They Called It Stonewalling. It Was Actually Shutdown.
A note before we begin: I am currently working on a much larger guide called What the Love Lab Missed: A Neurodiversity Justice Guide to Gottman’s Relationship Research and Its Application to Neurodivergent People. It covers the full architecture of Gottman’s framework, where it has real clinical utility, where its assumptions break down for neurodivergent people, and what amended research that actually includes us might find. That guide is coming to paid subscribers in the next few weeks. This article pulls out one issue from that guide because it is, I think, the highest-stakes single problem in the entire framework, the one that has produced the most direct and concentrated clinical harm, and the one I have heard about most from autistic people who have been in couples therapy. I wanted it to have its own space.
Here is what happens in a lot of couples therapy offices, probably more than anyone who trains therapists wants to acknowledge.
An autistic person and their partner are in a session. The conversation gets hard. The emotional intensity climbs. The autistic person goes quiet. Maybe completely quiet. No eye contact. No response. Possibly their body goes still in a way that reads, to their partner and to the therapist watching, like deliberate withdrawal. Like a door slamming shut. Like someone has decided, in the middle of a conversation that matters, to simply leave.
The therapist has a name for this. They learned it in training, it appears in the literature, and it is one of the four behavioral patterns that Gottman’s research identified as the strongest predictors of relationship dissolution. The name is stonewalling.
The therapist is wrong. What they are watching is autistic shutdown. These are not the same thing. They do not have the same cause, the same internal experience, the same meaning, or the same clinical implications. Treating one as the other, which couples therapists have been doing for decades, is a clinical error with consequences that can follow a person and a relationship for years.
What Stonewalling Actually Is
Gottman’s stonewalling construct describes a relational behavior: a listener who has become so physiologically flooded during a conflict conversation that they withdraw from the interaction, shutting down social engagement as a self-protective response to overwhelm. The stonewaller goes silent, avoids eye contact, offers no acknowledgment, may leave the room. Gottman’s research identified this pattern as a strong predictor of relationship dissolution, particularly when it becomes a chronic response style.
The clinical theory behind stonewalling is that flooding, the state of physiological arousal that precedes it, can be regulated through deliberate self-soothing: a mutually agreed-upon break, some form of physiological calming, and then a return to the conversation. The antidote is not silence as an ongoing state; it is a structured interruption of the conversation that allows the flooded partner to regulate enough to return and continue. The stonewalling itself is understood as a relational communication, a signal of overwhelm that has tipped into avoidance, often carrying with it some element of emotional withdrawal from the partner as well as from the conversation.
In genuinely stonewall-prone neurotypical couples, this framework is probably describing something real. A person who habitually shuts down during conflict as a learned avoidance strategy, who has come to use silence and withdrawal as tools for managing relational discomfort, who is in some sense choosing, consciously or not, to leave the conversation rather than stay in it: for that person, the intervention Gottman prescribes makes sense. You identify the flooding early, you signal the need for a break before you reach shutdown, you regulate, you return.
The entire framework depends, however, on the shutdown being volitional enough, and motivated enough by relational dynamics, that the person can learn to manage it differently. That dependency is exactly where the framework breaks down for autistic people.
What Autistic Shutdown Actually Is
Autistic shutdown is a neurological event. It is the autistic nervous system’s response to inputs that have exceeded its processing capacity: sensory, emotional, cognitive, social, or some combination of all of them at once. When the system is overwhelmed beyond a certain threshold, it reduces. It reduces output, reduces responsiveness to external stimulation, reduces language production, reduces social engagement. From the outside, the person may appear completely unresponsive. They may go still. They may stop speaking. They may stop making eye contact. They may appear to be entirely absent from what is happening around them.
This is not a choice. It is not a message being sent. It is not contempt expressed through absence, avoidance deployed as a power move, or a relational withdrawal from the partner or the conversation. The autistic person in shutdown is not deciding not to engage. They are in a neurological state in which engagement is temporarily unavailable, in approximately the same way that speech is temporarily unavailable to someone who has just had a stroke. The absence of speech is not a statement. It is a neurological condition.
Autistic shutdown is not a choice. It is not a message being sent. The autistic person in shutdown is not deciding not to engage. They are in a neurological state in which engagement is temporarily unavailable.
Shutdown can be triggered by many things, and emotionally charged conflict conversations are particularly effective at triggering it, for compounding reasons. The emotional content of conflict requires processing that is cognitively demanding. The social demands of a conflict conversation, tracking what the other person is saying, formulating responses, managing one’s own emotional state, reading the other person’s emotional state, maintaining appropriate body language and eye contact, are each individually costly for many autistic people, and they are happening simultaneously. Add sensory inputs, the therapist’s office lighting, the ambient sound, the partner’s voice at an elevated pitch, and you have a situation that may move an autistic person toward their threshold faster than anyone in the room realizes.
Shutdown does not always arrive with warning. Many autistic people cannot reliably identify the pre-shutdown state with enough clarity and enough advance notice to take any action before it arrives. Some describe a very compressed window between “functioning normally” and “in shutdown,” with almost no middle ground in which intervention would be possible. Others describe feeling the threshold approaching but being unable to communicate that information in the moment because the communication systems are already compromised by the approach of shutdown itself. The instruction “signal to your partner when you feel flooding coming on” assumes a level of interoceptive access and real-time verbal capacity that shutdown, by its nature, tends to foreclose.
Why the Misclassification Happens
The behavioral overlap between autistic shutdown and neurotypical stonewalling is, from the outside, nearly complete. Both look like the same thing: a person who was present in a conversation and is now, by every observable measure, absent from it. No eye contact, no vocalizations, no physical engagement, no responsiveness to what the partner is saying. A therapist who has been trained to identify stonewalling by its behavioral presentation has no observational basis, without neurodivergent-specific training, to distinguish between the two.
This is not an individual therapist failure. It is a training failure, and behind that, a research design failure. Gottman’s observational coding systems were developed without any mechanism for identifying autistic participants in the sample. The behavioral pattern of shutdown, when it appeared in autistic people who happened to be in the Love Lab studies, was coded as stonewalling, absorbed into the distribution, and used to help calibrate what stonewalling looks like as a predictor of relationship outcomes. The framework that emerged from this process can’t distinguish shutdown from stonewalling because the data it was built on never distinguished them either.
What gets transmitted through clinical training is the behavioral picture: this is what stonewalling looks like, this is what it predicts, this is what to do about it. The clinical picture is accurate for neurotypical stonewalling. For autistic shutdown it is accurate only at the level of surface behavior and wrong at every level that matters for clinical intervention: causation, internal experience, meaning, and response.
What Happens When Shutdown Gets Treated as Stonewalling
The harm is specific, it is documented in the experiences of autistic people who have been in Gottman-inflected couples therapy, and it follows predictably from the misclassification.
The first category of harm is what happens to the autistic partner’s self-understanding. If a credentialed therapist, using a respected evidence-based framework, tells you that what you do in conflict conversations is stonewalling, and stonewalling is one of the Four Horsemen, and the Four Horsemen predict relationship failure, you will tend to believe them. You are in a therapy office specifically because you want to understand what is happening in your relationship. The therapist has training and authority and a clinical language for what they are observing. You come away from that session understanding yourself to be someone who stonewalls, which means someone who, at the moment of conflict, chooses relational withdrawal over engagement, someone whose pattern of behavior is actively destroying the relationship.
This is a false understanding, and it is one that may take years to undo, if it gets undone at all. Autistic people who have internalized the stonewalling frame carry it into subsequent relationships, into subsequent therapy, into their own interpretations of why connection has always been hard. The clinical language gave the experience a name, and the name was wrong, and the wrong name produced a story about themselves that fit none of their actual internal experience but that carried the weight of professional authority.
