The Language of Stress and Global Workspace Theory: Where We Converge and Diverge

Global workspace theory is the most empirically productive framework in consciousness research. It holds that conscious contents are those broadcast widely across the brain, made available to reasoning, memory, and report, and it has decades of neuroimaging work behind it. This paper grants that record and argues that the theory answers a question adjacent to the one this framework asks. Its conclusion differs from the companion comparisons: the two frameworks are largely complementary.The one real disagreement is not about whether broadcast occurs. It is about whether explaining access explains consciousness. Dehaene holds that it does, which is a philosophical position rather than an omission, and the paper says plainly that imaging will not settle it. On this framework, a complete account of how information reaches the workspace still leaves open why the contents that arrive there should matter to the system receiving them.That gives three divergences. The first concerns whether an account of access is an account of experience. The second concerns what weights the competition for access. Workspace theory takes salience, relevance and reward as inputs and assigns their origin elsewhere, which is a division of labor, not a gap. This framework offers a candidate account of that weighting, in which attention follows what is staked on the content, and says that until rigidity can be measured it is a candidate rather than a rival. The third is the most testable: whether valence is one kind of content among others, or a medium in which all conscious content is presented. A fourth difference, concerning where in the brain any of this happens, is noted and deliberately not pressed, since the framework makes no positive anatomical claim.The paper is explicit about what does not divide the two accounts. It withdraws several claims an earlier version offered as tests, including the claim that workspace theory predicts attention follows information content, which it does not. Phenomena that both frameworks handle are listed as such rather than presented as support.Version 2.1 scopes the framework’s proposal to attention and access: attention follows the largest distortions, but what the organism does is not settled by the largest one. It limits the regress argument to deliberation, since reflex can settle competing demands without one, and it states the identity conditionally where it had been stated as fact.