Necessary and sufficient conditions in lexical causative verbs

Fabienne Martin, University of Stuttgart

As often observed, lexical causative verbs (henceforth LCs) can not only express ‘direct’ causation events (taking place between temporally adjacent causes and effects), but also indirect ones, pace Fodor 1970. For instance, (1a) can mean that Peter directly communicated an idea to Mary—‘d(irect)-reading’—but can also convey that Mary found this idea thanks to something Peter said  two weeks ago — ‘i(ndirect)-reading’, cf. Oehrle 1976. The same way, (1b) does not have to express a direct killing event (Neeleman & van de Koot 2012). It is also felicitous in a situation where, e.g., the gunsmith accidentally put real bullets instead of fake ones in the toygun used some days later by Sue, Ana’s co-player.


  1. Peter gave Mary the idea she needed for her paper.
  2. Eventually, the gunsmith killed Ana!

I argue that the difference between the two readings has a.o. to do with the way sufficient and necessary conditions for the event satisfying the predicate are distributed among its participants, and by the causal dynamics they project in discourse.

Roughly, under both readings, all conditions independent from the subevent involving the subject’s referent (or e’) are assumed to be satisfied in the context. Additionally, under  the direct reading, a LC statement asserts that e’ is both the necessary and sufficient cause for the occurrence of a P-event (i.e. an event satisfying the predicate), see also Lauer 2010:21. On the other hand, a LC statement under the i- reading is interpreted very similarly to two-way implicative verbs as analysed by Baglini & Francez 2016 and Nadathur 2016:  they presuppose some discourse-familiar event   e’  to be causally necessary, but not sufficient, for the occurrence of a P-event in the context of utterance, and assert that this event was sufficient for it (and therefore actually caused a P-event).

I then show how this analysis

(a) sheds a new light on the acceptability of the indirect reading of LCs, and

(b)  captures the similarity between the inferential profile of implicative verbs like manage to, the Tagalog/Salish limited/non-control morphology (cf. also Alonso Valle & Hsieh 2017), and the English LCs under their indirect reading.