Human thinking, like human action, is based on the use of tools. Just as we use fork-lift trucks, earplugs, and fire to manipulate the world around us, we use tools like arithmetic, reading, and reasoning to manipulate our thoughts.
Some tools are more effective than others; the aeroplane is a faster mode of travel than a horse, and a industrial press more precise than a blacksmith’s forge.
This has made me consider computer programs and their relationship to human thought. Arithmetic was once the unique preserve of human mathematicians, but now it’s almost entirely a machine task. I’d venture to guess that the majority of letters now resided on computer, too. I certainly get more email than post. Much that was once done by considering information is now done by loading a computer with the information and asking it to do the considering.
All of this is rudimentary, but I’d like you to see where I’m coming from; certain forms of thinking are given to machines where once they were human activities.
I’d like to suggest, further, that computers can assist humans in one of three ways; first, they replace humans, as happens in a computer-controlled manufacturing plant. (While humans might have an overview role, they don’t have a construction role) Second, they assist humans, as seen in spell-checkers in word processors. The human is relying on the computer’s fastidiousness or precision or fast recall to take part of the task away. Third, computers offer no help at all, which is often the case in a social situation.
Now, computers can assist to the extent that they can understand your information. A computer can add up a list of numbers, but it can’t add up an essay; the adding-up program must be fed a list of numbers, and nothing else. Because of this, programs are often written in a way that only lets you type in those numbers.
So, computers can assist in simple tasks, like arithmetic, and reference tasks, like spell-checking. While computers are capable of processing vastly complex tasks, they also tend to be very specialised tasks, often built up of lots of simple tasks.
To the extent that thinkers have invented different modes of thinking, to what extent can computers help automate these tasks? A clerk has his job vastly helped by an adding machine, but an abstract thinker often has to rely on pen and paper and the review of his peers, a system that hasn’t improved since ancient Greece.
So to what extent is it possible to improve thought by modelling complex and general thinking systems? That’s a topic rarely touched on.