Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add pronouns attribute to Person #3272

Open
wants to merge 1 commit into
base: main
Choose a base branch
from

Conversation

shaedrich
Copy link

Fixes #2935 and #2925
Inspired by #1112

@shaedrich
Copy link
Author

One thing that isn't covered: Some people have multiple sets of pronouns or use any of them. While you can use the :Text option, it would be nice to have an array for each set of pronouns. Should there be a :PronounsAny type as well?

@danbri
Copy link
Contributor

danbri commented Mar 1, 2023 via email

@RichardWallis
Copy link
Contributor

Is there a recognised authoritative source for such things (sets of pronouns) that could be referenced or linked to.

@shaedrich
Copy link
Author

shaedrich commented Mar 1, 2023

@RichardWallis I linked https://springfield.edu/gender-pronouns and https://pronoun.is/ for further information in the issue but of course, these are no official sources. Would you consider grammar an authorative source and if not, what would qualify as such?

@danbri I don't even think, gender works in each and every language. But I'm not enough of a linguist to provide you with several languages' culture and characteristics. What kind of information would you want as a precedent? And in case of exceptions from the rule, there's always the :Text type.

@zichy
Copy link

zichy commented Apr 4, 2023

I would love to see this move forward 🙏

@shaedrich
Copy link
Author

@zichy I'm afraid, my last answer was not enough to react to it :/

@WeaverStever
Copy link

I checked with the AP Style to see if they have an authoritative source. They are linking to the following.

GLAAD Media Reference Guide - 11th Edition
https://www.glaad.org/reference

@github-actions
Copy link

github-actions bot commented Jul 6, 2023

This pull request is being nudged due to inactivity.

@github-actions github-actions bot added the no-pr-activity Discuss has gone quiet. Auto-tagging to encourage people to re-engage with the issue (or close it!). label Jul 6, 2023
@shaedrich
Copy link
Author

Well, I answered the questions as good as I could. If that's not enough …

@github-actions github-actions bot removed the no-pr-activity Discuss has gone quiet. Auto-tagging to encourage people to re-engage with the issue (or close it!). label Jul 7, 2023
@github-actions
Copy link

github-actions bot commented Oct 6, 2023

This pull request is being nudged due to inactivity.

@github-actions github-actions bot added the no-pr-activity Discuss has gone quiet. Auto-tagging to encourage people to re-engage with the issue (or close it!). label Oct 6, 2023
@RaineAllDay
Copy link

I'm really confused why there's no movement on this addition. It's a pretty simple addition, recognized sources have been provided for references, and a large majority of websites currently implement these fields on profiles, would be nice to have schema support on this.

@zichy
Copy link

zichy commented Jan 6, 2024

I just realized that GitHub itself does already use itemprop="pronouns".

Screenshot of the pronouns in GitHub

@shaedrich
Copy link
Author

shaedrich commented Jan 6, 2024

After all, it's just a string, and it can say anything—but nice catch though 👍🏻

I checked, and it doesn't seem like schema.org secretly sneaked in this attribute via another PR or something.

@danbri
Copy link
Contributor

danbri commented Jan 6, 2024 via email

@shaedrich
Copy link
Author

what could a reasonable party conclude about me, based on the
proposed new definition?

If there wasn't any value, why would so many websites (including GitHub) implement this as profile information?

Am I communicating something about my gender
identity, or about how I prefer to be written about?

Pronouns might correlate with gender (being the case in most normative settings), however, they might just as well not, especially for non-binary people.

@danbri
Copy link
Contributor

danbri commented Jan 6, 2024

I am certainly not saying there is no value!

when we add new definitions to schema.org, a vocabulary used on many millions of websites, we have some responsibility to try to make it as clear as possible what the markup means; … what it communicates. That said, we have a long tradition of scruffy pragmatic definitions with various kinds of wiggleroom for alternative readings and evolving interpretations. In this current case it seems prudent to be as clear as possible since we touch on matters that are socially very weighty.

@shaedrich
Copy link
Author

So what are we exactly talking about at the current stage?

  • Whether we want to have that attribute or not? → I think, this is a no-brainer
  • The naming of the attribute and/or the possible values → I'm open for suggestions
  • The interpretation of the attribute and its values → if the first point is answered with "yes", this is merely a question for the docs rather than the implementation itself

@danbri
Copy link
Contributor

danbri commented Jan 6, 2024 via email

@shaedrich
Copy link
Author

I can only emphasize this once more:

what could a reasonable party conclude about me, based on the
proposed new definition?

If there wasn't any value, why would so many websites (including GitHub) implement this as profile information?

I just realized that GitHub itself does already use itemprop="pronouns".
Screenshot of the pronouns in GitHub

It's already common practice. So this would merely be catching up with real life.

Let’s start from what needs to be expressed. We also generally seek
implementation commitments from parties who will use (consume, process
etc.) the data in software, services, tools etc.

So, how exactly does this process work and what can we do about it?

@danbri
Copy link
Contributor

danbri commented Jan 6, 2024 via email

@RaineAllDay
Copy link

RaineAllDay commented Jan 6, 2024

Schema.org Proposal

Add "pronouns" property to "Person"

Background

Schema.org's person schema offers a way to identify the gender of an individual, however, one's gender identity does not automatically identify the appropriate pronouns on an individual. In some cases, especially with the increasing use of AI generated content, to be able to identify the pronouns of a person(s) to properly generate programmatic content, as well as generating metadata in search algorithms, etc. This feature would also be useful for any sort of database generation where someone, or some program may be scraping data to populate a database to be used when writing articles about an individual to reference specific data about them. This data would be valuable for storage in many systems including patient data management systems, employee tracking systems, identity providers.

The more we include pronouns into our meta-data for users as we do many other, less important, pieces of information on a "person" the more we can reduce the mistakes of people when referencing other people. As someone who is, themselves, transgender, I can see a world where in the metadata for sending an email within an organization, pronouns may be displayed in the editor to notify the author of the pronouns any person in the organization may use. In a patient data management system it would be a field displayed (and in some already is) to medical providers to ensure they are following standards of care by referring to patients with the appropriate pronouns, as some systems don't currently support this and results in having to hunt the information down.

I believe the debate over the value, or need of the addition of this relatively simple parameter is a bit reductive. The person type has the following properties, callSign, honorificPrefix, honorificSuffix. Why callSign would serve more use than pronouns in the schema would be a pretty interesting take to hear as I imagine very few of schema.org's users actually implement this field.

Implementation

To align with a use for programmatic content generation, as well as a general data collection perspective I propose the creation of a new intangible type. In this context "pronouns" would be a "thing" a person has to describe it in the same sense a person would have an "occupation".

Thing > Intangible > Pronouns:

nominativePronoun

examples: he, she, they, xe

possessivePronoun

examples: his, hers, theirs, xeirs

objectivePronoun

examples: him, her, them, xir

displayPronoun

examples: "he/him/his", "she/her/hers", "they/them/theirs", "xe/xir/xirs"

I personally would leave it up to every individual site to full define how these are filled, I imagine most sites would just use the "displayPronouns" option by default as in GitHub's implementation it's a free-form field with no per-designated options. This offers a way for websites that do offer pre-baked options, to have a more defined schema in the backend when it comes to generating content from these fields (ex: Facebook who generates notification strings based on pronoun fields).

Otherwise, the currently proposed commit would also work, however to keep with the programtic nature that schema wants to enforce, require a string consisting of 1-3 parts separated by "/". This separation format is widely used when identifying pronouns for an individual for easily parsing.

@RaineAllDay
Copy link

I know you've been responding from email, I'd like to note I made some additional updates to my previous message.

@RaineAllDay
Copy link

I also would like to make some additional points. I've seen in previous comments that have been made on this discussion that maybe it shouldn't be added because people may "miss" the subtext due to translation, etc. You used the example of a dating website that became popular in Iran where users misinterpreted a "looking for" field's subtext to include friendship, or business type relationship building. This would not be an error on the part of schema.org, as in a schema sense, "lookingFor" if it existed in the schema could mean many things depending on the website it's being used on. I guess "seeks" could be used in this context, but it would be solely up to each website to implement their own definition of this property. Pronouns, would have a singular use across the board. While some users may use different pronouns based on what website they're using, and the region their in, local laws, etc, the underlying purpose is the same, provide the pronoun definition for an individual.

We are evolving as a society, and with it comes new ways for people to describe and express themselves. Pronouns are used everyday, whether people want to believe it or not, and there is a very real need for a machine-readable field to be able to process this information and store it or distribute it elsewhere. I'm not sure why the argument so far has been a bunch of what ifs, or where the value is. It's beginning to look like schema.org is resistive to implementing the property, but is trying to avoid any negative public opinion on the push-back at the same time. This field is widely used, serves an obvious purpose, and as other fields, would be up to the developer implementing the field into their schema, to figure out the semantics used in their implementation.

I think the biggest thing that got me to comment and participate in the discussion is the fact that these were originally discussed in July and August 2021, a PR was made Feb of 2023, the last comments from contributors was March 1 despite several updates by those proposing this change. Many of the responses to the discussions have been rather defensive and borderline combative. While I can understand the need to gather information given the sensitive nature of the subject I believe it could have been done better if the goal is truly making sure it's done right. But simply asking for an "authoritative" source and then not re-engaging when people provide these resources is not really a great way to show you're trying to "get it right"

@shaedrich
Copy link
Author

Maybe the Solid project would be interested here?

There's only one way to find out. Let's at-mention @solid (not sure, whom to approach here: @VirginiaBalseiro /@csarven /@justinwb /@kjetilk /@KyraAssaad)

The meaning there seems to be “pronouns that some account
holder on a website prefers be used when mentioning (eg in UI elements)
that user to others on that site”.

In this situation it isn’t clear what value using a machine readable data
sharing system like Schema.org adds, since Github are both the publisher
and consumer of the markup. In fact they may not use the markup at all.

Just because something is used in the UI, doesn't mean, it can't also be used in some shape or form of structured data about an entity.

I believe the debate over the value, or need of the addition of this relatively simple parameter is a bit reductive. The person type has the following properties, callSign, honorificPrefix, honorificSuffix. Why callSign would serve more use than pronouns in the schema would be a pretty interesting take to hear as I imagine very few of schema.org's users actually implement this field. […] We are evolving as a society, and with it comes new ways for people to describe and express themselves. Pronouns are used everyday, whether people want to believe it or not, and there is a very real need for a machine-readable field to be able to process this information and store it or distribute it elsewhere. I'm not sure why the argument so far has been a bunch of what ifs, or where the value is. It's beginning to look like schema.org is resistive to implementing the property, but is trying to avoid any negative public opinion on the push-back at the same time. This field is widely used, serves an obvious purpose, and as other fields, would be up to the developer implementing the field into their schema, to figure out the semantics used in their implementation.

My thoughts exactly! Are there any plans to move callSign to a more specific type? Since, contrary to what some conservative people claim very loudly time and again, everyone does have pronouns—something that couldn't be said about a call sign.

@RaineAllDay Thanks for speaking up 👍🏻

nominativePronoun
examples: he, she, they, xe

possessivePronoun
examples: his, hers, theirs, xeirs

objectivePronoun
examples: him, her, them, xir

displayPronoun
examples: "he/him/his", "she/her/hers", "they/them/theirs", "xe/xir/xirs"

Some might even need

reflexivePronoun
examples: himself, herself, themself/themselves

There might be more in other languages.

Interestingly enough, schema.org already has a gender attribute, contributed in #1112, which I referenced as "inspired by"—however, it's more than that. Some people actually stored gender (or actually "sex" according to their definition) as a boolean in their database and didn't offer an option to change this via the UI. Furthermore, some websites started to use services like finnp/node-genderize or markus-perl/gender-api-client to guess the gender by given name. Even though, that might work in a normative world, this is an error-prone approach. The only solution can be self-declaration.

@RaineAllDay
Copy link

If there is opposition to the proposal I prepared from the maintainers, they need to come from a perspective of not matching current standards. As I have not participated in the creation of new schema properties I am not versed in the structure in which the proposal process usually goes. However, the continued discussion and argument over it's value, or need for authoritative sources for values, etc. needs to stop. I am happy to discuss the potential formats, or structure of the data, but it's obviously something that would be useful, unless of course the users of schema.org are only conservative websites that believe that pronouns aren't real, then maybe I can accept the argument that there is no value. Will every schema consumer use the pronoun property? Probably not, but then again, there are many properties not used from any given type, so it's not a valid argument.

This data can serve MANY purposes being shared in a structured way. 99% of the time the data will most likely be used to just display on a profile somewhere. But especially in medical systems which are becoming more and more integrated, ex: insurance data is shared between doctors prescription system and pharmacy, sharing pronoun data is really important, and by having recognized schema orgs include it in their schema, I think it can help continue the progression of implementing these fields into systems that would benefit from consumption.

@shaedrich
Copy link
Author

Second that.

Also, there's much value for research, when grouping data together. Something, that can hardly be said about a call sign.

@kjetilk
Copy link

kjetilk commented Jan 6, 2024

Maybe the Solid project would be interested here?

There's only one way to find out. Let's at-mention @solid (not sure, whom to approach here: @VirginiaBalseiro /@csarven /@justinwb /@kjetilk /@KyraAssaad)

Thanks for the mention.

Indeed, this has been discussed in the Solid context. I was opposed to add it there, but only because I felt that vocab is for technical stuff. I think it would be great to have it in schema.org, as I think that would be not only appropriate for it, but also a good place to have it for use in your Solid profile. I don't have any particular opinions on the details of how it should look in schema.org.

@danbri
Copy link
Contributor

danbri commented Jan 6, 2024 via email

@shaedrich
Copy link
Author

Eg. “pronouns”: “A textual representation of a person’s pronouns. The exact form used varies between sites, cultures, languages and communities.”

Sounds good to me.

3.) Having said that, a fuzzy by design “pronouns” property may have positive user safety characteristics. Eg if leaked a user could appeal to flexibility of definition. […] I would be tempted to add: “It is best to consider this information as a
hint for use when generating text and other information mentioning some
person, rather than as a proxy for other facts about them.” (considering
the issues raised by GitHub I mentioned earlier, and the Orkut/Iran case).

Pro

  • As with gender, there'll be people who don't fit into any common box, so a fuzzy value is needed
  • People often just want to enter one value instead of filling out a whole form just to define themselves
  • The given value can be displayed in profiles and such as is, so there's no debate over rendering implementations

Contra

  • This value could not be used for said notification string placeholders, as this needs to be a certain type of pronoun
  • Different languages might have different types of pronouns → i18n/l10n

So, ideally, we have both, as addressed by @RaineAllDay in #3272 (comment). There's even some implementation details up for debate because of the potential complexity: #2925 (comment)

Will every schema consumer use the pronoun property? Probably not […]

The point here is, that of course, not everyone will use it at all and there'll be preferences for one of the two options (one attribute vs multiple), respectively. As schema.org is a generalized/abstracted vocabulary to be widely applicable, it shouldn't necessarily be schema.org's task to decide for other sites, how to implement it. It's like when a site offers "basic settings" and "advanced settings"—they don't automatically contradict the need for the other.

@RaineAllDay
Copy link

2.) I am not demanding official lists of values. Rather I am concerned that we need to be clear about the meaning of the property.

The pronouns definition would be far easier to define than the currently existing "gender" definition. Gender has a wider spectrum of meaning depending on many factors. The current implementation refers to values of "biological sex", however, general social standards creates a clear segmentation of the two, and unfortunately this has not been helped by academia's interchangeable use of these terms.

(considering the issues raised by GitHub I mentioned earlier, and the Orkut/Iran case).

I'm really not sure why you keep bringing this up. Transgender individuals, or individuals who may use pronouns that are not directly directive from their gender or biological sex, are very aware of the safety concerns that face us, we typically know when it is safe or unsafe to use the pronouns we best identify with. From the sounds of the Iran thing, you had individuals who were heterosexual marking "both" who then may have been exposed to persecution from their government. This is a case of people unknowingly selecting options that may present a representation that could cause harm. I don't believe a pronoun property to offer the same misunderstandings, and potential conflicts as described in this scenario.

Please assume good faith here.

It is hard to assume good faith when this topic has been repetitively brought up and abandoned. As I've looked back through the previous discussions, the implementation has been delayed by asking for information, when information is provided no response is given further to explore. There is no discussion going on here. Just weird reasons why you feel it shouldn't exist, or the need for further definition despite the fact that generally acceptable definitions are provided. It all actively feels resistive, not exploitative. It may even just best be handled by having a free-text field as described in the IETF post. While this would limit the use of automated systems parsing the information, in a majority of the online space and social spaces using English as the primary and collaborative language it is understood what "pronouns" means, and how to use them, and what the format is.

I will be happy to entertain a discussion on further exploring the best way to describe the metadata. There are several organizations I'm sure we could reach out to for input. In the United States, GLADD is a pretty standard resource. WPATH, an NGO responsible for creating the Standards of Care (US Standard for Trans healthcare, currently being reviewed for adoption by the NHS) could be another resource. I'm happy to sort through my catalog of references for gender identity, etc to provide academic level resources on the matter. My biggest issue is the consistent abandonment of the discussion by the maintainers, and consistent deflection. If you're not willing to do the leg work to satisfy questions you may have to make sure you going the right direction, please let me know what information I can provide to satisfy any concerns or questions. I am transgender myself, I am decently involved within the community, and often spend some free time reading through research papers to better understand the science and articulate how science agrees with/supports the transgender community.

I can understand Solid's reservations on adding it, as from what I could tell, they don't have many other terms specifically describing a person. Schema.org however, that's literally the whole point of the Person type.

@danbri
Copy link
Contributor

danbri commented Jan 6, 2024 via email

@shaedrich
Copy link
Author

He'll probably be glad to hear back once we've reached a consensus.

@RaineAllDay
Copy link

The point about Solid was that TimBL enquired informally about adding such a property to schema.org or foaf a while back, but we didnt follow up the conversation yet

I was simply referencing that I had researched their discussion on the matter. In their case, it does make sense that it shouldn't live in their standard as they are not actively holding other information that would be categorically similar.

@shaedrich
Copy link
Author

We just got another suggestion: #2925 (reply in thread)

@RaineAllDay
Copy link

@danbri Can we get some feedback on what the plan is moving forward? Once again the discussion has pretty much halted from your side.

@danbri
Copy link
Contributor

danbri commented Jan 31, 2024 via email

@shaedrich
Copy link
Author

until someone steps up to say “we plan to build something using this data”?

In the spirit of asking questions:
Who is considered the authority here: Programmers who use schema.org or schema.org itself?

Should we just start with unstructured “pronoun-oriented” text

That would at least be some kind of start 👍🏻

@danbri
Copy link
Contributor

danbri commented Jan 31, 2024 via email

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
no-pr-activity Discuss has gone quiet. Auto-tagging to encourage people to re-engage with the issue (or close it!).
Projects
None yet
Development

Successfully merging this pull request may close these issues.

Add pronouns attribute to Person
7 participants