The Double Empathy Problem: When Two Minds Fail to Meet

There is a patient I see, a thirty-four-year-old software architect, who describes her social history as a series of "near misses." She understands, with crystalline precision, how her colleagues organize their code, how their systems interlock, how the architecture of their project holds together or fails. She does not understand why they become offended when she points out logical inconsistencies in meetings, or why her direct questions about their reasoning are read as hostile rather than curious. She has been told, repeatedly, that she lacks empathy. She does not lack empathy. She lacks neurotypical empathy, and they lack hers.

This is the double empathy problem, named by Damian Milton in 2012, and it is among the most important reframes in contemporary autism research. The traditional model, still dominant in clinical and educational settings, assumes that autistic people have a deficit in theory of mind, an inability to understand that others have thoughts and feelings different from their own. The double empathy problem proposes something more precise and more radical: the difficulty in communication between autistic and neurotypical people is bidirectional. Each group finds the other's social signals difficult to read, each struggles to intuit the other's internal states, each experiences the other's communication style as opaque or uncomfortable. The difference is that only one of these difficulties is medicalized.

The evidence has accumulated steadily. Studies using the Reading the Mind in the Eyes test, long cited as proof of autistic social deficit, show that autistic people perform similarly to neurotypical controls when the stimuli are designed without neurotypical cultural assumptions. More tellingly, research on autistic-autistic interaction shows fluent, rapid, mutually satisfying communication that looks nothing like the "impaired" sociality observed in mixed groups. Autistic people understand each other. The problem is not a general social deficit but a specific mismatch between cognitive styles.

I see this clinically almost daily. The autistic adult who has learned, through painful trial and error, to perform neurotypical social scripts but who cannot sustain the performance without exhaustion. The one who has given up, accepting the label of "cold" or "difficult" rather than continue the effort of translation. The rare one who has found autistic community and reports, with evident relief, that social interaction can feel effortless, even joyful, when the other person shares the same perceptual and communicative assumptions.

The implications for clinical practice are substantial. I have stopped using "social skills training" that teaches autistic people to mimic neurotypical behavior without addressing the bidirectional nature of the difficulty. It is not that these programs are entirely without value; some clients find them useful for specific professional contexts. But they are incomplete, and they risk reinforcing the pathologizing frame that locates the entire problem within the autistic person. The work I find more useful involves helping clients understand the specificity of the mismatch: what neurotypical people are actually responding to, why their directness is read as aggression, why their silence is read as withdrawal. Not to change them, but to give them a map of territory that has been consistently mislabeled.

For autistic clients, this reframing often produces a specific kind of relief. They have spent years internalizing the narrative of deficit, interpreting their social difficulties as evidence of a fundamental flaw in their capacity for human connection. The double empathy problem does not erase the difficulty, but it redistributes it. They are not broken; they are different, and they are trying to communicate across a cognitive gap that their interlocutors are not even aware exists. The grief that accompanies this recognition is real, but so is the possibility of self-acceptance, and of seeking out relationships where the translation is not required.

For parents and educators, the frame is more challenging. It requires abandoning the assumption that autistic social difference is a target for elimination, and instead asking what accommodation would look like in both directions. This does not mean abandoning all social expectations; autistic people, like everyone else, need to navigate a world that is mostly neurotypical. But it means recognizing that the burden of translation has been carried almost entirely by one side, and that this is neither fair nor necessary. It means teaching neurotypical children, and adults, that autistic communication is not failed neurotypical communication but a coherent system with its own logic and its own strengths.

The double empathy problem also illuminates why autistic people so often find community with each other, and why these communities are sometimes misunderstood from the outside. The rapid exchange of information, the directness, the shared tolerance for silence or for intense focus on narrow topics, these are not deficits finding their level. They are features of a matching cognitive style, the same fluency that neurotypical people take for granted in their own interactions.

I do not want to romanticize this. The world is structured by and for neurotypical minds, and autistic people pay a continuous cost for the mismatch, in employment, in relationships, in mental health. The double empathy problem is not a solution to this; it is a more accurate diagnosis of what is actually happening. But accuracy matters. The wrong diagnosis produces the wrong treatment, and decades of wrong treatment have left many autistic people believing themselves incapable of genuine human connection. They are not. They are capable of different connection, and the work ahead is to build a world where this difference is recognized, accommodated, and valued.

— Kenji Mizukami_
Humanari Specialist in Psychology (Neurodiversity), Arcosmia Psychology