On the linguistic encoding of the notion of inference

Abstract

In our everyday life we often talk about events that we have not witnessed but just inferred based on some evidence available to us. This dissertation investigates two types of linguistic expressions that people use when talking about inferred events: inferentials and epistemics. Chapter 2 investigates the meaning of inferential expressions through the case study of the Italian non-temporal future, e.g., Susanna sarà malata ora (‘Susan will be sick now’). I argue - based on introspection - that the semantic contribution of inferentials in a clause – shortly ‘INFER-p’ – is to convey as a not-at-issue proposition that the contextually defined individual who is understood as undergoing a commitment to the truth of ‘p’ (called here ‘the judge of p’) has obtained ‘p’ through an inference. Moreover, I argue that inferentials do not encode an epistemic evaluation of ‘p’: the perception that an inferential weakens the judge’s commitment to the truth of ‘p’ is due to the combination of an extralinguistic considerations pertaining to how people perceive the reliability of information obtained through an inference and the competition with bare ‘p’ which is the natural way to convey knowledge. Chapters 3 and 4 investigate the meaning of epistemic necessity auxiliary verbs through the case study of the English word ‘must’ and the Italian word ‘dovere’, e.g., Susan must be sick now. I argue - based on a combination of data coming from experimental findings, corpus-searched utterances, and introspection - that epistemic ‘must’ and ‘dovere’ are polysemous between two meanings: the meaning of an epistemic necessity operator (defining an event as certain based on some evidence) and the meaning of an inferential (defining an event as having been obtained by the relevant judge through an inference). Furthermore, I speculate that the inferential sense of the ‘must’ and ‘dovere’ is a derivative meaning of the words stemming from their overuse as a marker of epistemic necessity: the original meaning indicating certain conclusions is weakened to indicate just conclusions (i.e., opinions) of the judge, after reiterated and implausible exaggerated uses of the words in their original sense. Chapter 5 offers an assessment of the debate about the relevant perspective for defining the truth-value of bare epistemic possibility statements: the utterer or the assessor. I offer findings from two studies suggesting that in scenarios where the statement is appropriately asserted from the point of view of the speaker but does not correspond to how in reality things are English speakers disagree on the truth-value not only of ‘might p’ but also, surprisingly, of bare ‘p’. I argue that underlying this behavior is not a disagreement on the semantic content of the statements but a disagreement on how to interpret the meaning of the adjectives ‘true’ and ‘false’ when applied to a statement uttered by another agent. Specifically, I suggest that ‘true’ and ‘false’ have an inherent relative nature – to decide if a statement as uttered by A is true one has to first decide a perspective (A or the assessor).