This dissertation is a study of the conditions under which Clitic Left Dislocation, a type of topicalization, can violate island constraints of the strong kind in different Romance languages. Under the classical characterization by Cinque (1990), Topics are described as insensitive to locality constraints of the weak kind (Wh-islands, for instance) but sensitive to islands of the strong kind (Subject Islands, Adjunct Islands and Complex-NP Islands). Nevertheless, the empirical facts show how under the appropriate circumstances, such islands can also be violated by Clitic Left Dislocation: (1) a. A Pedro, que le hayas hablado de ese tema, me molesta muchísimo. To Pedro, that CL-him you have talked about that issue, CL-me bothers very much. Subject Island Violation b. A María, si le cuentas esa historia, Juan se enfadará. To María, if CL-her you tell that story Juan will get angry. Adjunct Island Violation c. A Pedro, el médico que lo vio, le dijo que volviera mañana. To Pedro, the doctor who CL-him saw, CL-him told to come back tomorrow. Complex-NP Violation The dissertation aims to examine under which conditions such violation is possible and to offer an analysis for such cases. In order to do so, in the first place a characterization of Clitic Left Dislocation as a process involving syntactic movement is carried out. The reasons provided for that include (i) the presence of reconstructions effects involving binding relations; (ii) the presence of Case assignment, which is assumed to be local; (iii) the systematic differences in behaviour with respect to Hanging Topics, which can be safely assumed to be generated in situ in the sentential periphery; and (iv) the very existence of locality constraints. The rest of the dissertation is devoted to the examination of each case of island violation. For Subject Islands, the crucial observation is that Clitic Left Dislocation from a clausal subject can take place with unaccusative verbs, which points to the fact that their subjects must have been generated post-verbally. Since it is generally accepted that object positions are more transparent for extraction than specifier positions, it is proposed that Topics escape clausal subjects before they move to a specifier position, thus becoming islands. Therefore, the timing of syntactic operations is crucial for this type of violation. Adjunct and Complex-NP islands are not generally violable, and the only case in which it seems to be possible happens when there is a pronoun in a permitted position which is co-referential with the pronoun in the banned position. A noteworthy exception happens with certain kinds of adverbial subordinate clauses, conditionals most notably. For conditional clauses, the crucial observation is that they allow topicalization from within as long as they have been topicalized in turn. The dissertation closes with a conclusion for each of the chapter and a few proposals to develop the research lines pursued.
- Romance languages
- Left periphery