1 Life Misconstrued: 3 Astrocytes Don’t Compute: The Brain as Relational Field, Not Machine

Astrocytes make up roughly a quarter of the brain, yet for decades they were dismissed as mere support cells — the silent audience to the neurons’ starring role. Recent research, however, has begun to highlight their influence: by modulating the chemical environment around synapses, astrocytes can shape whether neurons fire in response to stimuli. Headlines leap to metaphors: “astrocytes shape computation” or “silent cells become co-processors,” reinforcing the familiar narrative of brains as machines and neurons as processors.

Relational ontology demands we resist this temptation. Astrocytes do not compute. Neurons do not compute. Synapses do not compute. Brains are not machines. The entire metaphorical architecture collapses upon inspection. What these cells actually do is modulate conditions of potential, regulating the field in which neuronal activity — and ultimately behaviour — can be actualised.

Consider the dynamics:

• Astrocytes influence ion concentrations, neurotransmitter availability, and local metabolic states.

• These influences modulate the likelihood of neuronal firing.

• Neurons respond not to encoded “instructions” but to shifts in the relational landscape created by astrocytes and other regulatory processes.

At no point does this system “compute” information. There are no discrete units of data, no codified messages, no symbolic transformations. What occurs is value-oriented regulation: patterns of activity constrained by the system’s requirements for stability, responsiveness, and viability.

Yet, because neuroscience has long relied on computational metaphors, these processes are routinely described in symbolic terms. “Shaping computation” becomes the shorthand for a complex web of modulation and constraint. The language suggests agency, intention, and information-processing where none exists. Meaning is projected onto biological coordination, obscuring the distinction that is crucial for rigorous understanding: value ≠ meaning.

Relationally, we can recast the brain as a dynamic, multi-scale field of potentials:

• Neurons are local event sites whose firing is contingent, not determinate.

• Astrocytes and glial networks modulate the conditions under which these events can occur.

• Behaviour, cognition, and mood are emergent patterns arising from these interactions within a semiotic organism, but they are not the output of computation.

The temptation to see neurons as processors and astrocytes as co-processors mirrors the error we observed in the ant colony. Both cases involve interpreting relational, value-driven dynamics as semiotic or computational acts. In ants, chemical perturbations were misread as “communication” and “sacrifice.” In the brain, biochemical modulation is misread as “computation.”

Astrocytes are not co-computers. They are regulators of potential. Neurons are not processors. They are event actualisers within a relational field. Recognising this distinction restores clarity, prevents conceptual overreach, and allows neuroscience to describe what life actually does rather than what we wish it were doing.

In the next post, we will examine another pillar of the computational metaphor: the so-called “neural coding” paradigm, where patterns of activity are interpreted as messages and information — another instance of value mistaken for meaning.