Selection, idioms, and the structure of nominal phrases with and without classifiers

Abstract

It is common to hypothesize that in classifier and non-classifier languages alike the various functional heads (determiner/demonstrative, numeral, classifier) each head their own projection, so that the maximal projection of the nominal phrase is not NP but something like DP. We evaluate the predictions this makes regarding selection and verb-object idioms in English, a nonclassifier language, and in Korean and Vietnamese, two classifier languages. These predictions are not upheld, and nominals act very differently from clauses and PPs, which are headed by functional heads. Selection and idioms show that the maximal projection of the nominal must be a projection of the lexical N itself, not a functional element. We develop such an analysis of nominals, and show how it accounts for data that was taken to motivate the DP Hypothesis.

Idioms and the syntax/semantics interface of descriptive content vs. reference

“…30 Second, there are cases where something similar happens with just the idiomatically interpreted nominal. This is arguably the case of the indefinite nominal in to have a ball, for which Bruening et al (2018) found no determiner variation. Here, a ball, which refers literally to a dance, according to the Oxford English Dictionary, is analogically connected to the associated effect or state that the dance produces, namely fun (cf.…”

“…In a recent paper, Bruening et al (2018) discuss English (as well as Korean and Vietnamese) data from the internet, in which idioms do not occur with the canonical determiners they are commonly taken to occur with. Initial inspection shows considerable variety, and the change in determiner sometimes carries along with it a change in the number of the noun involved.…”

“…First, semantic incorporation is typically morphosyntactically signaled (for example, by special case marking, or by the absence of any determiner on the incorporated nominal); there is no obvious independent evidence that it should be applied in the case of idiomatic expressions. Second, there is the problem of how to combine the D higher up: If determiners denote either functions from properties to entities or, alternatively, to generalized quantifiers, when combined with a semantically incorporated V+NP the result will not, without additional ad hoc assumptions, be the sort of thing a sentence denotes, which is what would be required (see also Bruening et al 2018 for similar criticism).…”

“…The second family of analyses handles the locality of the selectional relation between V and N by rejecting the assumption (widely held since Abney 1987) that D is the head of the nominal syntactic structure and reverting to the classic analysis on which N, rather than D, heads the nominal phrase. Bruening (2015) and Bruening et al (2018) are recent examples. They observe, using examples from English, Korean and Vietnamese, that local selection and facts such as those discussed in the previous section follow directly from the basic nominal structure they assume, given in (12), where "Cl" stands for "Classifier", and "Num" stands for "Numeral".…”

“…In (42) we repeat from (3) Bruening et al's (2018) examples with determiner variability in idioms involving canonical indefinites.…”