A Syntactic Side of Word Formation

Asia Pietraszko

University of Connecticut

Friday, February 15, 2019
12:30 p.m.–2 p.m.



The study of word formation has addressed two questions. First, we ask about the nature of word building, i.e. about the mechanism(s) involved in putting two morphemes together. The second question concerns the choice between synthesis versus periphrasis: how is it determined whether two morphemes form a single word or two separate words, in cases when such an alternation is possible? In current Minimalist/Distributed Morphology models, the two aspects of word formation have been unified as two sides of the same coin. Specifically, synthesis is viewed as successful application of a word building operation, while periphrasis arises due to the absence or a failure of such an operation. Despite the conceptual appeal of this unified theory, I will argue in this talk that it is incorrect. Periphrasis cannot be seen simply as the absence of word building. The argumentation is based on a crosslinguistic study of the expression of V and T, which may be expressed synthetically or periphrastically (in English, periphrastic expressions include compound tenses and do-support). I present the following three arguments against the hypothesis that periphrasis is a failure of word building: i) successful word building and periphrasis can cooccur, ii) the failure of word building need not result in periphrasis, and iii) the units created by periphrasis and by word building mechanisms need not overlap. I further argue that periphrasis is a syntactic phenomenon, triggered by featural complexity of clausal syntax. This view derives crosslinguistic generalizations about periphrasis triggers — generalizations missed in the unified theory. The conclusion that periphrasis is independent of word building allows us to reconcile the evidence for its syntactic nature with different approaches to word building. It is compatible with word building mechanisms that apply in syntax (e.g. Head Movement), postsyntax (e.g. Lowering) or in a designated computational module, such as the Lexicon.