Isolation and tortures
Why does isolation – and, far more severely, torture – cause people to lose their sense of identity?
The common explanation is suffering: enough pain, enough deprivation, and the self-breaks down.
The evidence points somewhere more specific.
A stable sense of self isn't a fixed thing. It's maintained by a loop: you predict what will happen, check that prediction against what actually happens, and quietly correct your internal model when reality disagrees. Every stable relationship, every routine, every conversation feeds that loop. It runs constantly, and mostly invisibly.
Isolation breaks this loop passively – it simply removes the feedback. The self-model has nothing left to check itself against, so it drifts. How fast, and how far, depends on what's left: Robinson Crusoe still had a demanding physical world to answer to. A sensory-deprived prisoner in solitary has almost nothing.
Torture is different in kind, not just degree. It doesn't just remove predictability – it actively destroys it. Contradictory signals. Punishment with no connection to behavior. Engineered helplessness.
Seligman's classic research on learned helplessness (1975) – later, disturbingly, adapted as a basis for coercive interrogation – found that inescapable, unpredictable aversive experiences cause deeper psychological damage than pain intensity alone.
Survivor accounts confirm this from the other side. Herman (1992) and Kroó (2013) both document the same pattern: the devastation of identity, self-concept, and basic trust in torture survivors traces specifically to the unpredictability of what they went through – not simply to how much they suffered.
Put together, this suggests identity isn't something pain erodes directly. It's something that depends on a predictable world to correct itself against. Take away predictability – passively or, far worse, by design – and there's nothing left for the self to hold its shape against.
This can reframe matters beyond the clinical literature: it says something about what a "self" structurally requires to exist at all – for anyone, or anything, built to model itself in an uncertain world.
I wonder if this resonates with those working in trauma research, predictive processing, or computational psychiatry.
If you want to participate, please contact rozum.framework [at] gmail.com
