The Semantics and Philosophy of Language Workshop is pleased to welcome Scott AnderBois (Connecticut) for the fourth talk of the quarter.

DATE: May 4, 2012
TIME: 11:30-1:20pm
PLACE: Cobb 107

‘Alternative unconditionals in Yucatec Maya’

Unconditionals are sentences which intuitively serve to indicate that a proposition will hold regardless of how some other issue is resolved. Despite their connections to well-studied constructions like conditionals, questions, free relatives, and subjunctive mood, they are relatively understudied in English and even more so cross-linguistically. Focusing on alternative unconditionals like (1), I propose an analysis where unconditionality arises from the collision of two conflicting properties: (i)unconditional antecedents (underlined) are expressions whose only at-issue contribution to discourse is to evoke alternatives (i.e. they are purely inquisitive); (ii) topics are inherently `anti-inquisitive’ environments.

(1) xíiimbal-nak wáa áalkab-nak Maribel-e’ k-u k’uchul tu yora’ij
walk-SUBJ or run-SUBJ Maribel-TOPIC IMPERF-3S arrive on time
`Whether Maribel walks or runs, she will arrive on time.’

While property (i) is reflected directly in English by the interrogative form of unconditional antecedent (as argued by Rawlins (2008)), I argue that it nonetheless holds of YM as well. Property (ii), on the other hand, is more readily apparent in YM since the topic morpheme -e’ occurs in a wide variety of superficially unrelated constructions. Despite the absence of a robust topic construction in English, I argue that we can nonetheless provide independent evidence for this property in English based on two observations about if-conditionals.

The Semantics and Philosophy of Language Workshop is pleased to welcome Seth Yalcin (Berkeley) for our third talk of the quarter.

DATE: Friday, April 27, 2012
TIME: 11:00-1:00pm
PLACE: Stuart 102

“Alternatives in the analysis of epistemic and deontic modalities”

Various motivations have been offered for treating epistemic modal clauses as somehow sensitive to a space of alternatives. Various motivations have also been offered for treating deontic modal clauses as somehow sensitive to a space of alternatives. To what extent do the motivations in each case have a common source? To what extent can a single model of alternatives serve to explain the relevant data? When does alternative-sensitivity for these modals belong to pragmatics, and when to compositional semantics? I investigate these questions in this talk.

The Semantics and Philosophy of Language Workshop is very happy to welcome Will Starr (Cornell) for our second talk of the quarter.

DATE: Friday, April 13, 2012
TIME: 11:30-1:20pm
PLACE: Cobb 107

“A preference semantics for imperatives”

While there is a long tradition in philosophy dedicated to understanding the meaning of imperative sentences, e.g. ‘Dance!’, recent research by linguists has made its own advances. In this paper, I argue that three observations about English imperatives are problematic for approaches from both traditions. In response, I offer a new analysis according to which the meaning of an imperative is identified with the characteristic effect its uses have on the agents’ attitudes. More specifically, I propose that an imperative changes what the agents’ take to be preferred. Using preferences rather than previously proposed structures achieves a desirable theoretical unity. Work on rationality in decision theory and artificial intelligence relies heavily on the idea that preference is key to understanding how rational agents decide what to do. This unity is essential for bridging the gap between a semantics for imperatives and an explanation of how imperatives are used to inform what we do. It also provides a more precise way of articulating Grice’s fundamental insight that pragmatics and rationality are deeply interconnected. I will conclude by describing how this approach can be used not only to understand the relationship between imperatives and modals.

The Semantics and Philosophy of Language Workshop is happy to welcome Friederike Moltmann (ENS, Paris) for our first meeting of the Spring term.

DATE: Friday, March 30, 2012
TIME: 11:30-1:20pm
PLACE: Cobb 107

`The semantics of ‘cases’‘

The noun ‘case’ in English (and corresponding nouns in other European
languages) appears in a number of related constructions:

(1) a. In case it rains, we won’t go. In that case we stay home.
b. This is a case of a strange illness
c. The case of the stolen statue is a puzzling one.
d. In this country, it is rarely the case that it rains.

I will outline a unified semantic analysis of the constructions in (1a-c) on the basis of a distinctive ontology of ‘cases’ as ‘filtered’ objects. Moreover, I will argue that the construction in (1d) involves the relation of truthmaking, relating a sentence to a ‘case’.

The Semantics and Philosophy of Language Workshop is pleased to welcome Eva Csipak (Göttingen, grad student) for our fourth meeting of the term.

DATE: Thursday, Feb. 16, 2012
TIME: 3:30-5pm
PLACE: Cobb 107

`Pizza subjunctives, plans and alternatives‘

While many analyses of the subjunctive (e.g. Giannakidou 2009 for Greek, Schlenker 2005 for French) focus on its non-factive uses, I discuss factive uses of the subjunctive in German, such as in (1):

(1) Ich hätte noch Pizza im Kühlschrank.
‘I have.SUBJ PART pizza in-the fridge’

The sentence is odd if the speaker knows that there is no pizza, or even if she is uncertain – in fact, it is very close to its indicative counterpart. I argue that an analysis of (1) requires three crucial components: factivity, “plans”, and Villalta’s (2008) notion of alternatives: for (1) to be felicitous, it needs to be uttered in a context where there is pizza in the fridge, where a salient individual can make use of this information (because he is hungry and can eat the pizza), and where there are alternatives (such as ordering other food).

The Semantics and Philosophy of Language Workshop is happy to welcome Peet Klecha (Chicago, grad student) for this term’s third talk.

DATE: February 10, 2012
TIME: 11:30-1:30pm
PLACE: Gates Blake 321

“Modals, Conditionals, and Imprecision”

This paper proposes that the interpretation of modals is subject to (im)precision (a la Lasersohn 1999), explaining certain contextual domain shifting effects observed by Lewis (1979).

Bryan: This must be a pen – I’m looking right at it.
Alice: Not so; you could be the victim of a deceiving demon.

I argue that Alice raises the standard of precision in (1), thereby widening the domain of `have to’ to include otherwise ignorable worlds. This explains a number of similarities between domain shifting of the type in (1) and canonical cases of
imprecision, and simplifies the modal semantics. I will also extend the analysis to conditionals, accounting for so-called Sobel Sequences.

The Semantics and Philosophy of Language Workshop is happy to welcome Elizabeth Allyn Smith (Northwestern) for this term’s second talk.

DATE: January 27, 2012
TIME: 11:30-1:30pm
PLACE: Gates Blake 321

“Some observations, revisions, and puzzles in the semantics of comparative correlatives”

Abstract: Comparative correlative sentences (CCs) like ‘the more, the merrier’ or ‘the more a dog eats, the more it drinks’ present a number of puzzles for their analysis, including the question of whether you can use the same denotation for its comparative morpheme as the one found elsewhere in natural language. Like many before me, I will argue that CCs can be given a completely compositional analysis using the ‘regular’ comparative, but I will also present new data showing that certain revisions are necessary and that, for example, the semantics of English CCs has less in common with the semantics of conditionals than previously believed. I will also pose a number of new puzzles, both for the analysis of CCs in English, and also cross-linguistically, including some current work on CCs in Bosnian/Croatian/Serbian and Japanese.

The Semantics and Philosophy of Language Workshop is happy to welcome Bridget Copley (CNRS, Paris) for this term’s first talk.

DATE: January 13, 2012
TIME: 11-1pm
PLACE: Landahl Center Seminar Room

`Connecting events‘

There is a consensus that certain events—Vendlerian Accomplishments, most saliently—are in fact composed of two sub-events, chained together in a causal relationship: John opened the door, for example, has a causing sub-event e1, and a result sub-event e2 . Various proposals have in common an operation which chains events causally, allowing a straightforward expression of the insight that John is the Agent of only the first, causing, sub-event, e1; this event then is ‘chained’ with e2, of which the Theme is predicated (Pustejovsky (1995), Higginbotham (2000), Ramchand (2008)). While these proposals allow for insightful analysis of structural and semantic properties of complex events, it is important to note that such advances come at the cost of introducing a novel rule of composition; it would be more desirable to express the connection between an event e1 and its result sub-event e2 in a way that made use of existing rules of composition.

There are also empirical challenges to this event-chaining approach to complex events.The event-chaining hypothesis entails that e2 is an inevitable consequence of e1. However, there are many cases in natural language where there is an Agent doing something (e1) which would normally (ceteris paribus) be the causing subevent of a second happening subevent, but the happening (e2) is non-existent, or the wrong kind of happening (e.g. progressives, non-culminating accomplishments, etc.). Possible worlds can be recruited to make up for this deficiency, but such recruitment of course complicates the theory.

In this talk I will present a proposal (Copley & Harley 2011) that treats an event as a function from an initial situation to the situation that results ceteris paribus. States are treated as situations. This conception of events and states corresponds to the traditional (Comrie 1976, e.g.) idea that events involve energy or force (whose ceteris paribus effect can nonetheless be nullified by other forces) while states do not involve energy at all. This treatment of events is helpful in addressing the issues mentioned above. Firstly, because an event is a function, an event and its result situation can be linked without recourse to anything other than Functional Application. Secondly, since the ceteris paribus condition is part of the definition of an event, the empirical challenges to event-chaining can be accounted for without additional machinery; in effect, causal chains of forces (events) and situations (states) are used to construct possible worlds.

The Semantics and Philosophy of Language Workshop is happy to welcome Katerina Chatzopoulos (Chicago, graduate student) for this term’s final talk.

DATE: November 18, 2011
TIME: 11-1pm
PLACE: Wieboldt 130

`Redefining Jespersen’s Cycle‘

The goal of this talk is to provide a semantic refinement of the Negative Cycle, known as Jespersen’s cycle (Jespersen 1917, 1924), through a definition that spells out and formalizes a background assumption in current leading research (van der Auwera 2009, 2010, van Gelderen 2008, 2011, Kiparsky & Condoravdi 2007): that the cyclicity of the phenomenon is semantic in nature and independent from its morphosyntactic realization in each one of its crosslinguistic manifestations. Although the assumptions and findings of this study are in agreement with generative outlooks on grammaticalization (Roberts & Roussou 2003, van Gelderen 2004), the representations are influenced by Autolexical Grammar (Sadock 1991, Sadock & Schiller 1993), as this allows a better visualization of the diachronic processes involved (instances of leftward lexical micromovement) and enables us to capture the notion of multiple dominance without reference to checking or agreement.

The proposed definition, abstracts away from the exact realization of the Negative cycle in French (Bréal 1897/1900, Clarke 1904, Horn 1989) and other typical Jespersen languages (e.g. English, Horn 1989, Wallage 2005; Dutch, Hoeksema 1997, Zeijlstra 2004; Egyptian, Gardiner 1903; Old Norse, van Gelderen 2008; Arabic and Berber, Lucas 2007) and views Jespersen’s cycle as a diachronic process that targets intensified predicate negation and elevates it to propositional. Motivation comes from the history of Greek, where negator renewal took place through a former emphatic, yet non discontinuous form of negation: the negative indefinite udhen, that followed the path from negative indefinite, to negative adverb, to sentential negation. This definition is all inclusive and accommodates not only for Greek, but a number of other languages in which negator renewal deviates in one way or another from the prototypical case of French (Chinese, Semitic languages, Athabaskan, German and Bantu languages).

Intensified negation (referred to as ‘emphatic negation’ in the relevant literature) is viewed here as a scale evoking or alternative evoking operation that specializes on scalar predicates, predicates that are gradable or allow for some sort of quantification. Intensified negation, e.g. ‘John didn’t drink at all’, literally negates the endpoint of a scale and everything to its left by implicature (if a Horn scale). It is shown that once the intensified form of negation loses this specialization and applies to all sorts of predicates, quantifiable and not, then it can be safely diagnosed as plain propositional negation. Thus the particular sort of semantic bleaching found in Jespersen’s cycle involves loss of reference to a scale. The exact mechanism of Jespersen’s Cycle can be formalized as involving: (a) lexicalization of the standard of comparison (cf. Kennedy & Levin 2008) and (b) re-application of the measure function. Akin in this sense is the diachronic development of other scale evoking linguistic operations in which regular renewal has been attested: comparatives (Paradis 2003), diminutives (Savickiene 1998) and honorifics (Traugott & Dasher 2002). This approach unites typical and atypical Jespersen’s cycle manifestations and views the Jespersen processes as an instance of a broader tendency active in natural language in general: scalar endpoint lexicalization followed by degree reinforcement.

The Semantics and Philosophy Workshop is pleased to welcome Maria Aloni (University of Amsterdam) for a specially-scheduled session.

DATE: Wednesday, November 16, 2011
TIME: 12-1:30pm
PLACE: Wiedboldt 130

‘Modal inferences in marked indefinites’ (joint work with Angelika Port)

Uses of unmarked indefinites like English a boy can give rise to a large number of pragmatic effects. For example, when told that you may invite a boy, you normally conclude that every boy is a permissible option (free choice inference), or, on the specific reading of the sentence, that there is a specific boy you may invite, but the speaker doesn’t know which one (ignorance inference). Many languages have developed marked indefinite forms (often with a restricted distribution) where these modal inferences are no longer pragmatic effects, but have been fully integrated in the conventional meaning of the expression. Free choice indefinites exemplify cases where the free choice inference has been conventionalized (Dayal 1998, Giannakidou 2001, Men ́endez-Benito 2010, Chierchia 2010, among others). Epistemic indefinites, also known as modal or referentially vague indefinites, exemplify cases where the ignorance inference has been conventionalized (Jayez & Tovena 2006, Alonso-Ovalle & Men ́endez-Benito 2010, Giannakidou & Quer 2011).

In this talk we will focus on two marked indefinite determiners: Italian un qualche (Zamparelli 2007) and German irgendein (Haspelmath 1997, Kratzer & Shimoyama 2002). In the first part of the talk we will identify a number of functions (context-meaning pairs) for marked indefinites, and discuss the distribution of these two items with respect to these functions (Aloni & Port 2011). The most striking aspect of the observed distribution is the different behavior of the two indefinites under epistemic and under deontic modals. Under epistemic modals both indefinites are licensed and give rise to an ignorance inference; under deontic modals only German irgendein is licensed and gives rise to a free choice inference. In the second part of the talk we will give a formal account of these facts in the framework of a Dynamic Semantics with Conceptual Covers (Aloni 2001).