Disturbing factors in a linguistic usage test

Abstract

The particular inquiry commented on here concerns French liaison. The aim of this paper is to show that some disturbing factors can prevent the results of such an inquiry from being interpreted as exact measures of linguistic regularities. Particularly, the answers given to a specific question may be biased by the influence of the surrounding questions; moreover, the testees, when answering a series of questions on the same item, answer in an increasingly systematic manner. Finally, one should never arrange the questions for an inquiry by intuition, because this as well might well bias the results. Eventhough the answers obtained from speakers cannot be considered to exactly reflect their competence, these results still will be more credible than when a linguist relies on his own competence only. If the testees after ten minutes already answer less spontaneously than at the beginning, the linguist who deals with the same problem for a much longer time cannot keep his mind from being influenced by his own reflection.

Notes

Many colleagues, friends, and people who did not even know me have contributed to the data for this research: first of all the 800 testees. My colleagues at Strasbourg university: Astrid Schneider and Michèle Biermann‐Fischer agreed to spend time during their lectures to let the students answer my questionnaires. So did Mary‐Annick Morel at the Université de la Sorbonne nouvelle (Paris). At the Hewlett‐Packard plant of Grenoble, I am indebted towards the personnel manager, who allowed the test to be carried out, and especially towards Jean‐Yves Chatron and Isabelle Licoine, who prepared the practical conditions. Finally, my daughter Anne Laure Hug corrected the main part of the English text.

Address correspondence to: Marc Hug, 19, rue Oberlin, F‐67000 Strasbourg, France.