Between epistemic modality and degree: the case of really

This paper identifies and explains the emergence of three different readings of really in the schematic domains of epistemic modality and/or degree. The different readings arise through the interaction between the ontological notion of [REALITY] and the construal of really in relation to the context. The readings are: really as a marker of evidentiality (truth attesting), of subjective emphasis and of degree (reinforcement). This paper questions the view that the readings of really are positionally/ syntactically motivated. Instead, it proposes that the motivating factors are semantic/ pragmatic in nature. The argument is that really is conditioned by the speaker’s wish to qualify an expression epistemically with a judgement of truth as perceived by the speaker. This condition thus acts as a motivating force on the type of conceptual representations that really evokes and takes scope over as well as on the prosodic salience of really itself. It is shown that valence and intonation are the main clues to the interpretation of really on the occasion of use. 1. Background and major claims Research on the interpretation of adverbs such as really, just, only, rather, quite, apparently or absolutely shows that they are contextually sensitive and highly flexible. This paper takes a closer look at really in order to account for its various interpretations. Stenstrom’s (1986) work on really forms the starting-point for the investigation. The following examples are from Stenstrom (1986: 151), where she claims that the different readings of really are due to position and syntactic function: (1) this question is really surprising (2) this is a really surprising question (3) this is really a surprising question (4) this really is a surprising question (5) really this is a surprising question Stenstrom states that when really is placed next to the adjective, as in (1) and (2), it is a degree modifier which serves as an intensifier of surprising. But, the further really is moved to the left, the less is the emphasis on surprising and the more it is on the whole a surprising question. When really is placed in initial position, as in (5), it no longer intensifies a single clause element but is a comment on the whole proposition. Stenstrom remains vague about the intermediate positions and about the more exact interaction between syntactic position and interpretation. She concludes that what finally decides the function of really is the combined effect of position, prosody and the wider context. In contrast to Stenstrom, I propose that the motivating factors for the readings are semantic/pragmatic in nature rather than syntactic/positional. Clearly, position is an important clue to the interpretation of really, but it is not strictly predictive of differences in readings. Position is a linguistic reflex of the semantics and pragmatics of an utterance. I claim that all the above examples of really are epistemic in the sense that they make a comment on the degree of truth of the proposition as perceived by the speaker in the actual situation of use. However, they differ in scope and semantic interaction with their environment. In (1) and (2)