In a recent post over at Scientia Salon
Mark O’Brien asks a question and gives his own answer with:
Could a computer ever be conscious? I think so, at least in principle.
As O’Brien says, people have very different intuitions on this question. My own intuition disagrees with that of O’Brien.
After a short introduction, O’Brien presents two starting assumptions that he makes, and that he will use to support his intuition on the question.
Empirical assumption 1: I assume naturalism. If your objection to computationalism comes from a belief that you have a supernatural soul anchored to your brain, this discussion is simply not for you.
Personally, I do not assume naturalism. However, I also do not believe that I have a supernatural soul. I don’t assume naturalism, because I have never been clear on what such an assumption entails. I guess it is too much metaphysics for me.
Although O’Brien calls that an empirical assumption, I take it to actually be a metaphysical assumption.
Empirical assumption 2: The laws of physics are computable, that is, any physical process can be simulated to any desired degree of precision by a computer.
It is not entirely clear what that assumption means. Certainly, laws such as Newton’s f = ma are computable as mathematical functions. That this is made an assumption suggests that O’Brien means something else. My guess is that he is talking about a presumed but perhaps unknown set of laws that could be said to govern the universe. That there are such laws would be another metaphysical assumption, and one of which I am quite skeptical. In particular, QM would seem to pose a problem for those who believe that there are such laws.
My view of consciousness
I see consciousness as resulting from a biological adaptation that allows us to cope with unexpected contingencies. If we were to program a robot, say to explore Mars, we would have to program into it the ability to cope with every possible contingency that could arise. But we cannot know what contingencies might arise. That’s why NASA designs its Mars explorer missions so that they can be reprogrammed as needed. My view is that consciousness evolved because we also need to be deal with unpredicted and unexpected contingencies. In some sense, consciousness is there because we cannot have been born with innate knowledge of the kind of metaphysics that O’Brien wants to assume.
For me, the real problem solved by consciousness and by intelligence, is that of allowing us to learn how to cope with the unexpected.
The problem with simulation
You cannot just take some aspect of the world, and simulate it. Simulation requires a great deal of knowledge. So if we built systems that behave like humans, and do it with simulation, then we will have to program into them an enormous amount of innate knowledge.
If the role of consciousness is to allow us to learn, then preprogramming with innate knowledge would seem to undermine that role. For the programming would result is a system that had no need to learn, and thus no need for consciousness.
Some people argue that consciousness is an illusion. It seems to me that, at best, the kind of simulation proposed might generate an illusory consciousness, but not a real consciousness.
A note on dualism
I’m no expert on Descartes. However, it is my impression that he saw Newton’s mechanics as suggesting a kind of clockwork universe. And apparently he gave us his version of dualism because he did not see how a clockwork universe could account for the human mind. So, if my impression is correct, we would have to conclude that Descartes’ intuition did not support the kind of viewpoint that O’Brien is arguing for.
For myself, I do not believe that we live an anything like a clockwork universe. So I have no need to appeal to dualism. I see the universe as such that consciousness could evolve as a biological adaptation (and probably did).
In short, I do not find the simulation argument at all persuasive.