What Happens in Your Body When Someone Yells

A raised voice is processed faster than almost any other social signal, and the response has largely happened before you have understood a single word of what was said. This page follows that response in time order, because the timing is what explains the parts people find most confusing about it.

20 milliseconds: before you have heard the words

Sound reaches the auditory system and a fast, coarse route sends a copy of it to the threat-detection structures before the detailed version has reached the parts of the brain that handle meaning.

That is why the flinch precedes the comprehension. You are already reacting to a shout before you know whether it was aimed at you, whether it was anger or surprise, or whether the person shouting is on the television.

The system that does this is built for speed rather than accuracy. It generates false alarms constantly, and that is the design rather than a fault in it.

100 to 300 milliseconds: the startle

Shoulders rise. The head pulls back or down. The eyes blink. The whole thing is over before you could have decided any of it.

Sudden loud sound is one of a very small number of stimuli that produce a startle response in every healthy human being from birth. Volume alone does it. Anger is not required.

1 to 3 seconds: the decision that is not a decision

Somewhere in here your system commits to a strategy, and which one it picks depends heavily on what has worked in your history rather than on what would work now.

• Muscle tension rises, particularly jaw, neck and shoulders.
• Heart rate climbs. Blood is redistributed toward the large muscles.
• Peripheral vessels constrict, which is why hands go cold and pale.
• Digestion slows. That is the drop in the stomach.
• The visual field narrows onto the source of the sound.
• Hearing changes. Some people report that other sounds recede, or that the shouting itself becomes strangely distant.

3 to 30 seconds: language goes

This is the part people are hardest on themselves about.

Under a strong enough threat signal, resources shift toward fast, automatic systems and away from the ones that handle deliberate reasoning and speech production. The practical result is that the argument you could win on paper is not available to you in the room.

You are not failing to think of the right words. Access to words has been deprioritised by a system that has decided this is not the moment for a considered reply.

For someone with a freeze pattern, this is where the shutdown arrives: the throat tightens, the mouth will not open, and the body goes still. For someone with a fawn pattern, this is where the apology comes out before any assessment has been made of whether there is anything to apologise for.

30 seconds to a few minutes: the peak, then the slow part

The initial surge tends to peak fast and start falling if nothing keeps feeding it. Adrenaline clears reasonably quickly.

Cortisol does not. It rises more slowly and stays elevated far longer, which is why you can be shaking twenty minutes after everything has gone quiet and everyone else has moved on.

This mismatch, between how long the event lasted and how long the chemistry lasts, is the root of most of what follows.

20 minutes to 2 hours: the reply arrives

The words come back. Fluently, completely, and far too late.

They come back because the threat conditions have ended and language processing has returned, not because you have finally thought of something. The reply was never a thinking problem.

For many people this is accompanied by a wave of shame that is worse than the original incident, and by replaying the exchange with the better version inserted.

That evening: the crash

Tiredness out of proportion to anything you did. A headache. Appetite gone or enormous. Jaw ache from clenching you did not notice. A short fuse with people who had nothing to do with it.

Sustaining a threat response is metabolically expensive, and the bill arrives after the event rather than during it.

The next morning, and the days after

Sleep is often disrupted, particularly the early hours. Some people wake at the same time each night for several nights after a significant incident.

You may also find the threshold has dropped. A door closing loudly, a raised voice on the radio, someone saying your name sharply, all land harder than they did last week. Threat systems recalibrate upward after a threat, and they recalibrate down more slowly.

When yelling is repeated over years

Everything above describes one incident. Repetition changes it.

A system exposed to unpredictable raised voices over a long period tends to settle at a higher baseline: scanning faces on entering a room, reading tone for temperature rather than content, apologising pre-emptively, startling at ordinary sounds. This is particularly marked when the exposure happened in childhood, when the threat-detection system was still calibrating what normal looks like.

None of that is a character trait, and none of it is a choice. It is calibration, and calibration can shift, although it does so slowly and usually with help.

The thing that has to be said plainly

Understanding the mechanism is not the same as accepting the situation. A trauma response explains why you go quiet when someone shouts. It does not make the shouting acceptable, and it is not your job to become better at absorbing it.

If you are being frightened or controlled by someone close to you, the National Domestic Violence Hotline is available on 1-800-799-7233, 24 hours a day. If you are in crisis, please contact the 988 Suicide and Crisis Lifeline by calling or texting 988.

Related: what to say when someone yells at you, is yelling abuse, being shouted at as a child, and why your hearing goes when you freeze.