Cognitive science is the interdisciplinary study of mind and intelligence: psychology + philosophy + artificial intelligence + neuroscience + linguistics + anthropology
Understanding the human mind requires numerous methods and theories.
Psychological experiments, computational models, brain scans, etc.
My own history: philosophy -> philosophy of science -> psychology -> computational models -> neuroscience.
Theoretical: explain how thinking works
Applied: e.g. robotics, education, design, management, mental illnes.
Impossible to review everything in all 6 fields. Instead, I am focussing on a central topic in all these fields: What mental representations does the mind use(or can computers use) to develop thinking?
Major focus will be on human cognition, but we will also discuss machine intelligence.
Readings, 3 exams, essay. See syllabus.
Lecture notes will be updated on Web site weekly.
e.g. pictures, "horses have tails", "horses have lungs"
"goed"
We are going to look at various representational schemes that have been proposed and debated. We will also consider the more philosophical question of whether cognitive science is a mistake: consciousness, emotions, knowledge without representation.
Example: What do you know about the university?
How do you decide what classes to take?
Rules: e.g. avoid 8 a.m. classes.
Concepts: e.g. "gut"
Images: e.g. route from one part of campus to another
Cases: e.g. particularly good or bad classes.
Fundamental philosophical questions since Plato
1. What is the nature of mind? Metaphysics and nature of reality.
2. What is the nature of knowledge? Epistemology and nature of knowledge.
3. How do we think?
Psychology, 1890s. Behaviorism: can't study what is in the mind.
1950's. Miller, etc.: mind has structure.
Artificial intelligence, 1956. Minsky, Newell, Simon, McCarthy.
Linguistics 1956: Chomsky versus behaviorist view of language. Innateness.
Anthropology: social, cultural aspects of knowledge
Neuroscience: how does the brain make a mind?
How can all these fields, with different histories and methodologies, cooperate to produce an understanding of mind?
Central hypothesis:
Thinking = representational structures + 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.
data structures mental representations
+ algorithms +procedures
= running programs =thinking
Methodological consequence: study the mind by developing computer simulations of thinking.
1. Deduction
e.g. Descartes, I am a thing that thinks
2. Thought experiment
e.g. imagine brain transfer
3. General theorizing
e.g. develop theory of mind
dualism, materialism, functionalism
4. Case studies in history and philosophy of science.
1. Experiments:
e.g. mental rotation, analogy, concepts
2. Theorizing: postulate structures and processes
1. Write programs aiming to be intelligent.
evaluation can either be cognitive or engineering
e.g. SOAR: see what it can do, and see whether it can do it like people.
Psychologists and philosophers now also use this methodology.
2. Theoretical analyses, e.g. computational complexity.
1. Experiments, but biological, not cognitive.
e.g. record cell firing, PET and MRI scans, lesions
2. Also theoretical
3. Also computational: build computer models of mind, e.g. neural networks
1. Judgments of grammaticality: syntax
2. Semantics, pragmatics: data, theory
3. Computional models
Ethnography, but pay attention to how people think. Like psychology, but less experimental, more cross cultural.
Psychologists are also doing cross-cultural studies.
Everyone is theorizing, but there are different methods of empirical evaluation:
Each of these has limitations, but all can contribute toward evaluating theories of mind.
See Mind, p. 15, Box. 1.1.
Herbert Simon is one of the founders of artificial intelligence and cognitive science.
Computational Epistemology Laboratory.
This page updated Sept. 10, 2012