Commitments de lingua and assertoric commitments: the case of expressives

Abstract

This paper presents the results of two series of experimental studies concerning the interpretation of expressives (e.g., ‘the jerk’) and the sentences they occur in. While expressives are known for their strong speaker-orientation, Harris & Potts (2009) found that in the right context, i.e. when a different subject is introduced into the discourse as a reported speaker, it is possible to interpret the expressive from this subject’s perspective. In our first series of experiments we corroborated the systematic availability of non-speaker oriented readings of expressives, but we also found a strong correlation between the attribution of the expressive and that of the sentence content: participants who attribute the expressive to the subject rather than the speaker, also tend to attribute the sentence as a whole to the subject. In other words, shifted interpretations of expressives do occur, but tend to go hand-in-hand with a reportative reading of the sentence in which the expressive occurs. In our second series of experiments, we identified factors that influence such a reportative reading. Following Kaiser (2015), we found that when we made the subject more prominent as an anchor—by removing the reference to the actual speaker and by adjusting the tense to facilitate a free indirect discourse reading—the number of subject-oriented readings grew significantly. On the basis of these findings we argue for a pragmatic account in terms of commitment attribution with three constraints at work: (i) commitments de lingua for expressives need a salient anchor, (ii) commitments de lingua tend to be attributed in concert with assertoric commitments, and (iii) the main speaker is the most salient anchor by default. These three constraints jointly explain the observations in the experiments.