Search

The present paper aims to uncover the processes governing the rise of canonical case-markings. Experiencer verbs with the DATExp-NOMStim case frame must necessarily first acquire canonical case marking on their second argument in order to enable the acquisition of the nominative by their first argument. The present paper concentrates, thus, on the acquisition of the accusative case by the second argument. Lithuanian verbs of pain are taken under scrutiny. I examine the change that leads to the acquisition of canonical objecthood, namely, the change from the original DATExp -NOM BodyPart case frame to the more canonical DATExp -ACCBodyPart case frame. In the latter, the body-part argument not only acquires the canonical object marking, but also certain syntactic object properties as, for example, the obligatory change into genitive under negation. Strikingly, this change is only found with the verbs of pain skaudėti 'to ache' and dial, sopėti 'to ache' in Lithuanian, while other DATExp-NOMStim experiencer verbs do not undergo this change. I argue that the rise of the canonical object with DATExp -NOM BodyPart verbs of pain in Lithuanian is due to some analogical processes internal to the semantic class of verbs of pain and not to a general drift leading to the acquisition of canonical case assignments. Verbs of pain represent a more complex subclass of experiencer verbs in that they typically take three arguments, namely, experiencer, body-part and stimulus. Those verbs that encode all three participants of the pain event as core arguments typically have the following case frame in Baltic: NOMStim-DATExp -ACCBodyPart. I argue that the loss of the stimulus position by some of these triadic causal verbs of pain let them conflate semantically with the dyadic stative DATExp-NOMBodyPart verbs of pain. This semantic merger results in the redundancy of the morphosyntactic variation between DATExp-NOMBodyPart and DATExp-ACCBodyPart which, in turn, leads to a generalization of one particular case frame: DATExp-ACCBodyPart in the standard language and DATExp-NOMBodyPart in some dialects. [ABSTRACT FROM AUTHOR]

A surprising property of the Ancient Greek bare partitive genitive NP in the subject position is that it behaves morphosyntactically in many ways as a nominative-marked NP: (i) while being in the logical subject position, it triggers semantically driven verbal agreement and, (ii), it may be coordinated with nominatives. This is striking, since obliques typically do not have access to agreement in ancient Indo-European languages including Ancient Greek. To account for this discrepancy, I assume a covert head of the partitive genitive that is filled by pro with an arbitrary reference ( proarb) by default; else, this position can be filled by any quantifier or determiner. It is this proarb that gets Case and triggers verbal agreement. The presence of the proarb in Ancient Greek explains also why the 'independent' partitive genitive is not restricted to structural positions only (as are partitives, e. g., in Finnic languages) but can occur at any position including non-argumental adverbials. [ABSTRACT FROM AUTHOR]