NOMINALIZED CLAUSES AND DISCOURSE‐GIVENNESS: EXPERIMENTAL EVIDENCE FROM RUSSIAN*

In a seminal paper, Kiparsky & Kiparsky (1970) proposed a two‐way correspondence between presuppositionality of clauses and nominal structure. The proposal remains highly relevant to current research (a.o. Kastner 2015, Bochnak & Hanink 2022), despite the existence of counterexamples in both directions. In this paper, we examine Russian nominalized clauses to show that presuppositionality is indeed neither necessary nor sufficient for nominalization. However, instead of completely discarding the correspondence between presuppositionality and nominalization, we argue for a weaker “preference” hypothesis, whereby presuppositional, or discourse‐given, contexts are associated with a higher likelihood of nominalization compared to discourse‐new contexts. We provide support for the preference hypothesis based on four experimental studies, a forced‐choice and a givenness‐rating study using matrix negation as a proxy for givenness and a forced‐choice and a sentence completion study directly manipulating the discourse context. We suggest a tentative explanation for the preference hypothesis in terms of definiteness/familiarity marking.