It is well-known that the quantized vs. cumulative nature of a noun in incremental theme position can influence whether or not the event denoted by the verbal predicate is interpreted as bounded or unbounded, i.e. telic or atelic, as illustrated in (1) (Verkuyl 1972, Krifka 1992 a.o.)
(1) a. Sandy drank a glass of wine in 10 minutes/?for 10 minutes.
b. Sandy drank wine #in 10 minutes/for 10 minutes.
In the syntactic literature on inner aspect, a dedicated functional projection has been proposed in the clausal spine to capture this aspectual influence of the incremental theme NP; call it AspP. On these syntactic approaches, the NP either moves to Spec,Asp or Agrees with Asp (Borer 2005, MacDonald 2008, Travis 2010). For a relationship between a functional projection and an NP to obtain syntactically, the NP itself is standardly thought to have some related feature/functional projection which is involved in the Move or Agree operations. Borer (2005) proposes a Quantity Projection within the DP and MacDonald (2008) proposes a Q(uantized) feature.
I argue against this aspect of syntactic approaches to inner aspect, namely, that there is an NP internal dedicated aspectual functional projection/feature. A consideration of Slavic languages, Mandarin, Japanese or Korean, may make this conclusion less striking, since it has been observed that they do not pattern like English in this way (Filip 2005, MacDonald 2008, Travis 2010). This is illustrated below for Korean in (2), where the nature of the incremental theme, quantized (2a) or not (2b) seems to be irrelevant for the (a)telicity of the verbal predicate, given that both an in adverbial and a for adverbial are grammatical on the relevant interpretations.
(2) a. Mary-nun (sam-pun tongan) sakwa hank kae-rul mok-oss-ta
Mary-Top (three-minute for apple one CL-Acc eat-Past-Dec
“Mary ate an apple (for three minutes).”
b. Mary-nun (sam-pun man-e) sakwa-rul mok-oss-ta
Mary-Top (three-minute in) apple-Acc eat-Past-Dec
“Mary ate apple (in 3 minutes).”
I argue that the effects in (1) largely result from the lexical conceptual information of the NPs, that is, from real-world knowledge. Consider English in (3),the bare NPs dinner, lunch, and breakfast.
(3) John ate (dinner/lunch/breakfast) in 10 minutes/for 10 minutes.
Both a telic and an atelic interpretation is available, as indicated by the grammaticality of the in and for adverbials. Dinners, lunches and breakfasts are bounded in time in the real-world and thus, despite them being bare cumulative NPs, a bounded interpretation can result. I provide other examples where conceptual information appears to play a greater role than DP internal grammatical information, from which I conclude that there is no DP internal functional projection or feature dedicated to (a)telicity.