|Author: Peter Bradley|
In the 1940s, Alan Turing, a British Mathematician, originally proposed the structure of what has become called a 'Turing machine'. The idea is simple, yet provides the basis for all modern computers and is the inspiration for the physical symbol system hypothesis. In this module, we build simple Turing machines and investigate how they might implement a symbolic model.
Turing machines are idealized computers. They are simple symbol processors that follow a strict set of pre-defined rules. They are, however, extraordinarily powerful computational devices, which have provided the basis for much of cognitive science as well as the computing revolution of the past century.
Symbolic models are theories of human cognition that take the form of working computer programs. They are distinguished from other models insofar as they process explicitly defined symbols via a set of explicitly defined rules.
In order to create a symbolic model, we need two things: a set of symbols and a set of rules to operate on those symbols. In this module, we begin construction of the latter, and harder, of these two: the rules.
Now that we have our rules, we have to create a system of representations, and a machine that can follow the rules to operate on those representations. One such machine is called a 'Turing Machine'. In this module, we introduce these machines, and show how they can implement simple machine tables.
The Turing Machines we have considered so far can only implement those rules that are loaded into the machine table. We seek a machine that can read in its own machine table, and thereby simulate any Turing machine. Such a machine is called a 'Universal' Turing machine, and in this module, we demonstrate how to construct one.
In his book The Language Instinct, Steven Pinker suggests that we could build a Turing machine to parse simple English sentences. In this module, we see if he is right.
In 1950, Alan Turing hypothesized that a Turing machine could be built that would be indistinguishable from a real human in dialogue. In his paper, he raised, and then rejected, a number of different objections to that idea. In this module, we offer his arguments for consideration.