On some descriptive generalizations in Romance

In this work we intend to present a number of cross-linguistic descriptive generalizations concerning Romance languages and point out their theoretical relevance for syntactic theory. We will make extensive use of dialectal variation, viewing it as a way to shed light on diachronic processes on the one side and on the complexity of syntactic structure on the other. We will restrict the empirical domain considering in general only some areas of Romance languages and Italian dialects. The linguistic domain that we take into consideration includes three distinct areas of syntactic processes: wh-items and questions in general, personal pronouns and negation. The choice of the grammatical topics is due to both practical and theoretical reasons, as these three domains have been - and still are - central to the development of syntactic theory, and have been systematically explored during fieldwork in the last ten years. The geographical area we have chosen is the one whose micro-variation has been more extensively investigated, both with respect to modern and preceding stages (going back to the 13th century). The aim of this article is not to provide new analyses for a single phenomenon, but to show how cross-linguistic variation can direct our research towards a precise path and narrow down the number of possible analyses of a given phenomenon. As will appear in what follows, descriptive generalizations will be formulated in their strongest form: this does not mean that we are particularly sure that they cannot be falsified if the domain of languages studied is widened. We think that a generalization has an empirical side, which has the function of a challenge: it provokes further, more detailed, observations and possibly more accurate description. Even if a generalization ends up being falsified, we will have increased our empirical basis and, more generally, our knowledge of how languages work. In section 2 we examine the pattern of clitic wh-elements and illustrate some empirical generalizations that are valid both diachronically and cross-linguistically; in section 3 we do the same with respect to the emergence of pronominal clitics . A comparison between the two evolutionary patterns is presented in section 4, where we isolate some properties common to both wh-items and pronouns. In section 5, we present and discuss some empirical generalizations that lead us to analyze wh- in situ and wh-doubling in Romance as closely related phenomena. Section 6 illustrates the factors that influence the cliticization process of a preverbal negation marker, namely modality and verb movement. Section 7 deals with the common properties of cliticization phenomena. Although the factors favoring cliticization are different for the various classes of elements (wh-items, pronouns and negative markers) that undergo the process, it appears that the pattern of cliticization is essentially the same wherever it is manifested. The clitic elements appear in positions where the strong counterpart used to move to; this can either be hypothesized or attested for past stages of a language or directly observed in the present, in closely related dialects or inside the same one as an option with slightly different interpretations