• An interdisciplinary science of the mind
• What’s special : the computational / representational approach = a successful science of the mind must include the study of computational processes and representations in the brain
• Is this new? Aristotle, Behaviorism, Thomas Hobbes, Descartes, Charles Babbage

• Computation = syntactic operations on representations
• Representations = objects or properties that represent or stand for something
• Anything can be a representation, and a representation can represent anything

• Most mental processes involve information processing
• Sophisticated information processing requires computations
• Not all computations are mental
• Maybe not all mental processes are computational
• Maybe some mental processes require more than computations
• Strong AI : Having the right kind of program is necessary and sufficient for having a mind

• The task level : what it does
• The program / algorithm level : the computational procedure used
• The implementation / hardware level : how the computation is realized and constructed
• Same task, different program; same program, different implementation
• Different disciplines in cognitive science look at different levels and how they are related
• These are three types of levels, e.g. there can be more than one program level

• Clarify key concepts : representation, computation, consciousness, etc..
• Clarify methodological issues : central assumptions, etc..
• Provides a systematic interpretation of what cognitive science is about
• Participates in development of new theories
• Examines relevant philosophical issues
• Exporting results in cognitive science to philosophy

1. What is weak AI according to Searle? How is weak AI different from the claim that the mind has a computational level?
2. Show what is wrong with this reply to the Chinese room argument : the person might not understand Chinese at first but he can try to understand the rules and come to learn Chinese from them. So programs are still sufficient for understanding.
3. Some other questions to think about can be found at :
   http://www.nyu.edu/gsas/dept/philo/courses/mindsandmachines/assgt4.html

Readings

• The sections on "Language of Thought" in Ned Block's article, at http://www.nyu.edu/gsas/dept/philo/faculty/block/papers/msb.html
• Optional : on intentionality, see http://mitpress.mit.edu/MITECS/work/lycan1_r.html ;

Intentionality and Mental Representations

• Reasoning and thinking involves states which are intentional : these are states with content, and their content determines what they are about.
• Words and sentences are also intentional. We might suppose that their contents derive from the contents of thoughts and beliefs, which in turn derive from the content of those mental representations which constitute our beliefs and thoughts.

Some Basic Questions

• What are these representations like? How are they different from natural languages?
• How do these representations get their meaning?
• How do these representations interact in reasoning, in the formation of perceptual beliefs, and in the production of behavior?

Compositionality and the Exaplanation of Systematicity, Productivity and Inferential Coherence

• Systematicity is the fact that our thoughts are systematically related in their contents - the same concepts can be involved in different thoughts.
• Productivity is the fact that in principle we can entertain indefinitely many new thoughts.
• Inferential conherence is the fact that in reasoning we seem to make use of certain general principles.
• If mental representations are compositional then these facts can be explained easily. Compositionality means that (a) there are complex representations made out of basic representations, and that the contents of the complex representations depend on the contents of the basic ones and their organisation.
• Notice that imagistic representations can also be compositional.

The Semantics (Meaning) of Mental Representations

• We should not confuse a representation with whatever it is that it represents.
• We also should not confuse the question of (a) what is it that representation X represents (e.g. Y), and (b) why is it that X represents Y and not some other thing.
• The latter question is in part a philosophical question about what determines the contents of representations.
• The causal approach explains the content of a representation in terms of its causal relationship with what it represents, e.g. representation X is triggered only by Y.
  1. One problem with this proposal is that it is not clear how it deals with misrepresentation, e.g. sometimes non-Ys can also cause X.
  2. Some people who take this approach sometimes appeal to evolutionary or design function to deal with this problem. (See http://mitpress.mit.edu/MITECS/work/dretske_r.html)
• An alternative is functional role semantics : see http://mitpress.mit.edu/MITECS/work/block_r.html

On feature detection : http://mitpress.mit.edu/MITECS/work/barlow2_r.html

Hubel's homepage : http://www.neuro.med.harvard.edu/http/hubel/hubel.html

Lectures 5 & 6 : Explaining Consciousness

Readings

• Article on phenomenal experience at http://mitpress.mit.edu/MITECS/work/levine2_r.html, and on the explanatory gap at http://mitpress.mit.edu/MITECS/work/levine_r.html

Different Concepts of Consciousness

• Phenomenal consciousness : having mental states with qualia or phenomenal properties. These are mental states for which there is something it is like to have them.
• Self-consciousness : being aware of oneself and can think about oneself.
• Attention : the selection for a certain portion of our conscious experience for enhanced processing.
• An interesting and difficult question : how is phenomenal consciousness related to self-consciousness?
• The problem of explaining phenomenal consciousness is in part a philosophical problem.
  1. We might try to find the neural correlate of phenomenal consciousness.
  2. But it is very hard to see what kind of scientific theories can explain why there is something it is like to undergo a physical process. This is the explanatory gap.
  3. Is this gap one which is due to our current limited understanding, or does it mark the limits of scientific explanations?

The Absent Qualia Argument Against Computational Explanations of Consciousness

• Ned Block asks us to imagine organising the population of China to act collectively as a system that is computationally equivalent to a person. Question : does this system possess phenomenal consciousness?
• Block thinks that the system does not have qualia, and so computations cannot be sufficient for consciousness.
• But if so, what else is needed? Does it mean that qualia requires a particular material basis? Why?
• However, recall the Searle's Chinese Room Argument. It is hard to see how we can argue in this case that the system does not possess phenomenal consciousness.