Language and Metaphysics

I ended the last post by endorsing an account of abilities as more precise than the language we articulate them in, and by giving an account of how our language refers to specific abilities. This raises the question of how we should regard our ordinary use of terms when theorizing about the nature of what those terms are about.

There are two extremes here. On the one hand, one may follow Ordinary Language Philosophy and claim that, at least by and large, how we use terms is the best clue as to the nature of those terms. One would therefore claim that if a theory does not neatly fit with how we usually use language, then so much teh worse for the theory. On the other hand, one might take the opposite point of view and think that how we generally talk about things has no necessary connection to how things actually are. The fact that we talk about chairs existing would then have no bearing at all on whether or not chairs should be thought of as actually existing.

The first point I want to make is that both views have to take language as at least capable of articulating how things actually are, however. If one denies this claim completely, then one cannot write anything about metaphysics. I can imagine someone claiming that it holds in a particular domain, but such a claim would deserve argument, and would leave metaphysics generally articulable in language. Indeed, one must hold that metaphysics is articulable at the most general level, since any domain one argued was not articulable would be one where one was making a metaphysical claim about the domain as a whole.

My second point is that we begin theorizing from a linguistic position. When we start discussing the nature of some object, we start with what seems obvious to us, but what seems obvious to us is quite different from what seemed obvious to the ancient Greeks. Reading metaphysics from, say, before Darwin or before Einstein's relativity theories reveals how much of what we take as obvious is actually just scientific and metaphysical views which have percolated down to the general public. Some of what strikes us as obvious is actually early modern philosophy, some is particular interpretations of scientific theories, and some is oddly altered theology. All of it is simply "in the air," spoken about by everyone, often without any felt need to check or cite. This point is a somewhat Heideggerian claim.

Thirdly, how things are constrains our talk about them. Wittgenstein, among others, holds that if we understand how a word is used, that is usually sufficient to understand the meaning. How we use terms is constrained by what the terms are about (or, inversely, how we use terms constrains what someone may take the terms to be about). There are facts about our nature and that of other phenomena in the world which restrict what we can meaningfully say. Linguistic behavior thus reveals something about how we relate to phenomena, and thus provides hints about the phenomena and about ourselves. Linguistic behavior need not always get things exactly right, partly because it is vague about both ourselves and the world in virtue of revealing something about both. Common linguistic behavior is also imprecise: it only needs to be as precise as necessary for how we live.

Thus far, my points have been highly sympathetic to Ordinary Language Philosophy: I have claimed that language is a suitable medium for metaphysics, that metaphysics begins from language, and that language gives clues as to how things are metaphysically. However, my fourth point is that language prioritizes the immediately practical over the theoretical or only long-term practical. Language is therefore most prone to mislead when we use it to provide hints about phenomena which are distant from everyday life, or which only matter in the very long term. Unfortunately, a great deal of metaphysics falls into these categories. It is only when we are dealing with relatively practical metaphysics, such as freedom and ability, rather than less practical metaphysics, such as the disambiguation of natural kinds or the nature of time, that language is a helpful guide.

Fifthly, even in domains which are distant from immediately practical concerns, however, language is a helpful starting point and check on wild theorizing. Our theories ought to make sense of our language, either by showing what we were getting at, or showing the phenomena which led us astray. All language, even that which is metaphysically inaccurate, is motivated by how the world is, and thus a theory which leaves linguistic phenomena without motivation is suspect.