Abstract

This paper discusses the view that subjectifications are primarily motivated by speakers’ need for self-expression. Approaching the issue from the perspective of animal signalling, we propose that semantic subjectifications are at least equally likely to reflect evaluations and attitudes read into utterances by listeners who attempt to read speakers’ minds. We compare speaker-based and listener-based theories with regard to their predictions, sketch ways in which they can be tested and report findings from first attempts at doing so. First, we report evidence from diachronic corpora. Second, we describe a game-theoretic model that relates listener’s interest in speaker intentions to the average degree of speaker-honesty in a population. Third, we report preliminary results of an experiment in which we tested if listeners were more likely to interpret an utterance as indexing speaker subjectivity when they perceived speakers as more powerful. We conclude that the listener-based hypothesis of subjectification is solid enough to warrant further investigation.