Root Variation vs. Root Extension – a Putative -i-Infix?

Abstract

Many PIE roots containing have semantically adjacent counterparts without it. A close examination of these pairs suggests that the appearance of -i- is neihter incidental, nor does it arise from phonological constraints, but reveals an early stem-building process. In the following article I discuss the collected material, comprising 77 i-containing roots (plain phonological variants uncounted) opposed to their non-infixed counterparts (including secondary root variants), either of the same meaning, or showing contextual semantic narrowing. I further attempt to integrate the process of the presumably morphological intrusion of -i- into the root in a plausible morpho-semantic context in PIE. Some opaque cases left apart, at least two separate functions of i-infixation come to light. Firstly, it is found in intransitives or frequently clear states/inchoatives as opposed to non-infixed transitives or factitives, and it shows striking affinity both in the semantic operations and the lexical material to Caland networks; in other cases, it could have been linked to the hic-et-nunc -i-, which had left traces at different slots in the morphological system of PIE.