Phil/Psych 256, Week 12

WATERLOO THIS-IDEA-MUST-DIE DAY, DEC. 14, Arts lecture hall 208, 1-3:30 p.m.

The Future of Cognitive Science: Integrations

What is Cognitive Science?

Cognitive science is the interdisciplinary study of mind and intelligence.

Central hypothesis:

• Thinking = representational structures + computational procedures that operate on those structures.
• This is broad enough to encompass rules, concepts, images, cases, and distributed representations.
• This hypothesis may be wrong, but it is far more powerful and successful than any competing hypothesis to date.

Challenges to the computational/representational view can be met by expanding and supplementing it to take into account the importance of consciousness, emotions, embodiment, social relations, and dynamic systems.

Futures for Cognitive Science:

1. Reject because of challenges?
2. Hybrid models?
3. Theoretical neuroscience?
4. Multilevel explanations

Multilevelism

Reject reductionism (only one fundemental level of explanation) and anti-reductionism (higher levels are autonomous).

I propose multilevelism. This view takes seriously the fact that operation at different levels of understanding with different methodologies is essential to understand nature of mind. But these levels and methods are not independent of each other: the strength of cognitive science depends on integration of them.

Example 1: multilevel approach to consciousness:

• Social mechanisms: consciousness evolved so that people could understand and teach each other.
• Psychological mechanisms: consciousness requires representations of representations, e.g. your experience of yourself feeling happy about the term ending.
• Neural mechanisms: convolution and/or synchrony bind together patterns of neural activity into new, integrated ones.
• Molecular mechanisms: neurotransmitter such as dopamine and hormones such as cortisol affect conscious experience.

Example 2: multilevel approach to creativity:

• Social mechanisms: interactions among people spread ideas and produce new combinations.
• Psychological mechanisms: generate new ideas and hypotheses by conceptual combination, abduction, analogy, etc.
• Neural mechanisms: generate new patterns of firing in neural populations by convolving old ones.
• Molecular mechanisms: dopamine contributes to learning, emotion, and

Work at every level should appreciate of work at other levels, particularly those directly above and below it. No level is immune from revision, although no level is expected to be wiped out from above or below.

Advice to students and researchers:

 Pick the level and method that interests you most and best suits your capabilities, and pursue them cooperatively with others working at different levels. Stay aware of implications of cognitive science for design, intelligent systems, human-machine interaction, and education.

Some open questions

Theoretical questions

1. How do neural networks in human brains produce rule-based, concept-based, analogical, and imagistic thinking?

2. How do emotion and cognition interact to produce intelligence?

3. How does consciousness emerge from neural activity?

4. How does brain chemistry contribute to intelligence?

Practical questions

1. How can cognitive science help to improve learning and teaching?

2. How can more intelligent robots be developed?

3. How can intelligent computer systems enhance human performance?

4. How can cognitive science contribute to more interesting games by modeling more realistic opponents?

5. How can health, intelligence, and happiness be increased by neurochemical interventions?

Ethical questions

1. If brain scans can help to identify psychopaths and others likely to commit crimes, should they be used?

2. If cognitive science contributes to the development of artificial intelligence, would this make people obsolete?

3. Should there be limits on drug modifications of brain activity?

4. Is free will an illusion?

Phil/Psych 256

Computational Epistemology Laboratory.

Paul Thagard

This page updated Nov. 24, 2015