Diachronica 25:3 (2008), 386–409. doi 10.1075/dia.25.3.04vit issn 017–4225 / e-issn 15–714 © John Benjamins Publishing Company Te phylogenesis of hypotaxis in Vedic* Carlotta Viti University of Pisa Hypotaxis has been found since the earliest records of Vedic, especially for relative and adverbial functions. However, some adverbial relations more ofen resort to alternative structures such as clause juxtaposition, nominalization, or particles. Te principles underlying the inconsistent representation of hypotaxis remain unclear. My analysis of clause linkage strategies in the Rig-Veda shows that non-hypotactic constructions are used preferentially for relations, such as purpose or concession, that are considered complex in studies of the ontogenesis of hypotaxis in frst language acquisition. Tis suggests that the spread of hypo- taxis follows a cognitive path of increasing complexity through the diachronic stages of Vedic. Moreover, the diferent entrenchment of some adverbial rela- tions, particularly of conditionals, in Vedic with respect to ontogenetic studies allows us to refne the concept of cognitive complexity in the adverbial domain, and to consider it as a contrast to the speaker’s expectations rather than to the extra-linguistic world, as it is usually seen. Keywords: hypotaxis, adverbial, ontogenesis, phylogenesis, Vedic 1. Introduction Te present paper aims to contribute to the issue of the relation between the on- togenesis and the phylogenesis of syntactic patterns, that is, of the possible cor- respondences between the way a certain linguistic category is acquired by a single individual and the way it develops over many generations. Contemporary studies of child language acquisition, supported by synchronic fndings on cognition, can be related to the philological evidence of languages that have a long diachronic record. Although order of acquisition does not always overlap with patterns of * I would like to thank the editor, Joe Salmons, and the anonymous reviewers of Diachronica for their detailed and valuable comments. I am also grateful to Romano Lazzeroni, Alberto Nocen- tini, Anna Orlandini, Saverio Sani, Rosemarie Lühr, and Sabine Ziegler for help and support.