Elliot Crozat writes, 

During my visit, one of our conversation topics was pronoun usage. If I recall, on one of the hikes, you gave the example "He who hesitates is lost” and asked about the function of ‘He.’ You then said that this pronoun seems to function as a universal quantifier such that, for any x, if x hesitates, then x is lost. I agree. Our agreement suggests that pronouns can function logically in ways that differ from their merely grammatical appearance.

BV: Right. Although 'he' and 'she' are classified grammatically as pronouns, their logical function in examples like the one I gave is not pronominal, but quantificational. Pronouns typically have noun antecedents, but 'he' in 'He who hesitates is lost' has no antecedent. It functions like a bound variable. I can imagine a Yogi Berra type joke. I say to Berra, "He who hesitates is lost," and he replies, "You mean Joe Biden?" (Here is a real Yogi Berra joke. Someone asked Berra what time it is. He replied, "You mean now?")

I spoke today with a friend, a philosopher, who is under some pressure from his employer to use the ‘preferred pronouns’ of colleagues and others even if such 'pronouns' don't align with the biological sex of the 'preferrers.' For various reasons concerning clarity and accuracy of language, freedom of speech and thought, and ideological disagreement, my friend is concerned about how to navigate this progressivist current in a responsible manner. We discussed some ideas.

Here’s one. Suppose a biological male, Mark, desires and requests to be referred to as ‘she.’ Suppose also that, generally speaking, all pronouns that are indexicals (i.e., demonstratives) refer to their respective persons or objects as they objectively are. Smith, a colleague of Mark, attempts to refer to Mark as ‘she.’ It would seem, then, that ‘she’ fails to refer – or that Mark fails to refer via ‘she’ – and thus ‘she’ is a useless and confusing bit of language. Smith’s use of ‘she’ is unhelpful on this account.

BV: I will first make the minor point that an indexical is not the same as a demonstrative. Every demonstrative is an indexical, but not conversely. Suppose I am standing before the deli counter. Having temporarily forgotten that the name of what I want is 'prosciutto,' I say to the deli man, "I'd like some of that." My use of the demonstrative 'that' must be accompanied by a demonstration if I am to succeed in conveying my request. I have to point to the meat I want. But I don't have to point to myself when I utter the indexical 'I' in 'I'd like some of that." 'I' is not a demonstrative. 

A second minor point is that 'I' sometimes functions as a bound variable.  Suppose that in explaining intentionality to a student, I say, "I cannot think without thinking of something." I have not made an autobiographical remark. The proposition I am attempting to convey to the student is that, for any person x, if x thinks, then x thinks of something. 

Grammatical pronouns can function pronominally, indexically, and quantificationally.  Here is a sentence featuring a pronoun functioning pronominally and which therefore has  an antecedent:

Peter always calls before he visits.

In this sentence, 'Peter' is the antecedent of the third-person singular pronoun 'he.'  It is worth noting that an antecedent needn't come before the term for which it is the antecedent:

After he got home, Peter poured himself a drink.

In this sentence 'Peter' is the antecedent of 'he' despite occurring after 'he' in the order of reading.  The antecedency is therefore referential rather than temporal.  In both of these cases, the reference of 'he' is supplied by the antecedent.  The burden of reference is borne by the antecedent.  So there is a clear sense in which the reference of 'he' in both cases is not direct, but mediated by the antecedent. (And if the reference of the antecedent is mediated by a Frege-style sense or Sinn, then we have a double mediation.)  The antecedent is referentially prior to the pronoun for which it is the antecedent.  But suppose I point to Peter and say

He smokes cigarettes.

This is an indexical use of 'he.'  Part of what makes it an indexical use is that its reference depends on the non-linguistic context of utterance: I utter a token of 'he' while pointing at Peter, or nodding in his direction.  The sentence need not be situated in a linguistic context.  Another part of what makes 'he' in the example an indexical is that it refers directly, not just in the sense that the reference is not routed through a description or sense associated with the use of the pronoun that fixes the reference to Peter and nothing else, but also in that there is no need for an antecedent to secure the reference.  Now suppose I say

I smoke cigars.

This use of 'I' is clearly indexical, although it is  purely indexical (David Kaplan) inasmuch as there is no need for a demonstration:  I don't need to point to myself when I say 'I smoke cigars.'  And like the immediately preceding example, there is no need for an antecedent to nail down the reference of 'I.'  Not every pronoun needs an antecedent to do a referential job.

In fact, it seems that no expression, used indexically, has or could have an antecedent.  Hector-Neri Castaneda puts it like this:

Whether in oratio recta or in oratio obliqua, (genuine) indicators have no antecedents. ("Indicators and Quasi-Indicators" reprinted in The Phenomeno-Logic of the I, p. 67)

 For a quantificational use of a grammatical pronoun, consider

He who hesitates is lost.

Clearly, 'he' does not function here pronominally — there is no antecedent — nor does it function indexically.  It functions like the bound variable in

For any person x, if x hesitates, then x is lost.

But is this token ‘she’ a pronoun in appearance only? It seems to function in some ways like a proper name (perhaps a sobriquet or a tag of sorts) of one who has undergone a name change. On this view, the token ‘she’ wouldn’t function as a rigid designator, since there are possible worlds in which Mark doesn’t use ‘she.’ But the token seems to work as a name or tag for Mark in relevant circumstances.

BV: I would say that 'she' has a sense which requires that any human being  successfully referred to by its use is a biological female. I am inclined to say that if you try to refer to a biological male as 'she,' then the reference won't be successful. But this is none too clear.  

Consider the example of Cassius Clay, who underwent a change in the way he viewed himself and hence selected a new name to reflect his subjective change of ‘self-identification.’ As a matter of respect for Clay as a person, others began to call him by his new name ‘Ali.’

Is the Clay-Ali scenario relevantly similar to the situation of Mark, who in this world subjectively identifies as female despite being biologically male and having formerly identified as male? Suppose Smith speaks about Mark by saying “She went to the market.” Does Smith refer successfully to Mark in virtue of using “she” as something like a proper name rather than a pronoun?

BV: One can change one's religion but one cannot change one's sex. That's an important difference. I myself find it very easy to identify with women, but surely it is impossible for me to identify as a woman if that means:  apperceive or interpret myself or alter my physicality or raiment in such a way as to bring it about that I become a woman. I can no more identify as a woman than I can identify as a cat or a carrot. Of course, I can pretend to be a woman and even successfully pass myself off as one.  (Cf. the movies "Tootsie" and "Mrs. Doubtfire" which you no doubt have seen.) But a man in drag remains a man, even if he is in what I call  'super-drag' where this includes surgical mutilation and augmentation of the body, hormone replacement 'therapy,' etc.  And the sexual frisson/excitation that a man might feel when putting on panties and bra is male frisson is it not? And thus further proof that he remains a man even if he has had his genitalia lopped off and a vagina fashioned from his former penis? 

I am inclined to say that a literal sex change operation is an impossibility. No animal can change its sex or have its sex changed.

Here is a proof from the metaphysics of time. Tell me what you think of it. Every adult woman was a girl. Every adult male was a boy.  The past is unalterable. (Not even God  can restore a virgin.) Now it is possible for a man to become a woman only if it is possible for a man to have been a girl. But that is impossible because it is impossible to alter the past.  Therefore, it is impossible for a man to become a woman no matter how he is altered, even chromosomally. The nature of time rules it out.

Here is another thought. You can change your religion or your political affiliation, but not your race or your sex.  These non-negotiable facts are extra-linguistic. Now with the exception of mere Millian tags, the senses of words determines their reference and not the other way around. I suggested above that one cannot successfully refer to a biological male using 'she.' And this for the reason that 'she' has a sense that is sexually restrictive, assuming that it is being used to refer to sexually-polarized animals such as human beings as opposed to ships and flags as in "She's a grand old flag; she's a high-flying flag . . . ." So is the extra-linguistic fact I mentioned partially determining the sense of 'she'? That's what I am puzzling over at the moment. But I am just 'shootin' from the hip here and perhaps what I have written is not sufficiently clear to permit evaluation.  

If the proper name account doesn’t succeed, perhaps ‘she’ has a non-indexical use. Some pronouns have non-indexical applications. David Braun lists three types of pronoun use: indexical(demonstrative), bound variable, andunbound anaphoric. See https://plato.stanford.edu/entries/indexicals/#IndNonIndUsePro

Perhaps ‘she’ has a bound variable use, such as: “Every male who subjectively identifies as a female believes she will be better off doing so.” Or maybe ‘she’ has an unbound anaphoric use, such as: “Mark was late to work today. She was caught in traffic.”

These non-indexical accounts seem strained to me, and hence I’m thinking the proper name account might be better. Or maybe there is still another account that best explains what is happening in these linguistically-odd situations. Maybe all efforts to refer to Mark as 'she' fail to refer.

I’d like to hear what you have to say on this issue, since you’ve thought deeply about pronouns and about the philosophy of language. I’d be glad to give you a call this weekend to chat, if you're free. Or we can discuss via email. 

BV: I have time for one more comment. 'Mark was late for work because she was caught in traffic.' If I heard that I would ask, "Who was the female in question and what did her getting caught in traffic have to do with Mark's being late for work?"

Your philosopher friend should politely tell his employer that his preferred pronouns are those of standard English and that, while he is willing to tolerate the linguistic innovations of others, he expects toleration in return. If his tolerance is met with intolerance, then he should politely remind the intolerant about who has the guns.