Linguistic variation in the discourse of outsourced call centers

Abstract

This study explores linguistic variation in outsourced call centers involving Filipino call-takers and American callers engaged in various types of customer service transactions. The specific goals of this study are: 1) to establish the statistical co-occurrence of linguistic features in outsourced call center discourse; and 2) to examine how the speakers use these features and patterns of speech based on role (as agent or caller), gender, and the types of service transactions (or `accounts'). The data for analysis come from a corpus of call center texts collected in the Philippines (N of texts = 364, approximately 453,630 words). The research design follows a quantitative multi-dimensional framework developed by Biber (1988) for the extraction and interpretation of linguistic co-occurrence in the corpus. Three linguistic dimensions are extracted and interpreted microanalytically: 1) Addressee-Focused, Polite, and Elaborated Information vs Involved and Simplified Narrative; 2) Planned, Procedural Talk; and 3) Managed Information Flow. Results show that the discourse of agents and callers are different in linguistic and textual composition across these extracted dimensions. Similarly, accounts and participants' gender are both found to affect linguistic choices in the transactions.