Induced speech errors as a tool for language description: a case study from Xong “prenasalized consonants”

Abstract

AbstractThe task of language description often requires a linguist to make a choice between alternative hypotheses which cannot be satisfactorily decided by data obtained through traditional descriptive methods. This paper proposes induced speech error tasks as a means of extending the evidential base used to address descriptive questions. Much previous research has explored language-general cognitive processes underlying speech errors, but little work has investigated their potential as a tool for improving the description of language-specific phenomena. In part, this is because the collection of natural speech errors requires more time and a more certain knowledge of the structure of the language than descriptive linguists often have available. Induced speech error tasks, however, help overcome these limitations by allowing the rapid collection of speech error evidence which can aid in selection between competing analyses. Practical advice is offered for researchers considering implementing speech errors tasks in their own research. These points are then illustrated by an investigation of the segmental status of “prenasalized consonants” in Xong, an under-documented Hmongic/Miao language spoken in southern China. While traditional descriptive techniques were unable to resolve the question, a simple speech error-inducing task produced results which can be interpreted as support for a cluster analysis.