Home > Positioning > Subjects > Basal cognition

Basal Cognition

Basal cognition is the research field that studies cognitive capacities in organisms without nervous systems — and in the non-neural parts of organisms that have them. Its founding move is methodological: Pamela Lyon’s biogenic approach (2006) starts from what living systems demonstrably do and works upward, instead of starting from human cognition and working down. On the field’s working definition, cognition is the suite of capacities by which organisms become familiar with, value, and exploit their environment while avoiding harm — operationalised as a toolkit: sensing, memory, valence, learning, anticipation, decision-making, communication. The term itself is Lyon’s, coined around 2018; the field’s convening documents are the double Philosophical Transactions of the Royal Society B theme issue of 2021, guest-edited by Lyon, Fred Keijzer, Detlev Arendt and Michael Levin, with Lyon’s agenda paper “Reframing cognition: getting down to biological basics” at its head.

The cases

The field’s evidence base is uneven, and honest accounts grade it.

Solid. Bacterial chemotaxis is the textbook floor: an E. coli cell swimming a chemical gradient compares the concentration now against a molecular record of a few seconds ago, laid down in the methylation state of its receptors — a genuine record of the recent past, though “memory” here names delay chemistry, not psychological retention. The slime mould Physarum polycephalum solves mazes by shortest path (Nakagaki et al., Nature 2000) and, given oat flakes at the positions of thirty-six Tokyo-area stations, converges on a transport network comparable to the real rail system in efficiency, cost and fault-tolerance (Tero et al., Science 2010) — both results widely reproduced, and the second the seed of a literature of Physarum-inspired algorithms. The ciliate Stentor roeselii works through a hierarchy of escape responses under repeated irritation — bend, reverse cilia, contract, detach — a result of Jennings’s from 1906, dismissed for half a century on the strength of a non-replication that had used the wrong species, and confirmed in 2019 (Dexter et al., Current Biology).

Probabilistic. Physarum anticipates periodic events — slowing at the time the next scheduled cold shock was due, even when it is omitted (Saigusa et al., 2008) — but only in a portion of individuals across replications: a real effect, in many but not all.

Contested. The plant-learning results are the field’s open wound. Gagliano’s Mimosa habituation experiments (2014) are disputed on interpretation (habituation or motor fatigue) without a clean independent replication either way; her pea-plant Pavlovian-conditioning claim (2016) has a failed large-scale replication (Markel 2020) and a live methodological counter-reply. The two claims are separate and should be kept so; the field’s own stance is unresolved-and-replicate. Beatrice Gelber’s 1950s Paramecium-conditioning experiments sit differently: a 2021 re-evaluation (Gershman et al.) argues the classic dismissals were flawed and the work deserves re-running — a historical rehabilitation, not yet a modern result.

Cell collectives. Levin’s programme extends the field into tissues: bioelectric pattern memory (planaria whose target anatomy is rewritten without genomic change), and constructs — xenobots, anthrobots — that show the capacities in novel configurations. The full programme is on his page.

The dispute

Whether any of this is cognition is the field’s standing argument, and the objection deserves its strongest form. Adams and Aizawa’s line: cognition requires a “mark of the cognitive” — intrinsic, non-derived content — and without one, attribution inflates until the word picks out nothing. The deflationary reading grants every behaviour and denies the vocabulary: “memory,” “decision” and “learning” as metaphors for adaptation chemistry, with the burden on the field to show the cognitive vocabulary buys predictions the chemistry alone does not. The field’s answers vary by author — Lyon’s is that the biogenic toolkit operationalises the terms; Levin’s is that the question is an engineering one, answered by whichever level of description yields prediction and control. The dispute is live, and it is the same definitional question the word meets everywhere: if a cell or a colony does this, is it cognition or metaphor?

Adjacencies

The field’s conceptual ancestry runs through Maturana and the Santiago theory’s widest reading — living as such already a process of cognition — and through autopoiesis and enactivism’s life–mind continuity; the philosophical statement of the wide reading is at Cognition as a Property of Life. On the engineering side the field overlaps AI at the margins: the xenobots were designed by evolutionary algorithms, and a small “diverse intelligence” research community bridges reinforcement learning and cellular behaviour.

See also: Lyon · Levin · Maturana · Autopoiesis · Enactivism · Complex adaptive systems