Woman as a Politically Significant Term: A Solution to the Puzzle

What does woman mean? According to two competing views, it can be seen as a sex term or as a gender term. Recently, Jennifer Saul has put forward a contextualist view, according to which woman can have different meanings in different contexts. The main motivation for this view seems to involve moral and political considerations, namely, that this view can do justice to the claims of trans women. Unfortunately, Saul argues, on further reflection the contextualist view fails to do justice to those moral and political claims that motivated the view in the first place. In this article I argue that there is a version of the contextualist view that can indeed capture those moral and political aims, and in addition, I use this case to illustrate an important and more general claim, namely, that moral and political considerations can be relevant to the descriptive project of finding out what certain politically significant terms actually mean.

In this article, my aim is to examine a puzzle that the term woman gives rise to, and which has important consequences both for the philosophy of language and moral and political philosophy. This puzzle has been discussed recently by Jennifer Saul, whose main aim is not so much to defend a particular solution to the puzzle, but rather to show that the puzzle is significant and deserves serious attention (Saul 2012). In this article, I will try to offer one solution that is both satisfactory from a philosophical point of view and does justice (I hope) to at least some of the moral and political goals that feminist philosophers might pursue in this debate.

THE PUZZLE
The puzzle can be put as follows. At first sight (and for many ordinary speakers), the term woman seems to function mostly as a sex term, that is, the term woman is supposed to refer to those who are biologically female, in the same way that the term man is supposed to refer to those who are biologically male. But as many feminist theorists have argued, this is problematic for several reasons, such as those having to do with intersex and transgender people. Let's examine each case in turn.
First, the case of intersex people poses the following problem: Saul argues that the view according to which woman is a sex term seems to presuppose that the distinction between males and females is both exclusive (that is, those who fall under man cannot fall under woman, and vice versa), and exhaustive (that is, all human beings must fall under at least one category), and furthermore, that "people fall neatly and easily into these categories" (Saul 2012, 198). But, she argues, it can be shown that these assumptions are false by reflecting on the case of intersex people, who do not fall neatly and easily into either category. Furthermore, some intersex people might fail to fall under either of the two categories, making the distinction nonexhaustive, or, depending on how biological sex is characterized, they might fall under both categories, making the distinction nonexclusive. In either case, this seems to be bad news for the sex-term view, or so many feminists argue. 1 However, it is not entirely clear that this conclusively shows that woman cannot be a sex term. Rather, what these cases seem to suggest is that a human being's biological sex is not determined by a single criterion (nor by a class of necessary and sufficient conditions), but rather by a cluster of criteria, including chromosomes, hormones, internal and external sex organs, and maybe others that are not co-extensive, that is to say, someone might count as male or female according to one of these criteria but not the others. 2 Given this, it could be argued that there can be human beings who count both as male and female (so the distinction is not exclusive), and also human beings who count as neither (so the distinction is not exhaustive), compatibly with the extension of female being determined in virtue of a cluster of biological criteria. And mutatis mutandis for the view that woman works as a sex term. 3 Saul presents a second and more crucial problem for the view that woman is a sex term: she argues that this view fails to do justice to the claims of trans women, for whom it is very important that they be fully considered as women. For example, it might be very important for trans women who haven't had surgery or hormonal treatment that their utterances of the sentence "I am a woman" count as true. But if woman in English is understood as a sex term, then their utterances would turn out to be false. 4 A possible alternative would be to consider the term woman as a gender term, where woman is supposed to refer to those individuals who share a certain social role (to be specified). However, this view is also problematic, as many feminist philosophers have argued (see Mikkola 2008 for an excellent survey). The main problem is that there is no particular social role that all women share, just in virtue of being women. In particular, women instantiate very different social roles, depending on their race, class, nationality, sexual orientation, age, disability status, and many other factors. Given this, it seems impossible to figure out the criteria that would determine the extension of the term woman, if it were to work as a gender term. More crucially, it would be impossible for feminists to specify that class of individuals that feminism is concerned with, assuming that feminism is concerned with the oppression of women. 5 The puzzle so far is this: when it comes to the meaning of woman, neither sexbased views nor gender-based views are satisfactory. A possible solution, Saul argues, is to understand woman as a contextually-shifting term, that is, as a term that has different meanings in different contexts (sometimes it can work as a sex term, sometimes as a gender term, or even as something different on other occasions). To illustrate this idea, Saul introduces the example of Charla, a trans woman who hasn't had genital surgery. Let's consider the following sentence: (1): Charla is a woman.
According to the contextualist proposal, in some contexts self-identification seems to be the most relevant criterion, such as, for example, a context where we are talking about the rights of trans women; there could be other contexts, however, where other criteria are more relevant. In such a context, an utterance of (1) can be true, whereas in a different context (say, when we are evaluating who should be screened for vaginal diseases) (1) may turn out to be false. More precisely, Saul puts forward the following contextualist proposal, which seems to capture the intuitions above: (CP): X is a woman is true in a context C iff X is human and relevantly similar (according to the standards at work in C) to most of those possessing all of the biological markers of female sex. (Saul 2012, 201) Does this proposal give the intuitive results? At first sight, it seems to do so: for instance, when we focus on sentence (1) uttered in a context where the standards at work seem to involve self-identification, then (1) will turn out to be true, because it is true that Charla is similar in how she self-identifies, with respect to most of those possessing the biological markers of female sex (that is, they self-identify as women). But on the other hand, if we focus on utterances of (1) made in a different context, say, a medical context where what matters is who should be screened for vaginal diseases, then (arguably) it is not true that Charla is similar to most females with respect to what seems to be the relevant criterion at work in that context, namely, having a vagina, and therefore (1) would turn out to be false. 6 Hence, this contextualist proposal seems both intuitive and politically useful. But, Saul argues, further reflection shows that it is problematic as well (and this is what gives rise to the puzzle Saul points out). One of the main problems is the following: the contextualist proposal (CP) seems to do justice to the claims of trans women only in a very trivializing way, because it can also render true the claims of trans-misogynist speakers. For example, we can imagine an utterance of (1) made in the context of a trans-misogynist community, where most people believe that trans women should not be treated as women in legal and practical matters: for instance, members of that community believe that trans women should not be allowed to use women's restrooms, and so on. In this context, (1) turns out to be false because it is false that Charla is similar to most females with respect to the relevant criterion that is salient in that context, namely, being biologically female (or so Saul argues).
We seem to have a puzzle here: one of the initial motivations for rejecting the view that woman is a sex term was that we wanted to do justice to the claims of trans women, but we have seen that a contextualist proposal, which was formulated with that very aim in mind, does not seem to fare much better. This gives rise to a very important methodological question: should we include as a desideratum for a correct theory of the meaning of woman that it does justice to the claims of trans women? This might seem an unfamiliar kind of desideratum in the context of mainstream philosophy of language, and indeed looks very different from other, more familiar theoretical desiderata typically invoked in philosophical debates, such as simplicity, fitting the intuitions of ordinary speakers, explanatory power, and so on. Maybe it could be argued that if these familiar criteria for evaluating a theory about the meaning of woman do indeed favor a view according to which woman is a sex term, then it means that claims such as (1) above are typically false, and that is the end of the story.
This, of course, seems extremely problematic from a feminist point of view. The challenge Saul raises, I believe, is mainly this: how can we explain what is problematic with the strategy above in a way that is consistent with the general aims and methodologies of mainstream philosophy of language, and analytic philosophy in general? (Or alternatively, perhaps we have the materials here to show that the general aims and methodologies of analytic philosophy of language are misguided, but if so, how should we proceed instead when we theorize about meaning?)

A CONTEXTUALIST SOLUTION
In what follows, I want to suggest a possible solution to the puzzle that does not call for such an extreme revision of the methods of analytic philosophy, but still does justice to the claims of trans women. It will be useful to compare it to one kind of solution that Saul puts forward: On my [contextualist] view of "woman," I cannot argue that the lawmakers are making a mistake about how the word "woman" works. But what I can do is argue that they are morally and politically wrong to apply the standards that they do. On my view, then, disagreements over who counts as a woman are simply not to be settled by appeal to the facts of language. They are to be settled by appeal to moral and political principles. There may well be a single right answer about what standards should be applied for determining who satisfies the definition of "woman" in a particular context; but it will be right because it is morally and politically right. So I can coherently, and maybe even correctly, insist that the lawmakers are wrong and I am right. But we must recognize this claim for what it is: a moral and political, rather than merely linguistic, claim. (Saul 2012, 204) In response, I want to argue that we can see this point also as a genuinely linguistic point, that is, as a descriptive claim about the (actual) meaning of woman, not just as a revisionary proposal about how the term woman should be used, given our moral and political aims (that is, what Sally Haslanger calls an ameliorative project [Haslanger 2006]). I want to argue that it is also possible to provide moral and political considerations that are relevant with regard to the descriptive project of finding out the meaning of woman (in addition to the ameliorative project of finding out what the meaning of the term woman should be, where moral and political considerations are clearly already relevant). How is this possible?
To answer this, we should go back to Saul's contextualist proposal: (CP): "X is a woman is true in a context C iff X is human and relevantly similar (according to the standards at work in C) to most of those possessing all of the biological markers of female sex" (Saul 2012, 201). As we can see, what is crucial in order to determine the truth-value of an utterance of the form "X is a woman" is the criterion that is at issue in the corresponding context. According to this contextualist proposal, any utterance of "X is a woman" will be true when X is similar to most females with respect to the standards that are relevant in that context. On my view, this is where moral and political considerations become relevant in order to determine the truthvalue of utterances of that form, to wit: competing views regarding what standards are relevant in a particular context may yield different truth-values for the proposition. Therefore, questions about the relevant standards are crucial. But, arguably, questions about the relevant standards might involve moral and political considerations (and in cases that concern the rights of trans women, will very likely do so).
For example, we can go back to the case of a trans-misogynist community where people are discussing whether Charla should be allowed to use women's restrooms. In that context, is (1) true or false? According to Saul's proposal, (1) is false in that context (as the opponents of trans women's rights want), because Charla is not similar to most females with respect to the criterion that is salient in that context (namely, having a vagina). But, we may ask, salient in which sense? If what is required is that the standards at work in C are those that people have in mind, or that people would be disposed to consider relevant, or something along these lines, it seems that Saul is right and (1) will turn out to be false in that context, unfortunately for the contextualist proposal. But if we could formulate an alternative way of determining what are the relevant standards at issue in a context, and if we could modify the contextualist proposal accordingly, then maybe we can find a solution to the puzzle.
My proposal, then, is that we should understand the relevant standards at issue in a context as those that are relevant for practical purposes (where these are broadly conceived to include theoretical, prudential, moral, political, and even aesthetic values). On this view, then, when someone utters a sentence of the form "X is a woman," that utterance will be true if and only if X is similar to most females with respect to the standards that are relevant in X's context, given the practical purposes that are relevant in this context, where this should be determined by our best theoretical and normative considerations. 7 In particular, when we focus on a context such that what is at issue is whether Charla should be allowed to use women's restrooms, it is plainly false that the relevant standard in this case is whether Charla has XX chromosomes (or a vagina), because there are plenty of moral and political considerations in support of the view that this is not the relevant criterion here (again, given our best normative and evaluative reasoning, which includes relevant moral and political considerations). In contrast, the criterion that is relevant here is whether Charla selfidentifies as a woman, or something along these lines. Therefore, it will be plainly false that Charla is not similar to most females with respect to the relevant criterion at issue here, because the relevant criterion is how Charla self-identifies, not what kind of chromosomes or genitalia she has, and with respect to that criterion she is clearly similar to most females, so (1) will turn out to be true.

SUBJECT FACTORS VS. ATTRIBUTOR FACTORS
In order to make the contrast between my proposal and Saul's account clearer, I would like to draw a comparison with a familiar distinction in philosophy of language, namely, the distinction between contextualist views that appeal to attributor factors, and those views that appeal to subject factors. 8 The main idea here is that whereas some contextualist views about the meaning of a certain term (say, "knowledge") hold that the term changes its content from context to context depending on different features of the speaker in those contexts (such as her beliefs and intentions), some other views hold that the term changes its content (or at least, the corresponding utterance changes its truth-value) from context to context depending on different features of the subject, that is, the putative knower in the case of "knowledge" (or the relevant entity that is supposed to fall under the relevant term at issue). Some relevant subject factors in the case of "knowledge" (or "knowing that p") include features such as certain relevant alternatives to p becoming more salient to the putative knower, or "objectively" more probable, and so on. For instance, if what is at issue is whether a subject S knows that there is a zebra in front of her, then whether there are painted mules in her surroundings could be a relevant subject factor that could affect the truth-value of utterances of "S knows that there is a zebra in front of her" (regardless of whether S is aware of it or not). This view about "knowledge" is different from attributor-contextualism, according to which two different speakers in different contexts who are describing S's situation could both be right if one says "S knows that there is a zebra in front of her" and the other says "S doesn't know there is a zebra in front of her," assuming S's epistemic situation doesn't change, in case the former speaker is in a context where the possibility of there being painted mules has not been made salient, but the latter is in a context where such a possibility has been made salient. Here, the meaning of "S knows that there is a zebra in front of her" changes from context to context, depending on features of the speaker in each context. In contrast, according to subject-contextualism, what it takes for S to know that there is a zebra in front of her changes from context to context, depending on features of S herself, regardless of who utters the sentence. 9 Likewise, we could contrast a version of (CP) above understood along the lines of attributor-contextualism, with a version of (CP) understood along the lines of subject-contextualism. According to attributor-contextualism about woman, "X is a woman" is true iff X is human and relevantly similar to most females, where what counts as relevantly similar to most females depends on features of the speaker, including factors such as her beliefs about what is relevant in order to be a woman, her intentions of use, and so on. 10 On the other hand, according to subject-contextualism about woman, "X is a woman" is true iff X is human and relevantly similar to most females, where what counts as relevantly similar to most females depends on "objective" features of X's context, including instrumental, moral, and political considerations having to do with how X should be treated (regardless of who utters the sentence or what their beliefs are). Hence, we can put the contrast as follows: according to attributor-contextualism, it is possible that there are two speakers, say, Nikki and Nora, such that Nikki truly says "Charla is a woman" in context C1, whereas Nora truly says "Charla is not a woman" in context C2, assuming Charla's situation is the same. But according to subject-contextualism this is impossible: if Nikki is right when she says "Charla is a woman," then if Charla's situation does not change, Nora will say something false if she utters "Charla is not a woman," regardless of Nora's context of utterance. Therefore, if we understand the contextualist proposal (CP) above in terms of subject-contextualism, then we can avoid the problematic consequences that Saul points out. In particular, we no longer have to accept the claim that utterances of "Charla is not a woman" by trans-misogynist speakers will be trivially true, because it is just not true that what matters in Charla's context (when there is a discussion about whether she should be allowed to use women's bathrooms) is whether she has a vagina, say. The speakers of such utterances may believe that what matters in order to be a woman is to have a vagina, but according to subjectcontextualism about woman, this is not what determines which standards of similarity are relevant in order to fall under woman. Rather, what determines these standards of similarity in each context has to do with our best normative and evaluative considerations concerning the putative subject, including theoretical, moral, and political considerations.
It will be useful to compare how these two different versions of contextualism about woman fare with respect to a couple of additional objections Saul presents, having to do with "challenges from mixed contexts (where speaker and audience have different standards in mind)" (Saul 2012, 206), and from belief reports. Let's focus on mixed contexts first. As I understand it, the core of this objection is that mixed contexts are those where speaker and hearer have different standards of similarity in mind (for example, an opponent of trans women talking to an advocate of trans women), and this gives rise to the following question: what are the relevant standards of similarity in the context of that conversation? According to attributor-contextualism, we could perhaps say that when the opponent of trans women utters woman, the term does not apply to trans women who haven't had genital surgery, whereas when the advocate of trans women utters the term, it does apply to all trans women because they self-identify as women and this is what is relevant according to this speaker's beliefs. 11 In my view, this challenge seems to have an easy solution if we accept a subject-contextualist account along the lines I have suggested, because as I explained above, what determines the salient standards of similarity in a certain context according to this account is not a matter of what speakers have in mind, but rather a matter of which standards do in fact satisfy a series of practical and moral considerations. As Saul herself suggests, "[t]here may well be a single right answer about what standards should be applied for determining who satisfies the definition of 'woman' in a particular context" (204), and if so, this right answer should be enough to determine the relevant standards in that context. And in this way, when the advocate and the opponent of trans women are speaking to each other, they are using woman with the same reference, namely the one fixed by the relevant normative considerations in the context at issue, which plausibly will be one including all trans women. Therefore, according to this view, the opponent of trans women is wrong when she says "Charla is not a woman," regardless of how the speaker intends to use the term.
Saul has also discussed a different objection to contextualism, having to do with belief reports (Saul 2006). The objection can be put as follows: if an opponent of trans women (S1) says "Charla is not a woman," and an advocate of trans women (S2) wants to report that utterance and says "S1 says that Charla is not a woman" (for instance, in order to argue that what S1 said was false), the contextualist proposal, understood in terms of attributor-contextualism, would have the consequence that S2's utterance does not have a clear content. If the meaning of woman is fixed in part by what the speaker has in mind, then when S1 utters woman it has to do with being biologically female, whereas when S2 utters it, it has to do with self-identifying as a woman. Therefore, it is not even clear how S2 could report on S1's utterance. But this seems counter-intuitive: we take it that speakers can report what other speakers of the same natural language say without so much trouble. In my view, this problem can be avoided if we move to a version of subject-contextualism like the one I have been defending so far. To repeat, if the standards of similarity are fixed not by what the speakers have in mind, but rather by the relevant normative considerations having to do with the subject of the utterance (in this case, Charla herself, and some relevant features of her context such as the history of oppression of trans women, and so on), then it seems clear that the relevant standards of similarity at issue in both S1's and S2's utterances will hold that both uses of woman apply to Charla, regardless of whether S1 or S2 believe that or not. That is, what S1 says is simply false, and the content that S2 attributes to S1 in her belief report is exactly the same (false) content expressed by S1.

BETTCHER'S VIEW
To finish my discussion, I would like to compare my account of the meaning of woman with a different and very interesting account defended by Talia Mae Bettcher, as part of her excellent response to Saul. She says: On the multiple-meaning view, a trans woman can say that she is a woman in all legitimate contexts because those contexts in which she is not a woman occur in a dominant culture that has been rejected for the reasons mentioned above. . . . [T]here are actually two concepts and two meanings of "womanhood." The two concepts (and the two meanings) are related in that the latter is a result of changes performed on the former. . . . A preoperative trans woman might be a woman-R ("woman" in the resistant sense) but not as a woman-D ("woman" in the dominant sense). She would be a woman-R and fail to be a woman-D not as a matter of political decision but metaphysically speaking. The political question, instead, concerns which concept we should take seriously, and this is connected to the larger question regarding which gendered vision of the world (if any) we commit to. (Bettcher 2013, 243-44) As I understand Bettcher's view, the main idea is that there are several coexisting concepts of woman, and some of them are revisionary, more inclusive versions of other, more exclusionary concepts. This could sound just like a version of the ameliorative project of finding out which concepts we should use, for political reasons. But Bettcher's point is that some concepts embed worldviews that are more accurate than others, and because of this, our choice regarding which concepts to use is not merely ameliorative but also descriptive in part (or in other words, our choice regarding which concepts to use is not purely political or normative but also in part theoretical, having to do with which worldviews better represent reality). I am sympathetic to this account, but I am a bit worried about the following remarks: The point I have defended in this essay is that accounts that take for granted singular, fixed meanings of gender terms cannot plausibly provide a liberatory theory. Not only do such accounts go wrong by failing to square with the actual reality of the situation, namely the fact that central terms are used in trans contexts in multiple and contested ways; they actually undermine trans self-identifications by foreclosing the possibility of this multiplicity. (Bettcher 2013, 247) I have two worries about this view. First, I would want to defend the idea that subjectcontextualism can provide the materials for a liberatory theory that does justice to the claims of trans women, even if in a sense it takes for granted singular, fixed meanings, as opposed to a multiple-meaning view of the sort Bettcher advocates. Subject-contextualism does say that the referent of woman can change from context to context, but this will happen only depending on features of the subject matter of the corresponding utterance (and in particular, on normative considerations involving the subjects of which the speaker is predicating or denying womanhood), not just in terms of the beliefs and intentions of the speaker. So once the relevant subject is fixed, as in the case of utterances of "Charla is a woman," this sentence will have the same meaning and the same truthvalue, regardless of who utters it (assuming they are all speaking about the same person, and her self-identification and other relevant factors do not change). So there is a sense in which my view does take for granted singular, fixed meanings, but I do not think this is incompatible with a liberatory theory. Indeed, and this is my second worry, it could be argued that a contextualist view along the lines of subject-contextualism could be more liberatory than Bettcher's view. The reason is this: if we adopt a multiple-meaning account, we are accepting the existence of certain meanings that are exclusionary. That is, we will understand some speakers in some contexts as using woman in a way that is exclusionary. (For instance, Bettcher distinguishes between resistant and dominant uses of the term.) And this will have the consequence that whether (some) trans women count as women in some contexts will depend on which meanings are at issue. That is to say, there could be some speakers (of English) such that their utterances of "Charla is not a woman" turn out to be true, given that they are using the dominant concept instead of the resistant concept. And this seems problematic. My view avoids this problem. Relatedly, according to Bettcher's account, it will be in part a terminological issue whether some trans women count as women in some contexts: whether "Charla is a woman" is true or not will depend on whether the speaker is using the resistant or dominant concept, say. It is true that Bettcher thinks that there are objective reasons for choosing the resistant concept rather than the dominant concept, due in part to theoretical reasons having to do with the inaccurate worldview embedded in the dominant concept. But still, it seems to follow from this view that if one utters, "Charla is not a woman" with the dominant concept in mind, that will make that utterance true. And this seems problematic. Again, my version of subject-contextualism avoids this problem.
To conclude, we can say that Saul is right in thinking that discussions about the meaning of woman involve morally and politically significant considerations. But as I have argued, these moral and political considerations are not only relevant with respect to the question of whether we should change the meaning of our term woman (following the ameliorative approach) so that it can include trans women in its extension and therefore do justice to the claims of trans women. In my view, moral and political considerations can also be relevant with respect to the question of whether utterances of the form "X is a woman" are actually true or false. As I have explained, a modified version of Saul's contextualist proposal, namely, subject-contextualism about woman, can be used to show how moral and political considerations can help to determine the truth-value of such utterances, by determining in part what standards of similarity are relevant in a particular context, which are crucial in order to determine the truth-value of the utterance in that context, as my proposal makes clear. And as I have suggested, which standards of similarity are relevant in different contexts might depend on practical considerations, including moral and political considerations of the sort that feminist philosophers are concerned with. NOTES I am grateful to the Social Sciences and Humanities Research Council of Canada for financial support. I presented earlier versions of this material at the GRS reading group at the University of Barcelona, a session of the Society for Analytical Feminism at the Central APA in New Orleans, the MAP group at the University of Southern California, a session of the SWIP-UK at the Joint Session at the University of Cambridge, the XI Annual Estonian Philosophy Conference at the University of Tartu, and the Feminist Philosophy reading group at the University of Sheffield. I am grateful to the audiences for useful feedback. Extra thanks are due to Alex Davies, Marianne Janack, Katharine Jenkins, Meena Krishnamurthy, Shen-yi Liao, Dan L opez de Sa, Jos e Mart ınez, Rachel McKinnon, Sara Protasi, Jennifer Saul, and Pepa Toribio for helpful comments and discussion. I also wish to thank several anonymous referees for useful suggestions.
2. See Stein 1999 for an account of biological sex along these lines. 3. As I understand her, Saul does not really rule out this possibility, but rather suggests that the main worry here for the view that woman is a sex term seems to be that on this view it would be very difficult to establish which biological criteria would be relevant: "If "woman" is a sex term, then we need to know which of the biological markers associated with sex determine who is a woman. . . . It is hard to see any good, clear answer to this. What this shows is simply that using "woman" as a sex term is not a quick or easy solution: it carries with it its own problems" (198). The worry here seems to be that it is part of the folk concept of woman that there must be a good and clear response to the question of whether someone is a woman or not. In response, Dan L opez de Sa has argued (very convincingly, in my view) that our folk, pre-theoretical conception of woman, and the concept of woman that ordinary speakers actually employ, might come apart (L opez de Sa 2012). In other words, even if we accept that our ordinary conception of woman has it that whether someone is a woman or not must always have a good and clear answer, this might not be the case with respect to the ordinary concept (this divergence between concept and conception is a familiar phenomenon, he argues, which also applies to many other ordinary terms such as tree and chair). Therefore, L opez de Sa concludes that "the consideration of the intersex doesn't seem to tell against the suggestion that the concept ordinary users of 'woman' possess is one for a biological kind, as contended by the view that it is a 'sex-term'" (2). However, as we will see in what follows, this is not the only problem for the view that woman is a sex term.
4. See, for instance, Bettcher 2009. She says: "an FTM who identifies as 'trans man' may find himself represented as 'really a woman living as a man.' One obvious feature of this denial of authenticity is that transpeople are identified in ways that are contrary to or even hostile to our own self-identifications" (99). In contrast, she puts forward a view according to which "self-conception provides the criterial basis for category membership" (109). 5. Haslanger 2000 presents a definition of woman as a gender term, in terms of individuals who occupy a position of subordination along some dimensions in virtue of their (perceived or imagined) female role in reproduction, which can arguably provide a solution to this problem. However, it could be argued that this account does not do justice to the claims of trans women either. For an excellent formulation of this objection, see Jenkins 2016.
6. It could be argued that this is a problematic consequence of contextualism, since it seems wrong to say that trans women are not women in a medical context such as the one in this example. I am sympathetic to this worry, but I think the contextualist view is not committed to any particular answer here. As I will explain below, my version of contextualism appeals to the relevant normative and evaluative considerations at issue in that context. If it is true that our best normative considerations show that it is wrong to say that (some) trans women are not women in medical contexts where women are being screened for vaginal diseases (because, say, the harm caused to trans women overrides practical considerations of other sorts), then it will just not be the case that woman in this case refers to those with vaginas, but rather to those who identify as women. Therefore, contextualism (properly understood) is flexible enough to accommodate our best normative and evaluative reasoning here. Saul gives another possible example of a context where understanding woman as a sex term could be appropriate, namely, a forensic context where someone says, "This bone belonged to a woman." In this case it seems to me that understanding woman as referring to those who are biologically female is harmless. But if I am wrong about this, then the contextualist view can accommodate this, and then utterances of that sentence in forensic contexts could also be such that woman refers to those who self-identify as women.
7. Here I have in mind the kind of project that Alexis Burgess and David Plunkett have labeled conceptual ethics, that is, the project of finding out what are the best concepts to use in a particular context, given our best theoretical and normative considerations (Burgess and Plunkett 2013a;2013b). My point here is that this kind of normative reasoning is not only relevant vis-a-vis the ameliorative project of finding out which concepts we should use, but also with respect to the descriptive project of finding out what a certain term refers to, when we use context-sensitive terms whose reference depends on what the relevant standards are in each context, such as woman according to (CP) (and then it will be a normative question what the relevant standards are).
8. See, for instance, DeRose 1992 and Stanley 2005 for an elaboration of the distinction.
9. Keith DeRose claims that only attributor-contextualism is contextualism properly speaking, and argues that views that appeal to subject factors are not really contextualist, strictly speaking (DeRose 1992). I wish to remain neutral on this issue here, and I use the label "subject-contextualism" in that spirit. More recently, DeRose himself has also used the labels "'subject' contextualism" and "'attributor' contextualism" to distinguish these kinds of views (DeRose 2009).
10. Jason Stanley calls this the intention-based version of attributor-contextualism (Stanley 2005, 25). It would also be possible to formulate a version of attributor-contextualism where it is not the speaker's intentions, but rather "objective" features of the context of the speaker that determine the relevant criteria of similarity (26-27). In my view, this version of attributor-contextualism would still have the problematic consequence that Saul points out, namely, it would still be possible that a subject S1 says truly "X is a woman" (because X satisfies the standards at issue in S1's context of utterance) and another subject S2 says truly "X is not a woman" (because X does not satisfy the different standards that are at issue in S2's context of utterance), even if X's situation is the same. An interesting question here is whether this alternative view would have problematic consequences in cases such as the trans-misogynist community, where people say "Charla is not a woman" in the context of a discussion about whether she should be allowed to use women's bathrooms. According to this alternative version of attributor-contextualism that denies the intention-based view (which is similar to what Plunkett and Sundell 2013 call context-externalism), it is not just the intentions of the speaker that determine the relevant standards of similarity, so just the fact that the speaker is trans-misogynist will not automatically make it the case that what matters in this context in order to fall under woman is being biologically female or having a vagina. In what follows, I will put aside this version of attributor-contextualism, and I will focus on the intention-based version of attributor-contextualism, since it seems to better capture what Saul has in mind.
11. Saul suggests that this is especially problematic because it would have the consequence that although intuitively they seem to be genuinely disagreeing when one says "Charla is a woman" and the other responds "Charla is not a woman," this would be an illusion according to the contextualist proposal because they are just speaking past each other. However, I don't think this is a real worry (even for attributor-contextualism) because as David Plunkett and Tim Sundell explain, there can be some kind of genuine disagreement among speakers even if they do not mean exactly the same by their terms: in particular, we could understand the (appearance of) disagreement here as a case of metalinguistic disagreement, where people disagree about how the term should be used (Plunkett and Sundell 2013).