There are several subjects that have strong implications in mind-brain research. Most of them are apparently quite academic and should not be taken too seriously if you are not a professional philosopher. Others, on the contrary, need to be understood before starting putting down equations. I still remember that at CERN I was able to discuss at least three theories about a cold fusion effect that did not actually exist. This has been a very important experience for me.
Different beliefs about the nature and the limits of AI
Can a computer think ?
There are two different answers, depending on what we mean by "think".
1) If we consider the term in an operational way (question-answer scheme), a pocket calculator evidently thinks. We ask "3+2" and he correctly answer "5". No questions can be raised in this case on how the machine reaches the correct result. The pocket calculator can add 3 and 2 and it answers correctly. Whatever its internal structure is, it behaves well as far as elementary mathematical operations are concerned.
2) If we intend by "think" what we normally mean by referring this verb to a person, the answer does not exist for the moment as, we use this term without any clearly defined meaning. Does it look too simple? "Think" about it; you will discover that it's awfully complex.
If we concentrate on a specific task, it is very probable that we will be able to write a program capable of doing quite well by manipulating symbols (data and instructions) following well defined rules of manipulation. This shows that in most cases we are capable of inventing a program capable of doing something quite well. It's a statement about our mind's performance with tools we have invented (algorithms and computers).
Is the brain a computer ?
From the physical point of view, evidently not: No CPU, no ram, no hard disks in a brain. Then we must restate the question: "does the brain work as a computer?" The answer is evidently still "no": there are no programs, no algorithms identified in the brain, at least at this stage. We can be forced to answer positively by making a logic error. If a computer can implement programs that permit it to "think" in the question-answer scheme than what happens in the brain must be somehow the same that happens in the pocket computer. This is evidently false as there exists no logical demonstration of the above statement. One of the problems underlying the above question is due to the identification of the term "computation" with "Turing-computation". "Computation in the sense of Turing" is the correct and clear cut translation of the effective computation of a human computer (we should not forget that the word "computer" represented in times a profession). Researchers have developed non-Turing Machines computing Turing non-computable functions. One of the interesting problem going around is the following: as today's computers are working in local and global networs, so they are interacting with a lot of other computers, aren't we already running non-Turing machines? Peter Wegner and B. J. Copeland are two major researchers working on the scenario of interacting computers paradigm. A very understandable point they underline is that Turing Machines isolate themselves during their computation. All the communications with the external world are interrupted. So if a machine accepts input "during" its computation,it does not fit the Turing Machines standards. My personal impression is that this suggests some form of "relativity" in the concept of computation, as I evidently can partition the activity of such a machine in sub-intervals during wich it acts within the Turing Machine standard. Maybe I'm wrong, but I cannot see how.
Is the mind a computer ?
It's a nonsense as, due to the too large definition of computation, everything can compute from some point of vew, except abstract entities.
Who is computing? the observer or the observed system, or both?
A short discussion is given in my short discussion about Church-Turing Thesis