PhronesisThe kernel, granted — and it is a large one
Vision is anchored to the hand, and that anchoring drives movement. In the ventral premotor cortex, the parietal cortex and the putamen there are bimodal neurons that answer to touch on a body part and to visual stimuli entering the space just around it. Their visual receptive fields sit on the body part, extend a few centimetres into the surrounding air, and stay in spatial register with the tactile field — which means they travel with the hand rather than with the retina. Graziano and Gross described them; the space they define is what the literature calls peripersonal space. And their activity feeds neurons in motor areas, priming action. So there is a real sense in which the hand is watched from its own frame, and the watching is wired to the moving.
Sound and touch move the eyes, and the machinery for it is a set of aligned maps. The superior colliculus carries overlapping topographic maps of visual, auditory and somatosensory space in its intermediate and deep layers, and those maps are in spatial register with a motor map in the same structure. A visual, auditory or somatosensory target can trigger a saccade. The registration is what makes the transformation cheap: a location is a location in whichever modality reported it, and the motor map is already addressed in the same coordinates. Orienting of eyes, head and body falls out of that arrangement.
Temperature belongs here too, by the ordinary route — thermoreception is somatosensory, so a hot surface reports into the same somatosensory map that a touch does, and can pull the eyes the same way.
A location is a location, whichever sense reported it. That is what the maps being aligned buys.
So the felt claim — that several senses arrive as one signal, and that the body swings toward things without deciding to — is not a metaphor and is not contested. It has an address.
The correction
Three things the conjecture asserts that the evidence does not.
The inference to vision does not follow. That sensorimotor cortex can be lesioned without impairing action verb processing rules out one candidate. It does not elect another, and vision is not the one the field would elect. The incumbent alternative is the hub-and-spoke account: an amodal semantic hub in the ventral anterior temporal lobe, drawing on modality-specific spokes, with the hub itself not committed to any one sense. Whether that hub is truly amodal is argued over — multiple-representation accounts sit between the extremes — but it, not visual cortex, is what the lesion result leaves standing. Eliminating a suspect is not a conviction.
Visual cortex is not undercounted in the way the conjecture means. The commonly cited figure has around 30% of cerebral cortex given to visual processing, already the largest single allocation of any modality. So "larger than it has been denoted" is not quite the claim to make; the denoted figure is enormous. There is a real undercount, and it is of a different kind — described below — but the headline number is not hiding anything.
Olfaction does not join the stream. This is the specific place the picture breaks, and it breaks cleanly. Vision, audition and somatosensation are spatially mapped and mutually registered in the colliculus. Olfaction is not in those maps. It is the modality that skips the thalamic relay the others share, and in a primate it carries almost no usable spatial precision — you cannot saccade to a smell. Counting it as a fifth aligned source overstates the fusion by including the one sense that is not fused into that frame. It reaches the body by other means and on a slower clock.
The version that survives, and it is the better one
The undercount is real but it is not about acreage. “Visual cortex” is an anatomical label sitting on a functional process, and the process is wider than the label. Bimodal hand-anchored neurons in premotor and parietal cortex are visual by any functional test — they have visual receptive fields — and nobody calls that region visual cortex. Visually responsive neurons turn up inside auditory cortex, at something like 15–20% in primary auditory areas and around half or more in higher ones. The colliculus, doing the fusion this whole conjecture is really about, is not cortex at all.
So the sentence worth keeping is not that visual cortex is bigger than stated. It is that vision is not confined to the places named after it, and that a map of visually driven computation would not look like the map of visual cortex. That is a weaker claim than the original and it is the one that can be defended, which is the trade this site keeps making.
And it relocates where the body enters cognition. Embodied Displacement found the somatotopic account of action semantics failing its necessity test. This note is the constructive half: the place the body demonstrably enters is not meaning but orienting — the midbrain fusion of modalities into one spatial frame, and the hand-anchored visual fields that prime action from the body's own coordinates. That is a smaller claim about cognition and a much better attested one.
What this does not settle
Nothing here says orienting is where concepts live, and the temptation to slide from one to the other should be resisted precisely because the machinery is so satisfying. The colliculus builds a common spatial frame; a common spatial frame is not a meaning. The step from “several senses address one map” to “therefore one coherent signal underlies thought” is the step this note declines to take, and it is the same step the embodiment literature took and has been unwinding since.
The limits, plainly
- The bimodal-neuron work is largely macaque single-unit recording, with human evidence coming from neuropsychology and imaging rather than direct recording. The frames are well established; the extrapolation to human conceptual processing is not made here and should not be made from here.
- The 30% figure is widely repeated and I could not trace it to a primary source. Treat it as the conventional number rather than a measurement, and note that any such figure depends entirely on where the boundary of “visual” is drawn — which is the very thing this note says is unstable.
- The hub-and-spoke account is not established either. It is the leading alternative, actively argued over, and citing it as the incumbent is a statement about the state of the debate rather than about the brain.
- Olfaction's exclusion is stated for spatial orienting in primates. It is not a claim that olfaction is unimportant, and rodents orient to odour gradients in ways this note does not treat.
- No claim here belongs to this framework. As with the other applications, the vocabulary is describing results that belong to other people.
Sources. Aligned multisensory and motor maps: “The superior colliculus,” Current Biology, and “Alignment of multimodal sensory input in the superior colliculus through a gradient-matching mechanism.” Hand-anchored visual receptive fields and peripersonal space: “Peripersonal Space,” in The Neural Bases of Multisensory Processes, on Graziano and Gross's bimodal visuotactile neurons. The amodal alternative: “Mapping contributions of the anterior temporal semantic hub to the processing of abstract and concrete verbs,” Hum. Brain Mapp., 2025. Visual responses inside auditory cortex: “Visual influences on ferret auditory cortex.” The lesion result this note begins from is cited in Embodied Displacement.