The main contributions of the project to the state of the art stem from the qualitative, theoretical side of the project research, where I was able to bring forth new insights into the principles and mechanisms behind language change. In particular, I showed how broader typological processes may interact with more localized language-contact constraints in order to drive language change in different directions cross-linguistically. CompSubjInf proved to be an especially illuminating research area in this context, because it was shown to be driven both by broader typological forces as well as by local contact-induced constraints.
The general historical trajectory pertaining to CompSubjInf can be described in terms of a gradual retreat of Inf in favor of Subj. This development was observed to varying degrees in both Balkan Slavic and Romance languages. The tendency towards replacement of Inf by Subj was couched within the broader grammatical evolution affecting (at least) the Indo-European linguistic family. The evolution in question can be described in terms of a development of more articulated syntactic structures, which allowed for more efficient language processing thanks to the greater range of specialized functional projections they contained. This broader development favored the use of Subj over Inf, given that the former are more syntactically articulated than the latter. Nevertheless, Inf remained in relatively stable use cross-linguistically in those sentences where the subject of the main clause and the embedded clause is the same (so-called control complements). This was explained by referring to a competing typological principle, namely the tendency towards avoiding redundant linguistic representations. The use of finite Subj morphology in control complements would require that the person and number features associated with the main-clause subject be redundantly realized both on the main and on the embedded verb. Hence the use of Inf (which does not realize any person and number features) is favored in this type of clauses cross-linguistically.
In relation to languages under study, it was shown that the CompSubjInf patterns in Romance were mostly driven by broader typological mechanisms, as these languages largely lost the use of Inf in non-control but kept it in control contexts. Balkan Slavic developments, on the other hand, were partly driven by local language-contact pressures, specifically in those instances where Subj replaced Inf in control complements. These developments were explained as a function of the multilingual socio-cultural setting that characterized the Balkan region through history. In such a context, where speakers would often communicate in non-native languages, the typological constraint against redundancy in language was relaxed, because redundant repetition of certain linguistic information (e.g. the person and number features of the subject in case of Balkan control Subj complements) increased the likelihood that this information would be correctly interpreted by non-native speakers. This is just one illustration of how local pressures may interact with broader typological mechanisms in order to guide linguistic change in different directions.
