Modified: September 21, 2023
computational functionalism
This page is from my personal notes, and has not been specifically reviewed for public consumption. It might be incomplete, wrong, outdated, or stupid. Caveat lector.The view that performing the "right kind of computation" is necessary and sufficient for consciousness.
Chalmers (1995): if a person's neurons were gradually replaced by equivalent artificial prostheses, so that their behavior stays the same, it is implausible that the person would lose consciousness.
(in the speculative QRI view, an 'equivalent' prostheses might need to have the same effect on the EM field, which is the physical substrate of consciousness, so this argument doesn't fully rule out non-computational views)
Computational functionalism would allow that AI systems can be conscious; there is nothing inherently special about the biological substrate. More strongly, there is nothing special about any physical implementation:
- a silicon computer, or
- an appropriately designed assembly of pipes, or
- a person simulating a Turing machine through pen and paper, etc.
It's not clear to me why it would even matter for the computation to be physically executed at all. If the physical substrate doesn't matter, the consciousness is associated with an algorithm run on particular inputs, but this is just a mathematical object. Wouldn't that make consciousness directly a property of the mathematical object?