Antonio R. Damasio (1989), Cognition 33(1–2), 25–62 — doi:10.1016/0010-0277(89)90005-X (checked on Crossref; PMID 2691184)
Standing in the record
Filed on 2026-10-01 at the owner’s request for “some Antonio Damasio”. No anthology topic holds it: it is a systems-level theory of memory in the brain and carries no instruction for machine-learning practice.
Unavailable because no full text could be reached, so nothing beyond the
abstract is claimed and no NOTE is filed. What the abstract says, to scope
it until somebody reads it:
- The architecture has four parts: neuron ensembles in early sensory and motor cortices holding “feature fragments”; local convergence zones in single-modality cortex holding amodal records of which fragments occurred together; non-local convergence zones in higher-order association cortex holding records of which local zones occurred together; and reciprocal projections, many-to-one forward and one-to-many back.
- Recall is the time-locked activation of the fragment ensembles, directed from the convergence zones and mediated by the feedback projections.
- It rejects a single anatomical site where memory is integrated, and a
single store for meaning. “Meaning is reached by time-locked
multiregional retroactivation of widespread fragment records”, and only
those fragment records “can become contents of consciousness”. That
last claim is why the work is tagged
consciousness. - It is a proposal. A close reading should establish what evidence the paper offers (the abstract cites none) and what it predicts that a single-store account does not.
Damasio’s own later work leans on it. His 1996 statement of the somatic marker hypothesis, filed alongside it, cites it (as “1989b”) for “dispositional” knowledge and the convergence-zone framework: the ventromedial prefrontal cortex is proposed there as a set of convergence zones linking records of situations to records of the body states that went with them. So the memory architecture here is the substrate that hypothesis assumes.
In this record it sits nearest to the cross-modal unity account of LIT-190, which puts part of the unity of experience in a shared feeling of processing quality rather than in feature-level binding. This work is a feature-level, anatomical answer to the same question of how separately represented fragments come together. Whether that answer is a binding mechanism in the sense LIT-190 sets aside needs the full text.