[COCOA] Mitya Privoznov (Göttingen) / Léa Nash (SFL) & Rajesh Bhatt (UMass)

Join us for COCOA, Converging on Causal Ontology Analyses.
Merci de nous rejoindre à COCOA, Converging On Causal Ontology Analyses. 

Mitya Privoznov (Göttingen)
A note on lexical parameters in the meaning of causatives and deontic modals

Consider the following (all the judgments in the abstract and the talk come from elicitations with two native speakers of Canadian English, two native speakers of Karachay-Balkar and my own Russian judgments):

(1)
a. The weather situation let us run to the bookstore.
b. The weather situation made us run to the bookstore.

These two sentences both involve a causative construction and form a minimal pair (the only difference is make vs. let). In this talk, I will try to argue that, even though (1a) and (1b) may give rise to different implicatures, truth-conditionally, (1b) is strictly stronger than (1a). One piece of evidence comes from the following pair of sentences:

(2)
a. The weather situation let, but didn't make us run to the bookstore.
b. #The weather situation made, but didn't let us run to the bookstore.

While (2a) is fine, (2b) is considerably worse (doesn't make any sense). This can be accounted for, if one assumes that X make Y is strictly truth-conditionally stronger than X let Y (if implicatures are neutralized). In the talk, I will try to argue that this cannot be accommodated into the existing analyses of causation without additional stipulations. However, a type of account where let and make quantify over the same set of situations (let being existential and make being universal) offers a straightforward explanation for (2). This is the account that I will sketch at the end of the talk.

Taking a step back, the talk will present some evidence for the following claims. First, the interpretation of clausal embedding predicates similar to English make, force, require, order, give an opportunity, let, allow and permit, Balkar qoj 'allow' and bujur 'order' and Russian zastavljat' 'make', pozvoljat' 'allow', razrešat' 'permit' and prikazyvat' 'order' can vary based on three independent parameters: flavor (deontic vs. causal); force (mainly, existential vs. universal); and the descriptive characteristics of the situation associated with the main clause, aka s1 (agentivity, positive vs. negative situations). Second, the embedding predicates can lexicalize only force (while flavor and the descriptive characteristics of s1 are left unspecified -- e.g. English allow), or force and flavor (while the descriptive characteristics of s1 are left unspecified -- e.g. Russian razrešat''permit'), or force, flavor and the descriptive characteristics of s1 (e.g. English order). Third, there may be clausal embedding predicates that lexicalize flavor, but leave force and the descriptive characteristics of s1 unspecified (arguably, this may be the case of the Karachay-Balkar causative suffix -tIr).

Léa Nash (SFL) and Rajesh Bhatt (UMass)
Two routes to indirect causation

We hypothesize that there is a connection between the form of the linguistic item expressing causation: verb vs. morpheme.Verbs obligatorily introduce an event variable while morphemes don’t have to. Hence, the requirement that indirect causation involve event separation is satisfied even in cases of causative verbs embedding unaccusative structures but not in cases of causative morphemes embedding such structures. To get event separation in cases of morphological causation, the embedded event needs to have an agent. In order to express indirect causation, having two event variables is a necessary condition but there are at least two routes to getting two event variables. One is the FA(I)RE route where the causative verb freely introduces an event variable that is distinct from the event variable of its complement. The other is the morphological route where the causative head introduces a distinct event variable only when its complement has an agent of its own.