Proceedings of Sinn und Bedeutung
Not a member yet
1139 research outputs found
Sort by
A pragmatic solution to anankastic conditionals
Full text linkStandard accounts of modals and conditionals fail to derive the correct meaning of anankastic conditionals like ‘If you want to go to Harlem, you have to take the A train’, where it seems as if the modal in the consequent is restricted by the embedded complement of want (you go to Harlem), rather than by the whole antecedent (you want to go to Harlem). This has led to proposals for a special semantics for want (Condoravdi and Lauer, 2016) or a covert purpose clause associated with teleological (goal-oriented) modality (von Fintel and Iatridou, 2005). In this paper, we show that the apparent non-compositionality of anankastic conditionals is more general, and can be replicated with other modal flavors and attitude verbs: all can trigger what we call “harmonizing readings”. We offer a pragmatic account that generalizes across modal flavors and attitudes. Specifically, we argue that harmonizing arises when the meaning of the antecedent together with background assumptions gives rise to a modal inference that matches in flavor with the consequent modal. Our account predicts when harmonizing is possible and when it isn’t, without relying on any lexical or syntactic idiosyncrasies
Clefts can address wh-questions in referential anchoring contexts
Full text linkCleft constructions are considered infelicitous responses to wh-questions, compared to canonical focus declaratives. This study argues for a more fine-tuned generalization, by showing clefts are indeed a viable response to a question query when the discourse establishes an identifiable witness to the queried property. Using a judgment task, we tested the acceptability of cleft versus declarative answers across three conditions: unanchored (no identifiable witness), anchored (with an identifiable witness), and corrective (revises a misidentified antecedent in prior discourse). Results showed that clefts were rated on par with declarative focus both in an anchored condition and in a corrective condition. As predicted, lower acceptability ratings were found in an unanchored context
\u27Ja\u27 or \u27jaaaaa\u27? Scale interpretation in iconically lengthened response particles
No full textThe traditional view that language is arbitrary (Hockett, 1960) has become increasingly challenged recently (e.g., Blasi et al., 2016): Iconic mappings between form and meaning are found throughout language, as for example in prosodic modulations of length such as looooong (Fuchs et al., 2019). In German, it is possible to modulate the length of response particles in responses to polar questions containing a gradable adjective, as in jaaaaa for the response particle ja (‘yes’). We provide experimental evidence showing that prosodic modulations allow access to semantic scales and manipulate the degree to which a certain property has to hold for a given individual. Further, we analyze lengthening as a degree modifier similar to very, but with different effects on the level of non-at-issue meaning
Aggregate nouns in the plural are uncountable: Evidence from German
Full text linkNouns like rice and sand denote aggregates, i.e., pluralities of objects typically situated in great proximity. Aggregates may be encoded not only as a singular mass noun (rice), but also by nouns in the plural (onions, peas). It has been argued that the difference in grammatical form indicates a difference in meaning: Individuals in the denotation of a plural aggregate-denoting noun are more cognitively salient and disjoinable than individuals in the denotation of the singular mass form. In this paper, I present results of corpus studies of plural aggregate-denoting nouns in German which do not confirm this assumption. My findings rather suggest that there is no difference in meaning between aggregate plurals and aggregate mass nouns since both may occur in the same “mass” contexts, and the interpretation of plural in such contexts differs from the interpretation of the ordinary count plural. I use means of mereotopology to formally capture this difference
Propositional discourse referents and anaphora in dialogue
Full text linkThe article develops the notion of propositional discourse referents that contain dependent discourse referents. It explains the enablement of anaphora on the confirming uptake of assertions and questions by yes, and the blocking of anaphora by no, as well as the role of propositional anaphora like so. The explanation is embedded in a version of dynamic semantics that is based on dynamic context sets with their continuations
Another look at contrafactive predicates: The case of Spanish \u27creerse\u27
No full textVerbs that presuppose the falsehood of their complement are rare in natural language. Anvari et al. (2019) suggest that Spanish contains one such “contrafactive predicate,” creerse, formed from creer (‘believe’) and the reflexive pronoun. We examine the behavior of creerse under negation, where the mood of the complement affects the inferences that get drawn, in complicated ways. We propose an account on which the aspects of meaning that go beyond simple belief trace back to elements beyond the basic belief predicate creer. Our account is linked to a particular view of mood alternation in negative contexts
Question word distributivity
Full text linkThe Shan (Kra-Dai) question word for ‘which’ lǎj (or lǎɰ in some varieties) can also be used to have a distributive interpretation similar to each. This interpretation exists in addition to the indeterminate pronoun meanings, e.g., as a negative polarity item, question word, or universally quantified expression, such as are found in Japanese. This paper describes the distributive construction and indeterminate pronouns in Shan and develops a unified account for lǎj . The same Hamblin semantics that can account for indeterminate pronouns is extended to the distributive case with the incorporation of Neo-Davidsonian event semantics. This identifies a previously unseen connection between indeterminate pronouns and distributivity
Number marking in interrogative phrases: \u27What\u27 interrogatives in Farsi
No full textNumber inflection within interrogative phrases restricts the types of possible answers that these phrases allow for (Dayal, 2016). The paper focuses on Farsi what interrogatives. In Farsi, plural what interrogative phrases require answers naming plural individuals. In contrast, singular what interrogative phrases allow for answers naming singular or plural individuals. This is the case both for bare (what) and for complex interrogatives (what book), but only when these are not marked with differential object morphology: with the differential object marker -ro, only bare singular interrogatives allow for answers naming plural individuals. We derive this pattern from the following assumptions: (i) that interrogatives can range over generalized quantifiers (Spector, 2008; Xiang, 2016; Elliott et al., 2022), (ii) that singular marking is a default in the case of bare interrogatives, but not in the case of complex interrogatives, and (iii) that -ro restricts a set down to one of its singleton subsets. We note that extending this line of analysis to which interrogatives poses questions about their structure, which are left unanswered here
Conjectural questions in Ktunaxa as canonical questions
Full text linkConjectural questions (CQs) are a type of non-canonical question, where the addressee is not expected to know the answer to the issue raised. Such questions are formed in many languages using an inferential evidential. In this paper, we analyze CQs in Ktunaxa (isolate), which contain the modal xma. Although xma is not strictly an evidential, it is compatible with an epistemic reading that requires indirect evidence. We observe that xma in declaratives can be used parenthetically, where its prejacent contains the relevant information for the current issue in the conversation. Our proposal is that CQs in Ktunaxa are derived from the same semantic mechanisms that derive canonical questions, with the twist that CQs are formed on the parenthetical use of xma. We argue against an analysis for Ktunaxa that relies on non-canonical interrogative flip, in contrast with some recent analyses of CQs in other languages
Similative disjunction and strengthening
Full text linkMany natural languages have a simplex form and a complex form of disjunction, and it is commonly assumed that the complex form has stronger exclusive and exhaustive inferences. This study investigates Hindi disjunction, in its simple form and complex form. We report that Hindi disjunction follows the general tendency, but crucially, its simplex form can also give rise to a ”similative inference”. We propose an analysis treating Hindi disjunction as a ”similative disjunction”, and demonstrate how this similative disjunction can be strengthened into an exhaustive disjunction