Paolo Santorio

On a traditional view, the semantics of natural language makes essential use of a context parameter, i.e. a set of coordinates that representss the situation of speech. In classical frameworks, this parameter plays two roles: it contributes to determining the content of utterances and it is used to define logical consequence. This paper argues that recent empirical proposals about context shift in natural language, which are supported by an increasing body of cross-linguistic data, are incompatible with this traditional view. The moral is that context has no place in semantic theory proper. We should revert back to so-called multiple-indexing frameworks that were developed by Montague and others, and relegate context to the postsemantic stage of a theory of meaning.

Classical semantics for counterfactuals is based on a notion of minimal change: If ${\textsf {A}}$, would ${\textsf {C}}$ says that the worlds that make ${\textsf {A}}$ true and that are otherwise minimally different from the actual world are ${\textsf {C}}$-worlds. This semantics suffers from a well-known difficulty with disjunctive antecedents (see e.g. Alonso-Ovalle, 2009; Willer, 2018; Santorio, 2018, a.o.). In a recent study, Ciardelli, Zhang, and Champollion (Ciardelli et al., 2018b; henceforth, CZC) present new, related difficulties for the classical approach having to do with unpredicted differences between counterfactuals with De Morgan-equivalent antecedents, and related pattern of inferences. They propose a new semantics for counterfactuals, which builds on inquisitive semantics (see Ciardelli et al., 2018a) and gives up on minimal change. Building on this debate, we report on a series of experiments that investigate the role of overt negation in this data. Our results replicate CZC’s main effects, but they also indicate that those effects are linked to the presence of overt negation. We propose a novel account, based on three key assumptions: (i) the semantics for counterfactuals does involve a notion of minimal change, after all; (ii) the meanings of disjunction and negation are associated with alternatives, which interact with the meaning of counterfactuals; (iii) the alternatives generated by negation are partially determined by the question under discussion (QUD). We compare our account with other existing accounts, including CZC’s own proposal, as well as Schulz (2019) and Bar-Lev & Fox (2020) ones.

Some modal expressions in language—for example, “can” and “able”—describe what is possible in light of someone's abilities. Ability modals are obviously related to other modalities in language, such as epistemic or deontic modality, but also give rise to anomalies that make them unique. This paper develops a general theory of ability modals that is broadly compatible with standard modal semantics, while predicting their peculiar behavior. The central idea is that ability modals include reference to a notion of dependence. At the level of truth conditions, “ is able to ” simply says that there is a circumstantially accessible world where. On top of this, a not‐at‐issue element requires that whether As depends, in part, on intrinsic features of. Differently from most analyses, this account divorces ability modals from a notion of agency.

Suppose that, yesterday at noon, Maria considered flipping a fair coin, but didn't. What probability do you assign to "If Maria had flipped the coin, the coin would have landed heads"? Now suppose that, contrary to fact, Maria did indeed flip the coin. In that counterfactual scenario, what is the probability of "The coin will land tails"? The two questions sound strikingly similar. I argue that they sound similar because they are equivalent. The chance of a counterfactual "If A, would C" equals the chance of C, in the counterfactual scenario that A (and a similar principle holds for credence). This principle does better than similar principles that have been defended (like Skyrms' Thesis), avoids triviality, and gives us important clues for a semantics for counterfactuals.

We advocate and develop a states-based semantics for both nominal and adjectival confidence reports, as in "Ann is confident/has confidence that it's raining", and their comparatives "Ann is more confident/has more confidence that it's raining than that it's snowing". Other examples of adjectives that can report confidence include "sure" and "certain". Our account adapts Wellwood's account of adjectival comparatives in which the adjectives denote properties of states, and measure functions are introduced compositionally. We further explore the prospects of applying these tools to the semantics of probability operators. We emphasize three desirable and novel features of our semantics: (i) probability claims only exploit qualitative resources unless there is explicit compositional pressure for quantitative resources; (ii) the semantics applies to both probabilistic adjectives (e.g., "likely") and probabilistic nouns (e.g., "probability"); (iii) the semantics can be combined with an account of belief reports that allows thinkers to have incoherent probabilistic beliefs (e.g. thinking that A & B is more likely than A) even while validating the relevant purely probabilistic claims (e.g. validating the claim that A & B is never more likely than A). Finally, we explore the interaction between confidence-reporting discourse (e.g., "I am confident that...") and belief-reports about probabilistic discourse (e.g.,"I think it's likely that..").

A long-standing tension in semantic theory concerns the reconciliation of positive gradable adjective (GA) ascriptions and comparative GA ascriptions. Vagueness-based ap- proaches derive the comparative from the positive, and face non-trivial challenges with incommensurability and non-GA comparatives. Classic degree-based approaches effectively derive the positive from the comparative, out of sync with the direction of evidence from morphology, and create some difficulties in accounting for GA scale-mates with differing thresholds (e.g., cold ∼ warm ∼ hot). We propose a new reconciliation that capitalizes on recent proposals analyzing GAs as predicates of states. On our account, GAs lexically involve both a threshold property and a background state structure. Positive occurrences of GAs make use of the threshold property, while comparative occurrences make use of degrees representing elements of the background structure. Our approach preserves the virtues of classic degree-based approaches while offering a natural account of scale-mates, and without appeal to covert morphemes like POS or related devices. As we show, it is possible to inject our solution back into the classic degree-based approach, but we find reasons to prefer our states-based account.

Suppose you’re certain that a claim—say, ‘Frida is tall’—does not have a determinate truth value. What attitude should you take towards it? This is the question of the cognitive role of indeterminacy. This paper presents a puzzle for theories of cognitive role. Many of these theories vindicate a seemingly plausible principle: if you are fully certain that A, you are rationally required to be fully certain that A is determinate. Call this principle ‘Certainty’. We show that Certainty, in combination with some minimal side premises, entails a very implausible claim: whenever you’re certain that it’s indeterminate whether A, it is rationally required that you reject A. This is a surprising result, which requires abandoning at least some intuitive views about indeterminacy and cognitive role.

The literature on indicative conditionals contains two appealing views. The first is the selectional view: on this view, conditionals operate by selecting a single possibility, which is used to evaluate the consequent. The second is the informational view: on this view, conditionals don’t express propositions, but rather impose constraints on information states of speakers. Both views are supported by strong arguments, but they are incompatible on their standard formulations. Hence it appears that we have to choose between mutually exclusive options. But the options are not exclusive. This paper develops a theory of the semantics and assertability of conditionals that is both selectional and informational. The theory vindicates the signature inferences of both selectional and informational theories, including so-called Conditionals Excluded Middle and principles about the interplay between conditionals and ‘might’. It also predicts some interesting and puzzling facts about the assertion of conditionals.

On an influential line of thinking tracing back to Ramsey, conditionals are closely linked to the attitude of supposition. When applied to counterfactuals, this view suggests a subjunctive version of the so-called Ramsey test: the probability of a counterfactual If A, would B ought to be equivalent to the probability of B, under the subjunctive supposition that A. I present a collapse result for any view that endorses the subjunctive version of the Ramsey test. Starting from plausible assumptions, the result shows that one’s rational credence in a would-counterfactual and in the corresponding might-counterfactual have to be identical.

Know-how and expressivism are usually regarded as disjoint topics, belonging to distant areas of philosophy. This paper argues that, despite obvious differences, the two debates have important similarities. In particular, semantic and conceptual tools developed by expressivists can be exported to the know-how debate. On the one hand, some of the expressivists' semantic resources can be used to deflect Stanley and Williamson's influential argument for factualism about know-how: the claim that knowing how to do something consists in knowing a fact. On the other, expressivism provides the resources to formulate a nonfactualist account of know-how. On this account, know-how has a kind of nonpropositional content and plays the role of guiding performance of action, rather than recording information from the environment

How can expressivists make sense of the practice of communication? If communication is not a joint enterprise aimed at sharing information about the world, why do we engage in communication the way we do? Call this *the problem of communication*. Starting from basic assumptions about the rationality of speakers and the nature of assertion, we argue that speakers engaging in conversation about normative matters must presuppose that there is a unique normative standard on which the attitudes of conversational participants ought to converge. This gives the beginning of a solution to the problem of communication on behalf of expressivists.

This paper develops an information-sensitive theory of the semantics and probability of conditionals and statements involving epistemic modals. The theory validates a number of principles linking probability and modality, including the principle that the probability of a conditional If A, then C equals the probability of C, updated with A. The theory avoids so-called triviality results, which are standardly taken to show that principles of this sort cannot be validated. To achieve this, we deny that rational agents update their credences via conditionalization. We offer a new rule of update, Hyperconditionalization, which agrees with Conditionalization whenever nonmodal statements are at stake but differs for modal and conditional sentences.

This paper investigates what happens when we merge two different lines of theorizing about counterfactuals. One is the comparative closeness view, which was developed by Stalnaker and Lewis in the framework of possible worlds semantics. The second is the interventionist view, which is part of the causal models framework developed in statistics and computer science. Common lore and existing literature have it that the two views can be easily fit together, aside from a few details. I argue that, on the contrary, transplanting causal-models-inspired ideas in a possible worlds framework yields a new semantics. The difference is grounded in different algorithms for handling inconsistent information, hence it touches on issues that are at the very heart of a semantics for contrary-to-fact conditionals. Roughly, Stalnaker/Lewis semantics requires us to evaluate the consequent of a counterfactual at all closest antecedent-verifying possibilities. Causal-models-based semantics also does this, but in addition uses the information contained in the antecedent, together with background causal information, to shift what worlds count as closest. This makes systematically different predictions and generates a new logic. The upshot is that we have a new semantics to study, and a substantial theoretical choice to make.

This paper investigates the link between informational consequence and credence. I first suggest a natural constraint, namely that informational consequence should preserve certainty: on any rational credence distribution, when the premises of an informational inferences have credence 1, the conclusion also has credence 1. Then I show that the certainty‐preserving constraint leads to triviality. In particular, the following three claims are incompatible: (i) informational consequence is extensionally distinct from classical consequence; (ii) informational inferences preserve certainty; (iii) credences obey (a subset of) classical Bayesian constraints. The proof is straightforward, but the theoretical implications are substantial. Informational theorists need to either give up the idea that credence applies to epistemic discourse, or develop a nonclassical theory of credence and credal update. Moreover, the result also suggests that there is a connection between informational consequence and triviality results, including classical triviality results like Lewis's.

Suppose that you're certain that a certain sentence, e.g. "Frida is tall", lacks a determinate truth value. What cognitive attitude should you take towards it—reject it, suspend judgment, or what else? We show that, by adopting a seemingly plausible principle connecting credence in A and Determinately A, we can prove a very implausible answer to this question: i.e., all indeterminate claims should be assigned credence zero. The result is striking similar to so-called triviality results in the literature on modals and conditionals.

On a traditional view, the semantics of natural language makes essential use of a context parameter, i.e. a set of coordinates that represents the situation of speech. In classical semantic frameworks, this parameter plays two key roles: first, context contributes to determining the content of utterance; second, it is crucial for defining logical consequence. I point out that recent empirical proposals about context shift in natural language (in particular, context-shifting semantics in the style of Anand and Nevins 2004) are incompatible with the traditional view of context. At the same time, there is increasing cross-linguistic support for this brand of context shift. Hence I suggest that context has no place in semantic theory proper. We should revert back to so-called multiple-indexing frameworks that were developed by Montague and others, and relegate context to the postsemantic stage.

Natural language conditionals seem to be subject to three logical requirements: they invalidate Antecedent Strengthening, they validate so-called Simplification of Disjunctive Antecedents, and they allow for the replacement of logically equivalent clauses in antecedent position. Unfortunately, these requirements are jointly inconsistent. Conservative solutions to the puzzle drop Simplification, treating it as a pragmatic inference. I show that pragmatic accounts of Simplification fail, and develop a truthmaker semantics for conditionals that captures all the relevant data. Differently from existing truthmaker semantics, my semantics extends, rather than replaces, standard possible worlds semantics. The main innovation is the notion of a truthmaker in play: this notion is cognitive, rather than metaphysical, and can be defined in a purely syntactic way.

Disjunctions in the scope of possibility modals give rise to a conjunctive inference, generally referred to as ‘free choice.’ For example, Emma can take Spanish or Calculus suggests that Emma can take Spanish and can take Calculus. This inference is not valid on standard semantics for modals in combination with a Boolean semantics for disjunction. Hence free choice has sparked a whole industry of theories in philosophy of language and semantics. This paper investigates free choice in sentences involving a non-monotonic modified numeral, under which we embed a possibility modal scoping over disjunction. One example is Exactly one student can take Spanish or Calculus. As we point out, the presence of certain readings of these sentences is a key test for a prominent approach, which analyzes free choice as a kind of scalar implicature. We report on two experiments investigating the readings of such sentences, using an inferential task. Our results are challenging for the implicature approach. We sketch two possible solutions within this approach, either adopting a different recent implicature algorithm, or exploring a different meaning for modified numerals with exactly. Both of them suffer from a variety of problems. We then discuss a third solution, which exploits a recent account of free choice based on homogeneity. This approach can account for our results, in combination with plausible assumptions about homogeneity projection, though it too has open issues with related cases. Regardless of which solution is chosen, non-monotonic contexts turn out to be an important test case for theories of free choice, implicature, and modified numerals.

On a popular view dating back to Russell, descriptions, both definite and indefinite alike, work syntactically and semantically like quantifiers. I have an argument against Russell's view. The argument supports a different picture: descriptions can behave syntactically and semantically like variables. This basic idea can be implemented in very different systematic analyses, but, whichever way one goes, there will be a significant departure from Russell. The claim that descriptions are variables is not new: what I offer is a new way of defending it. The argument centers on attitude reports. I argue that we should recognize a new reading of descriptions under attitude reports, which I call ‘singular opaque’. The existence of this reading cannot be explained on the traditional Russellian view, and demands a switch to the variable view. Along the way, some interesting new facts about attitude reports come to light

According to the orthodox account developed by Kaplan, indexicals like I, you, and now invariably refer to elements of the context of speech. This essay argues that the orthodoxy is wrong. I, you, and the like are shifted by certain modal operators and hence can fail to refer to elements of the context, for example, I can fail to refer to the speaker. More precisely, indexicals are syntactically akin to logical variables. They can be free, in which case they work, roughly, on the Kaplan model. But they can also be bound: this happens, in a systematic fashion, when they are in the scope of epistemic modals or attitude verbs. The new view has two interesting philosophical consequences. First, it vindicates a broadly Fregean perspective on referential expressions, essentially refuting the idea that indexicals are rigid designators. Second, it suggests a new picture of the interaction between context and linguistic meaning: compositional semantics does not need to look at the context and hence has no need for a context parameter.

I argue that classical counterfactual semantics in the style of Stalnaker, Lewis, and Kratzer validates an inference pattern that is disconfirmed in natural language. The solution is to alter the algorithm we use to handle inconsistency in premise sets: rather than checking all maximally consistent fragments of a premise sets, as in Krazter’s semantics, we selectively remove some of the premises. The proposed implementation starts from standard premise semantics and involves a new ‘filtering’ operation that achieves just this removal. The resulting semantics is interestingly related to the semantics for counterfactuals emerging from Judea Pearl’s causal models framework in computer science: in particular, filtering is a possible worlds semantics counterpart of Pearl’s interventions

Statements about the future are central in everyday conversation and reasoning. How should we understand their meaning? The received view among philosophers treats will as a tense: in ‘Cynthia will pass her exam’, will shifts the reference time forward. Linguists, however, have produced substantial evidence for the view that will is a modal, on a par with must and would. The different accounts are designed to satisfy different theoretical constraints, apparently pulling in opposite directions. We show that these constraints are jointly satisfied by a novel modal account of will. On this account, will is a modal but doesn't work as a quantifier over worlds. Rather, the meaning of will involves a selection function similar to the one used by Stalnaker in his semantics for conditionals. The resulting theory yields a plausible semantics and logic for will and vindicates our intuitive views about the attitudes that rational agents should have towards future-directed contents.

Many recent theories of epistemic discourse exploit an informational notion of consequence, i.e. a notion that defines entailment as preservation of support by an information state. This paper investigates how informational consequence fits with probabilistic reasoning. I raise two problems. First, all informational inferences that are not also classical inferences are, intuitively, probabilistically invalid. Second, all these inferences can be exploited, in a systematic way, to generate triviality results. The informational theorist is left with two options, both of them radical: they can either deny that epistemic modal claims have probability at all, or they can move to a nonstandard probability theory.