From: Ian Wright (iwright@GMAIL.COM)
Date: Wed Jun 02 2004 - 17:05:36 EDT
Hi Howard I can't give you satisfactory answers to your questions. >the old ways of talking are likely to be around for a while since we still >find ourselves embroiled in controversies that pit substance on the one hand >against form on the other. To explore the ways in which form just is >substance may take us a bit beyond such disputes. I guess you're right about this. But sometimes theoretical progress is hindered by concepts that are inadequate for the subject matter. I didn't know there was a controversy regarding substance and form -- maybe you could fill me in. > I wonder if you have explored the connection of the engineering sense of > control systems that you mention to work that has been done on natural > kinds? I'm sorry no, but the notion of a natural kind is I think more general than that of a control system, which I think is a particular example of a natural kind. I've appended some thoughts below but they are not organised. > Does a control system so to speak individuate in that way -- I mean does one > organize a thing in such a way that it is the thing that it is? Or is that > kind of talk sort of not relevant? It is sort of relevant I think. A characteristic feature of a control system is the presence of negative feedback loops (the simplest everyday case is a thermostat). Traditionally in control engineering the systems deal only with quantitative information and the internal memory of the control system is quite limited. A more general notion of control system, more prevalent in artificial intelligence, includes more complex systems that deal with both quantitative and qualitative information, with very complex and enduring internal memory states. All these kinds of control systems are teleological systems in the sense that they have representations of goal states and they act to change the world so that it conforms to the goal state. Again, the simplest case is the thermostat. The goal state is its temperature setting which represents, in virtue of its causal connections to the world, a non-existent or potential state of the world. The thermostat autonomously acts to turn the heating up or down so the ambient temperature of the room conforms to its goal state. This a paradigmatic example of a naturalised representation, albeit of a very simple kind, and indicates how semantic states are ultimately causal relations between representations and referents. This is why I tend to mention the thermostat a lot. There are some philosophical positions, such as Wittgenstein's I think, that deny that semantics are ultimately explicable in causal terms. But moving away from control engineering and thermostats we see that control systems are ubiquitous in nature and society. For example, homeostasis in plants and animals, chains of command in organisations, etc. Homeostatic mechanisms maintain states-of-affairs, or material configurations over time, so in that sense are mechanisms that act to constitute a thing as it is. Generalising further, we can also view minimum energy states in physics, such as the settling of atoms into molecular arrangments at lower temperatures, as examples of homeostasic mechanisms, for once a molecule is in an energy minima it is resistant to change from small perturbations, that is it retains its coherence and causal properties. One property of natural kinds is that they are stable over the period of observation, and control systems produce stability. Consider your example of the natural kind H2O, which endures at certain temperatures due to its minimum energy configuration. > Is this or that control system readily distributed over a variety of tasks > or more or less specific to a given kind of social relation/activity. Depends on the sophistication of the control system -- I'm using the term in very general manner. Most artificial control systems are dedicated. Many natural control systems, like most organisms, are very flexible and sophisticated but confined to operate in restricted niches and fail outside of them. But a human seems to be very general-purpose. But that could be straightforward conceit. > And what do you have in mind (generally) by the reference to emergent novel > ontologies? Only that qualitatively new things with new powers can emerge from another ontology, for example the emergence of a mechanism to allocate labour in a community that exchanges its products. The question of how emergence occurs in general is very difficult, and I'm sure I cannot do it justice. > "At root, representation is implemented via causal links, and commodity economies > instantiate the right causal relationships to support the representation". > > My impression is that this sort of thing is not sufficiently well developed > either in the analysis of Marx or more generally. The purchase of commodities with money by economic actors is a feedback signal -- it is no accident that money flows in the *opposite* direction to commodity flows. Once the economy is identified as a kind of control system some questions naturally arise. What is the goal state in this type of control system? How is it represented? What causal relations support the representation? By what mechanism are the objects of control brought into conformance with the goal state? How does the system react to perturbations? And so on. I think this gives a different perspective on economic questions, for example those concerning value as substance or form, although the perspective has more in common with dialectical materialism than analytical approaches to economic theory. For example, in both the neo-Ricardian and Arrow-Debreu approach to economic modelling prices do not function as a control signal. This is fine as an abstraction, and I have no quarrel with different ways of slicing the cake. But once disequilibrium prices are considered in a dynamic setting it becomes clear that prices have causal effects on the configuration of the system, rather then simply being nominal outputs consistent with a particular unchanging configuration. In static models prices are epiphenomena -- they do not feed-back into the system. The total causal relations that support the semantics of price are missing in such models, and therefore the meaning of price in these theories is limited and incomplete; for example, thinking that price is only an index of scarcity. ATB -Ian.
This archive was generated by hypermail 2.1.5 : Thu Jun 03 2004 - 00:00:01 EDT