The Wikipedia style guide discusses how if a person states that they consider themselves a particular gender, selects a name for themselves in conjunction with a change in gender, and asserts particular gender-inflected words would correspond to them, that the article should not mention the name they had before their altered identity, and that it should refer to the person using pronouns which they have elected.
Refer to any person whose gender might be questioned with gendered words that reflect the person's most recent expressed gender self-identification... This holds for any phase of the person's life... If a...transgender...person was not notable under a former name..., it should not be included in any page...
I wish to ask a broader question of what general ethical principle this draws from, with regards to the extent to which people should comply with attested characteristics someone makes about themself, or, about someone else.
I believe a case can be made that before it became more common to think of personal pronouns as loose, variable, and customisable, a pronoun was taken to have intrinsic, semantic, propositional content. Just as the word "red" was (by many, if not all) assumed to be a signal, a token by which people could commonly understand what was being referred to (a certain colour they can see in their mind or recognise in front of them), a pronoun such as "I" was (again, commonly, yet not obligatorily) taken to have impartial semantic implicature - it meant that the person speaking was making an assertion about themself. Thus, I think previously, a word like "he" was simply a stand-in for the idea that "the person I am talking about is a boy".
If someone chooses a pronoun, it may be interpreted that they simply prefer this pronoun, in a simple sense that they enjoy it, just as some people enjoy ice cream. Yet, it seems there is to a degree an intermingling of humanity's creative liberty with words, and situations in which the veracity of words is checked by people hearing them to see if they correspond with the situation at hand, with regards to what they understand the word to mean.
Therefore, if someone creates a new pronoun for themself, and in the case that it seems to be implied that the pronoun is "semantic", that it is not a mere preference, but that it is the word that propositionally denotes what they are, it appears to be implied that they therefore have a particular gender corresponding to that pronoun, and that that gender is not "male" or "female", possibly, since otherwise, they may believe themselves to be relevant to those terms, instead.
I sometimes do not know if this distinction is particularly well recognised or articulated in contexts in which it is unflinchingly communicated that the ethically sanctioned conduct with regards to social interaction is to be defined as unconditionally obeying someone's publicly communicated "pronouns" - at least from my own experience, it does not seem common to discuss whether or not a stated pronoun is nothing other than a "preference", just as I prefer to be vegetarian and to only eat vegetarian meals, or if "preferring a pronoun" is really only a minor extension of a much more substantial, propositional claim, that of what your gender is.
I guess where I was headed with this is that if pronouns, as a class of words, are neutral with regards to human attitudes - not something you like or dislike, but merely semantic vehicles communicating what something is - that if there is ever disagreement about what pronoun should be used, it really should be reformulated as what gender you think someone is.
From this perspective, it does not seem logically self-consistent to make the blanket statement that it is highly morally incumbent to use a pronoun because someone says they think those pronouns correspond to what they are, which is ultimately an objective assertion, like saying "I am George Bush", or, "I am from outer space", etc. From this perspective, it could be seen as analogous that a person says that when in their presence, thou shalt refer to dogs as squirrels, because they think that dogs are squirrels, and you must not defy the order of the world as they see it, when you are with them.
I do not believe this is a morally consistent principle. Every human has the freedom of consciousness to assess what they think they are; a natural corollary is that every human retains the freedom of consciousness to judge and evaluate what they think anything is, as well.
Ultimately, if pronouns are something one merely prefers, as one prefers salad, it does not seem morally dire to have to agree with someone's preference. Whereas, if pronouns merely hint at an objective characterisation of what someone is, then I think the common practice in general is for it to be ok if you state disagreement with something somebody says, if you so happen to disagree.
To me, it does not make sense that people can enforce what kinds of assertions are made about them, gender-specific or not. If pronouns are customary, something that can be ordained, they lose deep semantic force. Whereas if pronouns are semantic, then they would need to be scrutinised and evaluated as such - as objective claims about the world.
Therefore, I believe that Wikipedia should not make it mandatory to comply with a person's self-attested characteristics, because if it fills the role of an information resource, I would hope one of its foremost aims would be accuracy. The editors of an article can instead discuss and try to reach a consensus about which pronoun is most corresponding to their personhood.