The particle 'em' in Gujarati

Abstract

This paper investigates the discourse particle em in Gujarati, which derives a specialized utterance-final discourse use across declarative and interrogative clauses from its origin as distal demonstrative in the language. I show that em contributes a unified discourse function: it foregrounds the prejacent proposition, marking it salient relative to a set of alternatives. However, its discourse effects vary depending on the clause types and contextual conditions it occurs in. In declaratives, em serves as a specification strategy that selects a particular alternative from a set of alternatives that over-answer the primary QUD. This move triggers and answers a related QUD to the primary QUD, thereby already answering the primary QUD. In interrogatives, em can be used to seek neutral confirmation or convey the speaker’s negative epistemic bias towards the prejacent. em declaratives cannot be felicitously used to answer the primary QUD and em interrogatives are felicitous only when the context provides direct evidence for the prejacent and the speaker is negatively biased towards it. The analysis is framed within a commitment-based model of discourse structure (Farkas & Bruce, 2010; Roberts & Rudin, 2024), which captures how em adjusts the interlocutors’ commitments to the propositions on the Table. I propose that em questions signal the speaker’s assessment of whether the contextual evidence for the prejacent is adequate to warrant committing to it. This is formalized by introducing an adequacy value within the representation of evidence while placing the issue {p,¬p} on the conversational table.