The problem with relativism

Relativism is, broadly, the idea that there are no absolutes within a certain domain. That what can exist is all "relative", depending on the frame or context within which it exists. As the cognitive scientist Jerry Fodor notes, the domains of relativism are many, mainly within the "social" sciences. There is cultural relativism in anthropology, which claims that values are determined uniquely by one's culture. There is relativism in sociology, where epistemic commitments are seen to be a product of class affiliation. And again, in my home discipline of psychology, in "empiricist" theories that claim the mind is infinitely malleable depending on life events and conditions.

What I'm going to argue is not that these theories are wrong for any sort of evidence-based reason, but that maintaining that they are "relativistic" is fundamentally incoherent, even when they are right.

The reason is this: In so far as these theories are scientific -- testable by the scientific method -- they must make systematic predictions. What I mean by systematic, in this case, is "given a set of circumstances or preconditions X, we expect a set of outcomes Y". There is no alternative to this system. Any inquiry is forced to group events into some coherent sets -- for example, differences between "men" and "women", "2-month-old babies" and "5-month-old babies", "electrons" and "protons", Culture A and Culture B, etc. It is impossible to have a science of idiosyncrasies. Even if we were somehow to have a science of one person, for example, "person" is already a grouping, a concept combining, say, a 10-year-old's identity as somehow "the same thing" as a 50-year-old's. I won't belabor this, but it can obviously be taken endlessly further to smaller and smaller scales and groupings.

The problem and the point is that a theory that says, "things are relative, depending on the context", has to specify what the things and the context are. And specifying those necessitates some smudging of the details, so that grouping the individuals can be coherent. In that way, "things" can never be totally "relative", because we are already dealing with abstractions and absolutes -- we want theories that, in the end, make universal statements -- "if the context is A, then the outcome will be B, for all instances of A and B". Relativism, no matter how extreme, still has to deal with universal rules whether it wants to or not. It's the only way to do science. At its best, relativism will provide us with exactly the same universals as non-relativism. Those universals will just have a lot if "if... then"-type rules. And the theory will only be coherent or testable as long as those universals make precise predictions without any further "relativistic" leeway. It wouldn't be science if it said, "Well, it turns out the data don't match my predictions, but... it's all relative anyway"

So imagine you have a simple relativistic theory, and a good one. For example, "Which language people speak depends on which language is spoken by the community in which they are raised". This is a good beginning. If we interpret it reasonably and charitably, it's even true. But the problem is, it's still far from being a theory. What this statement leaves out is what constrains the possibilities. Unless the theorist believes that what counts as "language" in this case is completely random -- anything goes, including monkey screeches, melodies, random flailing of the arms that signifies nothing, etc. -- the onus is still on the theorist to specify what the universals of language are. Why it is, that some children learn Japanese and other children learn English, while no children ever learn languages without syntactic structure, for example, or ones where word-boundaries blur (that is, words are not discrete and separable), or any of other countless possibilities that no one has ever observed and probably can't even imagine.

In other words, any relativistic statement, in order to have scientific meaning, must explain not just how things vary, but why they vary only as they do, and not in some other logically possible way.

This is the killer. In psychology, any empiricist who believes humans start out with a totally blank slate (maybe granting a few general learning mechanisms), must specify how it is that those mechanisms cause us to have the exact kind of minds that we have. Why we think of colors as blue and green instead of grue and bleen, and why we very quickly learn to associate "rabbit" with those hopping animals rather than with the particular configuration the furry limbs have when they're all stuck together the way they are. Similar issues are found in anthropology, sociology, and all the others.

In other words, whether they like it or not, relativistic theories must be theories of universals. In that way, they end up being no different than non-relativistic theories. Maybe non-relativists can say they're interested in what the constraints are for the range of possible things, and relativists are interested in which of the possible things turns out given which sets of conditions. But those are not rivaling perspectives, they're complementary halves of the same whole. And even those relativists will still need to group different circumstances and outcomes into describable sets, in order to provide any coherent explanations. These will generally be in a set of universal "if... then" rules, which can, of course increase in complexity as the analysis proceeds.