The conversation ended at four. It is now eleven and the mind is still in the room, running the exchange again, editing the one line, drafting the message that will never be sent. There is no new information. There has not been since four. The replay does not need any.
Psychologists have a name for this, post-event processing, and a large enough literature that the question is no longer whether it happens or why, but what it is doing and how to stop feeding it.
r = 0.45
pooled link between replaying social events and anxiety, across 35 studies
392
students followed for six months: the replay and the anxiety each predicted the other
137
students measured two weeks before an earthquake: the ruminators stayed low longest
1 in 8
Americans call themselves overthinkers who weigh nearly every decision
What the replay is
Susan Nolen-Hoeksema spent a career on the question of why some people stay low after a bad event much longer than others. Her answer, in the early 1990s, was not the event. It was what people did with it afterwards. Rumination, in her definition, is repetitive, passive thinking about something painful, its causes and its consequences, without the thinking ever turning into a decision or an action.
The evidence for the cost arrived with an accident of timing. Nolen-Hoeksema and Jannay Morrow had measured 137 students on mood and thinking style two weeks before the Loma Prieta earthquake hit the San Francisco Bay Area in October 1989. When they followed up ten days and seven weeks later, the students with a ruminative style beforehand were the ones still carrying symptoms at both points. The earthquake was the same for everyone. What people did with it was not.
The replay of a conversation is a specific and common form. The clinical name, post-event processing, comes from Clark and Wells's 1995 model of social anxiety, and Brozovich and Heimberg's 2008 review defined it as a detailed review of one's own performance after a social situation. The entry on rumination covers the general case. This piece is about the version with another person in it.
What 35 studies found
The single most useful number in this area comes from a 2024 meta-analysis in the Journal of Psychiatric Research. Elizabeth Edgar and colleagues at Yale pooled 35 studies and found post-event rumination correlated with social anxiety at r = 0.45, a moderate to strong association by the standards of the field, with a tight confidence interval.
- 35 studies pooled: 0.45
- Social anxiety predicting rumination six months later: 0.33
- Rumination predicting social anxiety six months later: 0.28
A correlation cannot say which came first. A 2024 study in BMC Psychology could, at least partly. Wu and colleagues measured 392 students twice, six months apart, and ran the two directions against each other. Social anxiety in October predicted more rumination in March; rumination in October predicted more social anxiety in March. The coefficients were similar. The replay is not only a symptom of anxiety. It is one of the things that keeps anxiety going, which is why the models call it a maintenance factor rather than a side effect.
That distinction changes what to do. A symptom is something to understand. A maintenance factor is something to interrupt.
Three loops that sound the same
Not everything that circles is the same loop, and the three most common ones respond to different things.
| The loop | What it sounds like | What it runs on | What closes it |
|---|---|---|---|
| Replaying | What I said, what they meant, the look at the end | A past event with no new data | Saying the sentence once, out loud, and adding that it happened |
| Rehearsing | The conversation that has not happened yet, in every version | A future event with too many branches | One dress rehearsal, then a written stop |
| Deciding | The same three options in a different order | A decision that has not been made | A deadline, written down, with a time |
Most people run mostly one. The Overthinking test separates them, and reports how loud the loop is rather than how clever it is, which is the honest measure. The story that overthinking is the tax on an intelligent mind is comfortable and unsupported: rumination correlates with distress, not with IQ.
The loop has a subject, and it is usually a person
Ask what a replay is about and the answer is almost never the economy. It is a person. What they said, what they meant, whether they are pulling away. The replay is a monitoring system, and what it monitors is a bond.
This is where attachment enters. Attachment anxiety, in plain terms, is a tendency to watch the bond closely and to treat silence as data. Give that tendency a quiet evening and it produces the replay: every version of what the pause might have meant, rehearsed in full. Saffrey and Ehrenberg showed in 2007 that anxiously attached people ruminate more after a breakup and recover more slowly because of it. Haydon and Salvatore's 2023 actigraphy study of 208 couples found attachment anxiety amplifying the link between a day's arguments and a night's broken sleep.
The other person is often replaying too. A 2026 systematic review in Frontiers in Psychology, covering fifteen studies and 2,417 participants, found rumination strongly associated with the silent treatment and with stonewalling, and found that ruminating during those silences predicted emotional exhaustion for the person doing it. Which produces a bleak symmetry: one partner goes quiet to think, the other reads the quiet as a verdict and starts the replay, and both are running loops about a conversation neither is having.
Reflection or rumination
The reflex defence of the replay is that it is diligence. Sometimes it is. The difference is not how much thinking but whether the thinking has an exit.
Edward Watkins's 2008 review drew the line between constructive and unconstructive repetitive thought, and the practical hinge is that the same mind that circles can be pointed at a concrete question and stop circling. That is what rumination-focused approaches in clinical practice do, and it is why a good therapist's first move with a ruminator is rarely to explore the content of the loop.
What it does at night
The replay has a schedule. It is quiet during the working day, because the day gives the mind other jobs, and loud in the gaps, above all at the pillow. Allison Harvey's cognitive model of insomnia describes the sequence: the day's leftover arrives, the mind begins monitoring whether it is asleep yet, and the monitoring is itself arousal, so it produces the thing it is checking for. The piece on tired but wired goes into the physiology; the piece on sleep after an argument is what it looks like when the replay's subject is a fight.
The sleep cost is the one people notice first, and it is the one that makes the replay expensive in the daytime as well, because a mind that spent the night running a conversation is worse at the next one.
Reading your own loop
None of this is treatment. Rumination that has run for months and taken the colour out of things belongs with a clinician, and the interventions with evidence behind them exist. What follows is smaller and is about reading.
The Overthinking test reads which loop is running and how loud. The Attachment Style test reads whether the loop has a person in it and why. The Sleep Quality test reads what it is costing at night. Together they tend to name a mechanism rather than a mood, and Luna reads all three when the replay needs somewhere to go at eleven at night.
Two more things that help more than they should. Give the loop an appointment: twenty minutes on paper, in the afternoon, at a fixed time, for every subject that would otherwise wait for bed. The mind releases a topic it knows has a slot. And say the replayed sentence once, out loud, to a wall if necessary, and add the words and it happened. It sounds too simple. It works more often than it should.
The first sign that any of it is working is not thinking less. It is a replay that ends in a question that can actually be sent.
Sources
- Edgar, E. V., Richards, A., Castagna, P. J., Bloch, M. H., Crowley, M. J. (2024). Post-event rumination and social anxiety: a systematic review and meta-analysis. Journal of Psychiatric Research, 173, 87-97.
- Wu, P., Cao, K., Feng, W., Lv, S. (2024). Cross-lagged analysis of rumination and social anxiety among Chinese college students. BMC Psychology, 12, 28.
- Dubey, A., et al. (2026). Antecedents and consequences of silent treatment in close adult relationships: a systematic review. Frontiers in Psychology, 17.
- Nolen-Hoeksema, S., Morrow, J. (1991). A prospective study of depression and posttraumatic stress symptoms after a natural disaster: the 1989 Loma Prieta earthquake. Journal of Personality and Social Psychology, 61(1), 115-121.
- Nolen-Hoeksema, S., Wisco, B. E., Lyubomirsky, S. (2008). Rethinking rumination. Perspectives on Psychological Science, 3(5).
- Watkins, E. R. (2008). Constructive and unconstructive repetitive thought. Psychological Bulletin, 134(2).
- Brozovich, F., Heimberg, R. G. (2008). An analysis of post-event processing in social anxiety disorder. Clinical Psychology Review, 28(6).
- Saffrey, C., Ehrenberg, M. (2007). When thinking hurts: attachment, rumination, and postrelationship adjustment. Personal Relationships, 14(3).
- Haydon, K. C., Salvatore, J. E. (2023). Relationship stress, arguments, and sleep quality: a causal process analysis. Journal of Social and Personal Relationships.
- Harvey, A. G. (2002). A cognitive model of insomnia. Behaviour Research and Therapy, 40(8).
- Talker Research for Traditional Medicinals (2025). Why 1 in 8 Americans are overthinkers. Survey of 2,000 US adults, July 22 to 25, 2025.