In this paper, we develop a new proposal about how the monotonicity of attitude verbs like believe should be modeled within the content-based approach to clausal embedding (Kratzer 2006, 2013). We pursue the idea that monotonicity is a consequence of how the part-whole structure of attitudinal eventualities relates to the part-whole structure of their contents and thematic participants. In order to cash out this idea, we rely on a non-monotonic, equality-based implementation of the content-based approach (Moulton 2009, Elliott 2020, 2017, Bassi & Bondarenko 2022), supplemented with tools and concepts from mereology. We apply the mereological account of monotonicity to a pattern noticed by Sharvit (2024) concerning Negative Polarity Item (NPI) licensing in nominal arguments to monotonic attitude verbs, which existing implementations of the content-based framework fail to capture. EARLY ACCESS

Mitrović & Sauerland (2014, 2016) claim that, across languages, DP-conjunction decomposes into [JP [MuP [DP1 ⟳] MU] [J’ J [MuP [DP2 ⟳] MU]]]. Their account, when combined with the independently motivated assumption that children better comprehend utterances which display a greater degree of 1-to-1 mapping between form and meaning (van Hout 2008, Guasti et al. 2023: a.o.), predicts that sentences where only J or MU are pronounced should be harder to comprehend relative to sentences where both J and MU are realized. We conducted an experiment testing this prediction by investigating children’s comprehension of conjunctive expressions in Georgian and Hungarian. While, for Hungarian, we did not find any differences between the types of conjunctive expressions, for Georgian, we found evidence that J-MU expressions were harder for children to comprehend than J or MU expressions. Our results challenge the account by Mitrović & Sauerland (2014, 2016) and cannot be captured by other existing accounts of conjunctive expressions either (Szabolcsi 2015, Haslinger et al. 2019). EARLY ACCESS

We develop a states-based semantics for nominal and adjectival confidence reports like Ann is confident/has confidence that it’s raining, and their comparative forms. Our account leverages a Neodavidsonian analysis of adjectival comparatives in which adjectives denote properties of states and measure functions are introduced compositionally. We hereby provide the first systematic semantics for confidence reports, in addition to providing a needed modal extension to the states-based semantics of comparatives. As we show, the flexibility accorded by the Neodavidsonian implementation supports analysis of grammatical constructions with confident/confidence that might otherwise be puzzling, and it lends itself to certain natural ideas about the semantics of cross-categorial probabilistic language using, e.g., likely and probability. In the end, we sketch some immediate connections between confidence-reporting discourse (e.g., I am confident that…) and belief reports about probabilistic discourse (e.g., I think it’s likely that…). EARLY ACCESS

We present evidence from three rating experiments showing that what we call occasion verbs (e.g., thank, criticize, congratulate) allow for the cataphoric resolution of projective content more broadly than trigger types discussed in previous research. Experiments 1 and 2 used methods established by Tonhauser et al. (2018) to show that occasion verbs do indeed pattern with other well-known triggers with respect to projectivity and at-issueness. Furthermore, these experiments provide evidence that occasion verbs — as opposed to a number of other triggers — allow for the cataphoric resolution of projective content in a separate clause in subsequent discourse. Experiment 3 compared the filtering behavior of occasion verbs with that of factive and aspectual triggers for conjunctions in the antecedent of conditionals (if p and q, then r; Mandelkern et al. 2020). The results show that while factive and aspectual verbs show left-to-right filtering asymmetry, occasion verbs display symmetric filtering, that is, right-to-left and left-to-right. Taken together, these results provide evidence that occasion verbs constitute a species not yet observed within the “zoo” of triggers of projective content. BibTeX info

This paper introduces a novel scope paradox. Providing data from Farsi, I show that indefinites in the surface syntactic scope of negated intensional operators yield a reading in which the indefinite appears to take wider scope over the negation, and narrow scope with respect to the intensional operator. Genuine generalized quantifiers, in contrast, do not yield such readings. The uniqueness of indefinites in giving rise to such wide pseudo-scope de dicto readings, which are also found within a simple clause, provides evidence that indefinites differ from generalized quantifiers, not only in their ability to take exceptional scope across clause boundaries, but also in their local scopal properties. I argue that the existence of such wide pseudo-scope de dicto readings not only poses a problem for the generalized quantifier view of indefinites, but also for any approach that takes indefinites to scope via syntactic movement. In-situ accounts of indefinites, on the other hand, can straightforwardly account for the new data, without over-generating wide scope de dicto readings (a.k.a. the "fourth readings") which are widely believed to be impossible (von Fintel & Heim 2011 , Keshet & Schwarz 2019 , Elliott 2023) . I argue that an account in terms of world-Skolemized choice functions is more successful in accounting for the full pattern of the wide pseudo-scope de dicto reading in Farsi, as well as cross-linguistic variation in the availability of such readings.

Current formal approaches to by-phrases in passives analyze the Agent preposition by as semantically vacuous: the denotation of by is merely such that its argument fulfills the same function as the external argument in the corresponding active sentence. This leads to a view of agentive by as essentially homonymous with spatial and temporal by. We argue, on the basis of work in the cognitive linguistic tradition and a new analysis of the French Agent prepositions par and de, that Agent markers do have non-trivial semantic content, and are polysemous rather than homonymous with their spatial counterparts. To formalize this we propose to model these prepositions with general schematic denotations of a polymorphic type ⟨𝜂, ⟨𝜃, 𝑡⟩⟩, which can be instantiated with a concrete type in a specific syntactic and semantic context, such as ⟨𝑒, ⟨𝑒, 𝑡⟩⟩ for the spatial meaning of by. The use as an Agent preposition is simply one of these instantiations, with type ⟨𝑒, ⟨𝑠, 𝑡⟩⟩ (where 𝑠 stands for events). The concrete meaning in context depends on both the general, polymorphically typed denotation and the specific type in the given context. In this way our proposal integrates a useful insight from cognitive linguistics in a semantic formalization of the passive, and opens up possibilities for similar accounts of other highly grammaticalized prepositions.

Intensifiers (e.g. horribly in horribly warm) are usually deadjectival adverbs. I show that the lexical content of the adjectival base, and in particular its evaluative meaning, is directly relevant for the degree intensifying function of these adverbs. In particular, I highlight two generalisations that have remained unaccounted for so far. First, evaluative adjectives with a negative evaluative meaning tend to turn into deadjectival intensifiers expressing high degree, while adjectives with a positive meaning make intensifiers of medium degree. Second, negative modal adjectives can form deadjectival intensifiers, but positive ones cannot. I will argue that a relatively simple intersective semantics for evaluative and modal adverbs accounts for these observations, but that we can only show this if we supplement that semantic analysis with a probabilistic pragmatic component. BibTeX info

Consistent Agglomeration says that, when φ and ψ are consistent, ⌜ought φ ⌝ and ⌜ought ψ⌝ entail ⌜ought (φ ∧ ψ)⌝; I argue this principle is valid for deontic, but not epistemic oughts. I argue no existing theory predicts these data and give a new semantics and pragmatics for ought: ought is an existential quantifier over the best partial answers to some background question; and presupposes that those best partial answers are pairwise consistent. In conjunction with a plausible assumption about the difference between deontic and epistemic orderings, this semantics validates Agglomeration for deontics but not epistemics.

Numerical noun phrases (NNPs) are ambiguous between at least two and exactly two interpretations. This ambiguity has been commonly discussed as a case of Scalar Implicature (SI), where meaning is enriched by exclusion of an alternative. But the SI approach to NNPs has also been widely challenged. We tested NNPs and scalar expressions in inference tasks and found that scalars were sensitive to a manipulation that altered the relevance of alternatives, whereas exactly readings for NNPs were not. Our findings provide a theory-critical challenge to the SI view of NNPs and support alternative views. EARLY ACCESS

While Horn (1969) proposed that [[only]](p) presupposes that the prejacent p is true, von Fintel & Iatridou (2007) showed that the expected prejacent inference is not observed when a necessity modal occurs in the scope of only: [[only]](□p) may convey that p is possible, rather than necessary. What is the mechanism behind the surprisingly weak inference? The approach in von Fintel & Iatridou 2007 is to revise the analysis of only itself to weaken its contribution. In this paper, however, we argue that Horn’s only is correct after all, and introduce a source of weakening separate from only. In particular, in von Fintel & Iatridou’s modal environment, a phonetically null operator (AT LEAST; Crnic̆ 2011, Schwarz 2005) occurs in the scope of only to weaken the presupposed prejacent. Much recent attention has been paid to covert operators which strengthen meaning, in particular a covert EXH with a meaning similar to only (e.g. Chierchia 2006, Fox 2007, Chierchia et al. 2012). A key consequence of our analysis is that natural language incorporates a covert weakening operator, as well. EARLY ACCESS