The TNGS through the Lens of Relational Ontology: 2. Neuronal Groups as Relational Fields

One of the most important innovations of Gerald Edelman’s Theory of Neuronal Group Selection (TNGS) was the rejection of the neuron as the fundamental unit of cognition.

Cognition, Edelman argued, does not arise from isolated neurons behaving like switches in a circuit.

It emerges from:

• dynamically interacting neuronal populations
• shaped by developmental variation
• undergoing selectional stabilisation
• within ongoing embodied activity

This was already a major departure from classical computational neuroscience.

But relational ontology pushes the shift further still.

Neuronal groups are not units.

They are relational fields.

The residual objecthood of neuronal groups

Even after abandoning individual neurons as primary cognitive units, TNGS can still appear to retain a subtle ontology of quasi-objects.

“Neuronal groups” are often described as:

• assemblies
• networks
• maps
• populations
• coordinated structures

This language remains useful, but it risks suggesting that groups are stable entities that exist prior to their activity.

Relational ontology reverses this assumption.

A neuronal group is not:

• a thing that activates

It is:

a transiently stabilised pattern of relational neural actualisation under recursive constraint dynamics

The group does not first exist and then perform cognition.

The coherence of the activation is itself the temporary existence of the group.

From entities to activation regimes

This shift changes the ontology completely.

A neuronal group becomes:

• dynamic rather than static
• processual rather than object-like
• relational rather than bounded
• temporally emergent rather than structurally fixed

What appears as a stable “group” is:

a recurring coherence regime within a continuously reconfiguring neural field

The stability is real.

But it is:

• enacted
• maintained
• and continuously renegotiated through recursive coupling processes

Neuronal groups are therefore not neural objects.

They are stabilised activation geometries.

Why boundaries become unstable

Classical neuroscience often assumes that cognitive functions can be localised:

vision here, language there, memory elsewhere.

Even distributed models frequently preserve identifiable modules beneath the surface.

But TNGS already weakens strict localisation through:

• degeneracy
• distributed participation
• overlapping functional involvement
• and dynamic recruitment

Relational ontology radicalises this.

The boundaries of neuronal groups are not fixed because:

the group is constituted by the relational coherence of activity itself

Its “edges” shift depending on:

• bodily state
• environmental coupling
• prior neural history
• ongoing reentrant coordination
• and current constraint dynamics

A neuronal group is therefore not spatially fixed in the classical sense.

It is dynamically constituted through transient relational closure.

Neural identity as persistence-through-reconfiguration

This raises a difficult question:

if neuronal groups are continuously changing, what makes them identifiable at all?

The answer parallels earlier discussions of species and traits.

Identity is not fixed composition.

It is:

persistence of relational coherence across reconfiguration

A neuronal group remains “the same” not because identical neurons participate each time, but because:

• similar relational activation structures
• recur under comparable constraints
• with sufficient continuity of functional coordination

The group is not materially identical across activations.

It is relationally stabilised across them.

Degeneracy and relational non-essentialism

Edelman placed enormous importance on degeneracy:

different neural structures producing similar outcomes.

This insight becomes philosophically explosive under relational ontology.

Degeneracy undermines:

• strict localisation
• one-to-one mapping
• representational encoding
• and functional essentialism

If multiple neural configurations can stabilise similar experiential or behavioural coherences, then cognition cannot depend on fixed structural representations.

Instead:

cognition emerges from the dynamic stabilisation of relational activation patterns across multiple possible neural configurations

There is no essential neural realisation of a cognitive function.

Only:

• families of constraint-compatible stabilisations

This is profoundly anti-essentialist.

Reentry as field coordination

TNGS emphasises reentry:

recursive signalling across distributed neural regions.

But relational ontology reframes reentry away from communication between modules.

Reentry becomes:

recursive mutual constraint actualisation across overlapping neural relational fields

This matters because it removes the image of:

• separate units exchanging information

Instead, neural organisation appears as:

• continuously self-modulating field dynamics
• within which local coherences emerge temporarily through recursive coordination

The system is not composed of communicating parts.

It is composed of partially stabilised relational regions within a unified dynamic field.

Why synchronisation is not enough

Some neuroscientific models attempt to explain cognition through synchronisation alone.

But relational ontology requires greater precision.

Coherence is not merely simultaneous firing.

It is:

structured relational compatibility across recursively interacting activation trajectories

Simple synchrony cannot explain:

• meaning differentiation
• contextual modulation
• flexible reorganisation
• or dynamic experiential coherence

What matters is not simultaneity itself, but:

• the topology of relational constraint coordination

Neuronal groups are not synchronised clusters.

They are dynamically stabilised relational geometries.

The brain as metastable field

The concept of metastability becomes central here.

The brain is neither:

• rigidly ordered
  nor
• chaotically disordered

Instead, it operates within:

metastable regimes capable of rapid reorganisation while maintaining temporary coherence

Relational ontology interprets this as:

• continuous movement through spaces of possible relational actualisation
• under biological, bodily, and environmental constraints

Neuronal groups are temporary attractor formations within this metastable field.

They arise.

Stabilise briefly.

Transform.

Dissolve.

Reconfigure.

There are no permanently fixed cognitive structures beneath this process.

Why cognition is not located

If neuronal groups are relational fields, cognition itself ceases to be spatially localisable.

Cognition is not “in” a region.

It is:

enacted through distributed relational coordination across dynamically coupled neural, bodily, and environmental systems

The brain is therefore not the container of cognition.

It is one participating layer within broader embodied relational actualisation.

This prevents a subtle but important relapse into neural internalism.

Memory without storage

This relational account also transforms memory.

If neuronal groups are dynamically reconstituted fields, then memory cannot be static storage.

Memory becomes:

the constrained re-actualisation of relational activation tendencies within historically stabilised neural dynamics

Nothing is retrieved from a fixed archive.

Instead:

• prior relational stabilisations bias future actualisations
• making certain coherence patterns more likely to recur

Memory is therefore:

• historical constraint sedimentation
  rather than
• informational retrieval

Why neuronal groups are not representations

The temptation remains to interpret neuronal groups representationally:

as encoding features of the world.

But this reintroduces the very ontology TNGS destabilises.

Neuronal groups do not contain representations of reality.

They participate in:

embodied relational coordination with environmental and bodily constraint structures

The world is not reconstructed internally.

It is enacted relationally through ongoing coupling dynamics.

This is not idealism.

Nor is it naive realism.

It is relational participation.

Consciousness as field coherence

The implications for consciousness are immense.

Consciousness is not produced by:

• isolated regions
• symbolic processors
• or central observers

It emerges through:

dynamically stabilised coherence across recursively interacting neural relational fields

No inner theatre exists behind the process.

The coherence itself is the event of conscious actualisation.

Experience is not displayed to consciousness.

The stabilised relational coherence is consciousness occurring.

The collapse of neural atomism

Classical neuroscience often searches for fundamental units:

• neurons
• circuits
• modules
• representations

But relational ontology suggests:

there may be no neural atoms of cognition at all.

Only:

• recursively coordinated relational fields
• continuously stabilising and destabilising
• across multiple temporal and spatial scales

Neuronal groups are therefore not building blocks.

They are transient patterns within an ongoing field of neural relational actualisation.

Closing the field

The concept of neuronal groups marked a decisive break from computational atomism.

But relational ontology completes the shift.

Neuronal groups are not neural objects carrying cognitive content.

They are:

transiently stabilised relational fields emerging through recursive constraint coordination within embodied neural dynamics

The brain does not compute representations through fixed structures.

It continuously actualises temporary coherence regimes within a metastable relational field.

What appears as cognitive structure is not the operation of neural things.

It is the ongoing emergence of relational order within living neural dynamics that never fully settle into objects at all.