There are certain questions that seem so ordinary in therapy that they can easily pass unnoticed. “Do you feel safe?” is one of them. It is usually offered with care, sometimes gently, sometimes almost in passing. And for many people, it works well enough.

But for some autistic clients, the question can feel oddly difficult. Not distressing, exactly. Not even necessarily confusing in an obvious way. More that it is hard to know how to answer it with any confidence.

This can be easy to misread. It might look like hesitation, avoidance, or uncertainty about feelings. Yet quite often, something else is going on.

What Is Being Asked Here?

If you slow the question down, it becomes less straightforward.

What does “safe” mean in this moment?

• Physically safe?
• Emotionally safe?
• Safe with the therapist?
• Safe in the room?
• Safe in general?

And how would you know?

For many autistic people, this is where things begin to loosen. The question assumes that “feeling safe” is a clear, recognisable experience that can be identified and named. But that assumption does not always hold.

Instead, the experience might be something like:

• a sense of being slightly tense, but not threatened
  
• a feeling of uncertainty rather than danger
  
• an awareness of the therapist’s presence without a clear emotional label
  
• or simply no obvious internal signal that corresponds to the word “safe”

At that point, the problem is not “I don’t feel safe.” It is closer to:

“I’m not sure what would count as feeling safe here.”

Language and Mismatch

This kind of difficulty sits within what Damian Milton (2012) describes as the double empathy problem. Rather than locating misunderstanding within one person, it recognises that communication can break down between people who are working with different assumptions.

In therapy, language is part of that.

Many common therapeutic phrases rely on shared meanings that are rarely made explicit. They often involve:

• metaphor (“how did that land?”)
  
• open-endedness (“what’s coming up?”)
  
• broad emotional categories (“safe”, “connected”, “held”)

For some clients, these are easy to navigate. For others, they are too open, too vague, or too dependent on inference.

Research into autistic communication suggests a preference for clearer, more explicit language, where the task being asked is easier to identify (de Marchena and Gurbuz, 2025). When that clarity is missing, the question itself can become the obstacle.

When Categories Don’t Quite Fit

There is also something about the word “safe” itself.

It is not a single sensation. It is a kind of summary word, built from different elements:

• how the body feels
  
• whether there is a sense of threat
  
• how predictable things are
  
• whether there is room for choice or control
  
• how the other person is experienced

To answer the question, all of this has to come together in a way that fits the category “safe”.

For people who experience differences in interoception—the perception of internal bodily states—this process can be less straightforward (Khalsa et al., 2018). Similarly, alexithymia can make it harder to identify and name emotional states in general (Kinnaird, Stewart and Tchanturia, 2019).

So the experience may well be there. It just does not organise itself neatly under that particular label.

A Different Way In

One way to work with this is to move away from the label and towards its components.

Instead of asking:

“Do you feel safe?”

it might be more helpful to explore something like:

• Can you breathe comfortably right now?
  
• Do you feel able to say no if you need to?
  
• Do you know you could pause or stop the session?
  
• Does anything here feel threatening or overwhelming?
  
• Do you feel trapped, or do you have a sense of choice?
  
• Is your body settled, tense, restless, or something else?

These questions do not require the client to decide whether they feel “safe”. They offer a way of noticing what is actually present.

Often, a picture begins to form.

A client might say:

“I don’t feel unsafe, but I do feel a bit on edge.”

Or:

“I think I trust you, but I’m not relaxed yet.”

Or:

“I’m okay to be here, but I’d like to know what we’re going to talk about next.”

These are nuanced, usable responses. They tend to be more informative than a simple yes or no.

Making the Question Work

Sometimes it can help to say this out loud:

“I realise ‘feeling safe’ can be a bit vague. Would it help if we broke that down into a few different things, like whether you feel able to pause, speak freely, or relax your body?”

This does a few important things.

It acknowledges that the original question might not fit. It makes the therapist’s intention clearer. And it invites collaboration, rather than placing the responsibility entirely on the client to interpret and respond.

Neurodiversity-affirming approaches emphasise exactly this kind of adjustment—shaping the therapy to fit the person, rather than expecting the person to adapt to the therapy (Therapist Neurodiversity Collective, n.d.; Jones, Hamilton and Kargas, 2025).

A Small Shift, A Meaningful Difference

None of this requires abandoning questions about safety. It is more a matter of how those questions are asked.

For some clients, “Do you feel safe?” will always be enough.

For others, it may need translating.

And sometimes that translation is not about simplifying the experience, but about making space for it to be described in a way that actually fits.

References

de Marchena, A. and Gurbuz, E. (2025) ‘Communication in autistic adults: An action-focused review’, Current Psychiatry Reports. Available at: https://doi.org/10.1007/s11920-025-01616-6

Jones, F., Hamilton, J. and Kargas, N. (2025) ‘Accessibility and affirmation in counselling: An exploration of neurodivergent client experiences’, Counselling and Psychotherapy Research. Available at: https://doi.org/10.1002/capr.12742

Khalsa, S.S. et al. (2018) ‘Interoception and mental health: A roadmap’, Biological Psychiatry: Cognitive Neuroscience and Neuroimaging, 3(6), pp. 501–513. Available at: https://doi.org/10.1016/j.bpsc.2017.12.004

Kinnaird, E., Stewart, C. and Tchanturia, K. (2019) ‘Investigating alexithymia in autism: A systematic review and meta-analysis’, European Psychiatry, 55, pp. 80–89. Available at: https://doi.org/10.1016/j.eurpsy.2018.09.004

Milton, D.E.M. (2012) ‘On the ontological status of autism: The “double empathy problem”’, Disability & Society, 27(6), pp. 883–887. Available at: https://doi.org/10.1080/09687599.2012.710008

Therapist Neurodiversity Collective (n.d.) ‘Neurodiversity-affirming therapy: Positions, therapy goals, and best practices’. Available at: https://therapistndc.org/neurodiversity-affirming-therapy/