Intentionality—the aboutness of mental states—has resisted naturalistic explanation partly because most theorists treat it as equivalent to mental representation. This paper argues instead that intentionality is a feature of whole embodied agents (organisms) that can be directed at objects and states of affairs, while representation belongs to mental states and their mechanisms. The distinction is applied to Metzinger's project of naturalizing phenomenal representation. The paper challenges enactivism's equation of cognition with biology, proposing that a theory of intentionality and representation should delineate cognitive science's subject matter and allow for artificial intelligence.
The predictive processing (PP) framework, which posits prediction error minimization as the brain's sole mechanism, aims to provide a complete, realistic, and naturalistic account of all mental phenomena, including consciousness. However, many PP proponents also endorse a Kantian transcendental idealism—the view that we experience only appearances, not the world itself. This philosophical commitment undermines PP's three ambitions: completeness, Bayesian realism, and naturalism. The argument draws on philosophy of science debates about realism, the nature of computational models, and consciousness's foundational role for science.
The predictive processing framework, popular in mind and brain research, is unlikely to deliver major breakthroughs in the search for the neural correlates of consciousness because it is too vague to pinpoint neural mechanisms at a fine scale. Despite its unifying goals, the framework includes diverse computational models with different assumptions that are under-constrained by neurological data. Additionally, it is ill suited as a unifying theory of consciousness, as attempts to explain consciousness within it rely heavily on external assumptions, despite claims of compatibility with leading neuroscientific models.
Finding the neural correlate of consciousness (NCC)—the minimal brain activity sufficient for a specific percept—faces a methodological dilemma. Conceptual decisions about how consciousness relates to attention, cognitive access, report, and other functions determine how correlation data are interpreted. Relying on subjective reports likely confounds the NCC with neural mechanisms for cognitive functions, because reports presuppose cognitive access. No-report paradigms risk confounding the NCC with neural mechanisms underlying unconscious processes. There appears to be no way to be sure that the neural correlate of conscious experience has been isolated.