Re: Turing Test
Terry McD writes: I would like to again raise a topic that continues to underlie the discussion on GE Theory. Sensibly rising to some bait, Gil contended in response to one of my postings that the Roemer type models actually produce rather than merely "simulate" radical results. I think this is worth pursuing. I remember an analogy from grad micro defending utility theory in the following manner. Imagine a sealed clock. One can observe the movement of the clock's hands in front of its dial. One cannot see the mechanism which drives the hands. The observer hypothesizes a mechanism inside the clock which is capable of producing the observed movements. The observer has now explained the movement of the clock's hands for all practical purposes. It may still be possible to present alternative mechanisms for driving the clock's hands but one cannot contend that they are any more correct than the first mechanism. This metaphor stands as long as the clock remains sealed. However, if we extend the metaphor in the following way, the matter changes drastically. Suppose you wanted to alter the way the clock's hands moved. The observer would then have to intervene in the working of the clock's mechanism. It would then be crucial to have not just a model of the hands' movements, but the correct model of these movements. Having somehow ascertained the working of the clock's mechanism, it would become clear that only one model produced the movements. The other model(s) merely simulated the mechanism driving the hands. Pen-llers will have seen where this argument is driving. The point is not merely to explain the world, the point is to change it. This is what gives urgency to the discussions surrounding bastard Marxism. Terry here gives an argument I have long made against Milton Friedman- ish type arguments to the effect that prediction, rather than explanation, is the true test of theory. So he's not telling me anything new. But more to the point, he's missing the point, especially in terms of the comparative relevance of Roemer's argument to Marx's. 1) Roemer's account of capitalist exploitation, even if it doesn't qualify as a general theory, is not just some half-ass story, but a logically coherent account with arguably strong historical support. 2) In light of Roemer's account, it is clear that Marx's account in Capital is both logically incorrect and empirically suspect. For example, Marx does not preclude that the elongations in the working day so painfully elaborated in Ch. 10 of Vol I are due to relative changes in labor supply perfectly in keeping with Roemer's account. 3) Granting that the point is to change the world (and who ever denied it?), changing the world on the sole basis of Marx's theory would lead, I bet, to some pretty discouraging results. For example, if a law were passed mandating that all firms be labor- owned, Marx's argument of Vol I, Ch. 5 and Vol III, Chs 21-23 would lead to the conclusion that the elimination of labor's subsumption under capital would lead to the elimination of exploitation. But it is *at least* plausible, following Roemer, that this would not occur as long as firms still relied on interest capital. For such reasons, I fully agree with Terry about the urgency of the discussions surrounding "bastard Marxism". But which is the bastard? Gil [[EMAIL PROTECTED]]
Turing Test
I would like to again raise a topic that continues to underlie the discussion on GE Theory. Sensibly rising to some bait, Gil contended in response to one of my postings that the Roemer type models actually produce rather than merely "simulate" radical results. I think this is worth pursuing. I remember an analogy from grad micro defending utility theory in the following manner. Imagine a sealed clock. One can observe the movement of the clock's hands in front of its dial. One cannot see the mechanism which drives the hands. The observer hypothesizes a mechanism inside the clock which is capable of producing the observed movements. The observer has now explained the movement of the clock's hands for all practical purposes. It may still be possible to present alternative mechanisms for driving the clock's hands but one cannot contend that they are any more correct than the first mechanism. This metaphor stands as long as the clock remains sealed. However, if we extend the metaphor in the following way, the matter changes drastically. Suppose you wanted to alter the way the clock's hands moved. The observer would then have to intervene in the working of the clock's mechanism. It would then be crucial to have not just a model of the hands' movements, but the correct model of these movements. Having somehow ascertained the working of the clock's mechanism, it would become clear that only one model produced the movements. The other model(s) merely simulated the mechanism driving the hands. Pen-llers will have seen where this argument is driving. The point is not merely to explain the world, the point is to change it. This is what gives urgency to the discussions surrounding bastard Marxism. Terry McDonough
Turing Test
I would like to again raise a topic that continues to underlie the discussion on GE Theory. Sensibly rising to some bait, Gil contended in response to one of my postings that the Roemer type models actually produce rather than merely "simulate" radical results. I think this is worth pursuing. I remember an analogy from grad micro defending utility theory in the following manner. Imagine a sealed clock. One can observe the movement of the clock's hands in front of its dial. One cannot see the mechanism which drives the hands. The observer hypothesizes a mechanism inside the clock which is capable of producing the observed movements. The observer has now explained the movement of the clock's hands for all practical purposes. It may still be possible to present alternative mechanisms for driving the clock's hands but one cannot contend that they are any more correct than the first mechanism. This metaphor stands as long as the clock remains sealed. However, if we extend the metaphor in the following way, the matter changes drastically. Suppose you wanted to alter the way the clock's hands moved. The observer would then have to intervene in the working of the clock's mechanism. It would then be crucial to have not just a model of the hands' movements, but the correct model of these movements. Having somehow ascertained the working of the clock's mechanism, it would become clear that only one model produced the movements. The other model(s) merely simulated the mechanism driving the hands. Pen-llers will have seen where this argument is driving. The point is not merely to explain the world, the point is to change it. This is what gives urgency to the discussions surrounding bastard Marxism. Terry McDonough
