Chapter 19

Noam Chomsky

-Trained as linguist, he wrote a review of Skinner's work on language, showed the limits of a behavioral explanation and beginning the "cognitive revolution." -Much of modern psycholinguistics centers around his theory of language.

Cognitive Science

An interdisciplinary approach to studying the mind and mental processes that combines aspects of cognitive psychology, philosophy, artificial intelligence, neuroscience, linguistics, and computer science.

Attribution

For Heider, the basic patterns of explanation we use to make sense of the world. The fundamental attributions include effort, ability, task, and luck.

Jerome Bruner

-Along with Miller and Bartlett, one of the first cognitive psychologists. -Among his contributions were the popularization of Piaget and Vygotsky.

George A. Miller

-Did pioneering research on information processing in the 1950s and 1960s that significantly enhanced the popularity of cognitive psychology.

Jean Piaget

-Focused on cognitive development, and how schemata evolve during maturation and through experience. -Posited a well-known stage theory of intellectual development in children from birth to adolescence.

Alan Turing

-He is considered the father of Artificial Intelligence in computer science and psychology. -Among his contributions was the Turing test.

Ulric Niesser

-Noted cognitive psychologist. -Authorized two classic textbooks and advocated for cognitive research that was both applied and ecologically valid.

Sir Frederic Charles Bartlett

-One of the first modern cognitive psychologists. -Noted for his use of schema to explain the reconstructive nature of memory.

John Searle

-With his famous "Chinese Room" thought experiment, sought to demonstrate that computer programs can simulate human thought processes but not duplicate them. -Computer programs, he says, can only manipulate symbols according to rules (syntax), whereas humans assign meaning to symbols (semantics). -Therefore, he accepts weak artificial intelligence and rejects strong artificial intelligence.

Artificial Intelligence (AI)

A branch of computer science that investigates the extent to which machines can stimulate or duplicate the intelligent behavior of living organisms

Neural Network

A system of input, hidden, and output units that is capable of learning if the mathematical weights among the units are systematically modified either according to Hebb's rule or by back-propagation.

Turing Test

A test devised by Turing to determine whether a machine can think. Questions are submitted to both a human and a machine. If the machine's answers are indistinguishable from those of the human, it is concluded that the machine can think.

Connectionism

The most recent type of AI that utilizes artificial systems of neurons called neural networks. As contrasted with GOFAI, which employed the sequential processing of information according to specified rules, new connectionism employs the brain as a model. That is, the processing of information within a neural network is distributed throughout the entire network. Like the brain, neural networks are capable of learning; this was not true of GOFAI.

Back-Propagation Systems

Neural networks that are programmed to learn by systematically reducing the discrepancy between their output and some desired output represented by a model or "teacher". Such systems learn by corrective feedback instead of by applying Hebb's rule.

Information Processing Psychology

The approach to studying cognition that follows in the tradition of faculty psychology and methodological behaviorism and typically employs the computer as a model for human information processing.

Strong Artificial Intelligence

The contention that machines (such as computers) can duplicate human cognitive processes.

Weak Artificial Intelligence

The contention that machines (such as computers) can simulate human cognitive processes but not duplicate them.

Hebb's rule

Hebb's contention that neurons within the brain that are simultaneously or successively active become associated. One type of neural networks applies this rule by adjusting the mathematical weights of units that are simultaneously or successively active. The result is that consistent input gradually produces consistent output