The inadequacy of the consonantal root: Modern Hebrew denominal verbs and output–output correspondence

Semitic languages such as Hebrew and Arabic are known for what has become characterised as their discontiguous or non-concatenative morphology. In the overwhelming majority of the literature, in both ‘traditional’ and generative grammar, semantically related words in such languages are described as sharing a common ROOT, usually consisting of three consonants. Such consonantal roots are viewed as actual morphemes with lexical status. Words are formed by affixation to roots; the most common type of such affixation is the interleaving of vowels between the consonants of a root. Within current phonology, the morphological status of roots was originally expressed through a multi-tiered representation, where a root occupied a distinguished tier (e.g. McCarthy 1979, 1981). More recently the notion of root has been challenged by Bat- El (1994a), who argues, based on properties of the process of denominal verb formation (DVF) in Modern Hebrew (MH), that the concept of root can be eliminated.In this paper, I present further arguments that there is no need to refer to roots in the process of DVF in MH. I also show that under such a view a unified, comprehensive treatment of DVF in MH is possible within Optimality Theory (OT; Prince & Smolensky 1993). This analysis goes beyond that originally presented in Bat-El (1994a), in that it has the power to predict the surface pattern of biliteral denominal verbs, whose outputs exhibit variation.