Alexithymia: Being unable to Identify Emotions due to Interception Issues
- 5 days ago
- 8 min read
You can dissect a novel's emotional architecture, talk a friend down from a spiral, and name every current running through a room, except the ones running through you. Alexithymia is not a missing heart. Sometimes it is a missing label, and sometimes the feeling never registered in the body at all. Both can be worked on, once you know which one you are dealing with.
Alexithymia is difficulty identifying and describing your own emotions while they are happening. The Greek translates to roughly “no word
s for feelings,” which is a little misleading, because the problem is not that you have lost the vocabulary or forgotten what an emotion is. It is that you cannot tell which feeling is running, or whether one is running at all. And for a lot of autistic people that failure starts earlier than language, down at the level of the body itself.

Think of recognizing an emotion as a two-stage job. First the body throws a signal: the heart lifts, the gut tightens, the chest goes heavy, the jaw sets. Then the brain reads that signal and stamps a word on it, that is fear, that is anger, that is grief. Alexithymia can break at either stage. In one version the signal fires clearly and never gets labelled, so the feeling is fully present and only the translation fails. In the other, common in autism, the signal is faint or does not register in the first place, which is interoception, your sense of your own internal state, running low. You cannot name a feeling you never detected.
This is why the usual verdict, that you are cold or detached, lands close to the opposite of the truth. Often the emotion is there and doing its work in the body; what is missing is either the readout or the reception. It is also why “just notice what you feel” is useless advice in both cases, because the noticing is the exact part that is broken.
Clinically, alexithymia has three recognized parts:
a) Trouble identifying feelings
b) Trouble describing them and
c) Externally oriented thinking, attention aimed outward at facts and tasks rather than inward at your own state.
It is a construct, not a diagnosis; neither the DSM nor the ICD lists it as a disorder. It runs at roughly one in ten in the general population and far higher in autism, where Kinnaird and colleagues (2019) put it near half. It also travels with trauma, depression and long stretches of overload, and can be lifelong or acquired.
You cannot name a feeling you never detected.

Five ways it actually shows up
The textbook example, someone answering “fine” on autopilot, is the least interesting version and probably not yours. Here are five that tend to ring truer.
1. You reconstruct the feeling after the fact
You work out what you felt the way a detective works a cold case, backwards, from the evidence. You snapped at two people, quietly rearranged your whole week, could not settle at anything, and so, three days later, you conclude you must have been frightened about the result. The reasoning is airtight. The trouble is that the feeling never announced itself in real time. You inferred its existence from the wreckage it left behind, like reading a storm from the fallen branches.
2. The feeling arrives dressed as a medical emergency
Chest tightness, a churning gut, a flat dread convincing enough that you book the appointment and half expect the scan to find something. Every test comes back clean. It was grief, or fear, presenting as a body event, because the emotional channel has no bandwidth and the signal takes the physical route instead. You can hold a detailed, accurate model of your own physiology and still not join these particular dots, because the part that was supposed to generate the label and association simply never fired.
3. Fluent in the third person, mute in the first
You can write with real feeling about someone else's inner life. You can walk a friend through a breakup with precision, naming exactly what they are going through before they can. Then someone asks what you feel and the screen goes to static. The vocabulary is all present and correct. It just does not index to your own interior. You are a translator who is fluent in every language but the one spoken at home.
4. “What do you want” returns nothing
Two genuinely good options, and you stall. Not because you cannot reason, you can list every trade-off cleanly and fast, but because the tiebreaker is supposed to be a preference, a pull toward one of them, and that reading comes back empty. So you default, or you decide by external logic, and then discover a month later that you resent the outcome you engineered. Wanting is a feeling too. If feelings do not read, neither does wanting.
5. The delayed detonation
Something lands, a slight, a loss, a betrayal, and you are calm, competent, almost clinical about it. People remark on how well you are holding up. Then a week later it arrives at full volume, apparently out of nowhere, and wildly out of scale with whatever small thing is actually in front of you at the time. The processing did happen. It ran on a delay you could not see.
Starting from the body
Since the feeling will not surface on demand, work backwards from the body instead. But there is a fork here, and which branch you are on decides where to start. If your body throws clear signals you simply cannot label, you can go straight to mapping them. If the signals themselves barely register, you have an earlier job first: learning to feel them at all.
If the signal barely registers, build detection before labelling
For the interoception-driven version, mapping sensations to feelings is premature, because there is not yet a reliable sensation to map. The prior skill is registering the body at all, and it is trainable, slowly. Start with the loud, unambiguous signals rather than the subtle emotional ones: notice your heartbeat after stairs, notice hunger, thirst, needing the toilet, cold, heat, tiredness. When you start to notice these sensations a little more, start by honoring them and attending to them. If you are hungry - eat, tired - go to sleep early, cold - don't push through your physical sensation like a warrior, put on a jumper.
Practice care and respect for your body in the small increments where you start to engage with your personal sensations more. Hopefully in time, through this practice, fainter emotional signals may gradually become easier to detect. External proxies help when the internal read is offline: a smartwatch showing your heart rate has climbed, catching that your shoulders are up around your ears, noticing you have gone mute.
Treat the body like a faint radio station you are slowly tuning into, not one you can already hear clearly.
Change the question
Once there is a signal to work with, change the question you ask. Not “what am I feeling,” which returns nothing, but “where in my body is something happening, and what is it doing.” Scan the usual sites: chest, throat, gut, jaw, shoulders, hands, face. Then note temperature, weight, speed, tension. You are collecting physical facts, not emotional words, and physical facts you can actually get hold of. Write down these physical sensations into a journal or the Notes App on your phone. This is help you collection this information a and identify patterns later.
Build your own legend
Emotions do have somatic signatures, but yours are specific to you, so treat this as fieldwork with yourself as the subject. Log three columns: the situation, the body sensation, and, once you eventually know it, the label. Over a few weeks a lookup table assembles itself. “Hot face, fast chest, urge to leave the room” resolves, for you, to anger. This suits a mind that trusts evidence over instruction: you are running a study, and you are both the researcher and the instrument.
Borrow the research as a first guess
Nummenmaa and colleagues (2014) mapped where people feel emotions in the body and found strikingly consistent patterns: anger and fear activate the chest and upper body, sadness drains the limbs and pools in the chest and throat, happiness registers almost everywhere at once, disgust concentrates in the throat and stomach. Use those as opening hypotheses, then correct them against your own logs, because your map may not match the average and it is your map that matters.
A ROUGH STARTING LEGEND
Heat in the face, fast chest, tight hands, urge to act | likely anger or frustration |
Tight chest, cold, a drop in the gut, shallow breath | likely fear or anxiety |
Heaviness, aching throat, tired limbs, pressure behind the eyes | likely sadness or grief |
Restlessness, churning stomach, cannot sit still | likely anxiety or anticipation |
Heat climbing to the face, the wish to shrink or vanish | likely shame or embarrassment |
Lightness, warmth, the urge to move or expand outward | likely joy or excitement |
A population-average starting point to test and overwrite, not a reading of your specific interior. It also assumes you can feel the sensation in the first place; if you cannot, that is the detection problem, and this legend is step two, not step one.
Get a baseline first
You cannot detect a spike without knowing your resting state, so read the body at neutral moments too, not only at the flashpoints. Boring data is still data, and half the value of the log is knowing what “nothing much” feels like so you can tell when it changes.
Work with the outward pull, not against it
Since attention wants to point outward, give it external scaffolding rather than relying on introspection you may not have. Set a fixed time to run a body scan. Keep a note on your phone. Use a checklist of sensations instead of a wheel of emotion words. Outsource the noticing to a system so it does not depend on a lookup that keeps returning empty.
Let the label lag, and leave the box blank when it will not come
The first win is “something is happening.” The second is “it has a shape.” The third, sometimes days later, is “this shape usually means X.” A blank entry is a real result, not a failure. It tells you the signal was below your threshold this time, which is itself worth knowing. And there is a payoff beyond self- knowledge: naming a feeling reduces its intensity. Affectlabelling research (Lieberman and colleagues) shows that putting a state into words quietens the brain's emotional response. The map is not only a dictionary. It is a way to turn the volume down.
What the map fixes, and what it does not
Straight version. The labelling stage responds well to this work: with a sensation-to-label map and practice, identifying and describing feelings both improve, and emotional granularity, the fineness of the distinctions you can draw, is trainable. The detection stage is slower, because you are rebuilding an input rather than just a translation, and for some people the signal stays quieter than average no matter how much they practise. And the third part, externally oriented thinking, barely moves at all; attention that points outward tends to keep pointing outward.
So the realistic goal is not to feel feelings the way the greeting cards promise. It is a working link between a body that produces signals and a mind that can register and read them, built slowly, from wherever your particular break sits. The quieter thing this retires is the story that you are cold, or broken, or missing the part everyone else was issued. You were not missing the feeling. You were missing the reception, or the readout, and both can be built.

Wired Differently Australia publishes educational content on neurodivergent experience. It is general information, not individual clinical guidance.

Comments