The Causal Passage of Thought: When Do Cognitive Representations Become Material Consequences?
This conceptual and methodological study asks under what conditions cognitive representations in human beings and artificial-intelligence systems may be assigned a defensible causal contribution to changes in the material world. It distinguishes the total effect of assigning control to a system, the mediating contribution of a particular representation, and the cross-carrier comparability of representations. Those claims differ in estimands, interventions, observability and evidentiary requirements. Existing work on structural causal models, mediation, abstraction, compound treatments, embodied cognition and brain–computer interfaces supplies substantial answers; accordingly, the paper does not propose a new causal identification theorem or an ontologically distinct mental field. Three structured tables and four constructed adversarial families illustrate where matching outputs conceal different control mechanisms, where shared autonomy creates redundant causes, and where evidence adequate for an operational effect remains inadequate for a representational claim. We propose a claim-relative qualification procedure whose value must be tested against equally informed conventional audit workflows. The findings are logical and comparative, not empirical. The strongest present conclusion is negative but useful: material success, an agent’s declared intention and a mechanistic account are not interchangeable witnesses. Field-oriented descriptions remain optional and must be reduced to conventional methods when they confer no independently demonstrable advantage.
Authors
- Rogério Figurelli (ORCID: https://orcid.org/0009-0002-5726-1765)
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-21
- DOI
- https://doi.org/10.5281/zenodo.24430895
- Primary Topic
- Philosophy of Biology and Scientific Explanation
- Type
- preprint