Wednesday, September 18, 2013

Strong Reductionism Failed in the Natural Sciences

This is noted by P. W. Anderson:
“The behavior of large and complex aggregates of elementary particles, it turns out, is not to be understood in terms of a simple extrapolation of the properties of a few particles. Instead, at each level of complexity entirely new properties appear, and the understanding of the new behaviors requires research which I think is as fundamental in its nature as any other.” (Anderson 1972: 393).
Macroscopic bodies in our universe that are subject to the same physical laws nevertheless display complex emergent properties in which the whole becomes more than just the sum of its parts, such as, for example, in superconductive materials, antiferromagnets, ferroelectrics, liquid crystals, DNA (Anderson 1972: 395) and, of course, human consciousness.

The same fallacy of strong reductionism and the equally flawed methodological individualism infect modern neoclassical economics, which seeks to reduce macroeconomic phenomena to the behaviour of isolated individual agents.

But macroeconomics cannot be reduced to microeconomics, and macroeconomics is an autonomous or semi-autonomous disciple in its own right, as John King has recently argued.

Anderson, P. W. 1972. “More Is Different,” Science n.s. 177.4047: 393–396.


  1. I don't think that idea of "emergent properties" is really right. I think that what people call an emergent property is something that is simple for small cases, but unexpected or difficult to calculate in larger cases the way we did in small: if we can approximate the calculation for large cases with something different but simple, we call that an emergent property.

    Take the Newtonian celestial mechanics for example. Calculating orbits with 3 or more bodies is difficult because simulating the interactions requires computationally chaotic calculations. But two body orbits, using the same sorts of computations, mathematically simplify to definite formulas. And single body calculations simplify even more to the idea that a body moves in a line until it is perturbed.

    Collisions between particles can be modeled pairwise, but in vast numbers we call them pressure and use pressure laws to save calculation. But there's no reason except computational difficulty to say there are "emergent" effects: instead the behavior converges on laws of pressure as numbers of particles increase.

    In all these cases, the exact same forces give rise to the simple and the complex results. The only difference is whether there is a shortcut to approximating the answer. The reductionism is only impractical for calculation: it still is correct. Properties do not emerge: they just can be more simply approximated for large scales than by using micro properties the hard way.

    IMHO, that's why microfoundations for macroeconomics is a fool's errand: even if you have them correct, you will not be able to use them because of impractical calculational requirements.

  2. What is reduction?

    Says very much the same thing with an extraordinarily good example at the end.