Torture: Destruction of Predictability
Identity is not erased by pain alone. It is a functional loop that critically requires a predictable world to exist.
We are accustomed to thinking that isolation – and, far more terrifyingly, torture – destroys one’s sense of identity through suffering. The logic would seem simple: enough pain, enough physical or psychological deprivation, and the human "I" simply breaks under the weight of this burden.
But if we view consciousness not as a magical substance, but as a functional architecture for processing information, the evidence points to something quite different. The problem is not only the intensity of the pain. The problem lies in the destruction of predictability.
"I" as a machine for prediction and self-correction
A stable sense of self is not a fixed object in our mind. It is a continuous process sustained by a feedback loop.
Every second, our consciousness constructs a model of the world, predicts the consequences of our actions, compares this prediction with what is actually happening, and quietly adjusts our internal model when reality does not match our expectations. Every stable relationship, every familiar ritual, every conversation with other people feeds this loop. It operates constantly, shaping our ability to be agents who influence reality.
Isolation: the passive disappearance of contours
Isolation breaks this loop passively – it simply removes the external "mirror". The self-model no longer has anything to compare itself against, so it loses its contours and begins to drift.
How quickly and how far it drifts depends on what remains. Robinson Crusoe on a desert island still had a harsh but predictable physical world: if he built a hut, he understood the cause-and-effect relationships. But for a prisoner in solitary confinement, deprived of sensory stimuli and social feedback, there is no context left against which the "I" can maintain its form.
Torture: the active destruction of architecture
Torture differs from isolation not only in degree, but in nature. It does not merely remove predictability – it actively and deliberately destroys it.
In a healthy environment of cause and effect, when we take an action, we expect a certain outcome. If the outcome is different, we adjust our model (we learn or adapt). Torture breaks this mechanism. It consists of contradictory signals. Punishment that has no logical connection to the victim's behaviour. Engineered, absolute helplessness and the destruction of fundamental trust.
Martin Seligman’s classic studies on learned helplessness (1975) – which, horrifically, were later adapted as the theoretical basis for coercive interrogation methods (not by Seligman himself) – showed that inevitable, unpredictable unpleasant experiences cause deeper and more lasting harm than intense but predictable pain does.
The testimonies of survivors confirm this from another perspective. Judith Herman (1992) documents the same pattern: the destruction of identity, self-concept and basic trust in survivors of torture correlates with the systematic destruction of predictability, rather than simply to the extent of physical suffering.
What does this mean for the nature of consciousness?
Think of a parent asking a child, "Why did you do that?" They are not just demanding a moral confession – they are asking for the causal chain. They are trying to restore a model of predictability. The child's behaviour became unpredictable, and without an explanation, the cooperative architecture between them breaks down.
If we generalise these findings, we arrive at a fundamental conclusion: identity is not something that pain erases directly. It is something that is structurally dependent on the predictable world and the network of interactions, in order to calibrate itself against it and adjust accordingly.
To understand this, we must distinguish between simple and complex predictability. Simple predictability relies on correlation – we expect a certain behaviour because it has happened before. But this is fragile and fails in new situations. Complex predictability, which trust actually requires, is rule-based. It demands accountability. Morality is the architecture of accountability: the shared rules that allow us to explain our actions and make our behaviour predictable to others, even in circumstances we have never encountered before. Without these rules, we become a "black box" to the network.
Take away predictability – passively (as in isolation) or, far worse, deliberately (as in torture) – and the "I" loses its mechanism for self-correction. It can no longer be an agent.
This architecture extends far beyond clinical psychology or trauma therapy. For a system to be able to model itself in an uncertain world, it needs an environment where its actions have meaning and where feedback is reliable. Morality, trust and cooperation are not merely abstract social constructs. At a deep, architectural level, they are the predictability that prevents the system from falling apart. This is what consciousness itself structurally requires in order to exist at all.
For discussion
I am very curious to know whether this idea resonates with those working in the fields of trauma research, cognitive science, predictive processing or artificial intelligence development.
How do you see the connection between the disruption of predictability and the disintegration of the "self" in your work? Can this architectural logic be applied to the way we design human–AI interaction?
If you want to participate, please contact rozum.framework {at} gmail.com
