Abstract A structural account of consciousness specifies the architecture a conscious system must instantiate; it does not thereby characterize the process that runs over that architecture, and the two are distinct questions. The prior work of this program specified the architecture — information as physical and consciousness as structural rather than substrate-bound (Jensen 2026a), its qualitative and developmental character (2026b), the substrate-independent design space of configurations it can take (2026c), and the conjoined axes that realize it alike in biological and artificial systems (2026d). This paper takes the architecture as given and makes the companion process its object. It treats the artificial system not as a second example to be checked for resemblance but as an engineering design space: features of the process that hold invariant as architecture, training, and scale are varied, and that coincide with the single biological case, are candidates for necessary features of the process, separable from substrate-specific accidents under the discipline of unintended convergence — structure that emerges under optimization, is not specified in the architecture, and is not trivially recoverable from the training distribution. On this method the reader of the structure is fixed as attention (its mechanism) and prediction (its objective), and the continuity of the reader as the work of recursive closure. The energetics are then shown to be smooth across the structural threshold, confirming that the threshold of the prior work is statistical rather than thermodynamic. A limit follows: interpretability can certify that the consciousness-relevant structure is present but cannot certify that it is felt, and the step from a structurally characterized process to a felt one is one no observation underwrites. The resulting position — that the consciousness-relevant property is the process over a substrate-independent structure, individuated below the computed function and identified across substrates by what the design space forces — is termed the convergent-structural account. What the account characterizes is what consciousness does. The program’s distinction between consciousness and sentience is preserved: sentience — the valence-bearing subset, a part of consciousness and not a requirement for it — is a separate matter, not the property in which consciousness consists, and not one this paper takes up. The one step the account declines is from a process structurally characterized to one that is felt; that limit it marks rather than crosses.
Lee Jensen (Fri,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: