|Institution:||University of Manchester|
|Keywords:||syntax; linguistics; semantics; argument structure; Welsh; impersonal; passive|
|Full text PDF:||http://www.manchester.ac.uk/escholar/uk-ac-man-scw:295940|
In this thesis I will explore the impersonal constructions in Modern Welsh. In doing so, I will follow the approach of the previous literature in comparing this construction with the analytic Welsh passive. The general linguistic literature on passivization assumes that both constructions involve passivization and despite some studies of Welsh concluding that the impersonal construction is not a passive, this thesis cannot support or deny this claim.I show that it is the definition of passive that obstructs a conclusive analysis for the Welsh impersonal morphology's syntactic and semantic effects, ultimately. Using the data described in detail throughout the thesis, I conclude with an assessment of the scope of our current theories of passive - be they typological or theoretical - that reveals problematic areas. Typological, prototypical and canonical approaches to the passive of course fail to include enough nuance to identify the relevant structural components of the Welsh impersonal, whilst theoretical approaches cannot account for the restrictions found on intransitive impersonals. LFG's mapping theory has the potential to accommodate the Welsh data according to current proposals and as such is examined in more depth.I have given an emphasis to using naturally occurring data whenever possible and this has led to a data-rich, descriptive work, in an attempt to expand the breadth of examples of Welsh found in discussions of linguistic theory. Additionally, this approach provide the basis for future work on Welsh verb classes by describing the behaviour of verbs in several constructions relevant to the work at hand.Along with a general synthesis of the previous literature on impersonals and analytic passives in chapter 2, I include an elaborated analysis of Welsh analytic passives and some problematic new impersonal data. The novel data on verb classes begins in chapter 3 with a study of verbs of psychological state. The lack of restriction on the impersonal contrasts with the analytic GET-passive's failure in a subset of these transitive verbs previously unobserved. In chapter 4, I investigate the availability of unaccusativity diagnostics in Welsh, in an attempt to prove that unaccusative verbs do indeed impersonalize, as suggested by previous literature. Chapter 5 then uses the diagnostic, amongst others to track down further restrictions on the impersonal and I show that whilst unaccusativity cannot be excluded as a potential restriction to impersonal morphology, the semantic restrictions are as plausible an account. This chapter uses different structural and semantic verbs and predicates and finds that animacy in intransitives is the only restriction needed to describe impersonals. In the following chapter (6) I examine this restriction in more detail and summarize the data on the impersonal in order to then probe the current linguistic theories in chapter 7. Advisors/Committee Members: PAYNE, JOHN JR, WALKDEN, GEORGE G.