‘More complicated and hence, rarer’

Emonds (1980) conjectured that in one limited domain there is a robust correlation between the complexity of a syntactic derivation and the rarity of the sentence type produced by that derivation. Other scholars, however, have hypothesized a more general correlation complexity and rarity which is not well motivated. In the three major historical periods of transformational syntax, we find typologically rare sentence types derived by means of simple operations and typologically common sentence types with complex derivations. By investigating syntactic complexity, I speculate that, in general, we cannot expect to see correlations between complexity and rarity, since implicational and frequency-based typological generalizations do not belong to the realm of I-language.