From: Ian Wright (ian_paul_wright@HOTMAIL.COM)
Date: Thu Jan 22 2004 - 16:40:48 EST
Andy, Thanks for being open-minded about this, despite my jibes at DM. I think that because Trostky once said that rejection of the dialectic leads to the "to the camp of bourgeois reaction", any criticism of DM can set alarm bells ringing, which has a tendency to shutdown conversation. >You make the distinction between formal logic and computation. >Indeed your point that formal logic is 'useless' for certain aspects of >robot building supports the above line of thinking. What is the >essential difference between computation and formal logic? >Clearly there is a need to relate the developments you >describe to materialist dialectics. A job of relating 'systems theory' >to materialist dialectics has been done by Levins in Science and >Society 1998, 62, 3. I wonder if this has any bearing on your >concerns? These two questions are related. I obtained a copy of Levin's article and read it. I think there is a widespread, but understandable, confusion here. The theory of computation is different from and more general than any particular formal logic, or systems theory (which is what Levin discusses), or dynamical systems theory (which, for example, Barkley-Rosser has related to DM in the Cambridge Journal of Economics). This can be understood in practical terms by noting that different kinds of formal logic, control systems and dynamical systems may in principle be implemented and run on a computer, a Universal Turing Machine (UTM). The theory of computation is more general than any of these particular formalisations. Compared to, for example, formal logics, such as the basic propositional, predicate and the lambda calculus, which can be interpreted as embodying ontological suppositions, such as the world is composed of discrete entities that can be uniquely identified and are not subject to change during the period of inference, the theory of UTM and computational complexity does not make such presuppositions. Again, computation is more general and abstract than any particular formalism. Another point, which Stephen Wolfram illustrates well in his book, is that many different kinds of machines, physical and virtual, from cellular automata, tag systems, Turing Machines, augmented finite state machines, to types of formal logical systems, dynamical systems, Diophantine equations, and computer programming languages, to name only a few, are *all* capable of implementing a UTM, and therefore in principle able to simulate any other, although not necessarily efficiently. This can be understood in practical terms by noting that all programs written in any computer programming languages can be translated to run in any other language. Choosing which to use is a matter of convenience, not necessity. Similarly, it is possible to implement a Von-Neumann type serial processing architecture on a neural network, and vice-versa. The important point to keep in mind is that universal machines are ubiquitous, and do not require much machinery to realise, and that the concept of computation is not tied to any particular formalism or machine embodiment. To make this very concrete, it would be possible to organise a group of people in a room to function like a UTM, and therefore in principle execute all kinds of programs (compare Searle's Chinese Room). I tend to think of computation as a the most general and abstract theory of causality that we currently have, although I do not think this is a widespread view, and I certainly am not a sufficient expert to pronounce with great confidence. As I understand it, DM aspires to a similar level of generality, although it does have a set of technical concepts, such as negation, interpenetration of opposites etc., that make ontological suppositions of a more concrete (although still highly abstract) character compared to those made under the heading of computation. For example, despite many misconceptions, computational processes are not only about quantitative relations and the manipulations of numbers, but may equally execute qualitative relations and manipulate symbols, a point that Ada Lovelace understood, perhaps more so than Babbage, when she wrote about new designs for his calculating machines. Another popular misconception is that computational processes cannot support semantic states, and are only syntactical. Given these considerations, I am not wholly satisfied that the relation between computation and DM has been studied with sufficient interest or depth. It seems that it has been confused with particular formalisms or models, such as formal logic, control system theory, or dynamical systems. >More substantially, whilst I think there are >great problems in critical realism it has at least popularised the >concept of 'emergence' [really the transormation of quantitative into >qualitative change], which is enough to establish that human society >and psychology are emergent from the micro-processes which >computation theory and AI etc. dwells upon. As such the study of >psychology and society does not require the study of such micro >processes; rather the study of society and psychology should entail >the study of the specifc mode of production. This puts the research >you refer to -- its appropriate goals and wider significance -- in a >different light to that which its practitioners see it. When I first read Bhaskar's "Realist Theory of Science" I had no difficulty understanding the concept of emergence. I think the reason was that I had worked with software for a long time. Software engineers work with emergent ontologies every day. For example, consider that I write a program that performs some matrix multiplications. I run it, and an answer pops out. For the duration of the execution of the program a matrix existed. I could not see it directly, nor could I touch it, and neither could I determine its existence from examining the "physical" machine, for example the play of electrons through the circuitry. This is an example of a real but abstract object that is implemented on a "lower" physical level. The point I wish to make is that the idea of computation can contribute in a major way to understanding the nature of emergence and a structured reality. The computer, in this sense, is an artificial physical arrangement for the controlled study of computational processes, but it is very natural to generalise and posit that computational processes are already "out there", always have been, and that the universe is computable. It is not entirely clear that this is a falsifiable proposition, but it does indicate the level of generality that the concept of computation can encompass. Again, this is a kind of panlogicism, which I believe relates to some of the philosophical ideas within DM. My aim here is to point out that those with a background in DM who wish to develop it further will find a lot of interesting, new technical material in the area of computation, which might benefit from a realist interpretation. -Ian. _________________________________________________________________ MSN 8 helps eliminate e-mail viruses. Get 2 months FREE*. http://join.msn.com/?page=features/virus
This archive was generated by hypermail 2.1.5 : Sat Jan 24 2004 - 00:00:02 EST