Among twentieth-century neuroscientists, few resisted computational orthodoxy as consistently as Gerald Edelman. In Neural Darwinism and later in The Remembered Present, he advanced a theory of neuronal group selection that displaced instruction, coding, and representation from the centre of brain theory.
The brain, he argued, is not programmed by the world. It does not receive structured inputs and translate them into internal symbols. Rather, it develops rich endogenous variation and selectively stabilises patterns of activity through reentrant coordination and value-modulated reinforcement.
This shift — from instruction to selection — is not merely biological. It is ontologically suggestive.
Selection Without Instruction
Edelman’s proposal rests on three interlocking claims.
First, the brain generates a primary repertoire: a richly varied population of neuronal groups formed during development. This repertoire is not sculpted by detailed environmental instruction; it is a product of differential growth and genetic constraint.
Second, experience acts not by imposing structure, but by selecting among this pre-existing variation. Patterns of correlated firing are differentially stabilised, forming a secondary repertoire.
Third, large-scale integration occurs through reentry: recursive, bidirectional signalling across distributed neuronal maps. Coherence emerges not from a central controller but from dynamic coordination across populations.
The crucial point is negative: the world does not write itself into the brain. There is no encoding of external form into passive matter. There is endogenous variation and differential stabilisation.
Selection does not create structure ex nihilo. It thickens some trajectories and lets others dissipate.
Thickening as Relational Stabilisation
In a relational ontology, a system is structured potential: a field of possible actualisations. Instantiation is not a temporal production but a perspectival cut — the actualisation of one trajectory among many.
What Edelman describes at the neural level can be read — carefully — as a local instance of relational thickening.
What persists is not imposed representation, but reinforced relation.
Importantly, this process is non-symbolic. The “value” operative in neuronal selection is biological coordination, not meaning. It belongs to a non-semiotic stratum. No conflation is required or permitted.
Thickening here is not semantic. It is structural.
Selection, in this sense, is the stabilisation of relational difference.
A Worked Illustration: Perceptual Differentiation
Consider perceptual categorisation.
On an instructional model, the world presents an object; the brain encodes its features; a representation is stored.
On Edelman’s account, no such encoding is required. Instead, repeated coordinations among distributed neuronal groups become selectively stabilised under value modulation. The organism’s history of interaction biases future actualisations.
What appears as stable recognition is the recurrent actualisation of a reinforced relational configuration within the neural field.
There is no internal picture. There is no symbolic mirror. There is differential stabilisation of coordinated activity.
Perceptual stability is the persistence of thickened trajectories.
Where the Resonance Stops
The parallel must not be overstated.
Edelman’s theory operates within biological evolution and neural development. Neuronal groups are material processes embedded in an organism. His claims do not extend beyond the domain of brain dynamics.
A relational ontology, by contrast, describes the structure of potential and actualisation as such. Thickening is not a biological mechanism but a structural description applicable across strata.
The resonance is therefore analogical, not reductive.
Edelman provides a concrete demonstration that stabilisation without representation is possible. He shows how coherence can emerge from selection among variation without instructional encoding.
That matters.
It undermines the intuition that structure must be imposed from outside. It suggests instead that structured persistence can arise from internal differentiation and differential reinforcement.
Meaning does not appear here yet. It will require construal. But before construal can operate, there must already be patterned persistence — a field whose trajectories are no longer indifferent.
Edelman shows how such patterned persistence can arise without instruction.
The rest is articulation.
No comments:
Post a Comment