>> The Sraffians think neoclassical economics from 1870 to, say, 1930 involved
>> the extension and adaptation of the theory of "market prices" to "normal
>> prices" - that is, the continuation of the Classical method, but the
>> adoption of a different theory. Normal prices, under competition, involve
>> a process of investment moving between sectors until the rate of profit has
>> been equalized in all sectors. The resulting neoclassical theory has never
>> been expressed coherently and in keeping with underlying conceptions of
>> substitution. Neoclassicals were forced to adbandon this method of
>> analysis based on comparisons of long run positions. Instead, they started
>> creating new equilibrium concepts, e.g. temporary and intertemporal
>> equilibrium. These new concepts do not have clear dynamics or admit of
>> clear predictions.

> And Sraffarian models have clear dynamics and admit clear
> predictions?


Well, I would have to admit that economists have alot of
equillibrium concepts ( Rational Expectations Equillibrium,
Limited Information REE, equillibriums based on some sort of
bounded rationality, ect, and of course various equillibrium concepts
common in game theory such as Nash Equillibrium, coordinated equillibrium,
iterated dominance, ect).  Which is perhaps why so much economic
theory at the moment is devoted to debate over which concept is
most useful.

But such concepts do have clear dynamics and admit clear
predictions.  True, sometimes one has multiple eq. or whatever,
but one can almost always still make predictions and statements
about the dynamics.  Why not?  Look at the enormous literature,
for example, using REE to model dynamic behavior and get clear
predictions?  Or the enormous literature on getting moment restrictions
from REE?  Or using experiments to see if the experimental dyanamics
are in accordance with bounded rationality ideas?

(of course, maybe I'm missing the point, since I haven't mentioned
long run positions or temporary equillibrium or whatever.  But
since I have never heard of such things I have no idea what to say about them.
I guess maybe if you took a dynamic model and imposed, say, REE,
and let time go to infinity, you would get something you
could call a "long run position", or if you looked at some finite
time period, you would get a  "temporary" thing).