Aspectual differences between agentive and non-agentive uses of causative predicates
Fabienne Martin
July 2019

This paper aims to provide an account for why, across languages, the zero-change (or failed-attempt) use of causative predicates is easier to obtain with agent subjects than with causer subjects, elaborating on the analysis developed in Martin (2015) and solving some of its shortcomings. The paper is structured as follows. Section 2 reports experimental studies suggesting that the degree of acceptance of the zero-change use of causative predicates at study varies across languages, focusing on Mandarin on one hand, and French and English on the other. This difference is accounted for in section 7. Section 3 identifies the source of the zero-change uses for these two sets of languages. Section 5 shows how the Voice head introducing agents vs. causers combines with causative VPs, and how the semantic difference between these two voice heads infuences the interpretationof the VP-event, and, in particular, the way the causing event type denoted by the VP is tokenized (i.e., mapped to event tokens in the model). It is then argued that the semantic difference between the two heads ultimately explains why typically, zero-change uses of causative VPs are acceptable with agents only, starting with standard lexical causative verbs in section 6.1, and then addressing defeasible causative verbs in section 6.2. Section 8 accounts for why the zero-change reading is occasionally accepted by some speakers even with a causer subject
Format: [ pdf ]
Reference: lingbuzz/004667
(please use that when you cite this article)
Published in: submitted to Bar-Asher Siegal, Elitzur and Boneh, Nora (editors), Perspectives on Causation. Springer.
keywords: telicity, perfectivity, non-culminating accomplishments, agentivity, defeasible causatives, sublexical modality, agents, causers, voice, mandarin, french, semantics
previous versions: v2 [July 2019]
v1 [July 2019]
Downloaded:75 times


[ edit this article | back to article list ]