search for specific locality AND event type together, which is problematic. For example, we want to find all the artworks in the collection that depict Denali. So for any artworks that do this, I am using a locality where the specific locality is recorded as "Denali" (10998105). The verbatim localities vary between these artworks; Denali, McKinley, and sometimes no verbatim, so, not recorded. So I can't search for depiction event type AND verbatim locality if I use verbatim properly. The only way I see to do this is to just use whatever I want to search for (Denali) in the verbatim for all the artworks depicting the mountain, even if this will be inaccurate for some artworks (ones where Denali is not recorded verbatim).

Does this make sense? Do you have any guidance or advice as to how I can best deal with these depiction events so that we can easily search for places depicted in works in the art collection?

[dustymc]

I think this is a new issue. https://github.com/ArctosDB/arctos/issues/755 is perhaps most-relevant. If we search by data from any locality, users get back results that may (from the table, unless they have "json locality" turned on) make no sense to them. If we don't, users can't find what they're looking for. Given a useful replacement, I'd get rid of the tabular results altogether - it's obviously not capable of representing our data. I don't have a useful replacement, and people would start getting out pitchforks if I tried to take their tables away so....

I'm up for better ideas. Whatever we ultimately do should at least be consistent and documented. It's currently neither.