How to Connect With Someone Who Communicates Differently

You have replayed it a few times now and you still cannot find the moment it went wrong. You asked a warm question and got a flat answer. You offered someone an easy out and they took it completely literally. Or you said the plain true thing, watched a face close, and you are still not certain which word did it.

There is probably a specific person this keeps happening with, or you may be the person it keeps happening to. Someone you like. Someone you are not trying to fix. And underneath it you have quietly settled on a story where one of you is bad at this, and on most days you have nominated yourself.

Here is a different way to hold it, and it has better evidence behind it than the story you have been using. When two people keep missing each other, the misreading tends to run in both directions at the same time, and neither person can see their own half of it.

The gap sits between you, not inside one of you

The cleanest demonstration of this comes from a study that worked a bit like a game of telephone.

Catherine Crompton and colleagues recruited nine groups of eight people. Three groups were made up entirely of autistic people, three entirely of non-autistic people, and three were mixed. One person in each group was told a story and asked to pass it along, person to person, until everyone had heard it. The researchers then counted how much detail survived each handoff.

The all-autistic chains passed information along as effectively as the all non-autistic chains did. The chains that lost substantially more detail were the mixed ones. People in the mixed pairs also reported lower rapport with the person they had been talking to.

That is worth sitting with, because it relocates the problem. If the difficulty lived inside one group of people, it would have shown up whenever those people were talking, including to each other. It did not. It showed up at the seam between two styles. This was a lab study with small groups rather than a recording of real life, so hold the size of the effect loosely. But the shape of the finding is hard to get around: what struggled was the pairing, not either person in it.

A note on where this evidence comes from. It was gathered by researchers studying autistic and non-autistic people, because that is where the mismatch is easiest to measure. None of it is a way to work out what anyone is, and you do not need a label on either of you for it to apply. What transfers is the mechanism: two sets of defaults, meeting.

It happens in seconds, and it may run on style

The second thing worth knowing is how fast the judgment forms, and what it runs on.

Across three studies, Noah Sasson and colleagues showed people brief clips of real social behavior and asked for their impressions. First impressions of autistic adults were markedly less favorable, and were associated with lower intentions to spend time with that person. The judgments formed within seconds and did not soften with more exposure.

Then the part that changes what you do with it. When the same people were judged on the content of what they said, stripped of audio and video, the unfavorable impressions largely disappeared. The authors read this as style driving the impression rather than substance.

So whatever was being reacted to was not what the person meant, or how interesting they were, or whether they were worth knowing. It was delivery: cadence, expression, the physical grammar of how something gets said. This is a finding about how observers behave. Depending on which side of it you have been on, that lands as an accusation or as a relief, and it is worth reading twice for the one that is not yours.

You are probably underestimating how much they are tracking you

The third piece tends to land hardest, because it inverts an assumption most of us carry without examining it.

Brett Heasman and Alex Gillespie asked 22 pairs, each made up of an autistic person and a family member, to rate a set of topics three ways: how they saw themselves, how they saw the other person, and how they thought the other person would rate them. That third rating is the interesting one: it measures whether you can model someone who disagrees with you.

The autistic participants were able to predict the lower ratings their family members would give, even while disagreeing with them. Meanwhile the family members tended to over-estimate how stuck their autistic relatives were inside their own point of view.

This was 44 people, so hold it as a signal rather than a settled fact. But the direction of the error is worth noticing. The person assumed to be missing the perspective of others was, in this data, tracking it accurately. The people confident they understood the situation were the ones carrying the distortion.

The same person lands differently depending on who is looking

If the first three findings suggest the gap is shared, this one shows where the leverage sits.

Kirsten Morrison and colleagues had 505 neurotypical raters form first impressions of 20 autistic adults. More of the variation in the ratings traced back to characteristics of the raters than to characteristics of the people being rated, particularly for questions about whether the rater wanted to interact. How much a rater knew about autism, and how much stigma they held, shaped the impression they formed. One rating was an exception. Judgments of awkwardness were driven more by the autistic adults themselves, which the authors report plainly rather than smoothing over.

And then a result that refuses to be tidy. Telling raters about the diagnosis generally improved impressions. Among raters holding high stigma, being told made their impressions worse.

That is a real limit, worth carrying rather than rounding off. Explaining yourself is not a tool that works on everyone. It works with people who are already willing, and it can cost you something with people who are not. This study measured strangers rating videos, so it does not tell us how disclosure plays out between two people who already know each other, which is most of the relationships you actually care about.

If you are the one who does the translating

Some readers are not the puzzled person in this article. They are the one being puzzled over.

If you are the person who rehearses messages before sending them, who runs a check on your own face, who has learned to produce the expected reaction slightly after the moment has passed, you already knew most of the above. You have been accommodating for years and it has gone unremarked, because work that succeeds is invisible.

Nothing here is asking you to drop that. For a lot of people it is doing a real job, keeping work survivable or a family peaceful, and an article cannot tell you which rooms are safe to stop in. What is worth saying is narrower. You are not obliged to be the only one adapting in every relationship you have, and it is reasonable to want one where you carry less of it.

A few things to try this week

Practical moves in the spirit of the research above, rather than ones that run on effort.

Say your own defaults out loud rather than decoding theirs. Something as small as "I am slow to answer, it is not me being cold" or "if I go quiet I am thinking, not upset". Most advice points you at reading the other person harder. This points the other way, and it is the half of the gap you actually have access to.

Treat their literal words as the message. If someone tells you they are fine and you have no specific reason to doubt it, take it, and if something in you is genuinely uneasy, that is a specific reason. A great deal of misreading comes from searching for a subtext that was never sent.

Be careful about who you explain yourself to. Not everyone earns the explanation, and the research is clear that it does not land well everywhere. Spend it on people who have already shown you some good faith.

Ask for the plain version, and offer yours. "Tell me straight, I would rather know" works in both directions, and with someone who has shown you good faith it often gives them permission they were waiting for.

Notice the impression you formed in the first ten seconds. Then set it aside deliberately. That reaction is well documented, it forms before you have any information, and you are allowed to overrule it.

Change the setting before you conclude anything about the person. A conversation on a walk, over a shared task, or in writing removes most of the cues the snap judgment runs on. If someone lands completely differently in text than in a room, that tells you about the channel, not their character.

How Murror helps you understand someone who is wired differently

The hardest part is that the confusion arrives mid conversation and dissolves before you can make sense of it.

Murror is an empathy practice built for that gap. You open up with a caring AI companion that helps you understand what you are feeling and what might be going on with the person you are trying to reach, before it is tidy enough to explain to anyone. Over time Murror gently surfaces insights about the people in your life and small, low pressure ways to show up for them, like a Moment to Care nudging you toward someone you have been missing, or a Connection that holds what you have learned about them, so you are not starting from scratch each time.

That last part matters here. A style mismatch repeats, and each time it feels like a fresh failure rather than a known pattern with a shape to it.

When something is worth saying out loud, a private reflection can become an optional takeaway you choose to share with someone you trust. Everything stays encrypted and private by default. It is a companion and a bridge toward people, never a replacement for them, and never therapy.

Two people, one gap

Nothing in this article says you are the one at fault, and nothing in it says they are. That is the point of it.

One thing this frame does not cover. If someone is dismissive of you, or you find yourself managing their reaction rather than talking to them, or the cost of being plain with them is that things get worse for you afterwards, that is not two styles missing each other. A gap is workable because both people are reaching. If only one of you is, the problem is not the seam.

The evidence keeps pointing at the space between two people rather than at either of them, and that space is workable in a way a character flaw never is. You can make your own defaults visible. You can distrust a ten second impression. You can change the channel and see who shows up in it.

And you can stop asking which of you is bad at this, which was never a question with a useful answer.