Face, knowledge and floor: Analyzing like usage through a combined variationist, discourse analytic and pragmatic approach

This thesis examines the distribution and functions of non-standard like in naturally occurring conversations among young Americans. Many of the past quantitative studies on like have focused exclusively on its correlations with age and gender. Meanwhile, within discourse analytic and pragmatics research, there has been considerable disagreement over like's core meaning and functions. As a result, there remain several unanswered questions about like. I propose that combining methods and theories from sociolinguistic variation, pragmatics, and discourse analysis will offer insights into answering the following: how is like distributed within and across interactions, and how does this distribution reflect its core meaning and functions?;The data for this study come from four spontaneous conversations among friends aged 19-29. Each utterance from these interactions was coded for two dependent variables -- presence of like in initial or medial position -- and six independent variables -- gender, turn length, epistemic rights to topic, face-threatening acts, narrative, and Speaker Involvement Index, which measures how much a speaker controls the conversational floor.;A multiple regression analysis of the 5656 utterances indicates that four of the above factors have a significant effect on like usage: epistemic rights to topic, face-threatening acts, turn length, and Speaker Involvement Index. In other words, speakers use like more often when they have exclusive knowledge of a topic, commit potentially face-threatening acts, speak in extended turns, and hold the conversational floor more throughout the entire interaction. Interestingly, gender had a relatively weak effect on like usage.;In light of these results and past work on discourse and epistemic markers, I argue that all of like's functions derive from a core meaning of inexactness and that the patterns of its usage reflect this core meaning.