Categories

6. Their epistemology

As theories of ontological categories are some of the things we know (at least if we had some exposure to metaphysical theorizing) we should also be able to come up with an account of how we acquire this kind of knowledge. There does not seem to be too much plausibility in assuming that we can somehow read off the ontological categories directly by looking at the world (even though some accounts, such as that of the Vaiśeṣikasūtra mentioned above at least leave open the possibility that specifically trained kinds of observation are able to perceive ontological categories). We cannot tell by simply looking at the world, for example, whether there are concrete and abstract objects, or only concrete objects.

One notion that is frequently referred to as a source of our knowledge of ontological categories are our intuitions about the results of substitution tests (see Westerhoff 2005: 40–59). For example, if I replace the term ‘book’ by the term ‘table’ in the sentence ‘The book is red’ I might turn it from a true into a false sentence (if the book is red, and the table brown, for example), but if I replace it by ‘the number five’ I will have turned it into a nonsensical sentence. One possible explanation of this difference is that books and tables belong to the same ontological category, while numbers have to be assigned to a different one. In particular, there are other sentential frames (such as ‘… is divisible by two without a remainder’ or ‘… is the direct successor of three’) where only numbers can be sensibly substituted for the blanks. It is therefore natural to argue that our intuitions about what can be substituted where constitute a guide to our knowledge of categories, such that whenever a given group of names, and only this group, can be sensibly plugged into a set of sentence frames then these names pick out members of a specific ontological category.

Of course accepting this does not commit us to claiming that ontological categories are only systematizations of our intuitions about substitution; it may be the case that these substitution patterns are a guide to the fundamental structure of the world, not just to the fundamental structure of the mind. Nor does this criterion have to be considered to be inextricably linguistic. If we can make sense of language-independent states of affairs, we can run a similar substitution argument for substitution in states of affairs (where one constituent is plugged in for another one, but not all of them can be plugged in for all others).

One difficulty for this immediately appealing theory, however, is the problem of reliably differentiating between whether a sentence is nonsense or just false. With sufficient ingenuity even a sentence like ‘colourless green ideas sleep furiously’ might be considered as meaningful (‘these ideas are somewhat insipid (“colourless”) but still alive (“green”) and not wholly conscious (“asleep”) but still ready to come to the front of the mind (“furious”) …’). We would then need some account of why such reinterpretations of nonsensical sentences are not allowed, and it is hard to see where this might come from.

A second, more fundamental criticism of this attempt to shed light on the epistemology of ontology is the question whether postulating an intuition about what can be substituted for what actually explains anything. For we have little idea of where this intuition comes from, and whether it is sufficiently invariant to serve as a foundation of an at least intersubjective, and possibly objective notion like that of an ontological category.