Abstract

The hard problem of consciousness asks why subjective experience exists at all and how it could arise within a naturalistic account of the mind. This paper approaches the problem through computation. It argues that a specific class of mental processes cannot be realized without consciousness, because they require recursive operations over intermediate states that must remain perception-like, evaluable, and re-enterable.

The central claim is that consciousness is not an optional addition to cognition. Given the existence of structurally unbounded mental processes such as mathematical reasoning, planning, and extended reflection, subjective experience becomes necessary for the computation of those processes under the constraints of bounded human cognition.

Main Idea

Many accounts of consciousness explain correlations between subjective experience and neural or computational structures. But they often leave open the relevant counterfactual: why could the same cognition not occur without experience? This paper argues that a sufficient answer to the hard problem must establish necessity. It must show that, given a certain class of mental processes, consciousness could not fail to emerge.

The strategy is to identify mental operations that humans actually perform and ask what computational structure they require. The paper focuses on structurally unbounded mental processes: tasks whose possible continuation is not fixed in advance, such as mathematical reasoning, long-horizon planning, proof construction, and recursive reflection.

Recursive Cognition

Structurally unbounded tasks cannot be solved by storing all possible input-output mappings. A bounded system must instead use recursion: it must preserve intermediate states, feed them back into later stages of processing, and evaluate partial results along the way.

This creates a constraint. Intermediate states cannot be arbitrary hidden objects. If they are to guide ongoing cognition and behavior, they must be evaluable by the system as they stand. The paper calls this requirement cognitive type consistency: recursive intermediate states must be of a type that can be reused, inspected, and acted upon by the same cognitive system.

From Computation to Experience

The paper argues that, for the relevant class of human mental processes, these intermediate states must be perception-like. They must be present in a form that can enter the same space as perception, memory, imagination, and anticipation. This produces an expanded perceptual field: not only what is currently sensed, but also simulated possibilities, counterfactual alternatives, remembered structures, and imagined continuations.

Subjective experience arises from this structure. A conscious state is determinate because it appears as this state rather than another within a field of possible alternatives. Experience is therefore not a mysterious extra layer placed on top of cognition. It is the perceptual availability of a structured space of counterfactuals required for recursive mental computation.

Necessity and the Hard Problem

The paper then turns from the “how” question to the “why” question. If recursive, structurally unbounded cognition requires perception-like intermediate states, and if such states generate determinate subjective experience, then consciousness is necessary for that class of cognition. A philosophical zombie with the same capacities would need an alternative implementation, but the paper argues that such an implementation would either require unbounded resources, fail to preserve the required intermediate states, or reintroduce an equivalent structure under another name.

This is the modal closure of the argument. Consciousness becomes necessary relative to the cognitive capacities being explained. The explanatory gap is not closed by adding a new ontological primitive, but by showing that subjective experience has a computational role: it enables a class of recursive mental processes that could not otherwise be realized in a bounded human mind.

Why It Matters

The paper reframes consciousness as a condition for flexible cognition rather than a passive accompaniment to it. It suggests that the relevant question is not whether consciousness can be correlated with computation, but whether certain computations require the kind of determinate, perception-like field that consciousness provides.

If the argument works, the hard problem becomes less an ontological mystery and more a problem of reconstructing the exact computational structures through which conscious determinacy emerges. The remaining challenge is empirical: to identify and test the mechanisms by which the mind generates, maintains, and uses this field of counterfactual experience.

Citation

Moullec, Matthieu. 2026. “On the Computational Necessity of Consciousness.” Working paper. Target journal: Journal of Consciousness Studies.

@misc{moullec2026computationalconsciousness,
  author = {Matthieu Moullec},
  year = {2026},
  title = {On the Computational Necessity of Consciousness},
  note = {Working paper. Target journal: Journal of Consciousness Studies}
}

consciousness hard problem computation recursive cognition philosophy of mind