On the right of being a comparative concept

We provide a critical review of the distinction between "comparative concepts" and "descriptive categories", showing that in current typological practice the former are usually dependent on the latter and are often vague, being organized around prototypes rather than having sharp boundaries. We also propose a classification of comparative concepts, arguing that their definitions can be based on similarities between languages or on differences between languages or can also be "blind" to language-particular facts. We conclude that, first, comparative concepts and descriptive categories are ontologically not as distinct as some typologists would like to have it, and, second, that attempts at a "non-aprioristic" approach to linguistic description and language typology are more of an illusion than reality or even a desideratum.