Why We Are Conscious
Answering this question yields counterintuitive insights about the mind and AI.
Posted Dec 07, 2017
When a scientist attempts to “reverse engineer” the human nervous system in order to figure out what consciousness is for – the most basic form of consciousness, such as awareness of pain, perceptual awareness of a red apple, or awareness of the urge to sneeze – the scientist is often reminded of something. One can easily imagine a system that does the same thing and can carry out the same function, without anything like consciousness.
To this point, which we will refer to as the “but I can imagine it without consciousness” criticism, the scientist might reply that, though it might well be that such a function could be carried out without anything like consciousness, in the evolution of human beings, consciousness seems to have been selected to carry out particular mental operations, for one reason or another. This conclusion is based in part on the logic that, when consciousness is absent, those operations can no longer be instantiated.
Such a scientist is adopting a “descriptive” approach to the products of natural evolution, which describes natural products as they are and not as we humans would design them or as we think they should be designed (which is a “normative” approach). Indeed, human locomotion (legs) is very different from artificial locomotion (wheels), just as the artificial heart is very different in nature from its biological counterpart.
The products of evolution are often suboptimal, counterintuitive, and very different from how we would engineer them. Thus, from a descriptive standpoint, that fact that one can imagine a creature doing all the things that humans can do without anything like consciousness might reflect more the human power of imagination (or our good sense of engineering) rather than what has transpired in the course of evolutionary history. In short, the legs, in humans, are for locomotion even if you could imagine a different or better way to locomote.
Recent developments provide insights that further challenge the “but I can imagine it without consciousness” criticism. It turns out that our intuitions regarding how such a nonconscious creature might operate in terms of the perception-and-action cycle, in which a human responds to one stimulus when there are other stimuli in the visual scene, are actually computationally impossible. That is, just as a computer cannot solve the traveling salesperson problem, many of the problems humans solve in the perception-to-action cycle are known to be, a priori, unsolvable through any known formal computational means. No matter how much computational power one has, these problems cannot be solved through computer logic. Nature is doing things differently in its peculiar and often happenstance ways.
To conclude, in response to the “but I can imagine that function without consciousness” criticism, a scientist today can reply that, first, evolution might not have carried out that function the way humans imagine it should be engineered, and, second, that our human ideas regarding how such functions could be carried out nonconsciously are actually impossible, a priori. For these reasons, reverse engineering the brain and consciousness might reveal clues, not only about what we are, but regarding how to deal with the limitations of our computer and robotic systems.
For more discussion regarding this issue, see this previous blog post.