Wikidata:Property proposal/sex

From Wikidata
Jump to navigation Jump to search

sex[edit]

Return to Wikidata:Property proposal/Person

   Under discussion
DescriptionBiological sex of an animal or person
Representssex (Q290)
Data typeItem
Domainhuman (Q5) or animal (Q729)
Allowed valuesmale (Q6581097), female (Q6581072), intersex (Q1097630), male organism (Q44148), female organism (Q43445)
Example 1Kitty Anderson (Q59160028)intersex (Q1097630)
Example 2Dolly (Q171433)female organism (Q43445)
Example 3Alex (Q24628)male organism (Q44148)
Planned useThis property would be used for all individual animals with a known sex and for individual humans when their sex is notable (see Motivation below). Note that the existing sex or gender (P21) is being changed to just "gender" per consensus.
Robot and gadget jobsa bot will be written to migrate male organism (Q44148) and female organism (Q43445) claims from sex or gender (P21).
See alsosex or gender (P21) (which is being changed to just "gender"), property that may violate privacy (Q44601380)

Motivation[edit]

After many years of debate and discussion, consensus was reached to split sex or gender (P21) into two separate properties. sex or gender (P21) will be renamed "gender" and a new "sex" property will be created to handle animal sex and the rare cases where a human's sex is notable separately from their gender. For example, Kitty Anderson (Q59160028) is a notable intersex (Q1097630) activist, but her gender is female (Q6581072). Previously, sex or gender (P21) had to be overloaded for both uses. In some previous discussions, concerns were raised about whether or not certain languages (especially Japanese and Chinese) had separate words for "sex" and "gender". It has been confirmed that both of those languages have separate words for the two concepts, and even if a language doesn't have separate words, they can still be disambiguated by the description. Kaldari (talk) 18:49, 6 December 2019 (UTC)

Discussion[edit]

  • A discussion from 2013 closed in 2019? I wonder if any of the participant are still active in Wikidata .. --- Jura 19:13, 6 December 2019 (UTC)
  • Symbol oppose vote.svg Oppose use sex or gender (P21) with qualifiers. --- Jura 19:13, 6 December 2019 (UTC)
  • Actually, there are some benefits to split this unrelated to the explanation given above. --- Jura 01:07, 10 December 2019 (UTC)
  • BA candidate.svg Weak oppose. On one hand it always made intuitive sense to me to split it for the sake of modularity and precision. On the other hand, it may become a "fun" battleground for people who will insist on setting the sex value of transgender people to what they were assigned at birth, and I'm not quite sure it's totally good. It may work well if a very clear and strongly enforced policy is defined for dealing with such cases in a way that presents precise information and is respectful to the subjects of the items. I'm not sure what this policy should actually be, though. I haven't seen a description of such a policy in the current proposals and discussions, but maybe I haven't searched well. --Amir E. Aharoni (talk) 20:55, 6 December 2019 (UTC)
    I think there should be a policy, that it should only be added to a person if and only if they have stated it themself publicly and unambiguously, otherwise we should only indicate gender. Robin van der Vliet (talk) (contribs) 15:23, 7 December 2019 (UTC)
    Yeah, that could be a possibility. As I said, the distinction generally makes sense, and I'll be happy to support it if there's a policy that will prevent the misuse of this property in unhelpful and transphobic ways. --Amir E. Aharoni (talk) 20:16, 7 December 2019 (UTC)
  • In my opinion we may keep current P21 for "(biological) sex" and create a new property "gender identity".--GZWDer (talk) 07:56, 7 December 2019 (UTC)
    • I think it makes more sense to do it the other way around, as the currently existing property is mostly used for persons to indicate their gender. Very few uses are for non-human animals in comparison to humans. Robin van der Vliet (talk) (contribs) 15:23, 7 December 2019 (UTC)
      • Agree with Robin, the current uses of P21 are mostly to indicate people's gender, so it would make the most sense to leave that one as the gender property. Kaldari (talk) 22:39, 7 December 2019 (UTC)
        Fully agree, P21 must be gender and not sex. --Lucas Werkmeister (talk) 00:55, 10 December 2019 (UTC)
  • GA candidate.svg Weak support I think it makes sense to seperate gender from sex, but there should be a proper policy to avoid harassment of trans people. Robin van der Vliet (talk) (contribs) 15:23, 7 December 2019 (UTC)
  • Symbol support vote.svg Support This distinction is needed. Amir (talk) 19:01, 7 December 2019 (UTC)
  • Symbol support vote.svg Support subject to the policy statements related on privacy of individuals stated in Wikidata:Living people. John Samuel (talk) 18:16, 8 December 2019 (UTC)
  • Symbol oppose vote.svg Oppose. I oppose if the proposal means that a human's sex will be classified as "gender." "Sex" and "gender" are sometimes distinguished for valid reasons, and I don't think distinguishing them should only apply to "to handl[ing] animal sex and the rare cases where a human's sex is notable separately from their gender." Flyer22 Reborn (talk) 22:16, 8 December 2019 (UTC)
    A human's gender will be classified as "gender". A human's sex will be classified as "sex", but if and only if it's publicly known and relevant, and in compliance with Wikidata:Living people. Robin van der Vliet (talk) (contribs) 00:14, 9 December 2019 (UTC)
    Kind of, but what concerns me is how do you define "relevant" in a way that won't be abused. The current situation at Sam Bettens (Q274154), for example, doesn't look optimal. It currently says:
    sex or gender:
    1. female / end time: 18 May 2019
    2. transgender male / start time: 18 May 2019; nature of statement: coming out
    I'm not transgender, so please do tell me if I'm saying something wrong and dumb, but from what I heard, transgender people usually feel that they've always been their gender, so saying that Sam Bettens was female until 18 May 2019 is probably not entirely correct.
    Should this be changed? Probably. Saying that he is male is probably correct, and saying that he's transgender is probably correct and relevant, too. Mentioning the coming out date is probably relevant, too.
    But here's the most important part for this discussion: Is it correct and relevant that his sex is female? I honestly don't know, but I suspect that it may not be. I'd love to have something better, but I'm afraid that without very, very clear policy this can be misused.
    Maybe for humans we can have a policy that only allows specifying gender and an indication of being transgender, we use "sex" only for non-humans, and statements that don't conform to this are speedily deleted. But there may be problems with this approach, too. Again, I just don't quite know. --Amir E. Aharoni (talk) 13:08, 9 December 2019 (UTC)
    I think some of this is overthinking the issue. The most important point is that sex is rarely a relevant attribute for a human, whereas gender is. That is true of transgender people as well. And, rather than attempting to generalize at all about how transgender people work, we should instead rely on the subject's own self-identification, or reputable sources. But the current usage of P21 has been to signify the subject's gender, not sex, and there is no suggestion that once this property is implemented we would need to go in and determine everyone's sex as well. Just because most humans have a sex does not mean Wikidata needs to concern itself with describing that for every subject—just like we do not describe all of their street addresses even though there is such a property. The only relevant property for almost all humans should be gender. Sex should only be appropriate in exceedingly rare cases, and that can be explained in the property's usage guidelines, or even enforced with a constraint. Dominic (talk) 14:48, 9 December 2019 (UTC)
    The most important point is that sex is rarely a relevant attribute for a human - this.
    there is no suggestion that once this property is implemented we would need to go in and determine everyone's sex as well - this, too. However, I suspect that some people will start saying that we do need to determine everyone's sex. They will be wrong, and this should be prevented as early as possible by policies.
    Sex should only be appropriate in exceedingly rare cases - yes, and maybe even never. This should be defined. --Amir E. Aharoni (talk) 15:30, 9 December 2019 (UTC)
    Does this answer your question Amir? Amir (talk) 17:05, 9 December 2019 (UTC)
    User:Ladsgroup - Yeah, mostly, thanks. It's not totally explicit there, but as far as I can see, it suggests using "sex" for non-humans and for intersex people. This makes sense, although maybe it should be more explicit.
    Can anyone think of reasons to use "sex" for non-intersex humans? Whatever the answer is, it should probably be explicit. --Amir E. Aharoni (talk) 17:22, 9 December 2019 (UTC)
    "The most important point is that sex is rarely a relevant attribute for a human." Not true, as is clear by health aspects such as sex differences in medicine. Regardless of how one identifies, sex (rather than gender) differences matter when it comes to medicine. As for "most humans have a sex", what humans do not have a sex? Flyer22 Reborn (talk) 07:40, 10 December 2019 (UTC)
    This is relevant for a private conversation with a person's own doctor and family, and not so much on this public website.
    If a person's being transgender is verifiable and relevant as public information, it can be stated in another appropriate property, as proposed in Wikidata:WikiProject_LGBT/gender. --Amir E. Aharoni (talk) 08:38, 10 December 2019 (UTC)
  • Symbol support vote.svg Support At Wikidata:WikiProject LGBT there is a directory of discussions at Wikidata:WikiProject LGBT/gender which demonstrate the ongoing difficulty of having a property for sex or gender. I support the split and also I support this proposed method for the split. I sympathize with the oppose votes who wish to avoid having sex and gender being repeated fields for biographies. Over years of discussion this has been a sticking point, and what is new right now is that our community has matured to the point of having enough people be able to have conversations about property proposals on Wikidata. I think now is the right time to make the switch, and even a few months ago would have been premature. The most challenging part of this change is not a Wikidata property proposal, but coming to understand that by doing this we are changing our language such that the terms "sex" and "gender" are going to have specific meanings in the Wiki community drawing from LGBT+ discussions on the topic. I encourage anyone to have conversations anywhere, but if anyone sees a conversation, please consider listing it at that LGBT+ page with the others. Also edit the LGBT+ gender page to develop the guideline. Blue Rasberry (talk) 14:42, 9 December 2019 (UTC)
I am thinking more about how we are going to assign either sex or gender to people, and when this will be original research. We have consensus to do gender assignments as original research on almost every human. We will do this based on some evaluation, presumably including whether their name is gendered, and their physical appearance often based on one photo, and then in much rarer cases when there is additional context, like the sort of information which would go into a Wikipedia article.
I am sure that I support a split of sex and gender. Sex seems natural to have; this is the property for marking what external sex organ a person has. We usually will not have that information, and instead probably will label most biographies by gender presentation which we determine by original research. Hmmm! Blue Rasberry (talk) 14:29, 11 December 2019 (UTC)
Yep, you get it. As a non-binary trans person I've been dealing with these issues in daily life as well as online for years. It gives me a real headache but it is important to get it right. Funcrunch (talk) 15:32, 11 December 2019 (UTC)
  • Symbol support vote.svg Support from a purely semantic position, these are two different concepts, which require two different properties to meaningfully describe them. This is fairly obvious with some simple thought experiments, such as "Are there any subjects where the values would not be the same?" or "Are there subjects where one property would be appropriate, but the other would not be?". As a general rule, we do not combine distinct attributes into a single property and then distinguish which one is meant with a qualifier. So, something must be done about it.

    I would also point out that the idea that sex and gender can be represented as a single property means Wikidata is itself using fairly loaded language on a huge number of items, including sensitive ones. We have a practice of referring to certain people as, for example "transgender female," but do not generally identify people as "cisgender male". A person's gender should be reflected in a property with a value consisting only of their gender—regardless of what that is—without reference to their genitalia or sex. The fact that we only take this approach of conflating distinct concepts with a single property for sex/gender is troubling. To me, this seems important not only out of respect for the subjects we describe, but also to be a welcoming editor community for all who would want to participate and could be rightfully put off by how we talk about gender currently. Dominic (talk) 14:48, 9 December 2019 (UTC)

  • Symbol support vote.svg Support per Dominic, though as a trans person I am concerned as others have mentioned that some editors will still unnecessarily add differing "sex" attributes to trans people, insisting that our birth-assigned sexes be listed even though our genders differ from these. ETA: Most trans people prefer to avoid the term "biological sex", which is often used to invalidate our genders. Though "assigned sex" might not be appropriate in the case of non-human animals, another term should be considered for the description of this property. Funcrunch (talk) 15:35, 9 December 2019 (UTC)
  • Symbol oppose vote.svg Oppose Too late. sex or gender (P21) is already used in hundreds of templates in 100+ languages, changing anything now is just a very bad idea.--Jklamo (talk) 16:10, 9 December 2019 (UTC)
    That's not a valid reason, We deprecated a property that was used in millions of places (P106 IIRC). This is definitely smaller change than that one. Amir (talk) 16:58, 9 December 2019 (UTC)
    @Jklamo: Only about 10,000 existing uses of sex or gender (P21) would need to be migrated. Kaldari (talk) 17:07, 9 December 2019 (UTC)
    It is. Are going to go through all templates using sex or gender (P21) (in 100+ languages) and distinguish if it is appropriate to use "genderfied" sex or gender (P21) on new property (even for languages, that do not differentiate gender and sex at all)? Widely used properties need to be as stable as possible.--Jklamo (talk) 13:36, 10 December 2019 (UTC)
  • Symbol support vote.svg Support per Dominic; this would correct a very visible departure from the general "items and properties should not conflate distinct entities" ethos. Also, this would make it easier for tools to accidentally do the right thing with regards to gender, handling transgender females as females and transgender males as males unless there was some need to specifically query sex. Vahurzpu (talk) 16:47, 9 December 2019 (UTC)
  • Symbol support vote.svg Support Sex and gender are different concepts, so we should have properties for both. MBH (talk) 16:53, 9 December 2019 (UTC)
  • Symbol oppose vote.svg Oppose So is the purpose of this to give harassers an avenue to call trans women "male organisms", or is that just an acceptable side effect? There is no reason to distinguish "biological sex" and "gender" for people unless you think documenting people's genitalia or chromosomes is relevant and not an extreme invasion of privacy. Please don't respond to this if it's just to tell me "but you are a male organism".--Alexandra IDVtalk 16:58, 9 December 2019 (UTC)
    User:Alexandra IDV, this is my concern as well, but see User:Ladsgroup's response above. It mostly addresses this problem, although I think it should be more explicit. --Amir E. Aharoni (talk) 17:22, 9 December 2019 (UTC)
  • "consensus was reached" That's not how consensus works. I'm surprised to see such an attempt. Hence oppose, at least until there is a meaningful discussion on the matter. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:03, 9 December 2019 (UTC)
  • GA candidate.svg Weak support But biological gender need to be at birth, and gender must be current.46.188.23.100 17:19, 9 December 2019 (UTC)
    What is “biological gender” supposed to be? Can you clarify what you mean? --Lucas Werkmeister (talk) 00:55, 10 December 2019 (UTC)
  • Symbol support vote.svg Support They are different concepts. As someone said above, following the recommendations laid here Wikidata:WikiProject LGBT/gender would be important. Scann (talk) 19:00, 9 December 2019 (UTC)
  • Symbol oppose vote.svg Oppose This concept is to complex to be described with such tools as our properties. To describe all the shades and tints of this, you have to use text. So no, out of scope! 62 etc (talk) 19:04, 9 December 2019 (UTC)
    Certainly it’s complex, but we can try to do the best we can with the tools we have. Do you think a split into separate “gender” and “sex” properties would be worse than the current situation where they’re conflated in a single “sex or gender” property? --Lucas Werkmeister (talk) 00:55, 10 December 2019 (UTC)
    "the best we can" will not even be close to good, as soon as we go outside the non-binary. Not even the biological thing is easy to describe here as soon as we reach outside of the binary. But in those cases we can at least describe them in terms of medical conditions. 62 etc (talk) 07:38, 11 December 2019 (UTC)
  • Symbol support vote.svg Support for split. sex or gender (P21) should stay for distingusih, if object is male or female sex (Q290) (at birth) - majority of uses. The new property should be for storing information, that someone feels as female, but according chromosomes is male, somebody is transsexual and other weird uses (gender (Q48277) (minority of uses). JAn Dudík (talk) 21:46, 9 December 2019 (UTC)
    What do you mean by “majority of uses”? All of the wiki projects I’m aware of primarily refer to trans and nonbinary people by their gender, not their sex. (Also, in the spirit of assuming good faith, I should let you know that your comment reads as very transphobic to me, whether intentional or not, and if that was not the intention I suggest you clarify it and avoid phrases like weird uses.) --Lucas Werkmeister (talk) 00:55, 10 December 2019 (UTC)
  • Comment - while we need to acknowledge "those" types of people, at this point I feel as if they are still debating amongst themselves with what the proper terminology should be. Therefore it would be premature for institutions to make changes. Quakewoody (talk)
After reading continual comments from others, I can't help but want to repeat myself - at this point I feel as if they are still debating amongst themselves with what the proper terminology should be. Therefore it would be premature for institutions to make changes. And I can't help but point out 2 things.
  1. A major US city decided on 31 pronouns which angered the activists who wanted at least 101.
  2. The entire point of being "fluid" instead of being binary is not just to prevent falling into a "category" (which exactly what we are trying to do here, forcing them to be a label), but it is also changeable at a moment's notice - like a river, the river itself will always be there but it will never be the same water because it is constantly changing, aka fluid
So, for us, what exactly are we trying to do? Label something that doesn't want to be labeled. Name something that can be renamed before we even finish naming it. Quakewoody (talk)
@Quakewoody: regarding that first point… would you mind clarifying how it’s at all relevant to this proposal? And which city is this supposed to have been – do you have a source for this story? --Lucas Werkmeister (talk) 22:35, 10 December 2019 (UTC)
  • Pictogram voting comment.svg Comment I’m confused by the Motivation section of this proposal. It correctly notes that there have been many years and debate and discussion, which I’ve sadly not been very active in (though I believe it picked up around Wikimania this year?) – but then why does it only link to a discussion that had no new comments since 2013? Other people here have already linked to Wikidata:WikiProject LGBT/gender – is there a reason why this was not mentioned in the proposal from the beginning? --Lucas Werkmeister (talk) 00:55, 10 December 2019 (UTC)
    • Mainly just because that page wasn't as fleshed out at the time. Kaldari (talk) 22:05, 15 January 2020 (UTC)
  • Pictogram voting question.svg Question How would trans people be modeled once this property exists and Property:P21 has been adjusted accordingly? Would a trans woman have gender female (Q6581072) and sex male (Q6581097)? Wikidata:WikiProject LGBT/gender assumes the existence of a “transgender status” property, but I haven’t seen a property proposal for that yet.
More specifically, I’m not convinced that using “sex” in that way would be correct. I believe some people argue that “biological sex”, as a collection of multiple features (not limited to chromosomes), also changes over the course of transition: hormone replacement therapy may change a person’s hormonal balance and, over time, their secondary sex characteristics (e. g. trans women growing breasts and trans men’s voice dropping), and gender confirmation surgery may alter their primary or secondary sex organs as well. The current form of this proposal doesn’t make it clear to me how (or whether at all) this would be represented. (Changing the property to “birth sex” would resolve that question, but would fail to account for intersex people who only discover their condition later in life, I think.) --Lucas Werkmeister (talk) 00:55, 10 December 2019 (UTC) (Same disclaimer as Amir E. Aharoni above – please let me know if I’ve written something stupid or insensitive, that was not my intention.)
@Lucas Werkmeister: Those are great questions. The currently proposed data model at Wikidata:WikiProject LGBT/gender (which this proposal is related to), suggests that a trans woman would have gender female (Q6581072), but no "sex" property set unless for some reason it was notable and verifiable. I admit that there is some uncertainty about how that would actually play out, and we will need to create some guidelines around it to make sure that it isn't used in ways that are disrespectful to trans folks. Also, you are correct that there is not yet a proposal for a "transgender status" property, mainly just because it's easier to manage one proposal at a time, but anyone is welcome to make such a proposal at any time. My hope is that if we can introduce a "transgender status" property, people will use that rather than trying to use the "sex" property to indicate that someone is transgender. Hope that answers some of your questions, and feel free to propose improvements to the data model. Any constructive input is welcome. Kaldari (talk) 23:01, 15 January 2020 (UTC)
  • GA candidate.svg Weak support since this is a generally sensible proposal, but I’d like to upgrade it to {{Support}} or {{Strong support}} once my questions above have been answered. --Lucas Werkmeister (talk) 00:55, 10 December 2019 (UTC)
  • Symbol strong support vote.svg Strong support new property “sex”, but Symbol oppose vote.svg Oppose renaming of sex or gender (P21). Pictogram voting comment.svg Comment A separate, new property “gender” should be created, with suggested uses in something like the following fashion:
    sex or gender (P21) to be kept or used for statements stemming from cultural contexts where there is or was no clear distinction between sex and gender in the modern-day anglophone Western sense (if there are, or should arise, other sex-and-gender-related properties that are more culturally relevant on a case-by-case basis, these could of course be used as well); as well as in the case of unreferenced statements that make it impossible to discern whether sex or gender was intended, or where references exist but have not (yet) been followed up to confirm.
    The new “gender” property to be used in contexts and with intents that clearly distinguish between sex and gender in the senses implied here.
    BlaueBlüte (talk) 08:37, 10 December 2019 (UTC)
  • Symbol oppose vote.svg Oppose My language (swedish) does not have separate words for "sex" and "gender", both are "kön" (unless "sex" is a verb, then it is "sex" in swedish too). We get along just fine without any distinction between the two. /ℇsquilo 12:11, 10 December 2019 (UTC)
    In Russian it's unclear too. "Пол" (sex) and "гендер" (gender) are believed to be synonymic. As the term of "гендер" is actively pushing by feminist activists as the replacement for "пол", "гендер" is sometimes perceived with some irony (I'm not sharing this). The language problem is important. Not all not non-English-speaking users are using Wikidata with their local languages, but some people do, and may have a confusion. --Wolverène (talk) 13:18, 10 December 2019 (UTC)
    Sex and gender are synonymic in every language, because for the many people, gender is how you represent your sex. The term ‘гендер’ is perfectly valid term in Russian social science to describe everything that ‘пол’ can’t. Transphobia associated with it in colloquial usage is nothing to be concerned about in Wikidata or any other encyclopaedic project. I agree that adding ‘пол’ (sex) property might be really bad if it is allowed for usage for most humans, however. stjn[ru] 16:33, 10 December 2019 (UTC)
    @Esquilo: I think we in Swedish have "kön" and "könsidentitet". But I doubt there is a 1:1-fit to the English words sex and gender here. 62 etc (talk) 07:33, 11 December 2019 (UTC)
    Judging by the description of "sex" and "gender" in English, the distinction is totally different. If we introduce this property as it is described, it will have to replace sex or gender (P21) in 99,99% of all objects where it is used. /ℇsquilo 07:39, 11 December 2019 (UTC)
    There is a Russian word that is forgotten in this discussion, and it is relevant for other languages, too. In English, "gender" has been used for a long time mostly for grammatical gender—a linguistic property that is sometimes, but not always, related to biological sex. The Russian language has a separate word for it: род.
    In recent decades, "gender" in English began to be used for the social and psychological concept of gender, which may be different from what a person is assigned at birth according to how their body looks like. This word is also used as a loanword neologism in some languages, including Russian, and then it may be associated with a certain narrative that not everyone understands, and with which not everyone agrees.
    In practice, I could find two actual uses for the sex/gender property, at least as far as it concerns humans:
    1. Using the correct word according to the gender, if the language requires is. This may apply to English, too, for example actor/actress, and in many other languages it is needed in many more words. This is totally related to the grammatical gender sense.
    2. Running queries for people who appear on Wikidata and have a certain occupation, live in a certain country, have or don't have a Wikipedia article, and sorting them by gender. An often-repeated example is "What is the biggest city in the world the mayor of which is a woman?". This is useful for statistics, editathons, verifying data integrity, etc. Here, a human's sex is also irrelevant, and a gender-only propoerty can do this job perfectly. If someone wants to look up transgender mayors, this can be covered by another property, as already proposed in Wikidata:WikiProject_LGBT/gender. (It makes sense to me to also cover intersex people in a separate property, so neither gender nor sex would be used for that, although I might be wrong.)
    There may be other use cases, but I cannot think of any at the moment.
    So, one solution to translate the name of the property unambigusly can be to use the word that is used for describing grammatical gender.
    For languages in which this doesn't work well, perhaps gender could be translated as something like "human sex" or "social sex" and "sex" could be something like "animal sex" or "biological sex". However, this would further strengthen the requirement to not use "sex" for humans at all. --Amir E. Aharoni (talk) 11:06, 11 December 2019 (UTC)
  • Symbol oppose vote oversat.svg Strong oppose if it will be something that would be applied to everyone, because it will just be a lot of WD:BLP violations, an area where Wikidata is already very lax right now and the open invitation to bigots wouldn’t discourage it one bit. Support the renaming of ‘sex or gender’ into ‘gender’ and splitting off a separate ‘sex’ property for animals. As for problems when dealing with intersex people, I think the more kind solution is to do something like gender: female / biological trait: intersex (and maybe even gender: female / biological trait: transgender) or use the same schema trans people currently use. stjn[ru] 16:33, 10 December 2019 (UTC)
  • Symbol oppose vote.svg Oppose It's not clear how do we define "sex" for someone without access to their medical records. What if they are intersex but aren't aware of that? People are not routinely karyotyped. If we become suddenly very interested in whether a person is intersex or not, every male confessing to have hypospadias will be assigned "intersex" sex without any practical reason whatsoever etc. It is also unclear why do we need to state the sex altogether since Wikipedia is not a medical data library. Le Loy (talk) 20:10, 10 December 2019 (UTC)
  • Wikidata is partly a medical data library. We host a lot of data that's of interest to the medical community. Wikidata generally works in a way to allow different people to use it for different purposes. ChristianKl❫ 11:17, 11 December 2019 (UTC)
What if they are intersex but aren’t aware of that? Then we have no statement. This property is, per the motivation section, for cases where the sex is notably different from the gender, which to me presupposes that the two are known to differ. If someone goes through their entire life without being aware of being intersex, I’d say their sex isn’t notable and we don’t need a statement for it. As for hypospadias, I’m not familiar with the condition, but enwiki says the presence of hypospadias alone is not enough to classify a person as intersex, so I don’t see why that would constitute a “sex” statement either. (I suppose it would be a medical condition (P1050), if publicly known, notable and not in conflict with WD:BLP.) --Lucas Werkmeister (talk) 22:46, 10 December 2019 (UTC)

Pictogram voting comment.svg Comment I am afraid, there is one big misunderstanding. In many languages is primary meaning of sex or gender (P21) in biological sense (male, female). And the word gender have the psycho-sociological meaning (trans- and others). English maybe now prefer word gender for the first case, but majority of non-english speakers means this word for second case.

It seems to me, there is support for splitting this property, but P21 should stay binary (and biological) (male, female, hermaphrodite) and the new one should be for other cases (trans*, neutral, intersex, genderqueer, etc..) Do you agree? JAn Dudík (talk) 08:49, 11 December 2019 (UTC)

  • Most Western languages do mean gender in the social sense. Whether you use "he" or "she" to refer to a person is about how you interact with them in a sociological sense. It's important that a data user like Siri can deduce the gender of a person to use the right grammatical forms when speaking about a person. This is an important usecase that any solution shouldn't break. ChristianKl❫ 11:29, 11 December 2019 (UTC)
  • Symbol oppose vote.svg Oppose This is clearly a well-intentioned proposal, but I think creating a mechanism to say that someone has sex A but gender B is actually a mistake and a step backward. This proposal fails the most important criterion for a a wiki system, it is not idiot-proof (Q12981895) and would invite abuse. See a much more amenable propsal at Property talk:P21#New proposal, which is I think 80% of where we want to get to.--Pharos (talk) 04:34, 13 December 2019 (UTC)
  • Symbol oppose vote.svg Oppose Same meaning in too many Asian languages, if there are some little languages who consider both as different, better to use qualifiers to handle. --Liuxinyu970226 (talk) 04:49, 13 December 2019 (UTC)
  • Symbol strong support vote.svg Strong support We can't stop to add data because some people thinks that it could be subject of harassment, or because some languages don't have right words for describing the property. --Tinker Bell 04:18, 14 December 2019 (UTC)
  • BA candidate.svg Weak oppose in favour of several more specific properties or community-recommended property-qualifier combination patterns, and possible restriction to nonhumans as discussed above. Both humans and nonhumans have many measurements that collectively constitute sex, and these can contradict or even have little to do with one another (karyotypes were mentioned, but there are also sex-linked genes on the autosomal chromosomes for which sequencing or genotyping is the test, and there are macroanatomical measures as well such as the configuration of genitalia). I concur with the sentiment mentioned that ultimately the number of entities for which any of those measurements would be useful/appropriate for publication is limited (which is not in and of itself a reason to deny property creation, since that presents an opportunity for exhaustive documentation; but is important to keep in mind during discussion). Arlo Barnes (talk) 22:12, 30 December 2019 (UTC)
  • Symbol oppose vote.svg Oppose. While we need to have a certain level of cultural sensitivity towards those cultures which place priority on gender identity and consider biological sex to be a private and mostly irrelevant matter, we can't reasonably duplicate the entire P21 dataset of millions of statements, nor can we just prohibit all data on gender identity or otherwise limit our work to one cultural viewpoint. We can't just take the approach of blood type (P1853) (where part of the world considers it a basic fact about the person, and the rest consider it a private medical matter). Situations where the proposed property would add differences are sufficiently unusual that which we're able to handle with one property without difficulty. The close of a small six-year-old thread does not make consensus, or override all the much larger discussions on the topic since. --Yair rand (talk) 12:49, 31 December 2019 (UTC)
    • @Yair rand: There would not be a need to duplicate the P21 dataset. P21 would simply be renamed "gender", and "sex" would only be used in rare circumstances (notable individual animals, intersex persons, etc.). What larger discussions on this topic would this proposal override? Kaldari (talk) 22:01, 15 January 2020 (UTC)
      • Just as the fact that some cultures consider gender to be significant means that we can't universally omit that data, the fact that many cultures consider sex to be significant means that we can't universally omit that data either. We've had endless discussion of this, including much from the discussions surrounding the original change from P21 being labelled just "sex" to "sex or gender". --Yair rand (talk) 00:23, 16 January 2020 (UTC)
  • Symbol oppose vote oversat.svg Strong oppose In Polish (Q809) there is no clear distinction between the "gender" and "sex" nouns when it comes to describing person's gender or sex. Simillar to @Esquilo, Wolverène: I do not see how this would help this community. Nadzik (talk) 15:03, 23 January 2020 (UTC)
    • I am also not sure if discussion with only few opinions (last comment being from 2013) and closed in 2019 counts as en:Consensus decision-making Nadzik (talk) 15:06, 23 January 2020 (UTC)
      • We should also remember that splitting "gender" and "sex" may allow trolls to harass intersex people on Wikidata. It would just require to assign to an intersex person this property with "male organism" or "female organism". Nadzik (talk) 15:37, 23 January 2020 (UTC)
    • @Nadzik: I don't think your assertion is entirely correct. While there is no native Polish term for gender, it seems the Polish language has borrowed the English term in order to establish such a distinction. Polish Wikipedia has separate articles for pl:Gender and pl:Płeć, and it's easy to find websites in Polish that discuss the difference between the two concepts, such as this. In fact, it seems that the University of Warsaw named "gender" the Word of the Year back in 2013: http://www.slowanaczasie.uw.edu.pl/slowo-roku-2013/. Kaldari (talk) 11:51, 28 January 2020 (UTC)