Séminaire de Recherche en Linguistique

Deux présentations.

1) Armenian reportative t'e: how to assert without assent

Within and across languages, speech reports form a distinct class of constructions with respect to their syntactic and semantic properties (Spronck & Nikitina, 2019). One of this properties is reportativity, that allows the speaker to assert a reported proposition p without fully assenting to its truth, as in "He told me that p, but I don't believe it". Modern Eastern Armenian possess a dedicated complementizer, t'e, whose evidential properties seem to hardwire this "reportative exceptionality" in the language. I show that one of the features of t'e is to introduce verbatim reports that resemble direct speech/quotation in English, without being fully identifiable with it. This has interesting consequences for the study of indexical expressions such as I and you, as well as for our conceptualization of the difference between oratio obliqua and oratio recta.

2) ‘OCL-for-SCL’ in Valdostan Francoprovençal: the Fenisan Variety

Roberts (1993, 2015, 2018) observes in some Valdostan Francoprovençal varieties (Aosta Valley, North-Western Italy, henceforth VF) a phenomenon that he calls OCL-for-SCL: a subject clitic (SCL) can occur in proclisis on the auxiliary only if the object clitic (OCL) appears encliticized to the past participle. When a proclitic object pronoun appears, the SCL cannot occur (compare (1a,b)). The OCL seems to take the place of the SCL. 

(1) a. Gnunc l’ a viu -me

noone 3SG.SCL= has seen =1SG.ACC

‘No one has seen me.’ 

b. Gnunc m’ a viu 

noone 1SG.OCL= has seen 

(Roberts, 2018, VF of Ayas) 

Roberts considers subject pronouns in VF as the morphophonological realization of fission (in the sense of Distributed Morphology, Halle & Marantz 1993) of the φ-features on T. He explains the OCL-for-SCL phenomenon by proposing that the presence of the proclitic OCL blocks fission.
However, by relying on data from a specific variety of Valdostan Francoprovençal spoken in the village of Fénis (Fenisan Francoprovençal, henceforth FFP), I show that this claim is problematic in several respects. In this variety, SCLs occurring with the verb have are similar to those occurring with the other verbs. Moreover, SCLs show the same behavior as the other clitics (OCLs and partitive clitic) regarding the contrast between have and lexical verbs, as they all undergo an obligatory elision of their rime with have. I show that the difference in cliticization between have and the other verbs is linked to the properties of the different types of verbs and cannot be linked to T. In consequence, SCLs in FFP cannot be generated by fission of T’s φ-features. 

Moreover, considering interaction between SCLs and OCLs, FFP undergoes an OCL-for-SCL phenomenon, but this phenomenon is more complex than what Roberts described as the simple impossibility for an OCL and an SCL to cooccur. Relying on the behavior of SCLs with the verb be, I argue that we can account for the whole paradigm by proposing that the rule applying, which elides the clitics rime, has two components, a syntactic one and a phonological one.