We explore argument realization in the resultative V-de construction under
the framework of the Theta System. We find that the theta grids of the
resultative V- de construction are of two types, i.e., ([+c-m], [-c]) and ([+cm], [-m]), depending on the (a-)thematic relation between the verb and
second/internal argument. Crucially, the external argument always has a /-
m feature (the minus value of the mental state), regardless of its animacy,
leading to a non-volitional Agent reading (animate) and a Cause reading
(inanimate). It is due to this [+c-m] feature cluster that the construction
yields an unintended result reading, which captures its preference for the
occurrence of dou ‘unexpectedly’ in the construction. In this sense, the
resultative V- de construction can be considered as a non-volitional
counterpart to its de -less counterpart. This contrast signals the
grammatical/morphological coding of volition, with the assistance of de ,
in Mandarin Chinese. As for the internal argument, it can be either [-c] or
[-m], and their syntactic and semantic differences lend credence to the
distinction between “outer objects” and “outermost objects”. The syntactic
distribution of [-c] is sensitive to animacy, with animate [-c] being allowed
to in both [NP1 V- de NP2 XP] and [NP1 ba-NP2 V-de XP], and inanimate
[-c] in [NP1 ba -NP2 V- de XP] only, due to the interactions between
animacy and case. As for [-m], it occurs in [NP1 V- de NP2 XP], in which
NP2 is not thematically related to V and ba -introduction is barred. This
study enriches the Theta System with Chinese data, in special regard to
how animacy, mental state, and volition may affect argument realization.