Proto-Indo-European verb-finality: Reconstruction, typology, validation

Although the reconstruction of Proto-Indo-European as verb-final is widely accepted, there continue to be dissenting opinions (e.g. Friedrich 1975). See e.g. Pires & Thomason (2008), who question the fruitfulness of Indo-European syntactic reconstruction. In this article I address two issues: First, the reconstructable subordination strategies, including relative-correlative structures, are perfectly in conformity with verb-final typology — pace Lehmann (1974) and Friedrich (1975) who considered relative clauses with finite verbs and relative pronouns incompatible with SOV. Second, verb-final reconstruction makes it possible to account for prosodic and segmental changes that single out finite verbs, such as the non-accentuation of Vedic finite verbs and i-apocope preferentially targeting finite verbs in Italic, Celtic, and Baltic-Slavic. Both developments find a natural, prosodically motivated explanation if we accept PIE as SOV, but not if we do not accept that reconstruction. These facts show that, pace Pires & Thomason (2008), the reconstruction of PIE as verb-final is a fruitful hypothesis.