Canonical complexity

This chapter proposes canonical complexity as a counterpart to enumerative/taxonomic/inventory complexity. While enumerative complexity measures complexity as the number of elements, features, values, etc., in a system, canonical complexity counts the number of departures from what is canonical, and provides a usable measure of non-transparency and inconsistency in that system. The chapter lays out the definitions, terms, domains, and criteria for measuring the canonical complexity of a representative sample of inflectional morphology in nouns, pronouns, and verbs. Applied to a 113-language worldwide sample, it shows that canonical and enumerative complexity are independent of each other and hence can function as distinct typological features; there are large-scale distributional trends of interest especially in the northern hemisphere; and canonical complexity levels appear to correlate well with the sociolinguistics of isolation vs. expansion.