We argue that the superlative modifiers at least and at most quantify over a scale of answers to the current question under discussion (and in this sense, resolve issues), and that they draw attention to the individual possibilities along the scale (and in this sense, raise issues for discussion). The point of departure is a simple analysis on which at least denotes what only presupposes, and at most denotes what only contributes as its ordinary atissue content. This analysis captures the truth conditions, focus-sensitivity, and distribution of superlative modifiers but leaves some pragmatic facts unexplained. We enrich the simple account with unrestricted inquisitive semantics in order to explain the fact that superlative modifiers give rise to ignorance implicatures while comparative modifiers like more and less do not, the fact that superlative modifiers do not give rise to scalar implicatures, and two puzzles concerning the interaction between superlative modifiers and deontic modals. We argue that this proposal provides the most empirically successful published account of superlative modifiers to date.

According to standard linguistic theory, the meaning of an utterance is the product of conventional semantic meaning and general pragmatic rules on language use. We investigate how such a division of labor between semantics and pragmatics could evolve under general processes of selection and learning. We present a game‐theoretic model of the competition between types of language users, each endowed with certain lexical representations and a particular pragmatic disposition to act on them. Our model traces two evolutionary forces and their interaction: (i) pressure toward communicative efficiency and (ii) transmission perturbations during the acquisition of linguistic knowledge. We illustrate the model based on a case study on scalar implicatures, which suggests that the relationship between underspecified semantics and pragmatic inference is one of coevolution.

This paper presents two experimental findings pertaining to the semantics and pragmatics of superlative modifiers ("at least", "at most"). First, in a scenario with N objects of a given type, speakers consistently judge it true that there are ‘at least N’ and ‘at most N’ objects of that type. This supports the debated position that the ignorance conveyed by superlative modifiers is an implicature, not an entailment, and contrasts with results obtained using an inference-judgment paradigm, suggesting that truth-value judgment tasks are impervious to certain pragmatic infelicities that inference-judgment tasks are sensitive to. The second finding is not predicted by any previous theory: In a scenario with N objects, it is not consistently judged true that there are ‘at most N + 1’ objects, even though it is consistently judged true that there are ‘at least N − 1’ objects. To explain this, we propose a novel pragmatic principle requiring that the scenario depicted by a sentence must be considered possible by the speaker (the Maxim of Depictive Sincerity). Put together, the two findings show that truth-value judgment tasks are impervious to some aspects of pragmatics, but not all.

No abstract

Communication involves a great deal of uncertainty. Prima facie, it is therefore surprising that biological communication systems—from cellular to human—exhibit a high degree of ambiguity and often leave its resolution to contextual cues. This puzzle deepens once we consider that contextual information may diverge between individuals. In the following we lay out a model of ambiguous communication in iterated interactions between subjectively rational agents lacking a common contextual prior. We argue ambiguity’s justification to lie in endowing interlocutors with means to flexibly adapt language use to each other and the context of their interaction to serve their communicative preferences. Linguistic alignment is shown to play an important role in this process; it foments convergence of contextual expectations and thereby leads to agreeing use and interpretation of ambiguous messages. We conclude that ambiguity is ecologically rational when (i) interlocutors’ (beliefs about) contextual expectations are generally in line or (ii) they interact multiple times in an informative context, enabling for the alignment of their expectations. In light of these results meaning multiplicity can be understood as an opportunistic outcome enabled and shaped by linguistic adaptation and contextual information. 1Meaning Multiplicity in Communication2Ambiguous Signalling through Pragmatic Inference 2.1Preliminaries2.2Signalling behaviour2.3Communicative success2.4Iterated interactions3Predictions for Single and Iterated Interactions 3.1Simulations3.2Exploration and past experience3.3Preemptive adaptation4Discussion5ConclusionAppendix