The interaction of linguistic, pragmatic, and social factors: The case of datives and ditransitives in Thai

This dissertation explores the relationships among three major types of properties associated with argument structure constructions, i.e., linguistic, pragmatic, and social properties, through a case study of datives and ditransitives in Thai.; Data were obtained from two behavioral experiments and a corpus study. An analysis of the results based on the frameworks of Construction Grammar and the Variationist Approach shows that the dative and ditransitive constructions in Thai are associated with linguistic (distribution and meaning), pragmatic (NP weight and information structure), and social properties (register, age, and sex of speakers). These different types of properties distinguish the two constructions from each other and from other argument structure constructions existing in the language.; Previous studies on argument structure treat linguistic, pragmatic, and social properties as separate from each other. By investigating the relationships among all these properties, I argue that they not only influence the use of a construction independently, but they also interact in two manners. First, they have motivating interactions: the presence of a certain property is motivated by the presence of another property associated with the same construction. These properties may be of the same type (e.g., meaning motivating distribution), or of different types (e.g., register motivating NP weight). Second, different properties simultaneously influence the choice of a construction. Speakers have knowledge of alternating constructions, and their relative differences in entrenchment typically under the influence of linguistic factors. These differently-weighted forms result in different likelihoods of selection, and these likelihoods are affected by pragmatic or social factors that come into play in a given context.; Two aspects of linguistic knowledge beyond those normally accepted as part of language capacity seem to be involved in the speaker's decision to choose the right variant: knowledge of interactions and of probabilistic constraints. Traditional theories place interactions and probabilities outside the realm of linguistic competence. However, I present empirical evidence that calls that distinction into question. By examining variation in the use of a construction in different contexts, I argue that speakers require these two aspects of knowledge as parts of their linguistic competence when making a choice between alternating constructions.