Adverbial constructions such as "delante mío" and their constraints : Probabilistic patterns, reflexes of grammaticalization and analogical chaining in diachronic and diatopic varieties of Spanish

The present thesis combines Labovian variationism, probabilistic grammar, and grammaticalization theory, and sets out to identify i) variable patterns underlying phenomena of morphosyntactic variation, ii) the probabilistic knowledge of grammar that speakers of Spanish possess and regulate, and iii) the conditioning factors, mechanisms and motivations that constrain morphosyntactic variation and change in Spanish diachronically and synchronically. The object of study consists in the morphosyntactic variation and change in Spanish locative adverbial constructions found at a) the level of the complement (e.g. delante de mí vs delante mío), and b) in adverbial possessives’ suffixes (e.g. delante mío vs delante mía). In order to disentangle the trajectory of grammaticalization that the tonic possessives have undergone, four quantitative independent studies are conducted. Study I compiles Twitter-data from the Spanish-speaking world and maps out what factors synchronically condition the variation between the prepositional and the possessive complements. Study II provides a diachronic probabilistic account of the same instance of variation in a European context. Study III extends this diachronic approach to involve Uruguayan Spanish. Study IV is dedicated exclusively to the morphological variation found in the possessive complements. The compilation thesis’ comprehensive summary not only introduces the respective papers, but also serves as a conceptual point of departure for a quantitatively-informed and principally theoretical discussion on the status of the studied phenomena and their relation to evolving probabilistic grammars. In linking the results from the respective studies with the introductory chapters of the thesis, it is argued that Spanish tonic possessive pronouns likely underwent reanalysis in their use with denominal adverbial locatives (e.g. al lado 'by the side', etc.). From there, they have expanded to switch contexts (e.g. cerca 'close') and lastly, in some Latin American Spanish varieties, they have conventionalized entirely and are now used within verb complements (pienso tuyo 'I think of you', habla mal mío 'he/she speaks badly of me') through analogical mechanisms. Concerning the variation found in the suffixes of the possessive pronouns (e.g. cerca mío vs. cerca mía), the dissertation identifies that the mechanism underlying the change in European Spanish appears to be an analogical snowballing effect caused by word-level analogy. In all, the dissertation draws on both quantitative and qualitative examinations and argues that the studied phenomena follow a trajectory that is predictable by theories of diffusional change in general and grammaticalization in particular. Grammaticalization firstly alters meaning through reanalysis and pragmatic inferencing, which gives rise to desemanticization and form changes at later stages. Importantly, these alterations lead to synchronous modifications in speakers’ probabilistic grammars during interaction. These small probabilistic adaptations diffuse in probabilistic community grammars and, in turn, bring about linguistic change. Taken together, the thesis not only contributes to a fuller understanding of the grammatical status of tonic possessive pronouns in Spanish world-wide, but its novel theoretical approach also contributes to advancing theories of linguistic variation and change in general