Reduction

Assume that physical facts are causally complete, meaning that every physical fact has a physical cause, and that, in addition, this physical cause completely explains the result. Saying that it explains the result of course implies that nothing else has a causal influence on the result, since otherwise that external influence would also be needed to explain the result. In such a world are all regularities, for example the regularities expressed by economic laws, reducible to physical laws?

It seems obvious to me that in such a world everything must reduce to physics simply because there is no room for other explanations of events. But some disagree. One reason put forward to defend the idea that the special sciences, like economics, cannot all be reduced to physics is that the laws of economics have exceptions. Logically if the economic laws were really a very complicated kind of physical laws then they couldn’t have exceptions, since by hypothesis the physical laws are prefect descriptions of the world, and thus they aren’t really physical laws, or at least aren’t not reducible to them. All this seems to indicate, to me, is simply that the so-called economic laws are not really laws at all, or if they are laws then they are false laws. Perhaps they are better seen as statistical predictions over general kinds of situations, so that if there is an economic “law” that asserts that situation A is followed by situation B then what it really means is that a certain large percentage of the situations that can be described as A result in situations that can be described as B. But so phrased there is no obstacle to reducing such laws to complex physical laws; A could be reduced to a class of physical descriptions, B to another class, and then the physical laws could be used to deduce that a certain percentage of the first class of situations resulted in members of the second class.

A better objection is to deny that the situations that economic laws operate on (such as “high demand for a product”) can be characterized as a certain class of physical descriptions. After all “high demand” only requires that there be a large number of rational agents who want something in comparison to the amount of that thing available to them. But there are a vast number of ways in which a rational agent could be constructed physically. Not only are there many possible organisms that could be the rational agents described in economic laws, but there could be rational agents made of non-organic substances as well, and the laws of economics are supposed to cover all of them. How could this class possibly be reduced to a physical class?

I think that this argument rests on a bit of a misunderstanding about what reduction is supposed to be. Reduction is not supposed to be an identification, but rather an explanation. When we say that a system of laws, X, reduces to a system of more fundamental laws, Y, all we are claiming is that anytime our X laws predict of a specific situation that it would result in another specific situation that our Y laws could have made the same prediction, and in addition that knowing the Y laws allows the complete prediction of the system.

So, in the case of economics, when we claim that it can be reduced to physical laws all we are claiming is that the changes of the prices of goods, say in a situation of high demand, can be in any specific situation predicted, in principle, by knowing all the physical facts and laws. And that is indeed the case, in virtue of the hypothesis that the physical word is completely causally closed. And we can also claim that we can completely predict what will happen in any economy by use of the physical laws alone, which again follows from causal closure of the physical world. So at least under this understanding of reduction the economic laws do seem in principle reducible to the physical laws.

Now in these terms the supposed difficulty for reduction come from our claim that the physical laws can be used to predict certain economic outcomes, like the prices of goods. My imaginary opponent will concede that I can predict all the physical facts, but they will deny that I can thus predict the economic facts, because I don’t have any further laws that tell me what economic facts are implied by a situation described in terms of the physical facts. And furthermore they will deny that such laws exist. The best way to counter this objection is simply by inquiring in virtue of what then a particular economic fact truly describes a situation. There are only two possibilities. One is that economic facts can only be defined in terms of other economic facts. If that is the case then no one knows when any economic facts hold, since we don’t have an economic sense analogous to our visual sense. That is clearly absurd, and thus we must concede that economic facts hold in virtue of some more primitive facts, say facts about rational agents. And then we can ask in virtue of what those facts hold. At every level the facts must hold by virtue of some more primitive facts until we arrive at facts that we do sense directly. But those facts can all be reduced to the physical facts (since what we sense directly are things like the position of bodies, how they change over time, what wavelengths they reflect, and so on). Thus if we are able to know economic facts about a situation they must reduce to the physical facts in some systematic way, otherwise we couldn’t know what the economic facts were, and it was just this connection which we needed in order to complete our reduction of economics to physics.

Of course no one actually knows ho to construct this entire chain of definitions. But that isn’t because it doesn’t exist, it is because we think about economic facts without ever formally defining them. At some level, probably at the level of rational agents and their behavior, we don’t consciously infer these facts from our individual observations. Instead we have learned what a rational agent is by demonstration, by learning in various cases what is and isn’t a rational agent, and then deciding further cases by similarity. This doesn’t mean that a definition of what a rational agent is in physical terms doesn’t exist, it just means that the definition we actually use is buried in our unconscious, and thus not readily available to construct the desired definitions.

So I conclude that if the facts of special sciences are really facts, meaning that it is a matter of objective reality whether they hold of a system or not, then they must reduce to physical facts, since if they didn’t we could have no knowledge of them (and if they aren’t facts then we needn’t concern ourselves with them, since no one has ever claimed that facts which hold only in opinion can be reduced to physics).