I started this series of essays with the claim that I wanted to channel Daniel Dennett’s style of philosophy to the study of unicellular cognition. Easier said than done: cognitive scientists and psychologists and AI researchers have a long history of working with one another, but the language and idiom of molecular biology and biophysics isn’t philosophical at all. Of course, that challenge is also an opportunity.
The prominent philosopher Jerry Fodor once said that every idea has two lives: the first, in philosophy, and the second, in cognitive science. To which I might add: the third, in biology. The third avatar is quite relevant for our investigation into unicellular cognition. In the 1950s, 60s, and 70s, there was a lot of debate over the substrate independence of intelligence, cognition and mind, with some people saying that you need a very specific kind of fleshy apparatus to be considered intelligent, while others saying that it doesn't matter what you're made of, as long as you're solving the same problem that a human being is, then you're as intelligent as a human in that respect.
This framework was called functionalism; the idea being that, let's say, you are a robot using two front of head-mounted cameras to recover depth from the visual field, well, then you are doing the same thing as a human being who's using their two eyes for that purpose, even though the physical substrate that's doing the information processing is quite different.
We can extend that argument to unicellular beings too! If there is a problem such as locomotion, that's being solved by bacteria using a flagellar motor in the same way that a multicellular creature with a tail would do so, the bacterium is performing the same function, even though the control structure used by the bacterium is radically different - a combination of chemosensory proteins and a histidine-kinase two component system instead of a network of neurons. The bacterium and the eel are (roughly) functionally equivalent, and therefore the bacterium is as cognitive as the eel.
Of course, skeptics might object that locomotion alone is too crude a standard for true cognition. But the beauty of functionalism - especially when paired with Pamela Lyon’s biogenic approach - is that it applies just as cleanly to internal information processing as it does to swimming or crawling. Traditional cognitive science was born out of an anthropogenic bias: it took human higher-order reasoning, symbolic manipulation, and cortical architecture as the gold standard of mind, asking which other creatures (or machines) were lucky enough to join the club. Lyon and colleagues reject this methodology. A biogenic approach begins not with human introspective paradigms, but with the universal constraints of biological existence: survival, growth, and reproduction in an unpredictable world.
In this view, cognition is an essential function (like metabolism) that helps organisms adapt to, survive and replicate in an uncertain environment.
When viewed through this biogenic functionalist lens, the single cell reveals an astonishing array of cognitive operations, i.e., computational problems solved by its cellular apparatus. Consider the classic case of bacterial chemotaxis in Escherichia coli. To navigate toward nutrient attractants like glucose or away from harmful repellents, a bacterium cannot simply measure spatial gradients across its micrometer-long body; it is far too small. Instead, it must measure gradients over time. The bacterium continuously compares its present chemical environment with the environment of a few seconds ago. This temporal comparison requires a functional analogue of working memory. Biochemically, this short-term memory is implemented through the dynamic methylation and demethylation of methyl-accepting chemotaxis proteins (MCPs) governed by the enzymes CheR and CheB. If conditions are improving, the cell suppresses tumbling and continues its smooth forward run; if conditions deteriorate, it triggers a tumble to randomly reorient. As Marc van Duijn, Fred Keijzer, and Daan Franken have argued, this two-component signal transduction system constitutes a functional sensorimotor coupling - a minimal cognitive apparatus that senses, buffers past states, computes, and generates adaptive behavioral outputs.
This functional equivalence extends into what Lyon and Franz Kuchling term biological valence: the capacity of a system to evaluate environmental conditions as intrinsically favorable or harmful relative to its own homeostatic needs. When a human experiences noxious heat, a fish encounters low oxygen, or a bacterium detects a sharp drop in pH, each system registers a negative valence state that triggers an evasive maneuver.
Note: the gazillion $ question: what is the relationship between the human experiencing, a fish encountering and a bacterium detecting? The first is very much a subjective, conscious affair; what about the other two?
The physical substrates could not be more distinct - neurotransmitters and nociceptors in the animal versus histidine kinases and response regulators in the microbe. Yet, functionally, the control loop is performing the exact same biological job: valuing what happens in the environment and altering behavior to preserve organizational integrity.
By freeing cognition from its neurocentric straitjacket, functionalism allows us to recognize that brains are not the sole origin of mind, but rather a late-stage, specialized implementation of biological information processing. As Lyon, Michael Levin, and others in the basal cognition paradigm emphasize, if cognition is functionally defined as the toolkit an organism uses to sense, value, remember, and interact productively with its world to meet existential needs, then cognition is not a rare privilege reserved for animals with brains. It is an ancient, fundamental property of life itself.