(De)coding Modality : The Case of Must, May, Måste and Kan

Abstract: This study investigates the mechanisms of (de)coding modality, focusing on the interpretation of utterances containing the modals must, may, måste, and kan. The main research question posed in this study is what enables the interlocutors to interpret modal expressions so that communicative goals are achieved. To answer this general question, Coates's (1983) study on modal contexts is initially replicated on data extracted from the English-Swedish Parallel Corpus. The data sample used in this investigation is considerably larger than that of Coates, and allows for contrastive observations. Coates's findings were confirmed to a large extent, and additional features were found to be associated with modal interpretations. It was also established that in most examples the features combine. Finally, in a number of indeterminate examples the indeterminacy was shown to be due to inconclusive or conflicting arrangements of the relevant features in these utterances. To explicate the systematic association of the said features with modal interpretations, the notion of Controllability underlying these features was introduced. In this study, Controllability is concerned with the ability of an agent to control the event described in the proposition (of a modal utterance). Controllability is central to the interpretation of modal utterances: utterances in which propositions are encoded so as to indicate lack of agent control are interpreted epistemically, whereas utterances where the relevant features indicate that the intended agent is in control of the situation are deontic. Type of subject, the distinction between states and events, time reference for modality and the proposition, explicit and implicit conditions, and situation type all facilitate the notion of Controllability in English and Swedish. A number of additional features, such as the presence of epistemic adverbials or particles, utterance type, and negation, were also discussed as being of importance for the interpretation of modal utterances. Controllability was linked to Transitivity as per Hopper and Thompson's (1980) article. It was suggested that these notions overlap. The last part of this study reports the results of a Data Mining analysis undertaken to uncover the statistically significant patterns of the relevant features with respect to the interpretation. It also serves to establish how the features related to Controllability combine to give rise to the intended interpretation. The analysis demonstrates that the features constituting Controllability are repeatedly implicated to systematically play a role in the interpretation of utterances containing must, måste, may, and kan.