MoonlessNight
Fides et Ratio
- Sep 16, 2003
- 10,217
- 3,523
- Country
- United States
- Gender
- Male
- Faith
- Catholic
- Marital Status
- Private
- Politics
- US-Others
kingreaper said:Oh and jubilationtcornpone I'm much more inclined to trust a conclusion arrived at through a series of connected and logical causal steps (determinism) than one generated out of some kind of odd acausal causal paradoxical free will, or one generated by randomness
This is quite a false dichotomy. The setup is that every argument can be performed by a computer, or else it is random and untrustworthy. But there needs to be an important distinction between mathematical logic and computation. According to Roger Penrose (and I agree with him) mathematical logic can't be reduced to the types of computations that a computer uses. This is suggested by Godel's incompleteness theorem, but there are other arguments for it. What they basically come down to is that either human beings can (often easily) deduce things that could never be solved by any algorithm, or our logical system is hopelessly flawed. What they suggest to me is that for artificial intellegence to truly exist it would at least have to operate in a way drastically different than what we see today. It flat out couldn't run on a modern computer, the entire system would have to be redesigned in a way that isn't clear to me.
EDIT: There are of course related arguments that have nothing to do with mathematics. The Chinese Room thought experiment, for example, is a fairly good argument that human understanding does not function in the same way that a computer program does. But I think the mathematical arguments are more interesting because they suggest that higher logic is something that can never be duplicated by a computer, even if we accept that the computer would have to be unthinking. In other words, you can't build a computerized mathematician.
Upvote
0