Temporal Access to Structure

Cognitive performance is often evaluated through observable reasoning, yet research on insight, associative retrieval, incubation, verbalization, and representation shows that observable sequence need not transparently reflect the process that produced an outcome. Temporal Access to Structure (TAS) distinguishes four analytically separable processes: structural formation, conscious access, representational transformation, and externalization. It proposes that sequential and non-linear cognition are different expressions of shared underlying cognitive mechanisms across a continuum defined primarily by the temporal accessibility profile of relational structure. This profile concerns both when relational organization becomes consciously available relative to structural processing and the level or grain at which access occurs. In more sequentially accessible cognition, intermediate relational states become repeatedly available as processing unfolds. In more non-linear cognition, an integrated relational configuration may become consciously accessible without equivalent availability of its constituent relations as a continuously accessible serial sequence. The framework further distinguishes solution generation, process report, justification, and post-hoc linearization, and proposes that representational transformation may sometimes alter or degrade relational information when integrated structure must be rendered serially. These distinctions constrain what can be inferred from observable reasoning and generate testable predictions concerning configuration-level access, serial decomposability, concurrent explanation, representational congruence, evaluative context, and neurodivergent cognition. The framework builds on several established component phenomena; its proposed contribution is their relation within a common access architecture and a reclassification hypothesis about sequential and non-linear cognition. TAS predicts that accessibility timing and grain should explain variation in phenomenology beyond competing variables such as restructuring, search dynamics, and metacognitive prediction error. Its central methodological implication is that externally produced sequence should not automatically be treated as a transparent record of the cognition that generated it.