Reconsidering the Measurement of Pragmatic Knowledge Using a Reciprocal Written Task Format

Abstract

Current influential models of communicative language ability represent language knowledge as a set of separate yet related components, including grammatical and pragmatic dimensions, which interact with background knowledge, metacognitive strategies, and contextual features (Chapelle, 1998) in language use. Although some researchers have attempted to measure aspects of pragmatic knowledge, the vast majority have not incorporated a clearly articulated pragmatic component into the test construct. The purpose of the current study is to investigate the extent to which scores from a test designed for the current study can be interpreted as indicators of test-takers’ grammatical and pragmatic ability. This study attempts to address some of the limitations of prior pragmatics assessment research, namely, the issues of construct underrepresentation, the lack of highly contextually constrained reciprocal tasks, and the use of less sophisticated statistical tools to support claims of validity. It examines the construct validity of a test based on Purpura’s (2004) theoretical model, which specifies the grammatical and pragmatic components of language knowledge. This model accounts for the way in which grammatical resources are used to convey a range of pragmatic meanings in language use. The analysis explored a broader range of features of pragmatic knowledge than had previously been investigated.