Tuesday, June 2, 2009

The Turing Machine as Context.

The following was first posted on the Biosemiotics Forum, operated by Fatima Cvrckova:

"Consider this example of the a priori nature of context vis-à-vis meaning. The Turing machine, of which modern computers are a close approximation, contains within itself the capacity for all possible computations. This is to say, the computer contains all possible computational meaning. Programs, on the other hand, are informational, and in particular, serve to whittle away from the total set of meanings contained within a computer, yielding the particular subset of meanings that define a specific computational goal.

"While programs can learn (for example, see the work of John Myhill 1963), computers cannot learn. Instead, computers are computationally omniscient. That is, the Turing machine embodies all possible computational meaning, a priori, just by its construction. The mere fact of construction provides the Turing machine with this omniscience."


1 comment:

Gordana Dodig Crnkovic said...

Turing Machine is just one of many possible models of computing. It is not a very good model of parallel computing systems such as Internet for example. It is not a very good model of information processing (computing) in living organisms. We are actually learning in the field of Natural Computing and especially in Organic Computing how to construct new computing machines able to perform subsymbolic computing such as analog and continuous computing.
Turing Machine is a logical model of symbolic computing in an isolated system. Real world physical information processing systems have a repertoire of behaviours for which Turing Machine model is not representative.

Gordana Dodig Crnkovic