X-Message-Number: 14713 Date: Tue, 17 Oct 2000 08:00:55 -0400 From: Thomas Donaldson <> Subject: more on brains and computers Hi everyone! I note the problems John K Clark has with my recent message about how we work. My very first impulse is to suggest that he read it again. To look at his last point, as to whether or not our brains can meaningfully be considered as computers (and thus following Turing's Theorem etc etc) I did NOT claim that they could not be, I only raised the question. AS further discussion of this issue, I will point out first of all that we're all aware of many machines that no one would claim were computers. At the same time many of these machines do react to things happening in the world outside them: consider a Model T car, or a lawnmower. A subsidiary but important question here is that of whether or not symbolic thought is the only kind we do. If so, I'd agree that there is a strong case that we can be considered as computers. However I doubt very much that symbolism is the only way we think. Underneath that symbolism we have other kinds of thinking which we share with most mammals: we learn to react to events in the world, and even think about them, without at root using symbolism at all. In particular, this kind of thinking takes place on a higher level than simply using nonsymbolic objects (like semiconductors or brain cells). Our brain cells connect to many others, and they send and receive impulses from them. At root, these impulses do not come from symbols but from impulses received from the world. Moreover, that other kind of thinking is essential to the way we work. We learn language not by getting other language but also by our experience of the world, which is not linguistic. As for our parallelism, it's not important simply that most of our brain works in parallel as the exact structure of that parallelism. Our brain grows and destroys connections between our neurons, constantly. This at least should raise questions as to whether we operate like computers; I'm not saying no computer will ever be built to do this, but if it were built that way it would behave differently. (And perhaps we would not consider it a computer). Closely linked with this feature is the close relation between our drives and our thinking, which fundamentally cannot be separated. It is not the use of symbolism, but the domination of our drives, which characterizes our thinking ... even though we have many drives such as curiosity which are far from primitive. If you wish to see your brain along the model of a Turing machine, then the first thing you should do is to work out whether or not such a machine can model one which grows and destroys its connections as part of its response to the world. The second is to see whether a Turing machine can work WITHOUT use of any symbolism. We all know that for a real machine operating in the real world, parallelism is necessary (otherwise if attacked it may take too long to work out that it's being attacked and respond). In itself parallelism only raises the problem of whether or not Turing machines dealt with an important real-world trait; but we aren't just parallel. We have a structure and mode of thinking which no present computer imitates. Best wishes and long long life to all, Thomas Donaldson Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=14713