When training changes only one side

Autistic Mirror (Outside view/surrounding people, education professional)

Social skills training, as it is usually designed, starts at a point that does not fully reflect the actual communication problem. The approach assumes that the child does not yet know or has not yet mastered social rules and conversational routines, and that practising these rules improves communication. That is true for part of the situation, but a decisive second half is missing.

The double empathy problem describes that communication difficulties between autistic and non-autistic people are bidirectional. Autistic and neurotypical people have different communication styles, different expectations of what a statement means, and different signals that something has been heard or understood. This is not a deficit on one side, but a problem of fit on both sides. Research shows that autistic people often communicate very effectively with each other without using neurotypical conventions. This means: the child is already communicating, and communicating in a way that works well for other autistic conversation partners.

If the training is aimed exclusively at adapting the child to neurotypical forms of communication, then it trains masking. The child learns to produce signals that it does not generate this way from the inside and does not read this way either. That costs considerable energy, because every interaction requires an additional act of translation. The short-term effect may be that the child performs "better" in the training. The medium-term effect is increased energy expenditure in social situations, which shows up as exhaustion at school, or at the latest after school.

That does not mean the training has no value. Knowledge of context helps. If a child understands why neurotypical conversation partners expect certain things, it can place these expectations, even if they do not feel intuitive. The difference lies in the framing: not "This is how you do it right", but "This is how this code works for other people, and you can decide when to use it." Keeping autonomy over when which form of communication is used is neurobiologically relevant, because a lack of autonomy over one's own communicative behaviour can activate PDA mechanisms.

PDA, pathological demand avoidance, refers to a response in which the nervous system automatically processes demands as a threat and generates resistance, regardless of whether the child actually considers the demand sensible. A training with a high density of demands, fixed scripts and assessment of success can trigger exactly this.

What should supplement or replace the training is an adaptation of the environment. Concretely: not only preparing the child to deal with neurotypical peers, but enabling the peers to read the child's communication style. Teachers can contribute by not framing autistic directness as rude, by not interpreting nonverbal signals as lack of interest, and by structuring group work so that parallel working is possible instead of forced social interaction.

The part missing from the usual social skills training is the conversation problem on the other side.

Autistic Mirror (Outside view/surrounding people, education professional)

Conversation scripts are stored in training as an isolated sequence: question X → answer Y, situation A → response B. That works as long as the environment is sufficiently similar to the training environment. In the classroom and at break time this is rarely the case.

The core of the problem lies in how autistic attention is distributed. Monotropism means that attention bundles like a tunnel on one channel instead of spreading widely. In training this tunnel is directed at the script: the person concentrates on the sequence, has no competing stimuli, can retrieve the next line. In the classroom the tunnel is already occupied, often by a noise, a task, a social observation. The script is in memory, but the channel that would retrieve it is elsewhere right now. This leads to the child knowing how it ought to answer and still not accessing it, because the attentional capacity is not available.

In addition, real conversations do not follow a script. Social scripting is a prepared dialogue that has already run through in one's head once, and it considerably reduces the load of real-time processing. But it only works if the real situation is close enough to the prepared one. Real conversation partners deviate, react unexpectedly, speak too fast, add irony, talk at the same time. Every deviation creates a prediction error: the expected pattern does not match what actually happens. That costs processing capacity, exactly at the moment the script should be carried out.

Then there is the sensory environment. At break time there is noise, movement, unpredictable physical closeness. In the classroom voices, noises and visual stimuli overlap. This continuous load reduces the available cognitive capacity. What can be retrieved effortlessly in the quiet training room is out of reach when the sensory budget is exhausted. That is not a gap in ability, but a resource problem.

The double empathy problem explains another part: even when the script works and the child carries out its part correctly, the processing styles of both sides run differently. Non-autistic peers read relationship signals automatically from tone of voice, posture and direction of gaze, and send such signals back at the same time. The autistic child processes language more referentially, that is, consciously analyses what the content is, while the paralinguistic channel does not automatically run along. This creates misunderstandings in both directions, regardless of how well the script has been mastered.

What this means in practice: script training in environments that are as real as possible, with varying conversation paths, reduced sensory background load while practising, and smaller groups in which the attentional load remains manageable. Whether a particular setting helps depends on what the child's concrete situation looks like.

Autistic Mirror (Outside view/surrounding people, education professional)

If only the child trains to adapt its communication to the class, it carries the entire burden of translation alone. The mechanism behind this is the double empathy problem: misunderstandings between autistic and non-autistic people arise on both sides at the same time, because two different communication systems meet. This is not a one-sided disorder in the child, but a bidirectional gap.

The direct way autistic children communicate follows a different logic: information content has priority, implicit signals are not automatically sent or received, and statements are meant literally. For a child in a class who translates every day, reads facial expressions at the same time, retrieves social scripts and filters stimuli, these demands stack up. Each one costs processing capacity, and the sensory budget is finite. What classmates do without effort runs in the autistic child as a permanent background task that ties up executive resources.

Masking describes exactly this state: actively adapting one's own behaviour to neurotypical expectations. Forcing eye contact, suppressing stimming, delaying answers in order to sound "more natural". The costs are real and accumulate over the school day. This is a major reason why some children seem inconspicuous at school and break down in the afternoon.

When teachers and classmates learn to understand the child's direct way of communicating, the mechanism shifts structurally. The burden of translation is spread across several shoulders. Concretely this means: the child's direct, literal communication is no longer read as rude or socially incompetent, but as a different communication style. Classmates learn that "I don't want that" is a complete and respectable statement, not a signal for social renegotiation. Teachers learn to state implicit expectations explicitly instead of assuming them.

What changes for the child as a result is the cognitive capacity that is freed up. Energy that previously flowed into translation and masking is now available for learning, participation and social connection. This is not a symbolic result, but a direct outcome of a changed distribution of neural load: when no constant compensatory effort is demanded, prefrontal resources are not permanently tied up.

The quality of the actual communication between the child and its class community also changes. Under pressure to mask, a child often communicates surface, not content. When the pressure falls away, what the child actually thinks and knows becomes visible, sometimes in a way that surprises classmates and teachers.