Degrees of Rootedness in Acquisition of Language: A Look at Universal Grammar in Homesigners and Late Learners of Libras

The primary goal of this dissertation is to investigate the relationship between Universal Grammar and the properties that Universal Grammar constrains, by investigating how language is created/acquired. The framework proposed in this dissertation provides us with tools for predicting what will and will not appear in linguistic systems of homesigners, late learners of a first language, and native signers/speakers of a given language. New data presented from the spontaneous production and experimental studies of Brazilian homesigners, late learners and native signers of Libras (Brazilian signed language) support the proposal with regards to the strength of rootedness of recursion, merge, hierarchical structural dependency, word order, and topic. If a particular property of language is ‘strongly rooted’, this indicates a high degree of innately specified guidance specifically for language development. Also, there are some properties that are constrained by UG, but with possible options, which are considered ‘somewhat rooted’ in my framework. The studies described in this thesis test hypotheses using elicited production, spontaneous production, and comprehension involving aspects of language, which fall into the categories of ‘strongly rooted’ and ‘somewhat rooted’ properties. The findings provide support for merge, recursion, and hierarchical structural dependency as ‘strongly rooted’ properties ‘Somewhat rooted’ properties, in the form of word order and topic, were also supported by the findings from the experiments with the participants. The proposed framework in this thesis sets the stage for future hypothesis-driven research on language development and language creation