It's not clear how many of these objections are to be persuasive to those who accept the respective definitions (presumably, that's the point of the objections).
For example, consider your objections to the self-id view, i.e. "A woman is one who identifies as a woman". I don't think these are going to be persuasive to the trans-inclusive side, i.e. liberals.
Objection (3) can be avoided by defining identification as some kind of mental state, rather than outward behavior. In fact, I'm pretty sure this is what most trans-inclusive people take identification to consist in. In my experience, trans-inclusive people are going to be fine with (2) and (4). They happily take "being a woman" to be constituted by some kind of mental state that is vaguely described as identifying as a woman. Yes, you look like a man, but so did Dylan Mulvaney and (as far as I can tell), most people who are already trans-inclusive are willing to accept that Dylan Mulvaney was a woman the second they came out.
(1) is probably the biggest problem, but the definition can be slightly modified to avoid this objection without losing the spirit of view. For example, one could just say a woman is someone who identifies with some set of roles, expectations, stereotypes, etc. that are associated with adult human females within their culture. The details are unclear, but it seems that the spirit of this view is that being a woman supervenes on having the right kinds of mental states.
------
On your objections to the biological view, i.e. "A woman is an adult human female". I don't think these are persuasive to the trans-exclusionary side, i.e. conservatives.
I don't see how (1) is persuasive at all. Conservatives who accept this definition will proudly admit that trans people who "pass" are still of the gender that aligns with their sex. I think conservatives will happily admit that men can look like women.
(2) and (4) are basically the same point, which is that many people will refer to trans people who pass in accordance with their preferred gender, and even conservatives will unintentionally do this sometimes even when they know the person is trans. I don't think these observations are persuasive evidence against the biological view. These observations are not unexpected even assuming the biological view is correct, for a few reasons: (a) trans people who pass appear like the opposite sex, and people tend to use terms based on their appearance, so it would be expected that people would refer to Blaire White as a "woman" even if "woman" was defined in terms of sex; and (b) for pragmatic reasons, people (including many conservatives) will refer to trans people based on their preferred pronouns, regardless of which definition of "woman" is correct. In fact, you yourself say "even if I were convinced that some transgender person was not a woman in a purely descriptive sense, I would use the pronouns that they would like me to use".
In fact, objections (1) and (4) can also both be levied against a biological definition of "female". You could say: "doesn't Blaire White seem like a female? The biological definition of "female" denies this, so it's implausible. Furthermore, most people would probably refer to Blaire White as "female" if they were at a bar". Do these observations provide good evidence that the biological definition of "female" is incorrect?
(3) isn't really relevant. As you say in your intro, "what a woman is is different from who we should treat as a woman for many practical purposes."
----
Some points on your proposed definition.
You said most people wouldn't consider you a woman if you came out. But it's not clear to me that this is true. If you stated with strong and seemingly sincere conviction that you're a woman, I think the number of people that would consider you a woman would be only slightly lower than the number of people that consider Blaire White a woman. If you limit to the subset of people that genuinely believe Blaire White is a woman (who are aware that Blaire is trans), I think basically everyone in that subset is going to accept you as a woman unless you're obviously trolling or something. Again, I think you would be treated the same as Dylan Mulvaney, i.e. conservatives would mostly reject and liberals would mostly accept your new proposed gender identity.
Lastly, I don't think the fact that most philosophers agree with the social view of "woman" is strong evidence. Firstly, it's not clear that philosophers are relying on any empirical research to figure how people are using these terms. More importantly, we should discount our trust in expert agreement very significantly if the following conditions are met:
* The agreement is on a topic that is highly morally/politically contentious, as this is when cognitive biases become most powerful.
* It's not clear that the experts are employing a reliable methodology to establish truths.
* We (non-experts) lack a clear method for assessing the reliability of the experts.
* The experts almost universally align with one particular political orientation that would bias them towards a particular answer.
One interesting finding in the link you posted is that whether an expert is a capitalist predicts very highly whether one adopts the social or biological view on gender. For example, of the experts that accept capitalism, only 52% accept the social view (140/(140+127)). But, of experts that reject capitalism, 78% accept the social view (416/(416+119)). But there's no rational reason for one's views on such disparate topics to be correlated in this way. Thus, this is strong evidence that philosophers' views on either politics or gender (probably both) are driven significantly by some kind of general cognitive bias.
-----
In sum, I think the vast majority of people's concept of "woman" and "man" will be captured by either the biological view (i.e. for conservatives) or some variation of the self ID view (i.e. for liberals) (I'm guessing this is the psychological view mentioned in the philpapers survey? I'm not sure). And I don't think your objections are persuasive to those adhering to either view.
Also, one issue I have with your definition is that it doesn't clearly differentiate itself from the alternative definitions. E.g. someone endorsing the biological view could accept your definition but give all the weights to various markers for biological sex. Someone who endorses some form of the self ID view could give all the weights to various psychological traits. Presumably, your definition differs from the alternative definitions in that it provides different verdicts about particular cases. But until you provide more details about the traits and specify their weights, it's not clear how it produces different verdicts. For example, when considering the case of Buck Angel identifying as a woman, you said wouldn't consider Buck to be a woman. But, then you said someone with your definition could just "adjust the weights accordingly". If anyone could take your definition and adjust the weights as they please, then it's not clear that your definition provides better verdicts than the alternative definitions in those cases where you think the alternatives go wrong.
The self-id definition need not be circular. A woman may be anyone who associates themself with the sequence of sounds "woman", much as a "Derek" is anyone who associates themself with the sequence of sounds "Derek". Of course, that leads to a problem if you don't speak english...
Hey Matthew, typically love your articles but I didn't really see the force of your arguments against the traditional view of the meaning of woman (adult human female). Responding to section 3 and stealing most of these arguments and examples from Bogardus' work (with all mistakes introduced being my own):
1) "Seeming like a woman" and "being a woman" are not the same thing, because how we define properties can be different from how we identify them. Gold is the element with 79 protons, but we commonly don't do atomic experiments to decide whether something to call something gold, instead we just look at how it appears (yellow, shiny, malleable) and we recognize that there exists pyrite (fools' gold) that imitates the appearance of gold but isn't, so appearances are not necessarily 100% reliable, while the definition is. Similar thing goes for sex: the sex you have is a disposition to produce a certain type of gamete. You might see a person whose body has the typical appearance of someone who has one kind of disposition, but they might actually have the opposite (a trans person), so they seem like one sex but are actually the other.
2) Yeah, but that's because conservatives are just trying to be rude to trans people. There are legitimate reasons to not use the "right" pronouns while talking to trans people (such as politeness), and it can be unnatural to not use the wrong pronouns since we usually choose the pronoun for a given person based on their appearance, not on whether they meet the definition, but this doesn't merit a reason change the definition, given that our appearances can be wrong (as explained in the previous answer).
3) That's not a reason to reject the view, it's just a reason to be polite. When I go to court and refer to the judge as "Your Honor", I don't have to actually think he's honorable or define my concept of honorability so as to include him for "pragmatic reasons", I'm just doing it to be polite.
4) "The vast majority of people wouldn’t call Blaire White a man." I think if you told them that they were trans and had a penis, they would, and the same goes for Buck Angel. This just seems to rehash the distinction between "seeming like a woman" and "being a woman".
5) I don't think that's the correct analysis of sexual orientation. A sexual orientation is disposition to being attracted to the typical sexual characteristics of a given sex, so a woman being lesbian means they have a disposition to being attracted to the typical sexual characteristics of females. Lesbians being attracted to transwomen who pass is what we would expect, and might also expect them to change their mind if they discover the transwoman is pre-op (given that genitals are a primary sexual characteristic). A lesbian would be attracted to Cortana from Halo or a realistic female-resembling robot, so sexual attraction is not attraction towards sex itself but towards the typical sexual characteristics of a given sex.
Unfortunately, I think this post makes an error by conflating the question "Who is a woman?" with the question "What is the definition of the word 'woman?'" And once you make that distinction, a lot of these objections don't make a lot of sense. For example, you point out that many people would look at Blaire White and think, "That person is a woman." But that doesn't necessarily imply that the definition of 'woman' relies on the sorts of features that would lead someone to make that distinction. After all, many people might also look at pyrite and think, "That mineral is gold" - that doesn't mean the definition of 'gold' is anything other than "the chemical element of atomic number 79." People can and do constantly misuse terms on the basis of their own sense of what the word means, or even without any coherent meaning in mind at all. That doesn't have some sort of transitive impact on the nature of the thing being referred to.
In general, I would say *most* natural kinds don't have definitions in the way you're thinking. Take chairs, for example - an object doesn't become a chair simply because it has an appropriate number of "chair-like traits" that can be listed out in order of importance. So attempting to sort women from non-women by reference to a checklist of necessary and sufficient criteria seems like a doomed project to me. It would be better to step away from definitions entirely and start considering the referent - the particular group of people in the world the word 'woman' was first determined to pick out - and ask if trans women are included in that group. And I think you can make a very good case that, for at least the vast majority of them, they aren't. But someone else might disagree with that judgment on empirical grounds, and could make an alternative case. Either way would be better than trying to build some canonical definition from the ground up.
(I think you sorta hit on this issue, btw, when you talk about restricting prisons to "only some women." That, in my opinion, is the natural kind poking out from the new definition layered on top of it. Most people think there should be men's prisons and women's prisons, so if you need to redefine the women's prisons to be "female people's prisons," then that's good evidence that most people's conception of "woman" is actually just "female people." You can modify the definition of 'woman' to cover all sorts of people, but it does seem like most people's judgments about how groups ought to actually be categorized in practical reality split along lines that new definition does not follow. It might be that the ultimate result of a trans-inclusive definition of 'woman' will just be most people moving to a new word that better captures the natural kind 'woman' previously picked out. Which is fine, I guess, but it seems better to me that you would just stick with the parameters previously established.)
Blaire White passes. She looks like a female, despite being biologically male. And I think the other, potentially plausible cases where biological males might be thought to be women are going to be similar to this - intersex cases where someone is biologically male but is raised and perceived by others in ways typical of females.
What would your account of what a woman is say about someone who doesn't pass? The feminine behaviours you mostly listed - wearing a dress, having long hair, etc - aren't really about how other people perceive us. I suspect a lot of people will think it is just as obvious that an adult male who doesn't pass, who wears dresses, and all those other things, is not really a woman.
Great article! "Woman" is a fuzzy family resemblance category - defined by a set of overlapping features/traits rather than any one single trait that all members of the category must share in common.
I'm curious what your take is on other potential "transes" e.g. could someone be transracial? Or what about transage? If someone sincerely insists that the age they feel internally is different from their chronological age, should we refer to them accordingly in order to be polite? Should we also think of these as fuzzy sets e.g. is an old person simply someone with a sufficient number of elderly traits (wrinkles, ailing health, wisdom, etc.)?
Imagine a scenario where a person’s sex was female, but every day, they became deeply distressed by another person, who's sex was male, wanting to be called a female from 1pm to 3pm. If the male insisting on being called a female drove the female into a deep depression, even if it’s not technically his real sex, you should still call him female. This is not lying or denying the truth—it’s just being nice.
Maybe I missed something, but you haven't actually defined what a woman is. You say a woman is a person with a sufficient number of feminine traits, but you don't say what a sufficient number is, nor do you give any comprehensive list of them, or weight them in any way, or what people will be in charge of judging whether people have the traits. (Of the nine traits you do list, 1-3 and 8-9 and arguably 5 are either entirely or significantly cultural.) In any case, even if people agreed with your definition, the debate over the list of traits would be endless.
Your "dating test" seems peculiar to me, since it implies that sexual orientation is somehow more basic and more essential than sex. In the comments you write "by definition, gay men are not attracted to women", but that definition is too strict, a man who is usually attracted to men but who was once attracted to a woman is still gay (and if you say he isn't, then you have a "true Scotsman" like problem of identifying the "truly gay men"; or you can make the ridiculous claim that the woman he was attracted to wasn't a woman).
I'm skeptical Blair White or Buck Angel would be very convincing if you saw them in the wild, although I agree broadly with your gender externalism. Any pronoun you like, as long as it's he or she and you pass.
On the objections to the biological view: I think (4) has some force, but I don't think (5) is persuasive at all.
(5) said: "If a woman is consistently interested in dating and having sex with other women, she would be either lesbian or bisexual. But one who is interested in dating and having sex with Buck Angel would not necessarily be lesbian or bisexual. Therefore, Buck Angel is not a woman. This is just trivial—a much larger portion of gay men would be interested in Buck Angel than of lesbian women."
I think heterosexual men can be heterosexually attracted to things that aren't women, but resemble women. A heterosexual man, for example, can be heterosexually attracted to a cartoon image of a woman. Further, it seems like a heterosexual woman could be heterosexually attracted to a phallically shaped fruit. A cartoon is not a woman and a banana is not a man, though. Defenders of the biological view should just say that heterosexual men are attracted to trans women because they resemble women, not because they are women.
My main response to this is that all three definitions are in fact pretty good. They probably agree in 99%+ of cases. All concepts are a bit fuzzy- if we disagree only in ~1% of cases, we're actually doing pretty well.
A further objection to the conservative definition which you'll sometimes see floating around- "female" itself isn't unambiguously defined. Some people have ambiguous sex.
What do you think of Moore's open question argument? Can we apply it here and thus claim that woman-like is a non-natural property, which can't be defined? It explains why people (rationally) disagree.
I’m a black gay guy and I can personally say that my life has been much better since reading this article
It's not clear how many of these objections are to be persuasive to those who accept the respective definitions (presumably, that's the point of the objections).
For example, consider your objections to the self-id view, i.e. "A woman is one who identifies as a woman". I don't think these are going to be persuasive to the trans-inclusive side, i.e. liberals.
Objection (3) can be avoided by defining identification as some kind of mental state, rather than outward behavior. In fact, I'm pretty sure this is what most trans-inclusive people take identification to consist in. In my experience, trans-inclusive people are going to be fine with (2) and (4). They happily take "being a woman" to be constituted by some kind of mental state that is vaguely described as identifying as a woman. Yes, you look like a man, but so did Dylan Mulvaney and (as far as I can tell), most people who are already trans-inclusive are willing to accept that Dylan Mulvaney was a woman the second they came out.
(1) is probably the biggest problem, but the definition can be slightly modified to avoid this objection without losing the spirit of view. For example, one could just say a woman is someone who identifies with some set of roles, expectations, stereotypes, etc. that are associated with adult human females within their culture. The details are unclear, but it seems that the spirit of this view is that being a woman supervenes on having the right kinds of mental states.
------
On your objections to the biological view, i.e. "A woman is an adult human female". I don't think these are persuasive to the trans-exclusionary side, i.e. conservatives.
I don't see how (1) is persuasive at all. Conservatives who accept this definition will proudly admit that trans people who "pass" are still of the gender that aligns with their sex. I think conservatives will happily admit that men can look like women.
(2) and (4) are basically the same point, which is that many people will refer to trans people who pass in accordance with their preferred gender, and even conservatives will unintentionally do this sometimes even when they know the person is trans. I don't think these observations are persuasive evidence against the biological view. These observations are not unexpected even assuming the biological view is correct, for a few reasons: (a) trans people who pass appear like the opposite sex, and people tend to use terms based on their appearance, so it would be expected that people would refer to Blaire White as a "woman" even if "woman" was defined in terms of sex; and (b) for pragmatic reasons, people (including many conservatives) will refer to trans people based on their preferred pronouns, regardless of which definition of "woman" is correct. In fact, you yourself say "even if I were convinced that some transgender person was not a woman in a purely descriptive sense, I would use the pronouns that they would like me to use".
In fact, objections (1) and (4) can also both be levied against a biological definition of "female". You could say: "doesn't Blaire White seem like a female? The biological definition of "female" denies this, so it's implausible. Furthermore, most people would probably refer to Blaire White as "female" if they were at a bar". Do these observations provide good evidence that the biological definition of "female" is incorrect?
(3) isn't really relevant. As you say in your intro, "what a woman is is different from who we should treat as a woman for many practical purposes."
----
Some points on your proposed definition.
You said most people wouldn't consider you a woman if you came out. But it's not clear to me that this is true. If you stated with strong and seemingly sincere conviction that you're a woman, I think the number of people that would consider you a woman would be only slightly lower than the number of people that consider Blaire White a woman. If you limit to the subset of people that genuinely believe Blaire White is a woman (who are aware that Blaire is trans), I think basically everyone in that subset is going to accept you as a woman unless you're obviously trolling or something. Again, I think you would be treated the same as Dylan Mulvaney, i.e. conservatives would mostly reject and liberals would mostly accept your new proposed gender identity.
Lastly, I don't think the fact that most philosophers agree with the social view of "woman" is strong evidence. Firstly, it's not clear that philosophers are relying on any empirical research to figure how people are using these terms. More importantly, we should discount our trust in expert agreement very significantly if the following conditions are met:
* The agreement is on a topic that is highly morally/politically contentious, as this is when cognitive biases become most powerful.
* It's not clear that the experts are employing a reliable methodology to establish truths.
* We (non-experts) lack a clear method for assessing the reliability of the experts.
* The experts almost universally align with one particular political orientation that would bias them towards a particular answer.
One interesting finding in the link you posted is that whether an expert is a capitalist predicts very highly whether one adopts the social or biological view on gender. For example, of the experts that accept capitalism, only 52% accept the social view (140/(140+127)). But, of experts that reject capitalism, 78% accept the social view (416/(416+119)). But there's no rational reason for one's views on such disparate topics to be correlated in this way. Thus, this is strong evidence that philosophers' views on either politics or gender (probably both) are driven significantly by some kind of general cognitive bias.
-----
In sum, I think the vast majority of people's concept of "woman" and "man" will be captured by either the biological view (i.e. for conservatives) or some variation of the self ID view (i.e. for liberals) (I'm guessing this is the psychological view mentioned in the philpapers survey? I'm not sure). And I don't think your objections are persuasive to those adhering to either view.
Also, one issue I have with your definition is that it doesn't clearly differentiate itself from the alternative definitions. E.g. someone endorsing the biological view could accept your definition but give all the weights to various markers for biological sex. Someone who endorses some form of the self ID view could give all the weights to various psychological traits. Presumably, your definition differs from the alternative definitions in that it provides different verdicts about particular cases. But until you provide more details about the traits and specify their weights, it's not clear how it produces different verdicts. For example, when considering the case of Buck Angel identifying as a woman, you said wouldn't consider Buck to be a woman. But, then you said someone with your definition could just "adjust the weights accordingly". If anyone could take your definition and adjust the weights as they please, then it's not clear that your definition provides better verdicts than the alternative definitions in those cases where you think the alternatives go wrong.
The self-id definition need not be circular. A woman may be anyone who associates themself with the sequence of sounds "woman", much as a "Derek" is anyone who associates themself with the sequence of sounds "Derek". Of course, that leads to a problem if you don't speak english...
Isn’t this pretty similar to Judith Butler’s ideas around gender performativity?
Hey Matthew, typically love your articles but I didn't really see the force of your arguments against the traditional view of the meaning of woman (adult human female). Responding to section 3 and stealing most of these arguments and examples from Bogardus' work (with all mistakes introduced being my own):
1) "Seeming like a woman" and "being a woman" are not the same thing, because how we define properties can be different from how we identify them. Gold is the element with 79 protons, but we commonly don't do atomic experiments to decide whether something to call something gold, instead we just look at how it appears (yellow, shiny, malleable) and we recognize that there exists pyrite (fools' gold) that imitates the appearance of gold but isn't, so appearances are not necessarily 100% reliable, while the definition is. Similar thing goes for sex: the sex you have is a disposition to produce a certain type of gamete. You might see a person whose body has the typical appearance of someone who has one kind of disposition, but they might actually have the opposite (a trans person), so they seem like one sex but are actually the other.
2) Yeah, but that's because conservatives are just trying to be rude to trans people. There are legitimate reasons to not use the "right" pronouns while talking to trans people (such as politeness), and it can be unnatural to not use the wrong pronouns since we usually choose the pronoun for a given person based on their appearance, not on whether they meet the definition, but this doesn't merit a reason change the definition, given that our appearances can be wrong (as explained in the previous answer).
3) That's not a reason to reject the view, it's just a reason to be polite. When I go to court and refer to the judge as "Your Honor", I don't have to actually think he's honorable or define my concept of honorability so as to include him for "pragmatic reasons", I'm just doing it to be polite.
4) "The vast majority of people wouldn’t call Blaire White a man." I think if you told them that they were trans and had a penis, they would, and the same goes for Buck Angel. This just seems to rehash the distinction between "seeming like a woman" and "being a woman".
5) I don't think that's the correct analysis of sexual orientation. A sexual orientation is disposition to being attracted to the typical sexual characteristics of a given sex, so a woman being lesbian means they have a disposition to being attracted to the typical sexual characteristics of females. Lesbians being attracted to transwomen who pass is what we would expect, and might also expect them to change their mind if they discover the transwoman is pre-op (given that genitals are a primary sexual characteristic). A lesbian would be attracted to Cortana from Halo or a realistic female-resembling robot, so sexual attraction is not attraction towards sex itself but towards the typical sexual characteristics of a given sex.
Unfortunately, I think this post makes an error by conflating the question "Who is a woman?" with the question "What is the definition of the word 'woman?'" And once you make that distinction, a lot of these objections don't make a lot of sense. For example, you point out that many people would look at Blaire White and think, "That person is a woman." But that doesn't necessarily imply that the definition of 'woman' relies on the sorts of features that would lead someone to make that distinction. After all, many people might also look at pyrite and think, "That mineral is gold" - that doesn't mean the definition of 'gold' is anything other than "the chemical element of atomic number 79." People can and do constantly misuse terms on the basis of their own sense of what the word means, or even without any coherent meaning in mind at all. That doesn't have some sort of transitive impact on the nature of the thing being referred to.
In general, I would say *most* natural kinds don't have definitions in the way you're thinking. Take chairs, for example - an object doesn't become a chair simply because it has an appropriate number of "chair-like traits" that can be listed out in order of importance. So attempting to sort women from non-women by reference to a checklist of necessary and sufficient criteria seems like a doomed project to me. It would be better to step away from definitions entirely and start considering the referent - the particular group of people in the world the word 'woman' was first determined to pick out - and ask if trans women are included in that group. And I think you can make a very good case that, for at least the vast majority of them, they aren't. But someone else might disagree with that judgment on empirical grounds, and could make an alternative case. Either way would be better than trying to build some canonical definition from the ground up.
(I think you sorta hit on this issue, btw, when you talk about restricting prisons to "only some women." That, in my opinion, is the natural kind poking out from the new definition layered on top of it. Most people think there should be men's prisons and women's prisons, so if you need to redefine the women's prisons to be "female people's prisons," then that's good evidence that most people's conception of "woman" is actually just "female people." You can modify the definition of 'woman' to cover all sorts of people, but it does seem like most people's judgments about how groups ought to actually be categorized in practical reality split along lines that new definition does not follow. It might be that the ultimate result of a trans-inclusive definition of 'woman' will just be most people moving to a new word that better captures the natural kind 'woman' previously picked out. Which is fine, I guess, but it seems better to me that you would just stick with the parameters previously established.)
Blaire White passes. She looks like a female, despite being biologically male. And I think the other, potentially plausible cases where biological males might be thought to be women are going to be similar to this - intersex cases where someone is biologically male but is raised and perceived by others in ways typical of females.
What would your account of what a woman is say about someone who doesn't pass? The feminine behaviours you mostly listed - wearing a dress, having long hair, etc - aren't really about how other people perceive us. I suspect a lot of people will think it is just as obvious that an adult male who doesn't pass, who wears dresses, and all those other things, is not really a woman.
Great article! "Woman" is a fuzzy family resemblance category - defined by a set of overlapping features/traits rather than any one single trait that all members of the category must share in common.
I'm curious what your take is on other potential "transes" e.g. could someone be transracial? Or what about transage? If someone sincerely insists that the age they feel internally is different from their chronological age, should we refer to them accordingly in order to be polite? Should we also think of these as fuzzy sets e.g. is an old person simply someone with a sufficient number of elderly traits (wrinkles, ailing health, wisdom, etc.)?
Imagine a scenario where a person’s sex was female, but every day, they became deeply distressed by another person, who's sex was male, wanting to be called a female from 1pm to 3pm. If the male insisting on being called a female drove the female into a deep depression, even if it’s not technically his real sex, you should still call him female. This is not lying or denying the truth—it’s just being nice.
Maybe I missed something, but you haven't actually defined what a woman is. You say a woman is a person with a sufficient number of feminine traits, but you don't say what a sufficient number is, nor do you give any comprehensive list of them, or weight them in any way, or what people will be in charge of judging whether people have the traits. (Of the nine traits you do list, 1-3 and 8-9 and arguably 5 are either entirely or significantly cultural.) In any case, even if people agreed with your definition, the debate over the list of traits would be endless.
Your "dating test" seems peculiar to me, since it implies that sexual orientation is somehow more basic and more essential than sex. In the comments you write "by definition, gay men are not attracted to women", but that definition is too strict, a man who is usually attracted to men but who was once attracted to a woman is still gay (and if you say he isn't, then you have a "true Scotsman" like problem of identifying the "truly gay men"; or you can make the ridiculous claim that the woman he was attracted to wasn't a woman).
I'm skeptical Blair White or Buck Angel would be very convincing if you saw them in the wild, although I agree broadly with your gender externalism. Any pronoun you like, as long as it's he or she and you pass.
On the objections to the biological view: I think (4) has some force, but I don't think (5) is persuasive at all.
(5) said: "If a woman is consistently interested in dating and having sex with other women, she would be either lesbian or bisexual. But one who is interested in dating and having sex with Buck Angel would not necessarily be lesbian or bisexual. Therefore, Buck Angel is not a woman. This is just trivial—a much larger portion of gay men would be interested in Buck Angel than of lesbian women."
I think heterosexual men can be heterosexually attracted to things that aren't women, but resemble women. A heterosexual man, for example, can be heterosexually attracted to a cartoon image of a woman. Further, it seems like a heterosexual woman could be heterosexually attracted to a phallically shaped fruit. A cartoon is not a woman and a banana is not a man, though. Defenders of the biological view should just say that heterosexual men are attracted to trans women because they resemble women, not because they are women.
One of your best articles imo
My main response to this is that all three definitions are in fact pretty good. They probably agree in 99%+ of cases. All concepts are a bit fuzzy- if we disagree only in ~1% of cases, we're actually doing pretty well.
A further objection to the conservative definition which you'll sometimes see floating around- "female" itself isn't unambiguously defined. Some people have ambiguous sex.
What do you think of Moore's open question argument? Can we apply it here and thus claim that woman-like is a non-natural property, which can't be defined? It explains why people (rationally) disagree.
Based