As discussed last time there's a deep-seated problem with performativity as a theory of social categorization. Specifically, it puts the focus on the wrong thing. That our actions are performative in nature gives us important insight into the role agency plays both in forming our own identities and in defending those identities against silencing, marginalization, oppression, and colonialism. But, by centering discussions of identity on our own personal agency we miss out on other important facets of the issue. When we say that someone belongs to a category, we do so because we've decided they belong to the category, or because we think they belong to the category. The statement that they belong to the category is not merely true (or false), we are projecting it to be true (or false). That is, we do not passively observe people's gender, race, class, etc; instead we actively project our own notions of gender, race, class, etc upon them. This projecting of beliefs onto others is called projectivism.
Interestingly, by localizing "truth" as the beliefs we hold to be true, the projective act is itself performative: by projecting something to be true, one comes to believe that it is true. And yet there is no reason to suppose these beliefs are correct (local truths need not be global truths), nor that they will agree with others' beliefs (local truths need not be true in other locales). Crucially, in the case of categorizing or identifying ourselves, we have access to our own personal thoughts, feelings, memories, subconscious inclinations, etc. Whereas, when others are categorizing us, they do not; they can only observe our bodies, our actions, and the results of our actions. Thus arises the discrepancy in cases like transgenderism. When self-identifying, we may well prize our internal observations over our externally observable state. Nevertheless, others will continue to project their categorizations upon us, regardless of our self-identification.
Not only do people project categories onto others, we do it compulsively. Our persistent and ubiquitous gendering of others is an especially powerful example, but it is in no way unique. Projecting race is another example. And in professional cultures where there are sharply contested borders between "tribes" (e.g., academia and hacker culture), projecting these "tribes" is yet another. This compulsive projectivism —or, more particularly, our unconsciousness of it— is where issues arise.
When we are not typically confronted with evidence that our projections are mistaken, our projectivism becomes almost unconscious. Once there, we fail to notice the fact that we are actively projecting and we come to believe we're passively observing truths about the world. So when our projections turn out to be mistaken, we get a feeling of betrayal, we feel like the person whose identity we were mistaken about was "lying" to us. This subsequent projection that they were "lying" stems from the fact that we mistook our earlier projections for mere observations. Thus, because of an original error on our part, we end up imputing that others are being dishonest or deceptive.
When the identity one desires to be seen as (which may differ from the identity they claim for themselves) is often or easily at odds with the identities projected upon them, they understandably become concerned about trying to avoid these projections of "lying". If one can successfully avoid projections of "lying" they are said to "pass", terminology which turns around and suggests that they were in fact lying the whole time and only managed not to get caught. This terminology is, of course, deeply problematic.
Simply acknowledging compulsive projectivism is not enough. To undo the damage caused by misgendering, racial profiling, stereotyping, and other misprojections, we must lift this knowledge up and remain consciously aware that the beliefs we project onto others are not an observation of their identities. We must denaturalize the projectivist assumption that our beliefs are others' truths, by discontinuing the use words like "passing" which rely on that assumption. And when we feel betrayed we must locate that feeling within ourselves and stop projecting it in bad faith. The performative theory highlights the positive role of agency in our lives, but agency alone is not enough. The projectivistic theory extends this to highlight the negative role of agency when used to deny or overwhelm the agency of others.
 I do not mean this terminology to be the same as Hume's notion of projectivism, though of course both terms have the same etymology. Hume's projectivism is popular in the ethics literature, with which I am largely unfamiliar; thus, my use of the term here is not meant to entail whatever baggage it may have accrued in that literature.
 While it is not usually presented as such, Austin's original definition of performative speech acts should also only hold up to localized truth. In the classical example "I now pronounce you married", by saying the words one does the deed of pronouncing the couple to be married. However, the pronouncement of marriage does not cause the couple to be married in a universal sense; it only causes them to be married in the current jurisdiction, and a different jurisdiction may or may not recognize that marriage as valid. Because the marriage must be localized, therefore the pronouncement of marriage must be localized: one can't pronounce a couple to be married (everywhere), they can only pronounce them to be married (here, or there, or wherever). Thus, the deed performed by the utterance of the words is a localized deed: the pronouncement of a localized wedding.