My doctoral dissertation was presented in April 2010, and examined by the following board of professors: Sonia Cyrino, Charlotte Galves, Ilza Ribeiro, Ian Roberts, and Lorenzo Vitral. It was written in Portuguese (some parts of it were published in English, though). It can be downloaded either from the Tycho Brahe Project Publications page, or from the Unicamp Digital Library (here a password must be previously set).
Portuguese Clitic Climbing:
A Study on the European Variety from the Sixteenth until the Twentieth Century
The dissertation addresses the occurrence of clitic climbing in the history of European Portuguese from the sixteenth to the twentieth century, combined with a formal characterization consistent with the observed data. The phenomenon consists of the manifestation of a dependent clitic pronoun on a non-finite predicate together with a governing verb, usually finite, in the context of a complex predicate. I adopt the separation between the concepts of clitic position and clitic placement, the first resulting from syntax, and the second handled by morphology. In order to describe the phenomenon, more than 4,000 tokens from corpora of Modern European Portuguese and Classical Portuguese were separated in terms of two constructions with distinct syntactic and semantic characteristics: “restructuring” and “clause union”. In the former the governing verb is an auxiliary or semi-auxiliary; in the latter, a causative predicate.
I claim that the infinitival complement has a defective character in both types of complex predicates, once it projects only up to vP; therefore, it is transparent to suffer Agree with non-interpretable features in functional categories of the higher domain of the sentence, according to the assumptions of the Minimalist Program. Crucially, clitic climbing is manifested by the presence of a feature associated with a category of the inflectional layer whose semantic effect is the pressuposition of the clitic referent, regardless of the intensional value of the infinitival domain to which the pronoun is linked. This formal characterization is consistent with the results of research on the variation in the occurrence of climbing which receive a unified approach from the connection between clitic climbing and information structure: the climbed clitic incorporates a salient element in discourse and functions as a secondary topic in the sentence.
I postulate that the change in the markedness pattern of the phenomenon is a reflection of how syntax organizes information. Therefore, the non-marked nature of climbing until the seventeenth century is a reflection of the strong use of syntactically marked topics and null subjects, which serve to obtain continuity of discourse topics. The change to a non-marked pattern of the phenomenon is manifested gradually from the sixteenth century due to instability in the use of informationally marked constructions syntactically expressed as the movement to a prominent position in the beginning of the sentence. Such a change in use is deemed responsible for a change in the Primary Linguistic Data, which causes a syntactic change identified as the loss of the V2 parameter around the year 1700.