The Ghost Constructed by the Machine

In the previous post, I noted that the phenomenal world as it appears to our senses is virtual. Contrary to our tacit belief that it is directly accessible, our world is constructed — computed — from a multitude of measurements that the brain carries out on the environment. It is only to be expected, then, that our embodiment, or the physical make-up of the brain's measurement devices and computational circuits, as well as the rest of the body's machinery, should affect the way we perceive the world. More surprising to many of us is the realization that the particulars of the brain's embodiment also shape the perceiver — the phenomenal self that is the ghost computed by the living machine.

I had some wonderful opportunities to appreciate this point all over again during my annual spring hiking trip to the Death Valley National Park, from which I returned yesterday. The first of these presented itself in the form of a bighorn sheep, halfway up the wall of a canyon through which I was traveling. This was a rare sighting: bighorn are very shy, and also very capable of avoiding being seen and of getting out of sight when spotted, no matter how steep or precarious the slope. Both those capabilities are illustrated by the photograph (see below) that shows the animal while it was getting away (it helped to have a good camera with 10x optical zoom and image stabilization).

A bighorn sheep on a canyon wall in Death Valley NP

Given the unobstructed line of sight between the camera and the left eye of the sheep in the picture, it is safe to assume that it sees a potential threat (which is why it is retreating up the scree). More interestingly, even though the sheep and I are facing at right angles to each other, it is also the case that it is actually looking at me — attending to me, rather than merely seeing a blur in its peripheral vision, which is all that I would experience, were I to face away from it. The sheep's auditory attention is also directed at me: its left ear is turned so as to point straight in my direction.

The differences between the sheep's phenomenal experience of the sensory world and mine are due to the distinct features of the physical layout and functioning of our respective sensory and motor systems. For one thing, my ability to move my ears is much more limited; in particular, I cannot swing my left ear independently of the right one to listen to something that is happening to my left. Furthermore, unlike my eyes, which are front-facing, the sheep's eyes are side-facing. This physical arrangement of its light sensors enables it to take in — look at, and see — simultaneously a 290-degree panorama of its surroundings. In the central 40 to 60 degrees of that wide field of view, which is the segment that the sheep can see with both eyes, it is capable of binocular stereopsis, or depth vision (Kendrick, K. M. (2008), Sheep senses, social cognition and capacity for consciousness. In: Dwyer CM, editor. The welfare of sheep. Amsterdam: Springer. pp. 135-157; these figures are for domestic sheep, not the bighorn, which are understandably more difficult to persuade to come to the lab). Moreover, the horizontal "visual streak" in the sheep's retina, which contains a specialized distribution of neurons, enables sheep to look around monocularly without head or eye movement (A. Shinozaki et al. (2010), Topography of ganglion cells and photoreceptors in the sheep retina. The Journal of Comparative Neurology, 518:2305-2315).

The different embodiment of the sheep's mind compared to mine affects action-related phenomenology too. Human subjects' perception of the slope of a hill is known to depend on many bodily factors, including the level of glucose in the blood (Schnall, S., Zadra, J. R., & Proffitt, D. R. (2010). Direct evidence for the economy of action: Glucose and the perception of geographical slant. Perception, 39(4), 464-482). It stands to reason, therefore, that a narrow ledge leading up the steep wall of a canyon that to me appears unscalable even after a good breakfast feels like much less of a challenge to a bighorn sheep that happens to be in top shape (not all of them are: on another hike last week, I came across an almost complete skeleton of a bighorn with its neck conspicuously broken in one place).

The effects of differential embodiment, of the kind that we just identified, have an important implication for the construction of the phenomenal self — the virtual entity computed by the brain so as to facilitate its control over the body. As Thomas Metzinger puts it in his book Being No One (MIT Press, 2003, p.553), the phenomenal first-person experience works like a "total flight simulator" — a virtual reality rig that simulates the entire world along with the pilot, the latter being a model (a simulation) of the system itself. If so, the perception and action capabilities of the embodied cognitive system, which are inexorably linked to the peculiarities of its embodiment, play a critical role in defining that system's phenomenal selfhood. To be a panoramically aware bighorn self — that is, to be generated by a brain+body that can see almost all around without turning — is bound to feel differently, compared to being a humanoid self that can basically look in one direction at a time. The bighorn's ability to scale seemingly vertical slopes and walk hair-raisingly scary ledges is definitely something I envy (the older I get, the more), but it is the bighorn-type panoramic vision that I wished I had when I was standing last week on top of Aguereberry Point in the Panamint Mountains, my view from which I tried to approximate with the photograph that appears below (taken with a fish-eye lens, which has a meager 180-degree field of view, about the same as my own visual system).

Death Valley as seen from Aguereberry Point

Of course, being human has its advantages. I doubt it that a bighorn standing on that peak would be particularly thrilled by its ability to take in the view, or be happy just for having gotten there, for appreciating what I experienced, and for anticipating what was still to come, like I was. Simply being a healthy ram, well-fed and watered, out of danger, and on good terms with a kindly ewe must feel nice, but, as I argued elsewhere, the best route to varied, profound, and open-ended happiness is the accumulation and assimilation of experience by a big mind that can only be supported by a big brain, such as ours, embodiment being merely one of its tools. Standing on Aguereberry Point, I was, for a while, happy to be me, because the virtual world that I construct for myself, of which the physical environment is only a part, is so much richer than that of a bighorn.

All embodied sentient beings — a category that includes us humans — do, however, have limitations to their native, intuitive creativity with regard to virtual world construction. These limitations are put in place by embodiment itself. It is not entirely impossible for us to imagine what it is like to be a sheep, because of our close evolutionary relationship and consequently similar body and brain sensor plans (close relatively to that between us and, say, the octopus). In any case, what will eventually settle such issues decisively is the collection of scientific practices that Daniel Dennett calls heterophenomenology. It seems, however, that for mere philosophizing or unscientifically creative flights of fancy, certain feats of far-fetched cross-species phenomenology will remain forever beyond reach (note the embodied language that I could not help but use just now). Consider, for instance, a Pierson's puppeteer — a science-fictional creature invented by Larry Niven, which has two heads and two eyes, one per head. In principle, a puppeteer can see itself without using a mirror, simply by turning its heads to face each other. Now imagine a puppeteer one of whose eyes is blue and the other green. What color do you think it will see when looking itself in the eye?