‘Autistic’ seems to be employed pejoratively in two main ways. One is meant to derogate ingroups, (i.e., the members of the group), the other meant to insult outgroups.

When it targets an ingroup, this use of ‘autistic’ derogates a particular individual because of their neurodivergence, thus aiming to derogate the whole category at once. This is what we observe in accounts like this on Reddit:

We find similar reports on blogs:

This kind of uses, according to some people, is reminiscent of slurs:

In addition to this ingroup-directed derogatory use, ‘autistic’ is also used as an insult for outgroups. When it is used to insult an outgroup, the underlying mechanism is different from the one we have observed for ingroups and is reminiscent of Jeshion’s (2013a; 2013b) “G-extending” uses of slurs.12 In such cases, the term is used to insult an alleged non-autistic subject by ascribing to her the negative properties stereotypically associated with autism.13 These uses assume and thus suggest that these properties are negative and undesirable, and that it is shameful to display them, at least for neurotypicals.14 This resembles the misogynistic practice of insulting men by ascribing feminine characteristics to them in an attempt to diminish their worth (Jeshion 2013a: 324; Bolinger 2021): e.g., when ‘like a girl’ is used as an expression that is supposed to demean a man. In both cases, it’s unclear whether the traits at stake are presented as negative simpliciter or as shameful for a specific group of people—e.g., neurotypical individuals or men. Just to give a sense of how ‘autistic’ is used to target (allegedly) non-autistic people, consider these exchanges from the gamer forum NeoGAF. In a comment to a post, a user dismisses something the original poster (OP) said by replying, ‘OP is autistic.’15

As we have remarked, in this example ‘autistic’ does not derogate its target because of their non-typical cognition and behavior; users rather attempt to insult presumed neurotypicals (the original poster) by calling them ‘autistic,’ ascribing to them the negative properties stereotypically associated with autism. When ‘autistic’ targets outgroups, thus, it conveys the idea that there is something wrong with the person and presents autism as a negatively connoted condition that people would not want to apply to themselves. In this dialogue, again from the NeoGAF forum, a guy who has just lost his mom shares his difficulties in handling and expressing his confusing feelings. He says, “Am I doing this whole bereavement thing incorrectly? I feel like I’m autistic or something.” Another user reassures him by saying “I think that you are acting perfectly normal considering your situation and that this time you aren’t being autistic.”16 Once again, ‘autistic’ stands for something from which people would want to distance themselves.

By distinguishing ingroup-directed and outgroup-directed uses, we don’t assume that speakers always know whether their interlocutors are autistic. People can very well be wrong about whether their addressee is neurotypical or not, and yet their use of ‘autistic’ as an insult would still be intended as against an ingroup or outgroup,17 thereby exemplifying the two distinct mechanisms we have just described. This sarcastic dialogue reported (or created) by a blogger exploits precisely the ambiguity between these two kinds of derogatory uses:

The first speaker means to use ‘autistic’ as an outgroup-directed insult, but their interlocutor reveals that she’s offended by this pejorative use qua ingroup.

After showing that ‘autistic’ is indeed used pejoratively in some contexts targeting either ingroups or outgroups, we proceed to spell out analogies and disanalogies with slurs, ESTIs, and derived pejoratives to shed further light on its distinctive features.

‘Autistic’ surely resembles prototypical slurs in that it targets ingroups because of their group membership and contributes to the group’s dehumanization. But how can we establish whether ‘autistic’ is a slur? Defining slurs, and distinguishing them from other pejoratives, is no easy task. As Stojnić and Lepore (2025: 5) put it: “(T)he line between slurs and general pejoratives might be more difficult to draw, and not so much linguistic, but rather socio-historical or cultural.” But one of slurs’ most distinctive feature is so-called projection.19 Take again a slur like ‘wop.’ The Italophobic content conveyed by unembedded predications such as ‘she’s a wop’ survives when the term occurs under semantic embedding, e.g., in utterances like ‘she’s not a wop,’ ‘is she a wop?,’ ‘she might be a wop,’ ‘if she’s a wop, she eats spaghetti’: all these utterances still convey some derogatory content towards Italians. That is, the semantic operators such as negation, conditional, question, modals, and so on cannot “seal” the pejorative content of slurs, which scopes out of its embedded position. Projection characterizes slurs since not all pejoratives display it. For instance, particularistic insults such as ‘asshole’ or ‘jerk’ seem to lack projection.20 Take a derogatory utterance like ‘He’s an asshole.’ If we embed the term under an operator like negation, the utterance ceases to be derogatory: ‘He’s not an asshole’ conveys no derogatory content because nothing bad is said of the referent of ‘he,’ nor does the utterance target anyone else.

How does ‘autistic’ behave with respect to projection? Assessing the question is quite tricky, for—unlike slurs—not all typical uses of ‘autistic’ are derogatory in the first place. If we restrict our focus to derogatory uses, we observe that the derogatory content of ‘autistic’ seems to project out of semantic operators. In fact, the pejorative content towards ingroups survives when this term occurs under negation, as in the example we mentioned above, ‘you aren’t being autistic.’ Similarly, it would have retained its disparaging flavor toward group members, had the original poster in the example put his thoughts in the form of a question (‘Am I being autistic or something?’). However, this projection pattern does not extend to all uses of ‘autistic,’ but only concerns the derogatory uses.

This leads us to the many respects in which ‘autistic’ departs from prototypical slurs. First, as we have said, it doesn’t seem to show the same projection patterns as prototypical slurs insofar as many of its standard uses are not derogatory in the first place, but seem merely extensional, e.g., in utterances like “[a] recent qualitative investigation into ‘camouflaging’ in autistic adults distinguishes masking and compensation as two sub-components of camouflaging” (Livingston & Happé 2017, our emphasis) (and of course, all the uses in this very paper). On the one hand, the labels that a linguistic community recognizes as derogatory do not usually crowd the pages of medical and academic books without quotation marks; on the other hand, slurs hardly have merely extensional uses and tend to be derogatory across contexts. Authors like Diaz-Legaspe (2020: 1403) would thus exclude that ‘autistic’ is a slur, for “[u]nlike non-offensive terms occurring in slurring speech, slurs are always offensive. They are so even when used in sentences not intended to belittle or disparage” (our emphasis).21 In contrast, ‘autistic’ can, but need not, be used pejoratively, which is what happens with many terms that refer to social groups without systematically conveying derogatory contents. Overall, the negative connotation of ‘autistic’ seems to arise from the context (encompassing widespread assumptions, background facts about the participants to the conversation, tone, prosody, or emoji and emoticons), rather than from the lexical item per se.22

Unlike standard slurs, it is not the specific term that is conventionally and systematically associated with pejorative contents, but rather it is how autism is conceptualized, collectively represented, and contextually used, that gives ‘autistic’ the potential to be employed as an insult.

Relatedly, ‘autistic’ lacks a neutral alternative, which slurs typically possess (‘Italian’ for ‘wop,’ ‘lesbian’ for ‘dyke,’ etc.), for the term itself can be used merely descriptively.23 When speakers choose to use slurs over their neutral counterparts, they convey a pejorative content that they would not have conveyed otherwise (Bolinger 2017; Nunberg 2018). When ‘autistic’ targets ingroups and conveys pejorative contents about them, it does not do so in contrast with a discarded non-charged alternative (Bolinger 2017).

All in all, ‘autistic’ seems to diverge from prototypical slurs on several grounds. Yet, the fact that ‘autistic’ is used to target outgroups reminds us of another class of pejorative terms, recently characterized by Castroviejo, Fraser, and Vicente (2021) as ‘Ethnic/Social Terms used as Insults’ (ESTIs). Castroviejo and colleagues make the case for a class of pejoratives that is to be differentiated from slurs, focusing on Spanish (‘portera’—literal tr., ‘doorwoman’—conveying the idea that the target is especially inclined to gossip, ‘camionero’—literal tr., ‘truck driver’—conveying the idea that the target is rude or crude). What characterizes ESTIs is that they have an extensional and neutral use, that is, they are directed at ingroups in a non-evaluative way: We can say ‘Hubert is a Nazi’ to describe the fact that he belongs to the Nazi movement, without insulting anyone. But in addition, ESTIs also have a pejorative use directed at outgroups: we can say ‘Hubert is a Nazi’ to mean that he exhibits the negative properties that are stereotypical of Nazis (e.g., he’s very strict, cruel, etc.). Take the sentence ‘Antonio è una portinaia’ (literal tr., ‘Antonio is a doorwoman’). When ‘portinaia’ is used pejoratively, it insults Antonio by ascribing him properties such as being especially inclined to gossip. But in addition, it also derogates doorwomen, by conveying the idea that doorwomen are especially inclined to gossip. By using ‘portinaia’ as if it were an insult, the speaker suggests that there is something bad in being a doorwoman.

The question then arises: Is ‘autistic’ an ESTI? On the one hand, it surely resembles ESTIs in that it has extensional uses that do not convey derogatory content, but, in addition, it can also be employed as an insult directed at outgroups, via the mechanism just described above. On the other hand, it departs from prototypical ESTIs in that it is also employed to abuse ingroups as such. In these cases, it derogates people because of their neurodivergence. Interestingly, the fact that ‘autistic’ is employed to demean ingroups may explain the fact that ‘autistic’—unlike prototypical ESTIs like ‘portinaia,’ ‘camionista’ and the like—is being reclaimed by ingroups.24

Finally, another promising candidate is the notion of ‘derived pejorative,’ introduced by Bolinger (2021) to account for how mental illness-based pejoratives like ‘schizo,’ ‘paranoid,’ etc., work.25 In her view, derived pejoratives (i) are typically used in a derogatory way when they are directed at outgroups, they (ii) represent them as temporarily deserving a negative treatment for manifesting the stereotyped property of a (typically marginalized) group to which they do not belong, and (iii) represent the invoked group as stably possessing these properties that justify their exclusion from (or subordination within) the relevant moral, political, or epistemic community. The outgroup-directed uses of ‘autistic’ resemble derived pejoratives in that they insult their target by ascribing them properties that are stereotypically associated with autistic people, assuming and suggesting that this is shameful for neurotypicals. However, one reason for us not to adopt Bolinger’s terminology to qualify ‘autistic’ is that she understands derived pejoratives to be used in a derogatory way when they are directed at outgroups, while we want to account for the way in which ‘autistic’ can be used to derogate ingroups too. A further reason is that Bolinger (2021) includes in the category of derived pejorative both a term like ‘retarded’ and a term like ‘paranoid.’ In our view, the two are very distinct: while ‘retarded’ lost its descriptive use and is nowadays systematically derogatory, ‘paranoid,’ just like our case study ‘autistic,’ still has one.

So, we find that Bolinger’s description of how derived pejoratives work fits well with the outgroup directed derogatory uses of ‘autistic,’ but we resist the adoption of the notion of ‘derived pejorative’ to classify ‘autistic.’ As we shall see in §3, what makes the case of ‘autistic’ interesting is that people are not trying to expunge the pejorative uses by “implor[ing] speakers to refrain from using the terms pejoratively,” as Bolinger (2021: 401) puts it. They are rather reclaiming it, i.e. owning it with pride, to sabotage the very possibility of using ‘autistic’ in a disparaging way, against ingroups or outgroups.

Is ‘autistic’ a slur? Is it an ESTI? Is it a derived pejorative? It seems reasonable to conclude that ‘autistic’ does not exactly behave like either of these classes of terms. Another upshot of our discussion might be that distinctions between these classes should not be drawn too strictly. The defining features of slurs, ESTIs and derived pejoratives, as well as the divide between them, is not that clear-cut, and there may well be terms that do not exactly fit the prototypical features of either, but rather display an interestingly mixed behavior. ‘Autistic’ is not the only term exhibiting such mixed behavior: other prima facie candidates are ‘Gypsy’26 in English (and ‘gitano’ in Spanish and ‘zingaro’ in Italian),27 ‘ebreo’ (i.e., Jewish) in Italian and many others.28 What characterizes these terms is that they have both pejorative and non-pejorative uses. Their pejorative uses are roughly of two kinds, against ingroups and outgroups. Both kinds of uses end up derogating the whole group, but do so via crucially distinct mechanisms, as we have illustrated in §2.1. When these kinds of terms are used as insults directed at ingroups, they are associated with pejorative contents concerning the targeted group, just like slurs and unlike prototypical ESTIs like ‘portera’ and derived pejoratives like ‘paranoid.’ Unlike most slurs, however, they lack a neutral alternative. When, on the other hand, these terms are used as insults directed at outgroups, they ascribe to their target negative properties that are stereotypically associated with ingroups.29

Wrapping up, we conclude that ‘autistic’ is not a pejorative term, but it has widespread pejorative uses. In such uses, it displays features that approximate those of slurs (especially when directed at presumed ingroups) and features that resemble those of ESTIs and derived pejoratives (especially when directed at presumed outgroups). ‘Autistic,’ just like other words that we do not analyze in this paper (such as ‘Gypsy’ in English, or ‘ebreo’ in Italian), displays a mixed behavior, falling in a gray area between different kinds of pejorative terms. How should we locate these terms in the landscape of pejorative language? We leave room for the possibility that, if there is a significant number of relevant cases, scholars will feel the need to postulate a new class of terms that behave along the same lines as our case study ‘autistic’ (i.e., having neutral extensional uses, but also derogatory ones, such that the latter can be either ingroup- or outgroup-directed, along the mechanisms we described). We have mentioned a few such candidates in a few languages, and we expect more to emerge.

Yet, going as far as postulating the existence of a new class of terms that display a mixed behavior between slurs and ESTIs might seem unwarranted, especially based on a few examples. If this is the case, then our discussion in this paper would call for a less radical move, which involves relaxing the boundaries of the classes of pejoratives currently discussed in social philosophy of language and linguistics, like slurs, ESTIs, and derived pejoratives. If we refrain from taking the distinctions between these categories too rigidly, we can more easily account for expressions that—like ‘autistic’—display mixed behavior. Moreover, this more fluid view of these classes can account for cases of transition from one category to the other. If the ingroup-directed derogatory uses of a term displaying a mixed behavior become more and more common, the term can easily become a slur, and a new expression would then be introduced to neutrally refer to the group.30

The way in which words are perceived changes over time, so whether an expression counts as a pejorative term is often relative to a given moment (and a place, or context, in general). For instance, the terms that display mixed behavior typically start as allegedly neutral expressions. We put forth the hypothesis that terms like ‘idiot,’ ‘moron,’ ‘retarded,’31 and so on, may have initially had a similar history to that of ‘autistic.’ Introduced as scientific descriptive terms since at least the 1910s (Doll 1936), they started being used (also) as pejoratives that could either target ingroups (i.e., people with developmental or intellectual disabilities), or outgroups (i.e., people with no developmental impairment, but regarded as being stupid, unintelligent, etc.). When they targeted outgroups, they could do so using the mechanism we illustrated for outgroup-directed derogatory uses of ‘autistic.’ After a time of displaying mixed behavior, their descriptive/medical use had been challenged and finally faded. The rise in pejorative uses can explain why the use of such terms is sometimes abandoned in favor of new and less problematic alternatives. The more the association between certain words and pejorative contents gets conventionalized (in the direction of slurs and insults), the more a speaker needs to find a new expression that does not carry pejorative content. This is in line with Pinker’s (1994) “Euphemism Treadmill,” i.e. the process through which an initially neutral term gradually takes on negative connotations (either because its referent is negatively valued or because the word is used as an insult), and is thus replaced with a new one. The new expression—called an ‘euphemism’—initially lacks negative connotations but will eventually undergo the same process as the previous word: it will acquire negative connotations and ultimately require replacement.32

Pinker suggests that this is a possibly infinite cycle where the new term is doomed to eventually become offensive and be substituted again, in a vain and endless search for non-pejorative expressions. However, while ‘moron’ or ‘retarded’ did not undergo a reclamation process and—as a result—their derogatory use is the only one that survives nowadays, the case of ‘autistic’ shows that reclamation processes can aim at breaking down the euphemism treadmill, by challenging the very need for alternative expressions (e.g., ‘person with autism’), as we shall see in the next section.