The unpaid second job

Autistic masking, and the hidden labour the word leaves out

I was sixty-six when a clinician finally used the right word, and my first feeling was not relief. It was closer to the feeling you get when an itemised bill arrives for a service you did not know you had been buying for sixty years. AuDHD. Autism and ADHD together, the pair of them, named at last after decades spent under labels that fit me about as well as a hire suit fits a man whose measurements were taken by someone describing him over the phone. Depression, certainly. Bipolar II, for a good long stretch. And underneath all of it, communicated in a hundred small institutional ways across my life, the quiet suggestion that I was tired because I had not organised myself properly, or had not tried hard enough, or harboured some defect of character that a sturdier man would have drilled out of himself before breakfast.

The bill, when it finally arrived, was for work. Sixty-odd years of it, unbilled until then. Work I had been doing every waking hour without being told I was doing it, without being paid for it, and while being gently faulted for finding it tiring. I want to make the case in this piece that the word we usually reach for to describe that work is the wrong word, and that the wrong word is not a harmless imprecision. It quietly sends the cost to the wrong account. Mine, as it happens. Possibly yours.

The word is ‘masking’. Sometimes ‘camouflaging’. The standard framing treats it as a behaviour, a set of social performances an autistic person mounts in order to pass as non-autistic. Eye contact rehearsed and rationed like wartime sugar. Scripts drafted in advance. A laugh deployed half a second late, because the genuine article would have arrived at the wrong pitch and frightened the room. There is nothing false in that description as far as it goes. It simply stops short, and it stops short at precisely the spot where the harm is standing.

I want to put a different word underneath it: ‘Labour’. To get there I need to borrow a concept from a man who never wrote a single line about autism, an anthropologist who spent his career thinking about bureaucracy, debt, and power, and who managed, without ever intending to, to describe what an autistic person does all day with a precision the autism literature took another decade to match.

A concept from the wrong field

The late David Graeber, in The Utopia of Rules, was trying to explain something about power. He noticed that in any relationship of structural inequality, one party ends up doing a great deal of imaginative work that the other party simply does not have to do. The person with less power has to understand the person with more power. They have to model them, anticipate them, read their moods, forecast their reactions like an anxious meteorologist, and adjust accordingly. The person with more power is under no such obligation. They can remain in a state of magnificent ignorance about how the world looks from below, and the world will politely decline to punish them for it.

Graeber called this imaginative work ‘interpretive labour’, and his point was structural rather than psychological. The asymmetry is not a charming quirk of personality, where some people simply happen to be more observant than others. It is built into the situation like a load-bearing wall. Interpretive effort runs uphill, from the powerless to the powerful, because the cost of misreading runs downhill and gathers speed. If a worker misreads the manager, the worker suffers. If the manager misreads the worker, it is Tuesday. So the worker studies the manager with the devotion of a scholar, and the manager need not learn the worker’s name. Graeber’s summary of where this leaves the powerful is worth keeping. Power, he wrote, is largely a matter of what one does not have to worry about, does not have to know about, and does not have to do.

Read that back with an autistic adult in mind, and something clicks into place with an almost audible noise. The neurotypical social world is not a neutral medium that autistic people happen to find difficult, the way some people find Tuesdays difficult. It is a structure of unequal numbers. Autistic people are something like one in fifty, surrounded by a substantial majority whose social conventions are not experienced as conventions at all, but as the bare grammar of reality, the way things simply are. And the cost of misreading that majority falls, almost in its entirety, on the autistic person. Miss a cue and you are rude. Hold eye contact a beat too long, or a beat too short, and you are shifty, or cold, or arrogant, the jury deliberating somewhere you will never be shown. Fail to perform warmth in the locally approved register and you are difficult. The majority is not fined for failing to understand you. You are fined, every single time, for failing to understand them.

So the autistic person watches. Constantly, granularly, exhaustingly. They build a working model of the neurotypical mind and run it in real time, the way a simultaneous interpreter runs a second language while the conversation, with no regard for anyone’s nervous system, refuses to slow down. That is not a performance. A performance is something you choose to mount and may choose to stop, and then go home. This is interpretive labour, extracted by a structural asymmetry the autistic person did not design, did not agree to, and cannot resign from, there being no front desk at which to hand in one’s notice. And once you see masking as labour rather than performance, three things follow that the performance framing keeps tactfully out of view.

What the labour framing makes visible

The first is that the exhaustion is not a side effect. It is the product. It is the predictable result of doing cognitively demanding work, unpaid, unrecognised, and unrested, for most of one’s waking life. Treat masking as a performance and the tiredness looks like a curious bolt-on, a thing to be tidied away with better self-care, eight glasses of water, and an app that chimes. Treat it as labour and the tiredness stops being mysterious. It is simply what labour does when it never clocks off. Nobody is baffled that a simultaneous interpreter is wrung out after a six-hour conference. We would only be baffled if they bounced out of the booth wanting to go dancing.

The research carries the weight. A 2024 systematic review and meta-analysis by Khudiakova and colleagues, pooling the available studies, found camouflaging consistently associated with higher depression and anxiety and lower mental wellbeing, the relationships holding regardless of study quality, participant age, or the proportion of women in the sample. The qualitative work points at the mechanism rather than merely the correlation. Autistic people in these studies describe monitoring social situations with a vigilance most people reserve for unexploded ordnance, worrying about whether they are doing it right, and bracing for the consequences of being found out. That is not the vocabulary of someone enjoying a costume. It is the vocabulary of someone at work, on a probationary contract, who suspects the audit is already underway.

The second thing the labour framing makes visible is autistic burnout, and why it is not simply depression wearing a high-vis vest. The clearest account we have comes from Dora Raymaker and colleagues, whose 2020 study did the unfashionable thing of building a definition out of autistic adults’ own testimony rather than out of a committee. They described a syndrome arising from chronic life stress and a mismatch between expectations and abilities without adequate support, marked by long-term exhaustion, a loss of skills, and a reduced tolerance to ordinary sensory input. What the participants called it has not been bettered since. It is, in their words, the state of

“having all of your internal resources exhausted beyond measure and being left with no clean-up crew.” (Raymaker et al., 2020, p. 132)

If masking is labour, autistic burnout is what arrives when the worker is finally, comprehensively, spent. Not sad. Spent. The distinction is not pedantry; it matters clinically, because the two call for different responses, and reaching for the wrong one wastes years. You do not recover from burnout by treating the low mood as the headline problem, any more than you resolve an overdrawn account by sitting the account-holder down and encouraging them to feel more positive about money. The thing that has run out is a resource. The resource was being spent on the labour. Raymaker and colleagues drew the conclusion that follows, and it deserves to be said without a cushion under it: there are real dangers in coaching autistic people to mask more skilfully, and burnout, they argued, belongs in the conversation about suicide prevention. That is not a small claim. It is not mine to soften or to inflate. It is theirs, and it sits in the peer-reviewed record where anyone can check it.

The third thing follows from the first two, and it reaches well past autism into territory every clinician should care about. If you ask why severed agency does damage at all, the answer is not a special autism fact. It is one of the oldest findings in the psychology of motivation, and it has been sitting in plain sight since before most of us were born. In 1959, Robert White proposed in Psychological Review that human beings carry an intrinsic drive he named ‘effectance’: the need to interact effectively with the environment, to act and to watch the action land. White gave the idea a phrase that has outlived most of the furniture of its era. A sense of competence, he argued, is

“the master reinforcer” for human beings — a motive that, unlike hunger or thirst, is never finally and fully satisfied (White, 1959).

We need, continually, to be a cause. Masking attacks that need at the root, and it does so quietly, which is the worst way. The autistic person spending the day modelling everyone else is not acting as themselves and watching the world respond to themselves. They are operating a borrowed self, a sort of well-rehearsed glove puppet, and watching the world respond to the puppet. Whatever the world sends back is addressed to the puppet, and arrives at a person who was never quite in the room. The loop White identified as basic human nourishment—act and see the effect—is precisely the loop masking quietly unplugs. This is why masking can hollow a person out while every external instrument reads normal. The job is kept. The friendships are kept. The performance, judged strictly as a performance, gets respectable reviews. And the person behind it feels less and less real, because the part of them that was supposed to be the cause of all that success never once got to put its hand on the result.

Where the picture is genuinely unsettled

I would be doing the wrong thing, and writing the wrong kind of piece, if I left it there looking tidier than the evidence is. The intuitive model, the one most autistic writing quietly assumes, is straightforwardly causal: masking comes first, the damage follows behind it like a consequence in a morality tale. A good deal of evidence is consistent with that. But a 2025 longitudinal study by van der Putten and colleagues, which followed autistic adults over time rather than photographing them once and theorising about the snapshot, found that initial mental-health difficulty did not predict whether camouflaging changed later. Other recent work suggests the masking-to-distress link is fairly modest for many autistic adults and strong mainly for a particular subgroup, those with heightened autistic traits or higher negative affect.

So the honest sentence is narrower than the satisfying one. Masking and poor mental health travel together, reliably, across cultures and study designs; that much is solid ground. The arrow between them is not yet pinned down, the strength of the link varies from person to person, and for some it may be a loop rather than a tidy line. None of this collapses the labour framing, because labour is perfectly capable of exhausting you and, on the same day, securing you something you genuinely needed, which is the entire reason anyone keeps turning up to a job. But it does mean that anyone who tells you masking simply causes the damage—and that includes me on a day when I am feeling rhetorically confident—has wandered a step ahead of the data. The framing earns its place by being clarifying and humane. It does not get to be more certain than the evidence underneath it.

What changes if we take this seriously

If masking is labour, the central question rotates on its axis. It stops being how do we help autistic people perform better, which, examined under decent light, turns out to be a polite proposal to make the unpaid second job more efficient, possibly with a lanyard. It becomes a harder and more useful question. Why is this labour being demanded at all, and from whom, and what would it take to demand rather less of it.

Look at who actually does the interpretive work in a mixed room: the autistic person, almost always, and almost entirely. They have learned the majority’s conventions, rehearsed the majority’s facial expressions, and internalised the majority’s preferred rhythm of conversation down to the pause. The majority, as a rule, has reciprocated by learning nothing whatsoever, not out of malice but because nothing in the arrangement ever required them to. That is the asymmetry. And the genuinely useful thing about naming it as an asymmetry, rather than as a deficit lodged inside one person, is that asymmetries can be rebalanced. They are arrangements, not laws of physics. Arrangements can be renegotiated by anyone who notices they are in one.

Rebalancing does not mean abolishing masking by decree, which would be its own small cruelty, since for a great many autistic adults masking is currently load-bearing, the thing quietly holding the income and the housing and the safety in place. It means moving some of the interpretive labour back across the gap, to the side that has been travelling light. A workplace that states its actual norms out loud, instead of leaving them as an unwritten code the newcomer must reverse-engineer like an archaeologist with a trowel and a deadline, has just done a share of the work itself. A manager who learns that flat affect is not hostility, that a delayed reply is not evasion, that a request for explicit instructions is not a confession of incompetence, has picked up part of the load. A clinician who recognises that the person across the desk has been working without a break for sixty years, and that not one hour of it appears in any history of presenting complaint, has at least stopped mistaking the labour for the patient.

That last one is not abstract for me, which is why I opened with the bill. For most of my life I did a second job nobody named, that I could not see clearly enough to put down, and that I was nonetheless quietly invoiced for finding tiring. Naming it now does not undo it. Sixty years are not refundable, and I have made my peace with the non-refund, mostly, on the better days. But naming it does move the fault, and the fault had been sitting in the wrong place the entire time, gathering interest. The tiredness was never evidence of a defect of character. It was evidence of work. It always was. Someone simply forgot to send the invoice, and I, obligingly, forgot to ask for it.

Work that has finally been named can, at least in principle, be shared, reduced, or refused outright. That is the whole reason the word matters, and the whole reason it is worth the argument. You cannot renegotiate a performance. Nobody renegotiates a performance. But a job, a real one, with hours and a cost and a second party who has been getting it for free—a job, you can renegotiate.

References

Graeber, D. (2015). The utopia of rules: On technology, stupidity, and the secret joys of bureaucracy. Melville House.

Khudiakova, V., Russell, E., Sowden-Carvalho, S., & Surtees, A. D. R. (2024). A systematic review and meta-analysis of mental health outcomes associated with camouflaging in autistic people. Research in Autism Spectrum Disorders, 118, 102492. https://doi.org/10.1016/j.rasd.2024.102492

Raymaker, D. M., Teo, A. R., Steckler, N. A., Lentz, B., Scharer, M., Delos Santos, A., Kapp, S. K., Hunter, M., Joyce, A., & Nicolaidis, C. (2020). “Having all of your internal resources exhausted beyond measure and being left with no clean-up crew”: Defining autistic burnout. Autism in Adulthood, 2(2), 132–143. https://doi.org/10.1089/aut.2019.0079

van der Putten, W. J., Mol, A. J. J., Radhoe, T. A., Torenvliet, C., Agelink van Rentergem, J. A., Groenman, A. P., & Geurts, H. M. (2025). Camouflaging in autism: A cause or a consequence of mental health difficulties? Autism. Advance online publication. https://doi.org/10.1177/13623613251347104

White, R. W. (1959). Motivation reconsidered: The concept of competence. Psychological Review, 66(5), 297–333. https://doi.org/10.1037/h0040934