Post date: 16-Feb-2012 23:36:36
Universal Turing machines have come to represent the definition of a computation and underpin the most basic architecture of modern computers.
The Turing machine was developed in response to Hilbert's tenth problem (the Entscheidungsproblem) which enquired as to the existence of a general algorithmic procedure for solving mathematical problems.
Universal Turing machines manipulate an ordered list of symbols which represent some 'problem' along with the required algorithm. Hopefully, a Universal Turing machine will halt and provide an ordered list of symbols which represent an 'answer' to the 'problem'. The manipulations are wholly deterministic.
Modern computers are a naive popular analogy of what brains do. Artificial Intelligence research employs computers in a far more enlightened manner. However, there are some aspects of the underlying Turing machine model that cause me disquiet.
Firstly, A Turing machine is intended to halt when it arrives at an answer, while life is characterised by its continuance. When life halts it takes its consciousness and its intelligence with it. Turing machines appear superficially to be dynamic in their operation. However, the Turing machine is completely isolated from the environment for the entire period between the start and the stop, at which time it delivers a static answer. It matters nought if the time is long or short or if the machine is fast or slow in its operation, the end result is, by definition, static. Intelligence is a dynamic phenomenon which is useful until it halts while a Turing machine is a device which is useful when it halts. These two seem poorly matched. This dichotomy is perhaps simply a different perspective on the distinction drawn by Dennett between the tenacious Cartesian Theatre model and the Multiple Drafts model.
Secondly, a Turing machine exists on an exclusive diet of symbols. In a more familiar setting, a function in a computer programming language is defined by the human programmer as a set of parameters and an algorithm. The human understanding of the 'problem' is in the validity of the algorithmic relationship between the particular parameters of the function. Consider the familiar formula f=ma in Newtonian physics which relates force, mass and acceleration. The 'understanding' is in the fact that if mass and acceleration are manipulated 'just so' then the result is something useful. Performing the identical manipulation with mass and temperature just isn't useful. The possibility for 'understanding' is simply withheld from the symbol manipulator, be it a person or a machine. The human supplying the parameters to a function (and deriving 'meaning' from the result) is able to appreciate the analogy between one force-mass-acceleration tupple and another in order to use the result.
"The idea that Turing machines do not capture all computation, and that interaction machines are more expressive, seems to fly in the face of accepted dogma, hindering its acceptance within the theory community."
Goldin and Wegner in "The Church-Turing Thesis: Breaking the Myth"