In many languages, the same particles that form quantifier words also serve as connectives, additive and scalar particles, question markers, roots of existential verbs, and so on. Do these have a unified semantics, or do they merely bear a family resemblance? Are they aided by silent operators in their varied roles―if yes, what operators? I dub the particles “quantifier particles” and refer to them generically with capitalized versions of the Japanese morphemes. I argue that both MO and KA can be (...) assigned a stable semantics across their various roles. The specific analysis I offer is motivated by the fact that MO and KA often combine with just one argument; I propose that this is their characteristic behavior. Their role is to impose semantic requirements that are satisfied when the immediately larger context is interpreted as the meet/join of their host’s semantic contribution with something else. They do not perform meet/join themselves. The obligatory vs. optional appearance of the particles depends on whether the meet/join interpretations arise by default in the given constellation. I explicate the proposal using the toolkit of basic Inquisitive Semantics. (shrink)

According to the `grammatical account', scalar implicatures are triggered by a covert exhaustification operator present in logical form. This account covers considerable empirical ground, but there is a peculiar pattern that resists treatment given its usual implementation. The pattern centers on odd assertions like #"Most lions are mammals" and #"Some Italians come from a beautiful country", which seem to trigger implicatures in contexts where the enriched readings conflict with information in the common ground. Magri (2009, 2011) argues that, to account (...) for these cases, the basic grammatical approach has to be supplemented with the stipulations that exhaustification is obligatory and is based on formal computations which are blind to information in the common ground. In this paper, I argue that accounts of oddness should allow for the possibility of felicitous assertions that call for revision of the common ground, including explicit assertions of unusual beliefs such as "Most but not all lions are mammals" and "Some but not all Italians come from Italy". To adequately cover these and similar cases, I propose that Magri's version of the Grammatical account should be refined with the novel hypothesis that exhaustification triggers a bifurcation between presupposed (the negated relevant alternatives) and at-issue (the prejacent) content. The explanation of the full oddness pattern, including cases of felicitous proposals to revise the common ground, follows from the interaction between presupposed and at-issue content with an independently motivated constraint on accommodation. Finally, I argue that treating the exhaustification operator as a presupposition trigger helps solve various independent puzzles faced by extant grammatical accounts, and motivates a substantial revision of standard accounts of the overt exhaustifier "only". (shrink)

Any counterfactual with a true antecedent and a true consequent is invariably predicted to be true by the standard Stalnaker–Lewis semantics. But many such true–true counterfactuals appear false to ordinary speakers, which is considered by many authors as evidence that the standard semantics should be revised. However, Walters and Williams prove that allowing true–true counterfactuals to be false would unacceptably invalidate some very plausible logical principles. The objective of this paper is to provide a pragmatic account of seemingly false true–true (...) counterfactuals, which explains why such counterfactuals seem false to ordinary speakers while being true semantically. In particular, a pluralistic pragmatic account is pursued: I argue that some true–trues seem false because they conversationally implicate some falsehood, and some others seem false because the focus (intonation stress) involved in them invokes some pragmatic re‐interpretation that affects speakers' truth‐value judgments. (shrink)

This paper offers a new unified theory about the meaning of the imperfective and progressive aspects that builds on earlier of analyses in the literature that treat the imperfective as denoting a universal quantifier (e.g. Bonomi, Linguist Philos, 20(5):469–514, 1997; Cipria and Roberts, Nat Lang Semant 8(4):297–347, 2000). It is shown that the problems associated with such an analysis can be overcome if the domain of the universal quantifier is taken to be a partition of a future extending interval into (...) equimeasured cells. Treating the partition-measure (the length of each partition-cell) as a contextually dependent variable allows for a unified treatment of the habitual and event-in-progress readings of the imperfective. It is argued that the contrast between the imperfective and the progressive has to do with whether the quantifier domain is a regular partition of the reference interval or a superinterval of the reference interval. (shrink)

With this paper I aim to demonstrate that a look beyond the Aristotelian square of opposition, and a related non-conservative view on logical determiners, contributes to both the understanding of Aristotelian syllogistics as well as to the study of quantificational structures in natural language.

The distribution of the focus particle even is constrained: if it is adjoined at surface structure to an expression that is entailed by its focus alternatives, as in even once, it must be appropriately embedded to be acceptable. This paper focuses on the context-dependent distribution of such occurrences of even in the scope of non-monotone quantifiers. We show that it is explained on the assumption that even can move at LF Syntax and semantics, 1979). The analysis is subsequently extended to (...) occurrences of negative polarity items in these environments, which mirror the abovementioned distribution of even and which invalidate standard characterizations of NPI licensing conditions in terms of downward-entailingness. The idea behind the extension is that NPIs denote weak elements that are associates of covert even. The paper concludes by discussing two comprehensive theories of NPI licensing and how our proposal relates to them. (shrink)

Sentences with disjunction in the scope of a universal quantifier, Every A is P or Q, tend to give rise to distributive inferences that each of the disjuncts holds of at least one individual in the domain of the quantifier, Some A is P & Some A is Q. These inferences are standardly derived as an entailment of the meaning of the sentence together with the scalar implicature that it is not the case that either disjunct holds of every individual (...) in the domain of the quantifier, \Every A is P &\Every A is Q. As we show, this derivation faces a challenge in that distributive inferences may obtain in the absence of plain negated inferences. We address this challenge by showing that on particular assumptions about alternatives, a derivation of distributive inferences as scalar implicatures can be maintained without in fact necessitating plain negated inferences. These assumptions accord naturally with the grammatical approach to scalar implicatures. We also present experimental data that suggest that plain negated inferences are not only unnecessary for deriving distributive inferences, but might in fact be unavailable. (shrink)

Questions embedded under responsive predicates and definite descriptions both give rise to a variety of phenomena which can be grouped under the term plurality effects: quantificational variability, cumulativity, and homogeneity effects. This similarity has not gone unnoticed, and many proposals have taken inspiration in theories of definite plurals to account for these effects with embedded questions. Recently these phenomena have received less attention, as the field has focused on the so-called intermediate exhaustive reading of embedded questions instead, after Spector called (...) into question the traditional dichotomy between weak and strong exhaustive readings. As a result, the intermediate exhaustive reading has been accounted for at the expense of empirical coverage in other areas. In this paper, I propose a modular theory which derives the currently much discussed exhaustive readings without giving up the rich semantics necessary to account for plurality effects. My account of quantificational variability, cumulativity, and homogeneity effects builds on recent work on these phenomena in the nominal domain by adopting a categorial approach to embedded questions, while the strong and intermediate exhaustive readings are implemented using an independent strengthening mechanism suggested in Klinedinst and Rothschild :1–23, 2011). The resulting theory not only recovers important results on plurality effects; it offers new, simple solutions for some puzzles presented in George :166–177, 2013) and Paillé and Schwarz Proceedings of WCCFL 36, vol 36, Cascadilla Proceedings Project, Somerville, 2018), naturally derives readings that had been postulated in previous literature, makes correct predictions in many unexplored cases, and is compatible with recent results in psycholinguistics. In the last sections I justify my assumptions and show how possible limitations I inherit from the theories I build on can be accommodated under standard assumptions. (shrink)

We present a series of arguments for logical nativism, focusing mainly on the meaning of disjunction in human languages. We propose that all human languages are logical in the sense that the meaning of linguistic expressions corresponding to disjunction (e.g. English or , Chinese huozhe, Japanese ka ) conform to the meaning of the logical operator in classical logic, inclusive- or . It is highly implausible, we argue, that children acquire the (logical) meaning of disjunction by observing how adults use (...) disjunction. Findings from studies of child language acquisition and from cross-linguistic research invite the conclusion that children do not learn to be logical—it comes naturally to them. (shrink)

How do why-interrogatives work? How do they express the questions they express, in the contexts in which they express them? In this essay, I argue that, at a fundamental level, why-interrogatives work just like other wh-interrogatives, particularly other adjunct wh-interrogatives, and they express the questions they express, in the contexts in which they express them, by the same means that other wh-interrogatives do. These conclusions go against a trend in recent work on why-interrogatives, which holds that they are syntactically and (...) semantically unlike other wh-interrogatives. Since the claim that why-interrogatives are unlike other wh-interrogatives has been taken to support various philosophical theses about the nature of why-questions and explanation, showing that why-interrogatives are just like other wh-interrogatives undermines this line of support for these theses. (shrink)

I argue for a subsumption of any version of Grice’s first quantity maxim posited to underlie scalar implicature, by developing the idea of implicature recovery as a kind of explanatory inference, as e.g. in science. I take the applicable model to be contrastive explanation, while following van Fraassen’s analysis of explanation as an answer to a why-question. A scalar implicature is embedded in such an answer, one that meets two probabilistic constraints: the probability of the answer, and ‘favoring’. I argue (...) that besides having application at large, outside of linguistic interpretation, these constraints largely account not only for implicatures based on strength order, logical and otherwise, but also for unordered cases. I thus suggest that Grice’s maxim and its descendants are expressions of general explanatory constraints, as they happen to be manifested in this particular explanatory task. I conclude by briefly discussing how I accordingly view Grice’s system outside of scalar implicature. (shrink)

This paper studies the distribution of ‘list readings’ in questions like who does everyone like? vs. who likes everyone?. More generally, it focuses on the interaction between wh-words and quantified NPs. It is argued that, contrary to widespread belief, the pattern of available readings of constituent questions can be explained as a consequence of Weak Crossover, a well-known property of grammar. In particular, list readings are claimed to be a special case of ‘functional readings’, rather than arising from quantifying into (...) questions. Functional readings are argued to be encoded in the syntax as doubly indexed traces, which straightforwardly leads to a Crossover account of the absence of list readings in who likes everyone?. Empirical and theoretical consequences of this idea for the syntax and semantics of questions are considered. (shrink)

Approaches to anaphora generally seek to explain the potential for a DP to covary with a pronoun in terms of a combination of factors, such as the inherent semantics of the antecedent DP, its scope properties, and its structural position. A case in point is Reinhart’s classic condition on bound anaphora, paraphrasable as A DP can antecede a pronoun pro only if the DP c-commands pro at S-structure, supplemented with some extra machinery to allow indefinites to covary with pronouns beyond (...) their c-command domains. In the present paper, I explore a different take. I propose that anaphora is governed not by DPs and their properties; it is governed by predicates and their properties. To use a metaphor from dynamic semantics: discourse referents can only be ‘activated’ by predicates, never by DPs. This conceptually simple assumption is shown to have far-reaching consequences. For one, it yields a new take on weak crossover, arguably worthy of consideration. Moreover, it leads to a further general “restatement of the anaphora question”, in Reinhart’s words. (shrink)

The presupposition triggered by an expression E is generally satisfied by information that comes before rather than after E in the sentence or discourse. In Heim’s classic theory (1983), this left-right asymmetry is encoded in the lexical semantics of dynamic connectives and operators. But several recent analyses offer a more nuanced approach, in which presupposition satisfaction has two separate components: a general principle (which varies from theory to theory) specifies under what conditions a presupposition triggered by an expression E is (...) satisfied; and an ‘incremental’ component specifies that the principle must be checked on the basis of information that comes before E. Several researchers take this incremental component to be a processing bias, which can be overcome at some cost. If so, it should be possible, though costly, to satisfy presuppositions ‘symmetrically’, i.e. by taking into account linguistic material that comes both before and after the presupposition trigger. We test this claim with experimental means. Using inferential (and to some extent acceptability) tasks involving the anaphoric trigger aussi (‘too’) in French, we argue that symmetric readings are indeed possible (albeit degraded) in environments involving the connectives if, or, and unless. (shrink)

In non‐literal uses of language, the content an utterance communicates differs from its literal truth conditions. Loose talk is one example of non‐literal language use (amongst many others). For example, what a loose utterance of (1) communicates differs from what it literally expresses: (1) Lena arrived at 9 o'clock. Loose talk is interesting (or so I will argue). It has certain distinctive features which raise important questions about the connection between literal and non‐literal language use. This paper aims to (i.) (...) introduce a range of novel data demonstrating certain overlooked features of loose talk, and (ii.) develop a new theory of the phenomenon which accounts for these data. In particular, this theory is motivated by the need to explain minimal pairs such as (2)-(3): (2) Lena arrived at 9 o'clock, but she did not arrive at 9 o'clock exactly. (3) ?? Lena did not arrive at 9 o'clock exactly, but she arrived at 9 o'clock. (2) and (3) agree in their truth conditions. Yet they differ in felicity. As such, they constitute a problem for any account which hopes to predict the acceptability of the loose use of a sentence from its truth conditions and the context of utterance alone. Instead, it will be argued, to explain loose talk phenomena we must posit an additional layer of meaning outstripping truth conditions. This layer of meaning is shown to exhibit a range of properties, all of which point to its being semantically encoded. Thus, if correct, the theory provides a new example of how semantic meaning must extend beyond literal, truth‐conditional content. (shrink)

A construction is found in American Sign Language that we call a Question–Answer Clause. It is made of two parts: the first part looks like an interrogative clause conveying a question, while the second part resembles a declarative clause answering that question. The very same signer has to sign both, the entire construction is interpreted as truth-conditionally equivalent to a declarative sentence, and it can be uttered only under certain discourse conditions. These and other properties of Question–Answer Clauses are discussed, (...) and a detailed syntactic, semantic, and pragmatic account is provided. Question–Answer Clauses are argued to be copular clauses consisting of a silent copula of identity connecting an interrogative clause in the precopular position with a declarative clause in the postcopular position. Pragmatically, they instantiate a topic/comment structure, with the first part expressing a sub-question under discussion and the second part expressing the answer to that sub-question. Broader implications of the analysis are discussed for the Question Under Discussion theory of discourse structuring, for the analysis of pseudoclefts in spoken languages, and for recent proposals about the need for answerhood operators and exhaustivity operators in the grammar and the consequences for the syntax/semantics/pragmatics interface. (shrink)

This paper presents a comprehensive pragmatic theory of contrastive topic and its relation to focus in English. In discussing various constructions involving contrastive topics, it argues that they make reference to complex, hierarchical aspects of discourse structure. In this, it follows and spells out a proposal sketched in Roberts (1996, p. 121ff),using the formal tools found in Büring (1994,1997b). It improves on existing accounts in the accuracy with which it predicts the non-occurrence of the accent patterns associated with focus and (...) contrastive topic, and locates the analysis of contrastive topicswithin a broader picture of discourse and information structure. (shrink)

Previous processing studies have shown that constituents that are prosodically marked as focus lead to an activation of alternatives. We investigate the processing of constituents that are prosodically marked as contrastive topics. In German, contrastive topics are prosodically realized by prenuclear L*+H accents. Our study tests a) whether prenuclear accents (as opposed to nuclear accents) are able to activate contrastive alternatives, b) whether they do this in the same way as constituents prosodically marked as focus with nuclear accents do, which (...) is important for semantic modeling and c) whether the activation of alternatives is caused by pitch accent type (prenuclear L*+H as contrastive accent vs. prenuclear L+H* as non-contrastive accent) or by differences in F0-excursion (related to prominence). We conducted two visual-world eye-tracking studies, in which German listeners heard declarative utterances (e.g., The swimmer wanted to put on flappers) and watched displays that depicted four printed words: one that was a contrastive alternative to the subject noun (e.g., diver), one that was non-contrastively related to it (e.g., sports), the object (e.g., flappers), which had to be clicked, and an unrelated distractor. Experiment 1 presented participants with two naturally produced intonation conditions, a broad focus control condition with a prenuclear L+H* accent on the subject and a contrastive topic condition with a prenuclear L*+H accent. The results showed that participants fixated more on the contrastive alternative when the subject was produced with an L*+H accent, with the same effect size and timing as reported for focus constituents. Experiment 2 resynthesized the stimuli so that peak height and F0-excursion were the same across intonation conditions. The effect was the same, but the time course was slightly later. Our results suggest that prenuclear L*+H immediately leads to the activation of alternatives during online processing, and that the F0-excursion of the accent lends little. The results are discussed with regard to the processing of contrastive focus accents and theories of contrastive topic. (shrink)

We propose an analysis ofonly in terms of event semantics. This approach allows a unified treatment of a wide range of cases in whichonly is associated with focused expressions of different categories. Section 1 is devoted to a preliminary discussion of some problems that a good analysis ofonly should solve. In section 2 we concentrate on sentences in which the focused expression is a NP. In section 3 we show how our analysis can be extended to other categories. Finally, section (...) 4 contains some remarks on related topics, such as scalarity and exhaustiveness. (shrink)

I extend theories of nonmonotonic reasoning to account for reasons allowing free choice. My approach works with a wide variety of approaches to nonmonotonic reasoning and explains the connection between reasons for kinds of action and reasons for actions or subkinds falling under them. I use an Anderson–Kanger reduction of reason statements, identifying key principles in the logic of reasons.

Crosslinguistically, causative constructions conform to the following generalization: If the causal relation is syntactically concealed, then it is semantically direct. Concealed causatives span a wide syntactic spectrum, ranging from resultative complements in English to causative subjects in Miskitu. A unified type-driven theory is proposed which attributes the understood causal relation—and other elements of constructional meaning—to type lifting operations predictably licensed by type mismatch at LF. The proposal has far-reaching theoretical implications not only for the theory of compositionality and causation, but (...) also for the underlying theory of events, space, and time, in natural language discourse. (shrink)

In this paper we argue for a unified semantics for hypothetical conditionals, hc s, e.g. _if it rains, we’ll cancel the picnic_, and biscuit conditionals, bc s, e.g., _if you are hungry, there are biscuits on the sideboard_. We side with recent literature in proposing that differences in the interpretation are related to (in)dependence between antecedent and consequent, but we move beyond current accounts in spelling out a characterization of independence that is actually predictive. We further establish a systematic link (...) between _if_-constructions and discourse structure, providing a dynamic update model that integrates the QUD, and thus the intentional discourse-structure. We argue that in bc s the antecedent sets up the question that is addressed by the consequent, and show that rescuing (Gricean) relevance in face of independence gives rise to implicatures corresponding to the different flavors associated with bc s. Crucially, we argue, this is the same mechanism responsible for our understanding that in the hc above, for example, it is the rain that will cause the cancellation of the picnic. Along the way we notice how the phenomena observed in _if_-constructions are also replicated in other quantificational structures. Ultimately, there is not much that is biscuit-specific. Their interpretation is the result of a conspiracy among semantics, dynamic update and intentional discourse-structure. (shrink)

I outline a discourse-based account of presuppositions that relies on insights from the writings of Peter Strawson, as well as on insights from more recent work by Robert Stalnaker and Barbara Abbott. One of the key elements of my account is the idea that presuppositions are “assertorically inert”, in the sense that they are background propositions, rather than being part of the “at issue” or asserted content. Strawson is often assumed to have defended the view that the falsity of a (...) presupposition leads to catastrophe, in the sense that a false presupposition “wrecks the assertive enterprise”. I argue that the discourse-based account in terms of assertoric inertia can explain how cases of presupposition failure can sometimes be non-catastrophic; there are cases in which the assertive enterprise operates smoothly, despite presupposition failure. The chief problem facing this line of argument is to account for cases in which presupposition failure is catastrophic. If presuppositions are assertorically inert, then how can their falsity ever wreck the assertive enterprise? I offer a principled account that delineates the circumstances in which false presuppositions are, and those in which they are not, catastrophic. (shrink)

We combine two recent probabilistic approaches to natural language understanding, exploring the formal pragmatics of communication on a noisy channel. We first extend a model of rational communication between a speaker and listener, to allow for the possibility that messages are corrupted by noise. In this model, common knowledge of a noisy channel leads to the use and correct understanding of sentence fragments. A further extension of the model, which allows the speaker to intentionally reduce the noise rate on a (...) word, is used to model prosodic emphasis. We show that the model derives several well-known changes in meaning associated with prosodic emphasis. Our results show that nominal amounts of actual noise can be leveraged for communicative purposes. (shrink)

When used next to a predicate at the extreme of a scale such as perfect, the exclusive modifiers just and simply convey a distinctive intensifying effect, presenting a puzzle for theories of exclusivity and alternative-based meanings more broadly. In this article, I develop an analysis of these modifiers as a special kind of alternative-targeting operator, whereby the speaker signals that more specific descriptions than the one they just asserted—modeled here as granularity-based alternatives—are not assertion-worthy in the context—i.e., they need not (...) be asserted in order for a complete representation of the facts to be provided. On this analysis, the intensifying effect of these modifiers are derived from the interaction of exclusivity, granularity, and the distinctive semantic properties of predicates denoting the extreme of a scalar extreme. More broadly, the proposal affords a jumping-off point for further work exploring the interaction between scalarity and exclusive operators, highlighting a number of intriguing lines of future research across semantics and pragmatics. (shrink)

The paper develops a uniform compositional analysis of the various readings of the scalar particle still and its German counterpart noch. Noch/still is a presuppositional scalar particle that gives rise to implicatures. Interpretive possibilities arise through different choices for the scale that the particle associates with, different attachment sites in the syntax, and interaction with focus. These interpretive parameters allow for a wide range of possible sentence interpretations, which overlap, but do not coincide for still and noch. The contrastive perspective (...) allows us to examine the role of scales in the grammar. The implicatures triggered by the scalar item open an interesting perspective for the generation of implicatures in general. (shrink)

The paper provides a semantic analysis of intervention effects in wh-questions. The interpretation component of the grammar derives uninterpretability, hence ungrammaticality, of the intervention data. In the system of compositional interpretation that I suggest, wh-phrases play the same role as focused phrases, introducing alternatives into the computation. Unlike focus, wh-phrases make no ordinary semantic contribution. An intervention effect occurs whenever a focus-sensitive operator other than the question operator tries to evaluate a constituent containing a wh-phrase. It is argued that this (...) approach can capture the universal as well as the crosslinguistically variable aspects of intervention effects, in a way that is superior to previous approaches. Further consequences concern other focus-related constructions: multiple focus data, NPI licensing, and alternative questions. (shrink)

This paper examines the effect that focus has on repetitive versus restitutive again. It is argued that a pragmatic explanation of the effect is the right strategy. The explanation builds largely on a standard focus semantics. To this we add an anaphoric analysis of again’s presupposition and a detailed analysis of the alternatives triggered when focus falls on again.

The semantic literature takes degree operators like the comparative, but also measure phrases, the equative, the superlative and so on, to be quantifiers over degrees. This is well motivated by their semantic contribution, but leads one to expect far more scope interaction than is actually observed. This paper proposes an alternative-semantic analysis of certain degree constructions, in particular constructions with little and other negative antonyms. Restrictions on scope can then be explained as intervention effects.

A semantics for interrogatives is presented which is based on Karttunen's theory, but in a flexible manner incorporates both weak and strong exhaustivity. The paper starts out by considering degree questions, which often require an answer picking out the maximal degree from a certain set. However, in some cases, depending on the semantic properties of the question predicate, reference to the minimal degree is required, or neither specifying the maximum nor the minimum is sufficient. What is needed is an operation (...) which defines the maximally informative answer on the basis of the Karttunen question denotation. Shifting attention to non-degree questions, two notions of answerhood are adopted from work by Heim. The first of these is weakly exhaustive and the second strongly exhaustive. The second notion of answerhood is proven to be equivalent to Groenendijk and Stokhof's interrogative semantics. On the basis of a wide range of empirical data showing that questions often are not interpreted exhaustively, it is argued that a fairly rich system of semantic objects associated with questions is needed to account for the various ways in which questions contribute to the semantics and pragmatics of the utterances in which they appear. (shrink)

We discuss focus sensitivity in English, the phenomenon whereby interpretation of some expressions is affected by placement of intonational focus. We concentrate in particular on the interpretation of always and only, both of which are interpreted as universal quantifiers, and both of which are focus sensitive. Using both naturally occurring and constructed data we explore the interaction of these operators with negative polarity items, with presupposition, with prosodically reduced elements, and with syntactic extraction. On the basis of this data we (...) show that while only lexically encodes a dependency on the placement of focus, always does not. Rather, the focus sensitivity of always results from its dependency on context, and from the fact that focus also reflects what is given in the context. We account for this split using an analysis couched in event semantics. (shrink)

This paper contributes to the debate over the so-called “easy argument for numbers”, an argument that uses evidence from natural language to support the metaphysically significant claim that numbers exist. It presents novel data showing that critical examples in the literature are ambiguous between two readings, contrary to previous assumptions. It then accounts for these data using independently motivated linguistic theory. The account developed rescues the easy argument from the primary challenges leveled against it in the literature and sets the (...) agenda for future work to determine whether or not the argument is valid. (shrink)

The success of deep learning in natural language processing raises intriguing questions about the nature of linguistic meaning and ways in which it can be processed by natural and artificial systems. One such question has to do with subword segmentation algorithms widely employed in language modeling, machine translation, and other tasks since 2016. These algorithms often cut words into semantically opaque pieces, such as ‘period’, ‘on’, ‘t’, and ‘ist’ in ‘period|on|t|ist’. The system then represents the resulting segments in a dense (...) vector space, which is expected to model grammatical relations among them. This representation may in turn be used to map ‘period|on|t|ist’ (English) to ‘par|od|ont|iste’ (French). Thus, instead of being modeled at the lexical level, translation is reformulated more generally as the task of learning the best bilingual mapping between the sequences of subword segments of two languages; and sometimes even between pure character sequences: ‘p|e|r|i|o|d|o|n|t|i|s|t’ → ‘p|a|r|o|d|o|n|t|i|s|t|e’. Such subword segmentations and alignments are at work in highly efficient end-to-end machine translation systems, despite their allegedly opaque nature. The computational value of such processes is unquestionable. But do they have any linguistic or philosophical plausibility? I attempt to cast light on this question by reviewing the relevant details of the subword segmentation algorithms and by relating them to important philosophical and linguistic debates, in the spirit of making artificial intelligence more transparent and explainable. (shrink)

In the so-called reverse proportional reading :53, 1997), the truth conditions of statements of the form many/few \\ appear to make reference to the ratio of the individuals that are in the extensions of both \ and \ to the individuals that are in the extension of \. The analysis of such readings is controversial. One prominent approach assumes they are a symptom of many and few making reference to a context dependent standard of comparison. We observe that this initially (...) attractive approach systematically undergenerates, failing to capture pervasive reverse proportionality in environments that remove context dependency of the standard. Instead, we propose that reverse proportionality in such cases can arise from the underspecification of the measure function underlying the meanings of many and few. (shrink)

In the so-called reverse proportional reading :53, 1997), the truth conditions of statements of the form many/few \\ appear to make reference to the ratio of the individuals that are in the extensions of both \ and \ to the individuals that are in the extension of \. The analysis of such readings is controversial. One prominent approach assumes they are a symptom of many and few making reference to a context dependent standard of comparison. We observe that this initially (...) attractive approach systematically undergenerates, failing to capture pervasive reverse proportionality in environments that remove context dependency of the standard. Instead, we propose that reverse proportionality in such cases can arise from the underspecification of the measure function underlying the meanings of many and few. (shrink)

Fox and Spector use multiple instances of the exhaustivity operator EXH to derive the correct meaning of utterances that include pitch-focus marked disjunction in downward-entailing environments. They argue that the \ operator evaluates alternatives to be used by EXH. Though the method is sound and gets the right result, we argue that the way in which EXH would need to interact with other instances of EXH, as well as other focus-sensitive elements, is at odds with how EXH is used to (...) explain other phenomena. Specifically, the analysis in Fox and Spector predicts intervention effects for cases where EXH interacts with other focus-sensitive elements. This is problematic for many cases in which EXH is used to derive the desired inferences. We propose a different way of focus association for EXH that would work for the approach introduced in Fox and Spector as well as elsewhere. In addition, our account does not require a covert element to be focused. (shrink)

Intonational focus can be observed on parts of words that appear to lack intrinsic meaning, and triggers alternatives that are similar in form. In order to provide a unified treatment of focus above and below the word level (they do, after all, behave the same in most respects), I develop a theory of denotations for arbitrary word parts in which focused word parts denote their own sound and the unfocused parts are functions from sounds to word meanings. This allows focus (...) theories to generalize below the word level; any differences with focus above the word level are located in the semantics of word parts. The paper also explores phonological constraints on focus placement, and shows that the focusability of a word part depends solely on its prosodic status, not on any semantic factors. (shrink)

Differences in the interpretation of would-conditionals with simple (perfective) and perfect antecedent clauses are marked enough to discourage a unified view. However, this paper presents a unified, Lewis–Stalnaker style semantics for the modal in such constructions. Differences in the interpretation of the conditionals are derived from the interaction between the interpretation of different types of aspect and the modal. The paper makes a distinction between perfective and perfect aspect in terms of whether they make reference to or quantify over Lewis-style (...) events. In making reference to Lewis-events, perfective aspect is shown to be incompatible with counterfactual would-conditionals. The so-called ‘epistemic flavor’ of perfective conditionals about the future is derived from the use of diagonalization as an interpretive strategy called upon to resolve reference. (shrink)

In this paper, I offer a novel solution to the well-known problem concerning two polarity items in Korean, amu-(N)-to and amu-(N)-rato, that show a complementary distribution within the set of typical NPI-licensing contexts. I present a uniform analysis of the distribution of these NPIs, where the complementary distribution follows from the opposite scope properties of the emphatic particles to and rato contained in the NPIs in question. As the- oretical background, I adopt Karttunen and Peters’s (1979, Syntax and Semantics 11: (...) Presuppositions (pp. 1–56). New York: Academic Press) and Wilkinson’s (1996, Natural Language Semantics, 4, 193–215) scope analysis of even, Lahiri’s (1998, Natural Language Semantics, 6, 57–127) analysis of Hindi NPIs, and Guerzoni’s (2002, Proceedings of NELS 32 (pp. 153–170); GLSA, University of Massachusetts, Amherst, Mass.; 2004, Natural Language Semantics, 12, 319–343) analysis of the negative bias of yes/no-questions containing minimizers. The current analysis also lends further support to Guerzoni’s approach in that it bears out the prediction that in certain environments, a yes/no-question can be positively biased, i.e., only the positive answer is allowed as a legitimate answer in the context; this prediction was left unconfirmed in Guerzoni’s work. (shrink)

We review Potts' influential book on the semantics of conventional implicature , offering an explication of his technical apparatus and drawing out the proposal's implications, focusing on the class of CIs he calls supplements. While we applaud many facets of this work, we argue that careful considerations of the pragmatics of CIs will be required in order to yield an empirically and explanatorily adequate account.

This paper focuses on the interpretation of the Italian approximative adverb quasi 'almost' by primarily looking at cases in which it modifies temporal connectives, a domain which, to our knowledge, has been largely unexplored thus far. Consideration of this domain supports the need for a scalar account of the semantics of quasi (close in spirit to Hitzeman's semantic analysis of almost, in: Canakis et al. (eds) Papers from the 28th regional meeting of the Chicago Linguistic Society, 1992). When paired with (...) suitable analyses of temporal connectives, such an account can provide a simple explanation of the patterns of implication that are observed when quasi modifies locational (e. g. quando 'when'), directional (e. g. fino 'until' and da 'since'), and event-sequencing temporal connectives (e. g. prima 'before' and dopo ' after'). A challenging empirical phenomenon that is observed is a contrast between the modification of fino and da by quasi, on the one hand, and the modification of prima and dopo by the same adverb, on the other. While quasi fino and quasi da behave symmetrically, a puzzling asymmetry is observed between quasi prima and quasi dopo. To explain the asymmetry, we propose an analysis of prima and dopo on which the former has the meaning of the temporal comparative più presto 'earlier', while the latter is seen as an atomic predicate denoting temporal succession between events (Del Prete, Nat Lang Semantics 16: 157-203, 2008). We show that the same pattern of implication observed for quasi prima is attested when quasi modifies overt comparatives, and propose a pragmatic analysis of this pattern that uniformly applies to both cases, thus providing new evidence for the claim that prima is underlyingly a comparative. A major point of this paper is a discussion of the notion of scale which is relevant for the semantics of quasi', in particular, we show that the notion of Horn (entailment-based) scale is not well-suited for handling modification of temporal connectives, and that a more general notion of scale is required in order to provide a uniform analysis of quasi as a cross-categorial modifier. (shrink)

The exclusive component of unembedded disjunctions is standardly derived as a conversational implicature by assuming that or forms a lexical scale with and. It is well known, however, that this assumption does not suffice to determine the required scalar competitors of disjunctions with more than two atomic disjuncts (McCawley, Everything that linguists have always wanted to know about logic* (But were ashamed to ask). Chicago University Press, Chicago, 1993, p. 324; Simons, “Or”: Issues in the semantics and pragmatics of disjunction. (...) Ph.D. thesis, Cornell University, Ithaca, NY, 1998). To solve this, Sauerland (Linguist Philos 27(3): 367–391, 2004) assumes that or forms a lexical scale with two otherwise unattested silent connectives ( ${\mathbb{L}}$ and ${\mathbb{R}}$ ) that retrieve the left and right terms of a disjunction. A number of recent works have proposed an Alternative Semantics for indefinites and disjunction to account for their interaction with modals and other propositional operators (Kratzer and Shimoyama, In: Otsu Y (ed) The Proceedings of the Third Tokyo Conference on Psycholinguistics. Hituzi Syobo, Tokyo, pp. 1–25, 2002; Aloni, In: Weisgerber M (ed) Proceedings of the Conference “SuB7—Sinn und Bedeutung”. Arbeitspapier Nr. 114. Konstanz, pp. 28–37, 2003; Simons, Nat Lang Semantics 13: 271–316, 2005; Alonso-Ovalle, Disjunction in alternative semantics. Ph.D. thesis, University of Massachusetts, Amherst, MA, 2006). We note that the McCawley–Simons problem does not arise in an Alternative Semantics, if we assume that the set of pragmatic competitors to a disjunction is the closure under intersection of the set of propositions that it denotes. An adaptation of the strengthening mechanism presented in Fox (In: Sauerland U, Stateva P (eds) Presupposition and implicature in compositional semantics. MacMillan, Palgrave, pp. 71–120, 2007) allows for the derivation of the exclusive component of disjunctions with more than two atomic disjuncts without having to rely on the ${\mathbb{L}}$ and ${\mathbb{R}}$ operators. The proposal extends to the case of disjunctions with logically dependent disjuncts. (shrink)

The article proposes an analysis of imperatives and possibility and necessity statements that (i) explains their differences with respect to the licensing of free choice any and (ii) accounts for the related phenomena of free choice disjunction in imperatives, permissions, and statements. Any and or are analyzed as operators introducing sets of alternative propositions. Free choice licensing operators are treated as quantifiers over these sets. In this way their interpretation can be sensitive to the alternatives any and or introduce in (...) their scope. (shrink)

The natural interpretation of counterfactuals with disjunctive antecedents involves selecting from each of the disjuncts the worlds that come closest to the world of evaluation. It has been long noticed that capturing this interpretation poses a problem for a minimal change semantics for counterfactuals, because selecting the closest worlds from each disjunct requires accessing the denotation of the disjuncts from the denotation of the disjunctive antecedent, which the standard boolean analysis of or does not allow (Creary and Hill, Philosophy of (...) Science 43:341–344, 1975; Nute, Journal of Philosophy 72:773–778, 1975; Fine, Mind 84(335):451–458, 1975; Ellis et al. Journal of Philosophical Logic 6:335–357, 1977). This paper argues that the failure to capture the natural interpretation of disjunctive counterfactuals provides no reason to abandon a minimal change semantics. It shows that the natural interpretation of disjunctive counterfactuals is expected once we refine our assumptions about the semantics of or and the logical form of conditionals, and (i) we assume that disjunctions introduce propositional alternatives in the semantic derivation, in line with independently motivated proposals about the semantics of or (Aloni, 2003a; Simons, Natural Language Semantics 13:271–316, 2005; Alonso-Ovalle, Disjunction in Alternative Semantics. PhD thesis, 2006); and (ii) we treat conditionals as correlative constructions, as advocated in von Fintel (1994), Izvorski (Proceedings of NELS 26, 1996), Bhatt and Pancheva (2006), and Schlenker (2004). (shrink)

It has long been observed that the superlative construction, exemplified by John climbed the highest mountain, has two readings. On the absolute reading, the heights of the relevant mountains in a relevant context are compared; on the comparative reading, relevant climbers’ achievements of mountain climbing are compared (Szabolcsi, Comparative superlatives, MIT Working Papers in Linguistics, 1986). Two theories have been proposed regarding this ambiguity. One theory holds that it results from movement of the superlative morpheme -est (movement theory) (Heim, Association (...) with focus, Doctoral Dissertation, 1985, Notes on superlatives, 1999; Szabolcsi 1986).The other theory holds that the ambiguity is derived by assignment of different values to the context variable C, keeping a single LF structure where -est stays in situ (in-situ theory) (Farkas and Kiss, Nat Lang Linguist Theory 18:417–455, 2000; Sharvit and Stateva, Linguist Philos 25:453–504, 2002). As is pointed out by Heim (1999), a choice between these theories is hard to make based solely on English. Through an investigation of Japanese superlative constructions, this paper argues that, in Japanese at least, a movement theory is required. (shrink)

The central idea behind this paper is that presuppositions of soft triggers arise from the way our attention structures the informational content of a sentence. Some aspects of the information conveyed are such that we pay attention to them by default, even in the absence of contextual information. On the other hand, contextual cues or conversational goals can divert attention to types of information that we would not pay attention to by default. Either way, whatever we do not pay attention (...) to, be it by default, or in context, is what ends up presupposed by soft triggers. This paper attempts to predict what information in the sentence is likely to end up being the main point (i.e. what we pay attention to) and what information is independent from this, and therefore likely presupposed. It is proposed that this can be calculated by making reference to event times. The notion of aboutness used to calculate independence is based on that of Demolombe and Fariñas del Cerro (In: Holdobler S (ed) Intellectics and computational logic: papers in honor of Wolfgang Bibel, 2000). (shrink)

Some presuppositions are easier to cancel than others in embedded contexts. This contrast has been used as evidence for distinguishing two fundamentally different kinds of presuppositions, ‘soft’ and ‘hard’. ‘Soft’ presuppositions are usually assumed to arise in a pragmatic way, while ‘hard’ presuppositions are thought to be genuine semantic presuppositions. This paper argues against such a distinction and proposes to derive the difference in cancellation from inherent differences in how presupposition triggers interact with the context: their focus sensitivity, anaphoricity, and (...) question–answer congruence properties. The paper also aims to derive the presuppositions of additive particles such as too, also, again, and of it-clefts. (shrink)