When scientists make the bold statement that all computers are based on the infinite Turing Machine they are referring to the work of Alan Turing, an English mathematician who was a brilliant code breaker during the 2nd world war.
All machines follow instructions of some kind. As these instructions are often embedded in the nature of the machine we don’t always notice. Let’s consider a simple thermostat:- when the mechanical switch trips the thermostat 'ON' it ‘reads’ the same instruction set each time. If the line is completed then control goes to the next line of the same level unless the entire process ends [END]. If the answer to a question is YES then control passes to the next indented line.
if the temperature sensed and the set temperature are the same then .....if the air conditioner is on then ..........turn the air conditioner OFF: END .....if the heater is on then ..........turn the heater OFF: END if the temperature sensed is higher than the set temperature then .....turn the air conditioner ON: END if the temperature sensed is lower than the set temperature then .....turn the heater ON: END
I’ve capitalised the actions the thermostat takes. Note that the thermostat switches off whenever it actually does something. An infinite Turing machine is similar, accept a central process loops. That is, when it gets to the end of its instruction set it starts again at the beginning. For a simple thermostat we have:
Check to see if the temperature has changed if temperature NEW does not equal temperature OLD then .....RUN tolerance check program let temperature OLD equal temperature NEW Wait one minute Return to first instruction
The tolerance check program will check to see if the temperature has fallen to far below or is now to far above the set temperature etc and will run the switch ON/OFF devices program.
Thus the infinite Turing machine runs simple programs. In effect it can turn other Turing machines off and on.
This idea assumes that there is a central machine that controls all other machines. Is this how the brain works? There are some loops that seem to be essential, such as the ‘memory loop’, but the idea of a central processor does not work well for consciousness although some peripherals, such as sensors and effectors, could work this way.
The Turing machine is a non-changing device with non-changing code. If you start a Turing machine fresh each morning it will run in exactly the same way.
Neurons are not like that at all. They get tired, they respond to some changes in the blood, they fire at random, and they age and die. They don’t even retain the same connections throughout life. Dendritic growth is greatest between ages 40 and 60 (approximately).
Thus the hardware is nothing like the Turing machine. The ‘software’ also changes. And there is no central processor.
So what are the processes in the brain like? We can rule out the Turing machine model as the most basic organic unit has variable hardware and software parameters.
The key word in understanding human brain processing is ‘redundancy’. For any process there must be at least an alternative path or a non-critical loss with loss of processing units eg a loss of speed, loss of resolution, reduction in number of parallel processes and so on.
The Turing machine is simply from another planet when brain processing is to be considered. But that doesn’t mean that computers can’t be used to emulate consciousness, just that the computer must take an entirely different approach to that of the brain.
© Robert Karl Stonjek 2002