Proceedings of Sinn und Bedeutung
Not a member yet
1139 research outputs found
Sort by
Simultaneous readings of past-under-past in Russian
Full text linkThis paper examines simultaneous readings of past-under-past in Russian. Based on the pattern of cessation inferences, I argue that these readings can only be derived through a de re construal. I demonstrate that the pragmatic competitor for this reading is the double access interpretation of the embedded present tense, which is also derived via a de re construal.Under the grammatical approach to implicature calculation, the structure of the competitor provides insight into the structure of the original statement. I propose a straightforward test for identifying languages in which simultaneous readings can only arise through a de re construal
Diagnosing the presuppositional properties of scalar implicatures
Full text linkExisting accounts of scalar implicatures (SIs) treat them as post-semantic inferences (Grice, 1975) or as part of the asserted meaning (Chierchia et al., 2012), but these views have recently been challenged based on evidence that in certain environments, SIs behave like presuppositions (Bassi et al., 2021). In this paper, we provide additional evidence for the presuppositional approach by showing that SIs are sensitive to constraints on presupposition accommodation (Heim, 2015; Doron and Wehbe, 2022). In addition, we show that even when the implicature trigger is embedded in a downward-entailing environment andthe implicature does not strengthen the overall meaning, we can detect the contribution of a presuppositional implicature. We take this to provide evidence against both the Gricean perspective and the assertive exh perspective
Lying and commitment: The case of pointing gestures
Full text linkIn this paper, we examine pointing gestures that accompany underinformative definite articles and pronouns in German. Ongoing discussions on the definition of lying reveal a contrast between traditional, narrow definitions that focus on what is said and more recent definitions that emphasize commitment. Through two experiments, we demonstrate that first, deceptive pointing gestures are judged as lies to the same degree as speech only. This finding can be interpreted as support for a commitment-based definition of lying, or be explained by assuming that the pointing gestures in our experiment are part of the assertion. Furthermore, our results show that speakers are perceived to be committed to the content of their pointing gestures, although in comparison to speech only, ratings on commitment were lower. However, none of the two commitment-related properties, accountability and restricted deniability, accurately predicts the similarity between pointing gestures and speech for lie ratings. In conclusion, our findings present empirical challenges to both traditional and commitment-based definitions of lying
Do we need two uniqueness conditions? A reply to Lewis (2022)
Full text linkThis paper revisits the long-standing issue of how uniqueness and anaphoricity interact in definite descriptions, particularly the disappearance of the uniqueness effect when a definite description is used anaphorically. Over years of research, two types of uniqueness conditions have been proposed to account for this effect: the individual uniqueness condition and the discourse uniqueness condition. Recently, Lewis (2022) has argued that both are necessary for a comprehensive explanation. However, this paper challenges the necessity of positing two distinct uniqueness conditions. Instead, we argue that grounding the individual uniqueness condition within dynamic semantics (Beaver and Coppock 2015; Köpping 2020) provides a more adequate account. While we disgree with the core premise of Lewis’ dual-uniqueness analysis, we concur with her view that dyanmicness is essential for understanding the interaction between uniqueness and anaphoricity
Postmodern quantification with stuff
Full text linkModern semantic theories of count and mass nouns rest on mereologically structured domains (i.e., domains that have parthood relations). Such domains aim to capture cumulative and divisive reference. However, the mereological domains of modern theories presuppose ontologically troublesome atoms of stuff. Furthermore, they predict parts of “atomic” individuals (e.g., handrails) as well as composite individuals (e.g., cars), and they therefore need to define post-hoc properties of reference (e.g., mereotopological constraints) to get the right atoms back.It seems, therefore, justified to assume that modern theories primarily reflect artifacts of their initially chosen starting point (i.e., mereological domains). This assumption is exacerbated by the fact that cumulative and divisive reference is an immediate consequence of a “naïve” domain that assumes sets of individuals (for plural count) and quantities (for mass). Accordingly, we introduce a more recent theory of pluralities and quantification, namely Referential Transparency Theory (RTT), extend it with a quantitative type stuff, and show how issues of count and mass are discussed in a postmodern way. We argue for a unification of mereotopological and lexical approaches, refine earlier taxonomies by hypothesizing six classes of entities denoted by nominal expressions, and explain complement anaphora with mass noun antecedents
Semantic triviality leads to ungrammaticality through iterated learning
Full text linkA major line of research in semantics concerns meaning-driven explanations of combinatorial restrictions of various operators. Such explanations rely on a link from semantic triviality to ungrammaticality, standardly explicated in terms of LOGICALITY. According to the logicality approach, the grammar contains a natural deductive system that contributes to speakers’ grammaticality judgments, besides the purely syntactic combinatorial system. However, this involves a non-trivial architectural assumption about the interaction between syntax and semantics, and there is no consensus on the exact specification of the natural deductive system. In this paper, we provide an alternative account of the link between semantic triviality and ungrammaticality based on ITERATED LEARNING, an independently motivated model of language evolution. Within this model, it can be shown that for certain trivial sentences, a population of speakers possessing a grammar that in principle generates them is overtaken by learners who induce a grammar that rules them out after several generations. Crucially, our account does not need to postulate an additional natural deductive system
How to bind into alternatives
Full text linkPointwise abstraction has proven elusive to define, though language that requires the operation is commonplace. I show that modeling alternatives as sets of choice functions allows them to be bound into with a simple abstraction rule
Cross-world reference
Full text linkOne way for semanticists to analyze truth in fiction is by relativizing propositions to fictional worlds. However, some sentences in natural language refer to individuals, events, or states from multiple worlds. These include referential cross-world sentences, where the name of an actor from the real world appears to replace the name of the fictional character they play (downward), or vice versa (upward). The conventional analysis claims that in these cases, the name of the actor refers to the fictional individual, or vice versa. I present several empirical asymmetries between upward and downward cross-world reference that make this analysis problematic. Instead, I propose a covert operator FICT, which takes a predicate P, and returns a predicate that roughly means ‘plays a character of whom P is true in a fictional world.’ FICT can be applied to sentential predicates to form downward referential cross-world sentences, or within DPs to form upward referential cross-world sentences. In addition, this accounts for apparent differences between English and languages like Italian and Russian in how gender is handled in referential cross-world sentences, as well as for some previously studied facts about dream reports, which I analyze as a type of referential cross-world sentence. Finally, I explore how this analysis might be expanded to include account for facts about the treatment of fictional times and events
QUD-mediated redundancy
Full text linkWe propose a novel theory of redundancy capable of accounting for informational oddness observed in varieties of Hurford disjunctions. Drawing on Simons (2001), we relativize a sentence’s informational content, which encompasses inquisitive content, to some salient question under discussion, from which we formulate a novel non-redundancy constraint
Lying with at-issue and not-at-issue emojis
Full text linkDrawing on the commitment based definition of lying, we hypothesize that emojis can be used to lie if and only if they cause speaker commitment. We test the three-way relation between emojis, commitment, and lying in two experimental rating studies. In experiment 1, we ask participants whether a writer is lying when they use a pro-text, at-issue emoji, a message-final, not-at-issue emoji, or no emoji in the control. In experiment 2, participants rate the writer’s commitment, operationalized as accountability and restricted deniability, for the same message types. The results show that participants rate messages with final, not-at-issue emojis significantly lower on the lying scale and higher on the deniability scale than messages with pro-text emojis or without emojis. We conclude that writers are seen as similarly committed to their pro-text, at-issue emojis, as to their textual content. In contrast, message-final, not-at-issue emojis trigger less commitment and are less likely to be seen as lies. Our findings indicate that emojis can indeed be used to lie, and that their usage (pro-text vs. message-final) matters for the extent to which emoji content is identified as a lie. These findings are compatible with a view where different emoji uses map onto different types of semantic content, such as at-issue assertion (pro-text emojis) vs. not-at-issue supplemental meaning (message-final emojis). They closely align with a commitment-based definition of lying