From: Ian Wright (iwright@GMAIL.COM)
Date: Thu Jun 03 2004 - 12:47:58 EDT
Hi Ajit Thanks for your reply. > Ian, I think what I'm saying holds even for a causal > theory. A theory of price or for that matter a theory > of anything, say X, is supposed to explain the > phenomenon of X. To say that the value of X in time t > is determined by the given value of X in time t-1 is > not a causal theory that explains the phenomenon of X. > Because in your formulation, it is logical to say that > X in time t depends on the value of X in time (t-2), > since the value of X in time t-1 is explained by the > value of X in time t-2. This logically leads us to > infinite regresson. One will have to answer, how did X > come into being in the first place. And here your > causal explanation must identify a cause other than X. > That's why a theory of X that explains X on the basis > of X is not a theory. If you are doing forcasting etc. > then, of course, it makes sense to take into account > the previous values of the variable. But then > forcasting and theory are two different animals. I haven't read Freeman's response to this criticism, but the infinite regression argument is confused. Before explaining the confusion let's apply your argument to another domain, the theory of evolution. The purpose of the theory is to explain the development of species. Organisms generate copies with random variation. Selection occurs non-randomly. Therefore organisms adapt over time. But evolutionary theory explains the change in organisms by reference to previous organisms. By your criterion it is therefore subject to infinite regress and neither a theory nor an explanation of the development of species. > One will have to answer, how did X come into being > in the first place. And here your causal explanation > must identify a cause other than X. And the origin of life is a topic in the theory of evolution. But biologists did not reject the theory of evolution because its result is also a premise. The infinite regress argument gains its plausibility from the correct stipulation that the explanans should not refer to the explanandum. But the flaw arises from the failure to distinguish between a type and instances of a type. The unstated premise is that the X at time t is identical to X at time t+1 and therefore the explanation is circular. But in fact they are different instances, although of the same type, hence the explanation is non-circular. An explanation of price at time t+1 can refer to price at time t, just as an explanation of the evolution of horses can refer to earlier horses in history. The explanans and explanandum are different, although they happen to be of the same type. TSS models do not need to explain the origin of prices to qualify as an explanation of prices. They can assume price at a time t as a parameter. Not only is this a legitimate modelling approach, it allows prices to have a causal, not solely a nominal role, just as they do in reality. I am similarly unimpressed with Steve Keen's infinite regression argument against Marx's theory of value in "Debunking Economics", which he raised on OPE-L some time ago, e.g. OPE-L:5648. In general, this particular argument seems to be a misuse of static logic applied to historical processes. I do not think there is need to spend any more time on this objection. I'll reply to your interesting objections to the MELT when I get the chance. ATB, -Ian.
This archive was generated by hypermail 2.1.5 : Sat Jun 05 2004 - 00:00:01 EDT