De-Phrase Extractability and Individual/Property Denotation

This paper argues against a hypothesis commonly assumed in Romance linguistics, namely, that the thematic hierarchy Possessor ⧼ Agent ⧼ Theme regulates de-phrase extraction out of the (French) noun phrase. I propose that de-phrase asymmetries (in the domain of extraction and in other domains) correlate with an independently motivated semantic partition - one between individual-denoting phrases (IDPs) that pick out an entity in discourse, and property-denoting phrases (PDPs) that determine a type of entity. I put forward the Nominal Denotation Hypothesis (NDH): at most one de-phrase of each type can be associated with a given noun (N) head. I subsequently demonstrate (a) that a number of problematic exceptions to the thematic hierarchy can be directly accounted for under NDH, (b) that further facets of de-phrase behaviour bearing on matters of general distribution, aspect, linear order, definiteness/specificity, and the potential/lack of potential for exhibiting scopal ambiguity (which were not accounted for in previous works relying on the thematic hierarchy) also originate in NDH, and (c) that NDH is nonetheless compatible with certain fundamental aspects of previous approaches (in both derivational and lexicalist frameworks).