Pattern 28 — Double Empathy

Intent

Design for mutual incomprehension rather than one-sided deficit. Treat a failed interaction as a gap between two people, and put the work of bridging it on both sides and on the environment, not on whoever is outnumbered.


The Pattern

When an Autistic person and a non-Autistic person misunderstand each other, both of them have misunderstood.

This is the double empathy problem, named by Damian Milton in 2012 — though the experience had been described in Autistic community spaces long before it had a citation.

The ‘double empathy problem’ refers to the mutual incomprehension that occurs between people of different dispositional outlooks and personal conceptual understandings when attempts are made to communicate meaning.

Damian Milton, From finding a voice to being understood

The word doing the work is mutual. Both parties experience the disjuncture. What differs is whose experience gets called a disorder. Traditionally the whole problem has been located in the Autistic brain, which is how a two-way breakdown became a one-way diagnosis of social communication deficit.

The research points the other way. Crompton and colleagues ran information along chains of eight people and found transmission was equally good in all-Autistic chains and all-non-Autistic chains — and degraded sharply in mixed chains. The failure is not in a person. It is in the crossing.

Autistic sociality is largely interest-based; the dominant form is built on social alignment and group identification. Neither is deficient. But when an interest-led person is asked to operate in an alignment-led room, the mismatch reads as the person going wrong, and the exclusion that follows is recorded as evidence about them.

This is not a claim that Autistic people automatically understand each other. Two strangers are two strangers. It is a claim about where the disjuncture widens, and about who has been made responsible for closing it.

The double empathy problem decentres the idea of normalcy. It replaces a deficit with a gap, and a gap has two sides.


Why It Matters

When one party is held solely responsible for mutual misunderstanding, people may:

  • be diagnosed for the shared failure of an interaction
  • be judged incompetent in job interviews, classrooms, and clinical assessments built around one social style
  • be read as cold, rude, or uninterested by observers who never doubt their own reading
  • be taught to perform an alien sociality and to call that progress
  • experience thwarted belonging, which is associated with the very worst outcomes
  • be misjudged in the criminal justice system, where the cost of being misread is highest

Interpretations of Autistic sociality drawn from observation alone are frequently wrong. The observer’s confidence is not evidence.


Forces

  • Misunderstanding between differently disposed people is normal and goes both ways.
  • Only one side’s confusion has historically been pathologized.
  • Within-neurotype communication is measurably more efficient than across it.
  • Autistic sociality is interest-based; dominant sociality is alignment-based.
  • Numerical majority gets mistaken for correctness.
  • Non-Autistic people form negative impressions of Autistic people within seconds, on thin-slice judgments.
  • Believing you have understood someone feels identical to having understood them.

Problem

How do we design interactions so that the cost of a two-way misunderstanding is not charged to one side?


Solution

Assume misunderstanding is mutual and that your own reading may be the one that is off. Build in checking rather than inferring. Distribute the translation work instead of assigning it to whoever is in the minority. Where possible, let people communicate within their own neurotype rather than across a gap nobody is bridging.

Approach difference with humility rather than remediation.

This works in concert with Masking Pressure, Social Energy, and Exposure Anxiety.


What This Looks Like

In physical spaces

  • Make room for interest-led connection, not only mingling and small talk.
  • Do not treat eye contact, posture, or facial expression as evidence of engagement or honesty.
  • Create spaces where Autistic people can be together, and do not treat that as self-segregation to be corrected.
  • Let people ask clarifying questions without it reading as challenge.
  • Avoid interview and assessment formats that measure conformity to one social style.

In digital spaces

  • Prefer explicit text over inferred tone, and supply the tone you mean.
  • Let people state plainly how they communicate and what they need checked.
  • Do not read brevity as coldness or length as aggression.
  • Allow written contribution where speech would force a crossing.
  • Make it normal to say “I am not sure how you meant that” without it being a confrontation.

In social and organizational practice

  • Say what you mean rather than signalling it and expecting it to be decoded.
  • Check your interpretation out loud before acting on it.
  • Treat “that is not what I meant” as information, not as a dispute.
  • Share the translation work; do not assign it permanently to the same people.
  • Stop using social-skills training as the default response to a two-way problem.
  • Include Autistic people in judging whether an interaction went well — the observer is not the authority.

Implementation Notes

Start by noticing where your environment assumes one social style is neutral.

Ask:

  • Who is expected to adapt here, and who is expected to be understood?
  • Are we assessing competence, or conformity to a communication style?
  • When an interaction fails, whose account of it is believed?
  • Is anyone doing constant, unpaid translation work?
  • Could this exchange have happened within a neurotype instead of across one?
  • Do we check interpretations, or only act on them?

A good test: when a misunderstanding happens, count how often the non-Autistic party is asked to change something. If the answer is never, the environment has not accepted that the problem is double.


Signals of Misuse

This pattern is being violated when people hear:

  • “They lack theory of mind.”
  • “He just doesn’t pick up on social cues.”
  • “She needs social skills training.”
  • “I could tell from how they were sitting.”
  • “Everyone else understood me fine.”
  • “You need to meet people halfway.” — said only to one side

The last one is the pattern in miniature: halfway, measured from where the majority already stands.


Consequences

When this pattern is honored:

  • misunderstandings get repaired instead of diagnosed
  • competence becomes visible because it is no longer being assessed through a mismatched channel
  • the translation burden is shared, and so is the fatigue
  • Autistic-to-Autistic connection is recognized as connection, not withdrawal
  • assessments measure what they claim to measure
  • belonging stops depending on passing

When it is ignored, an entire population is held responsible for the confusion of everyone who meets them.


Anti-Patterns

  • Social-skills training as the default remedy for a mutual problem
  • Theory-of-mind deficit framing
  • Eye contact, handshakes, and posture as proxies for competence or honesty
  • Interview formats that score social alignment and call it culture fit
  • Observational assessment without the Autistic person’s own account
  • Treating Autistic community space as a problem to be integrated away
  • “Indistinguishable from peers” as a goal

Design Moves

Design strategies that share the work of being understood.

See recipes:


Related Environments

Environments where the double empathy problem shapes access and participation:


Related Experiences

Experiences shaped by whether misunderstanding is treated as mutual:


Patterns Below


Patterns Above

Masking is what the double empathy problem costs when only one side is required to close the gap.


Related Patterns


Related Principles

  • Nothing About Us Without Us
  • Multiple Communication Channels
  • Design for Range
  • Consent Over Compliance

See Also