Discuss: relation to chain of animals.
Are computers getting smarter?
Moravic's & Kurzweil's predictions: within a few decades, people will be obsolete.
Turing: computers are intelligent if and only if they cannot be distinguished from humans. Use a behavioral test.
Problems with the test:
Current computers obviously fail the test, but what about the future?
Functionalism = the view that mental states are functional states, not dependent on any particular physical states
Argument from multiple instantiation
Problems with this argument.
1. You are locked in a room.
2. Chinese symbols come in a slot.
3. You look up the symbols in a book and then put other symbols out the slot.
4. Unknown to you, you are holding a conversation in Chinese, but you clearly do not understand Chinese.
5. Analogously, a computer is just a symbol processor and does not understand anything.
Searle claims that only human brains have the causal powers to possess "intrinsic intentionality". Computer symbols have syntax but no semantics.
Objections to Searle
1. Process described is computationally infeasible: slow, does not reflect nature of language.
2. Robot reply: hook robot up to the world via vision and other senses. Program the ocmputer to learn. The robot would then have intrinsic intentionality.
3. System reply: Searle in room doesn't understand language, but whole system does. Searle's argument depends on a bad analogical mapping.
Searle's thought experiment fails, but AI is very hard. Many difficult problems to be solved. Maybe the body is needed: embodied cognition.
However, note successes of AI: expert systems, machine translation, chess.
If there is a god, could we understand her?
Medieval arguments from analogy: God's mind is like ours, only more so.
Does God have beliefs? How many?
Does God have emotions?
Computational Epistemology Laboratory.
This page updated Sept. 26, 2005.