Keywords attribute for Fine Arts

[krgomez]

Issue Documentation is http://handbook.arctosdb.org/how_to/How-to-Use-Issues-in-Arctos.html

Table https://arctos.database.museum/info/ctDocumentation.cfm?table=CTATTRIBUTE_CODE_TABLES

Value keywords

Definition Terms used to represent the content, subject matter and ideas involved in works of art.

Collection type UAM:Art

Attribute data type categorical

Attribute value We currently have about 250 values with documentation that we would like included in this table. Here is a work in progress draft for the keyword terms. https://docs.google.com/spreadsheets/d/1o_IG0Yr5diYdpB-HeYoR3WURjgZLJaXEYtQb151gsO0/edit?usp=sharing

Context Describe why this new value is necessary and existing values are not. Currently, we do not use keywords for cataloging artworks, with the exception of inconsistent term use in the description attribute field. The use of standardized keywords is an essential component in cataloging artworks, allowing users a means to find objects in more complex ways in which they may not otherwise be found. We wish to begin using a standardized list of keywords, drawn mostly from the Getty Art & Architecture Thesaurus (http://www.getty.edu/research/tools/vocabularies/aat/), and curated to include terms relevant to the UAM:Art collection.

The only way we see to implement this is as an attribute that uses a controlled vocabulary list of terms. The only other way we can see is to just record the terms as a free text list in the description field or in a keywords free text field. However, using free text for a concise, standardized set of terms will lead to errors and incompleteness.

One major drawback is that it doesn’t appear possible to search a controlled vocabulary attribute with the dropdown list of terms. Looks like this issue came up in #1466. By using keywords, we are specifically trying to enable more concise searching for users. It is not ideal to require the user to have to guess what keywords might be available to search. It would be much better to provide the dropdown list that includes all available keywords to search. Is this possible?

This is our best idea for implementing the use of keywords, but maybe someone can think of a better way?

Priority Please assign a priority-label. Normal?

[dustymc]

Getty Art & Architecture Thesaurus

Is this information not best implemented as metadata of taxon names?

https://github.com/ArctosDB/arctos/issues/2057

[krgomez]

Probably? If the plan is to potentially use AAT/Nomenclature in Arctos for classification of art and cultural collections, then it would make sense to also use it for keywords. I'm having trouble envisioning exactly how this would look. Would we add separate identifications for the object classification (painting, print, etc) as well as keywords (landscapes, climate change, glaciers, etc)? For any given artwork we would probably have around 5 or more keywords. Does it make sense to have both classification and keywords in identifications together like that? Or maybe we would instead add keywords as "associated taxa"?

[dustymc]

separate identifications

You can, or the A {string} ID formula allows for linking an ID string and any number of taxa.

classification and keywords in identifications

Some example data or further explanation would be useful - I might be way out in left field again - but I'm thinking common names are analogous.

associated taxa

That's more "this moose was hanging out in birch forest."

[Jegelewicz]

I think that using attributes here makes sense but I cannot access the Google Doc, so it's hard for me to decide. Can you make that document public or create a public version?

[krgomez]

Oops sorry - the spreadsheet is now public.

I talked with Mareca and while we originally saw keywords as attributes, we are now interested in the idea of using keyword terms directly from the Getty AAT. It does make sense if we will already have access to it for object classification purposes. We simply hadn't pictured keywords as part of the identification, but it actually does seem to make sense. I am presently seeing them as needing to be associated taxa though, but maybe I'm still not picturing it right. I'm imagining that maybe the keywords could be associated taxa much like the object materials are in this EH record: https://arctos.database.museum/guid/UAM:EH:UA99-018-0063

Some example data or further explanation would be useful - I might be way out in left field again - but I'm thinking common names are analogous.

Here are some screenshot examples of the ways other institutions are using variations of keywords as well as the AAT itself in their online collections: https://docs.google.com/document/d/1wbGB3F5cu6P0r1Q5ScEv9xXT18WFktEmH9mzSnLvAw8/edit?usp=sharing

[dustymc]

If keywords are attributes of taxa, then you won't be able to avoid using them consistently. SomeTaxon will or won't be associated with SomeKeyword, and users will be able to move smoothly between them and the stuff that uses them (eg catalog records). If you're consistent in identifications - and that's generally the most fundamental chunk of data we have, so I think most of us are - then you're consistent with your keywords. (And they add no extra work in this structure, which is nice.)

As attributes, I can find things you assigned some keyword to for some reason. That's it. Knowing the ones you could have assigned is probably better than not knowing, but I'm still left hoping you've clicked the same buttons I'd have clicked. Being consistent - avoiding the situation where a user searches "SomeKeyword," finds 3 of your 897,634 "SomeTaxon," assumes they've found everything you have, and leaves - is just not realistic. (And the fix involves somehow finding and editing 897,631 records, versus one under a taxonomy-based approach.)

Associated species is the same situation as attributes, but it's free-text and (should be) associated with an event, not cataloged item. (I think it's still hanging out in an old table because it doesn't get used by the same people who have any complexity in events, but we'll fix that at some point.)

I can't say how any of that lines up with whatever you're trying to do, but this is not an arbitrary decision. If keywords are not attributes of taxa (or if I'm confused in some other way) then specimen-attributes-or-whatever may be a better solution. If they can be normalized via taxonomy (or something else), then they should be.

[Jegelewicz]

I'm imagining that maybe the keywords could be associated taxa much like the object materials are in this EH record:

AHHHH, I see and I think that makes sense - it would mean that the keywords would BE names in the name table and would have associated classifications in the taxonomy table. Then you could associate as many of the keyword "taxon names" with the object identification. Which points out that identification is more than biological taxonomy and the wording there, like specimen elsewhere, may need to change.

I think this goes along with #1732 and that this would be a good issue to place before @ArctosDB/taxonomy

We have put Nomenclature 4.0 off until the beginning of next year. Can we lump this in with it? Also, it would be great if you could present this to the Taxonomy Committee at our next meeting! November 20th at 2PM mountain time in the Arctos Zoom Room.

[marecaguthrie]

Thank you Dusty and Teresa! I'm very excited that you are open to adding something like the Getty ATT to Arctos and I agree that it makes sense to lump it in with Nomenclature 4.0 in January 2020. That will give us time to think this over a bit more. We would be happy to present at the Taxon Committee on the 20th.

What we are trying to achieve with the keywords concept is to meet the Getty's CDWA (Categories for the Descriptions for Works of Art) core standards for documenting the "subject" of a work of art. I love the idea of Arctos being able to "talk" with other art databases the way it does with science databases. Thank you for being so open to including the needs of art collections in Arctos!

On Fri, Nov 8, 2019 at 8:45 AM Teresa Mayfield-Meyer < notifications@github.com> wrote:

I'm imagining that maybe the keywords could be associated taxa much like the object materials are in this EH record:

AHHHH, I see and I think that makes sense - it would mean that the keywords would BE names in the name table and would have associated classifications in the taxonomy table. Then you could associate as many of the keyword "taxon names" with the object identification. Which points out that identification is more than biological taxonomy and the wording there, like specimen elsewhere, may need to change.

I think this goes along with #1732 https://github.com/ArctosDB/arctos/issues/1732 and that this would be a good issue to place before @ArctosDB/taxonomy https://github.com/orgs/ArctosDB/teams/taxonomy

We have put Nomenclature 4.0 off until the beginning of next year. Can we lump this in with it? Also, it would be great if you could present this to the Taxonomy Committee at our next meeting! November 20th at 2PM mountain time in the Arctos Zoom Room.

— 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/2362?email_source=notifications&email_token=AJKSRR5UPX7W552YUA6TTCTQSWQSLA5CNFSM4JKLFTAKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEDS3EGI#issuecomment-551924249, or unsubscribe https://github.com/notifications/unsubscribe-auth/AJKSRR2TQ2VWL5ZV7KE3TFDQSWQSLANCNFSM4JKLFTAA .

-- 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]

@dustymc Mareca gave me access to the UAM:Art collection but I don't see the collection when I am logged in. Any idea why?

[dustymc]

Not sure what that means, but I think they don't have the public collection-stuff turned on yet.

[Jegelewicz]

Yeah - they have everything encumbered? How do we change that and @marecaguthrie do you want to?

[marecaguthrie]

Sorry for the radio silence- I was out sick last week. We are ready to make the collection publicly viewable (while keeping the encumbrances that @krgomez created on Value, Specimen/catalog remarks, and Part attribute location).

I've sent an e-mail to @dustymc and @amgunderson who both helped us hide the collection from public view in December of 2018 to enlist their help removing the collection-wide encumbrances that I asked for their help creating.

@krgomez and I have been working to pull together the documents we said we'd sent to you @Jegelewicz - we are just running a little behind from me being out of the office most of last week.

[Jegelewicz]

No worries! We decided to cancel the December Taxonomy Committee meeting, so we have until January to play around.

[marecaguthrie]

As part of our larger plan (See issue #2478 Planning for Art Collections in Arctos) We have a few updates on this issue:

Updated table to propose: We have an updated our subject terms index and have links to sources. Overall, the terms are mostly from the AAT, some are from Wikipedia, and a couple are from somewhere else.

A recap of what SUBJECT TERMS are: Subject terms (also known as keywords) are a way to categorize what is depicted in and by a work of art. The subject matter of a work of art (sometimes referred to as its content) is the narrative, iconographic, or non-objective meaning conveyed by an abstract or a figurative composition. It is what is depicted in and by a work of art. It also covers the function of an object or architecture that otherwise has no narrative content.

A recap of why it is important: Indexing subject through controlled vocabulary is critical for discoverability. Subject is a Getty CDWA core category, an important best practice, and one of the most common methods by which we and our users will want to search for something in the collection.

A recap of discussions of how to implement it: At this point there are two ways that have been discussed for implementing the use subject terms (identifications and attributes), and maybe there’s a third way that we haven’t thought of?

1. One way is through Identifications. As we understand it, we would draw our subject terms from the same table that we would draw our object/work (“taxonomy”) terms and the subject terms would be associated taxa of the object/work type identification. Maybe there’s another way through Identifications? I still don’t understand exactly how an art and culture taxonomy/AAT and Nomenclature would work in Arctos (See #1732 and #2057), but our object/work types we would use terms exclusively from the AAT and for subject terms we would need to use some from outside the AAT (maybe 85% from AAT - though some of these terms we can maybe get contributed to AAT).
2. The other way is as an attribute. We have talked about how this would probably work best as a hierarchical attribute, but my understanding is that this is not possible before the PG move. A con for using an attribute is something Karinna mentioned above at the beginning of this issue:

   One major drawback is that it doesn’t appear possible to search a controlled vocabulary attribute with the dropdown list of terms. Looks like this issue came up in #1466. By using keywords, we are specifically trying to enable more concise searching for users. It is not ideal to require the user to have to guess what keywords might be available to search. It would be much better to provide the dropdown list that includes all available keywords to search. Is this possible?

[dustymc]

I think I need about a week, which I don't feel I have right now, of deep exploration to say anything very pointed, so more surface comments will have to suffice for now.

First and foremost, please don't ever let any UI drive how you model data. That's just not going to end up in a place where anyone's happy.

Second and not far behind, the most central question here is what these data ARE. Are they more like identification hypotheses which might be backed by near-infinite potentially complex/hierarchical/etc data, eg https://arctos.database.museum/name/Canis%20lupus, or categorical assertions like https://arctos.database.museum/info/ctDocumentation.cfm?table=CTSEX_CDE, or something else entirely?

I'm not sure that stuffing complex data into a pick list is going to be workable, but it's very likely that an appropriate data structure will suggest a good UI. I strongly suggest first focusing on what sort of information these data hold, which is very likely to suggest an appropriate model, which is in turn very likely to suggest an appropriate UI.

[dustymc]

post-AWG call:

• these do seem like Attributes
• suggested path: create a new free-text attribute, be paranoid about how it's used, we'll clean up and formalize with a CT in PG.

[dustymc]

AAT is taxonomy: need to figure out https://github.com/ArctosDB/arctos/issues/2499 so we can add names, then we can look at setting up scripts to pull from getty (into a new classification).

[krgomez]

create a new free-text attribute

We would want this to be controlled entry as opposed to free text. Can we start with the list we compiled for the art collection? https://docs.google.com/spreadsheets/d/12myDP-lv6h5em8kWE-38hag_inHAqqNhRg75p4aUdwU/edit?usp=drive_web&ouid=101757478210691645860

[dustymc]

To be clear, I'm suggesting free-text ONLY as a temporary solution which doesn't require writing the code in two languages, adding it to the migration scripts, testing, etc. Doing all of that now is certainly possible, but I believe that would need prioritized by the AWG.

[krgomez]

Okay sorry I didn't understand that. Free text would work fine as a temporary solution if that's the most practical thing to do currently. There is already a keywords attribute that exists but is currently unused, so maybe we just use that? I don't know if there was consensus at the meeting about what this should be called (subject, subject terms, keywords, art keywords, etc) but it seems like keywords would be fine, in my opinion anyway.

[dustymc]

Perfect - yes please, use that, leave this open, once we get to PG we can finalize the vocabulary, clean up any inconsistencies that have found their way in, create a code table, then close.

I vote for definitely not "art ..." - this seems like something that'll inevitably find its way into other collections.

I'm still slightly suspicious that there might be some useful distinction between "subject" and "keywords" (and maybe some other stuff that could be split out, then dynamically pulled together as "keywords") but I think that conversation can wait until PG as well - it'll probably be easier to have once there's some data to dig around in.

[Jegelewicz]

I say go with keywords - it's generic enough that we can all potentially use it!

[Jegelewicz]

I think we are ready to pick up this issue again and UAM:Art has code table terms defined here.

I plan to set the Category terms as catalog item attributes:

term	definition
agents	Terms describing artworks from the Getty AAT Agents Facet excluding living organisms.
associated_concepts	Terms describing artworks from the Getty AAT Associated Concepts Facet
built_and_natural_environments	Terms describing artworks from the Getty AAT Ojects Facet
disciplines	Terms describing artworks from the Getty AAT Disciplines subset of the Activities Facet
events and activities	Terms describing artworks from the Getty AAT Activities Facet excluding disciplines
materials	Terms describing artworks from the Getty AAT Materials Facet
objects	Terms describing artworks from the Getty AAT Objects Facet NOTE: the taxonomy from Getty AAT comes from this facet
physical attributes	Terms describing artworks from the Getty AAT Physical Attributes Facet
plants_and_animals	Terms describing artworks from the Getty AAT Living Organisms subset of the Agents Facet
regions	I am unable to really define this and I sincerely believe it should come from locality associated with a depiction event
styles and periods	Terms describing artworks from the Getty AAT Styles and Periods Facet
visual_works_by_subject_type	Terms describing artworks from the Getty AAT Visual Works subset of the Objects Facet NOTE: the taxonomy from Getty AAT comes from this facet

and the terms will be controlled vocabulary for each of the categories. This means we are going to need some code tables for the controlled vocabs. I suggest the following:

ctAAT_agents ctAAT_associated_concepts ctAAT_built_and_natural_environments ctAAT_disciplines ctAAT_events and activities ctAAT_materials ctAAT_objects ctAAT_physical attributes ctAAT_plants_and_animals ctAAT_regions ctAAT_styles and periods ctAAT_visual_works_by_subject_type

However, as I went through these, many of them are already recorded elsewhere:

1. All of the "object" terms belong in taxonomy with the rest of the Getty AAT object facet terms
2. regions could be set with locality if the event type was "depiction".
3. plants and animals could be set with identification as long as appropriate common names are applied to the taxonomic names
4. visual works by subject type also belong in taxonomy as they are object facet terms

@krgomez what do you think of moving the items above to their appropriate places? How would we need to facilitate searching those terms for art? Are the category definitions sufficient?

I'd like everyone to ruminate on this for a bit and then I think we should meet up to discuss. @dustymc your thoughts on this are appreciated as well.

Note:

Most of the keywords are Getty AAT terms, but there is a small chunk of the terms we are using that are not currently in the AAT. Same goes for the materials and techniques, where most that we are using are in the AAT but some are not.

This might have implications for how we set up code tables? I have called them "AAT", but that will probably not be entirely true. Are we OK with some non-AAT terms in there?

[dustymc]

thoughts

It's a huge amount of complication that won't be used reliably, which will lead to frustrated users. Users will search on one of the very many isolated attributes, find some stuff, and leave having never found the stuff they actually want because other relevant material didn't get categorized in the same way.

What's the objection to just pulling this information from taxonomy (and maybe adding a single attribute to fill in any gaps, but that will inevitably lead to spotty usage as well)?

[Jegelewicz]

What's the objection to just pulling this information from taxonomy (and maybe adding a single attribute to fill in any gaps, but that will inevitably lead to spotty usage as well)?

I don't really think all of it IS taxonomy? Some of it clearly is, but things like styles and periods or activities don't seem right in taxonomy (the painting shows butchering - how does that fit?). I could perhaps see it in taxonomy if Art collections could have many accepted IDs - perhaps at least one for each facet of the AAT but even these might have to be A x B or even more than that because a painting may depict butchering, dancing and hunting, three different "event and activity" terms.

[dustymc]

periods

Events, if you mean "when it was done" rather than "what [when, whatever] it depicts." (Or maybe that's Events too - "who cares?" is generally the question that leads to the best answer.)

shows butchering

That may in fact be a new Attribute, and I certainly have no objection if that's the best place for those data. That should be discussed by itself first.

Dumping agents, places, taxa, existing attributes, etc. in here is an unavoidable path to lots more work for low-quality data. Maybe we can't entirely avoid (or even recognize) that, and it can be a useful step in figuring out what we should be doing, but I certainly can't recommend purposefully replicating data. That's just advice, not some technical limitation - if we must then it should be very clearly documented so that users (particularly including the next relevant Arctos collection) know what they can expect before wandering off into the minefield.

many accepted IDs

New Issue. Not much problem on the technical side, but we'd have to rethink some UIs.

Depending on the details, A {string} IDs might provide the right path to taxa as well.

[krgomez]

I plan to set the Category terms as catalog item attributes

We definitely don't want to have separate attributes for each of these categories. The purpose of the keywords is to document the subject of an artwork (core CDWA category) and enable better discoverability of artworks in the collection. We need to be able to search this in one place. This had seemed like the best option, but we are totally open to doing this in an entirely different way if this way will not work. If this were a simple code table attribute, we could eliminate the categories and all of the terms could just be alphabetized. There are certainly terms where having a way to use a hierarchy within the attribute would be very helpful, but I don't see those broad categories currently in that spreadsheet to organize the terms as being helpful at all for searching.

We had brought up in the past that if we were going to use the Getty AAT as a taxonomy source in Arctos, we might consider thinking about how this source could also be used for our subject keywords (as well as materials and techniques). Is there some way to do this?

[Jegelewicz]

We definitely don't want to have separate attributes for each of these categories.

is different than my understanding of

There are certainly terms where having a way to use a hierarchy within the attribute would be very helpful,

which is why I started out with the categories. I am open to the idea of a single attribute - "keyword" or something with a controlled vocabulary of ALL of the terms in the Google Sheet, but the first thing we need to sort out is what you consider to be "taxonomy", because a lot of the terms in "keywords" seem like they also belong in "taxonomy". All of the "objects" terms fall into this category and it seems like you will end up duplicating terms if we put those into keywords.

As an example - Boat exists in taxonomy - https://arctos.database.museum/name/Boat If we add "boats" to keywords, we could be saying the same thing twice, however, I could also see how "boats" could be applied to a painting of a boat, so maybe having the terms in both places makes sense, as long as we are very clear about how they are to be used?

[Jegelewicz]

Dumping agents, places, taxa, existing attributes, etc. in here is an unavoidable path to lots more work for low-quality data. Maybe we can't entirely avoid (or even recognize) that, and it can be a useful step in figuring out what we should be doing, but I certainly can't recommend purposefully replicating data. That's just advice, not some technical limitation - if we must then it should be very clearly documented so that users (particularly including the next relevant Arctos collection) know what they can expect before wandering off into the minefield.

See above. Eliminating the "categories" I think would also eliminate some confusion as ""elderly" would be a "keyword" attribute and not designated anywhere as an "agent" or an "age class".

[krgomez]

a lot of the terms in "keywords" seem like they also belong in "taxonomy"

There is definitely a lot of overlap in the terms, but object type (what we are using the taxonomy for right now) is very different from subject. We don't want to confuse what an object physically is with its subject matter.

[dustymc]

subject matter.

AHA! (again, maybe....) That seems a much more useful approach/term than "keywords" (which I'd assume to potentially apply to just about anything).

"subject term"?

[AJLinn]

Collection type UAM:Art

The subject attribute (and I agree Dusty is right here - this a more clear description of what we're talking about and is consistent with DublinCore) should also be made available to EH and Arc as we also have objects that illustrate subjects.

[Jegelewicz]

OK - so we add the attribute:

subject = A topic of the resource. Dublin Core

then we need a new code table:

ctsubject_terms

to which I will add the terms from the Google sheet.

Good?

[dustymc]

subject ctsubject_terms

Nothing overly critical, but these are generally singular and matching - "subject term" using "ctsubject_term" is my tentative vote. That'll also help disambiguate from eg http://test.arctos.database.museum/info/ctDocumentation.cfm?table=ctauthor_role&field=subject when they're all crammed in a streadsheet or something.

[krgomez]

I think "subject terms" works great! I will say that the spreadsheet I made is a work in progress and was created with the UAM:Art collection in mind. Other terms would need to be added for other collections. I see it as a starting place.

[Jegelewicz]

YAY! As soon as @dustymc creates the ctsubject_term code table, I will get this stuff loaded.

[Jegelewicz]

@marecaguthrie It looks like your Arctos accounts have been locked due to inactivity. I am going to send you an unlock email from Arctos - please log in to re-activate your account.

[krgomez]

I have been chipping away at this project and now have a spreadsheet ready to import for subject terms. We are really excited to have the ability to search by subject, as this will substantially increase discoverability of our collection items. The terms used to index the subject matter of works in the UAM:Art collection are from this list shared with you all.

Last we communicated about this issue, you guys seemed good with the attribute “subject term” - hoping that’s still the case. This should be a controlled entry attribute. Will this be possible, or will this need to be free text for the time being? Of course, other collections wishing to document the subject matter of their items with this attribute would also want to use other terms. The terms I have used are almost all from the Getty AAT, but some are from Wikipedia and a few are from elsewhere. The sources of the terminology are indicated in the shared spreadsheet.

Please let me know your thoughts and thanks for your help!

[mkoo]

Following this with interest... keywords as subject_terms would be relevant and potentially useful to the content of images in our Archives (which we already have from CalPhotos) but a lot of our terms are not on the current list (non-Getty terms). Some like 'amplexus' or 'parental care' could potentially be relationships. Others like 'habitat' -- maybe? What about 'blue' (the color)? The subject list has the seasons so that at least fits! Having our current keywords in subject would at least allow for easy discoverability from one field as the Art folks indicate

[Jegelewicz]

YAY! As soon as @dustymc creates the ctsubject_term code table, I will get this stuff loaded.

[krgomez]

Sweet! Thank you Teresa!

[dustymc]

creates the ctsubject_term

Done.

[Jegelewicz]

attribute created and associated with Art and EH collections

[Jegelewicz]

associated subject attribute with subject_term code table

[Jegelewicz]

@dustymc I prepared a file for you to load all of the terms for the subject_term code table. They should all be added for both Art and EH collections - I separated those in the file with a comma. If I should use something else just let me know, otherwise, here is the file - I hope all of the links are formatted properly! If anything looks nuts - just send it back to me.

subject_term.csv](https://github.com/ArctosDB/arctos/files/7737724/subject_term.csv)

[dustymc]

done - I didn't wire this up for colleciton_cde, figured that should be the same for anyone who wants to use this, that can be adjusted if necessary.

[Jegelewicz]

Yeah that makes sense! Closing YAY!

[krgomez]

Thank you Teresa and Dusty! I loaded the over 16k terms and they are almost all there. While bulkloading, I realized there were a few terms that somehow didn't make it to that master list I created and which you used to make the subject code table. I added a second sheet there called 'terms to add' that has five more terms. Also, the terms 'children's art', 'Steller's eider', 'devil's club', and 'Payne's gray (color)' didn't load due to the apostrophes. How do I get around this?

[Jegelewicz]

@krgomez I added the terms on the new sheet. Make sure that your file is saved as csv (UTF8) then the apostrophes should load. If that doesn't work, let me know!

[krgomez]

I did save the file as a csv UTF8, but Arctos still won't load the terms with apostrophes. Also, I misspelled 'potlatches' in the sheet you used for the code table (oops), so that would need to be corrected. Sorry about that. Can you see something wrong with this? subject terms that won't load.csv