Workshop on
Part-whole structures in natural language
26-27 May 2022
Workshop on
26-27 May 2022
The part-whole relation is an important notion in human mind that plays a crucial role in how we perceive and categorize objects in the external world as well as mind-internal representations. This workshop is dedicated to various expressions of this concept in natural language. It focuses mainly on semantic (but also syntactic) properties of part-whole structures within partitives, different types of pluralities, the mass/count distinction and event structure both from a theoretical and empirical perspective. The talks will investigate a variety of expressions referring within distinct ontological domains such as concrete individuals, eventualities and other abstract entities and offer mereological considerations based on a wide variety of linguistic evidence including data from typologically diverse languages, language acquisition and diachronic research. We hope that this event will contribute to the search of universal properties of the module of language faculty dedicated to the notion of part-whole.
All times below are according to the Central European Summer Time (CEST Time Zone, UTC +2).
Thursday 26/05/2022
12:45-13:00 Opening remarks
Chair: Marcin Wągiel
13:00-13:45 Viola Schmitt (Humboldt University of Berlin): Modal operators, orderings and the lack of world pluralities [slides]
13:45-14:30 Pavel Caha (Masaryk University in Brno): The (pseudo-)morphology of the pseudo-partitives [slides]
14:30-15:00 Break
Chair: Pavel Caha
15:00-15:45 Sarah Zobel (Humboldt University of Berlin): The German adverb of quantity größtenteils: Quantification over parts [slides]
15:45-16:30 Artemis Alexiadou (Leibniz-ZAS Berlin): Numeral-noun combinations and the hidden partitive hypothesis revisited [slides]
16:30-17:00 Break
Chair: Peter Sutton
17:00-17:45 Nina Haslinger (University of Göttingen): Contextual constraints on 'weak' plural predication: A unified perspective on cumulativity and non-maximality [slides]
17:45-18:30 Lucas Champollion (New York University): Negative events and truthmaker semantics [slides]
Friday 27/05/2022
Chair: Nina Haslinger
13:00-13:45 Michelangelo Falco & Roberto Zamparelli (University of Trento): Covert partitives [slides]
13:45-14:30 Kurt Erbach (Saarland University): Towards a history of the English countability system [slides]
14:30-15:00 Break
Chair: Sarah Zobel
15:00-15:45 Mathieu Paillé (McGill University): Inclusion of parts, or exclusion of predicates: Comparing two exhaustivity accounts of homogeneity [slides]
15:45-16:30 Peter Sutton (UPF Barcelona): Countability, copredication and polysemy [slides]
16:30-17:00 Break
Chair: Viola Schmitt
17:00-17:45 Scott Grimm (University of Rochester): Counting, relations and context [slides]
17:45-18:30 Marcin Wągiel (Masaryk University in Brno): Event-internal and external quantification and the mereotopology of events [slides]
The workshop is an online event organized by the Department of Linguistics and Baltic Languages at the Masaryk University in Brno. It is part of the research project "Part-whole structures across languages" (GA20-16107S) funded by the Czech Science Foundation (GAČR).
If you are interested in participating as an audience member, please fill in the registration form. The link to the talks will be sent to the provided e-mail address shortly before the workshop.
Artemis Alexiadou: Numeral-noun combinations and the hidden partitive hypothesis revisited [slides]
In this talk, I will revisit the idea that the structure of numeral-noun combinations contains a hidden partitive (Jackendoff 1968, Chomsky 1970) by looking at 'superfluous' partitive marking in child language, e.g., French une de voiture 'one of car', Karmiloff-Smith (1979) in comparison to data from a variety of languages, where partitives and numeral-noun combinations are similar. The non-canonical realization in child language is a so-called commission error, where children produce more material than adults. Such patterns support an analysis of numeral-noun combinations as hidden partitives, assuming, as in Sauerland, Alexiadou & Guasti (2021), that errors in the process of acquisition provide evidence for an underlying conceptual representation which the adult grammar in certain languages leaves unpronounced.
Pavel Caha: The (pseudo-)morphology of the pseudo-partitive [slides]
In the talk, I present a set of data showing that nouns in pseudo-partitive constructions enter into interesting morphological relations (syncretism and containment) with nouns in counting constructions and with plurals. I argue that in order to capture the patterns, we need to assume that the categories under discussion decompose into semantic ingredients smaller than morphemes. The implication is that morphemes (paradoxically standardly defined as the smallest meaning bearing units) actually have no inherent meaning, acquiring different meanings in different contexts depending on the semantic ingredients that they realize ("spell out").
Lucas Champollion: Negative events and truthmaker semantics [slides]
I develop an exact truthmaker semantics in the style of Kit Fine’s unilateral semantics, with special attention to the semantics of negation, which I base on a binary exclusion relation that represents a kind of incompatibility. My truthmakers are Davidsonian events, including negative events, which turn out to play a useful role in accounts of perception reports and other constructions. I discuss connections to possible world semantics and situation semantics. I argue for treating propositions as sets of events. I present evidence from counterfactuals with complex antecedents for this idea and for the cumulative nature of the exclusion relation. Parts of this project are joint work with Timothée Bernard (CNRS / Paris 7).
Kurt Erbach: Towards a history of the English countability system [slides]
The goal of this talk is to resolve seemingly conflicting claims from Toyota (2009) and Marckwardt (2019) about the history of countability in English. Toyota investigates numerical constructions and classifiers in the history of English and, finding few classifiers before 1100, claims that Old English did not differentiate between count and mass nouns, therefore it is a classifier language like Mandarin, despite having no overt classifiers. Marckwardt reviews the history of many, much, and fele ('many/much') and concludes that the present day distribution of many as a count quantifier and much as a mass quantifier reflect a continual development from their Old English forms. Taking a more complete look at the English countability system over time, including more than 20 morphosyntactic environments that can uncover a noun's countability, this talk presents evidence that undermines Toyota's claims, namely that Old English had a rich countability system including object mass nouns. While the presence of object mass nouns would rule out the possibility that Old English is a classifier language according to Chierchia (2021), analyses like Erbach et al. (2021) allow classifier languages to have object mass nouns and countability sensitive quantifiers, meaning the remaining question is whether Old English nouns all denote kinds as is assumed in classifier languages since at least Krifka (1995) or whether mass and count nouns differ in denotation, the former denoting kinds and the latter being predicates.
Michelangelo Falco & Roberto Zamparelli: Covert partitives [slides]
Consider the sentence in (1) and its possible continuations in (2a-d).
(1) We are a large industrial vehicle workshop and customers come and go all time. [Four trucks] came in yesterday evening and...
(2) a. [five] left the shop this morning.
b. [three] left the shop this morning.
c. [three trucks] left the shop this morning.
d. [three of the trucks] left the shop this morning.
The trucks that left could be a subpart of the trucks that arrived, a reading forced in (2d), or could at most overlap with them (2a). (3b) and (3c) are potentially ambiguous between a subset reading or a non-subset reading akin to that of (2a). In this study we present evidence that the subset reading of (2b) is indeed a "covert partitive", i.e., a covert version of (2d), and present the results of an on-line survey over 25 native speakers in two languages which shows which factors influence the availability of the subset reading.
Scott Grimm: Counting, relations and context [slides]
Nouns such as branch, twig, sequence, fence, or wall present a fundamental challenge in nominal semantics (Krifka 1989, Zucchi & White 2001, Chierchia 2010, Rothstein 2010). These nouns violate the typical analysis of countable nouns as "atomic" or "quantized": proper parts of a sequence (or a twig or a fence) may indeed also count as a sequence (or a twig or a fence). Other variants of this problem, such as the Pope's crown (which is constructed from three crowns), are known from the philosophy literature (Wiggins 1967) and demonstrate the problem's generality. This talk assesses a set of these problematic nouns, focusing on showing that they are a type of relational noun, albeit different from the most prominent relational nouns such as kinship terms (sister). That is, a piece of wood is a 'twig' by virtue of its relation to the branch it comes from. Further complexity enters as many of these nouns are ambiguous between multiple related senses, some relational and some not; yet, once the sense and, when relevant the relational variable, is fixed, the countability properties follow.
Nina Haslinger: Contextual constraints on 'weak' plural predication: A unified perspective on cumulativity and non-maximality [slides]
This talk investigates the pragmatic constraints on cumulative interpretations of sentences with multiple plural expressions. My starting point will be two sets of examples in which cumulative construals are unexpectedly absent. The first phenomenon are the "paired-cover contexts" discussed by Schwarzschild (1996) for examples like The fiction books complement the non-fiction books. The second phenomenon are cumulative interpretations of predicate and sentential conjunctions (Link 1984, Krifka 1990, Schmitt 2013), which have been used to argue that conjunctions of any type are plural expressions (Schmitt 2013, 2019), but are subject to contextual restrictions that are surprising under this view. I will take a fresh look at these data from the perspective of the recent literature on non-maximal construals of plural definites, e.g., a construal of The neighbors are asleep that does not require all of the neighbors to be asleep. It has been argued that non-maximal construals require a context with a special kind of goal or QUD (Malamud 2012, Križ 2015, Križ & Spector 2021, Bar-Lev 2021), and that this should be modeled by letting the QUD directly influence the truth conditions of plural sentences (Križ 2015). I argue that the same holds for cumulativity, and that the pragmatic restrictions on cumulativity are observed independently for non-maximality. Therefore, they do not provide an argument against a semantics that generates a cumulative construal for Schwarzschild's paired-cover examples, and treats predicate and sentential conjunctions as plural expressions. I conclude by sketching a formal implementation of this view within a version of the supervaluationist framework of Križ & Spector (2021).
Mathieu Paillé: Inclusion of parts, or exclusion of predicates Comparing two exhaustivity accounts of homogeneity [slides]
Discussion of homogeneity effects usually focuses on examples involving pluralities, but the effect is often taken to hold within atoms as well (e.g., Löbner 2000, Spector 2013, Križ 2015). In this talk, I compare two accounts of homogeneity, namely the account by Bar-Lev (2018, 2021) made for plural homogeneity, and the account proposed in Paillé (2020, 2021) for subatomic homogeneity. These accounts are similar in a number of ways: both understand the homogeneity paradigm in terms of weak lexical meaning paired with strengthening in positive sentences, both claim that this strengthening is specifically the result of an Exh(aust) operator, and both must posit constraints on the syntactic distribution of this Exh operator. But the accounts differ in the nature of the alternatives and in the specific property of Exh that leads to strengthening. Bar-Lev’s account posits the inclusion of proper parts, while Paillé's account uses the exclusion of conceptually related predicates. I show that neither theory can be carried over to the homogeneity effects they were not intended to capture: Bar-Lev’s account faces difficulties with subatomic homogeneity, specifically due to conjoined predicates (e.g., The flag is white and green), while Paillé's account is not translatable to plural homogeneity at all. As such, I suggest that plural and subatomic homogeneity effects are underlyingly united in only some of their properties (being the result of local exhaustification in positive sentences) while differing in the nature of the alternatives and whether these are included or excluded.
Viola Schmitt: Modal operators, orderings and the lack of world pluralities [slides]
In this talk, I will argue that worlds represent an exception to an otherwise apparently robust pattern, namely, to the generalization that all semantic domains contain pluralities. More specifically, conditionals and constructions involving neg-raising predicates have been discussed in the literature as candidates for configurations that might involve world-pluralities, as they exhibit two traits also connected with definite plurals: homogeneity and non-maximality. However, I show that they lack core traits of pluralities witnessed for all other domains, namely, cumulativity and the ability to be targeted by a certain class of quantifiers. I probe an explanation for this gap – partially building on previous claims in the literature, I try to link the use of worlds in these constructions to the use of degrees in constructions with gradable predicates that relate to non-additive scales.
Peter Sutton: Countability, copredication and polysemy [slides]
This first part of this talk presents a new situation-theoretic approach to polysemous nouns such as lunch, book, and statement in which they denote situations that contain multiple entities. For example, the thick and interesting book denotes a situation that contains a physical book and an informational entity (its contents) such that the meaning of book specifies that the latter is the contents of the former. Unlike other analyses (e.g., Asher 2011, Asher & Pustejovsky 2013, Gotham 2017), this analysis does not require a bespoke type- or object-constructor, but instead follows from standard situation theoretic assumptions. The second part of this talk addresses Gotham’s (2017) observation that numeral constructions such as three thick and interesting books force a ‘double distinctness’ reading under which both the physical and informational books must be distinct. I show that this reading can be accounted for by combining the above analysis with my previous work on countability (e.g., Sutton & Filip 2021), along with Gotham’s claim that modifiers can restrict individuation criteria.
Marcin Wągiel: Event-internal and external quantification and the mereotopology of events [slides]
Kim knocked on the door three times is ambiguous between an event-external interpretation (quantification over "occasions") and an event-internal interpretation (quantification over "acts" within a single occasion) (e.g., Cusic 1981, Andrews 1983, Cinque 1999). Polish and Mandarin Chinese mark the distinction formally via dedicated multiplicatives and distinct verbal classifiers (e.g., Donazzan 2013, Zhang 2017). Still, the relationship between occasions and acts remains unclear since both seem to fall into the ontological category of eventualities. In this talk, I argue that the event-external/internal distinction receives a straightforward explanation once the mereotopological notion of connectedness (Casati & Varzi 1999) is extended to the domain of events. Building on the theory of time by Mazzola (2019), I propose that event-internal interpretations concern quantification over simplex singular eventualities, whereas event-external readings concern counting clusters, i.e., structured configurations, thereof (see also Landman 2004, Henderson 2017).
Sarah Zobel: The German adverb of quantity größtenteils: Quantification over parts [slides]
In this talk, I analyze the semantic contribution of the German adverb of quantity größtenteils, which is comparable to that of English for the most part in that both adverbs intuitively quantify over parts of a totality. I discuss which sorts of entities and their parts größtenteils can intuitively quantify over, and how it accesses these totalities. In connection with this question, I also touch on the question whether größtenteils can quantify over parts of non-eventive entities directly, or whether any quantification over non-eventive entities arises indirectly via größtenteils quantifying over sub-events (as proposed for English for the most part by Nakanishi & Romero 2004). I argue that the former is the case and propose a formal analysis of größtenteils based on the proposal for English mostly/for the most part by Beck & Sharvit (2002).
marcin + dot + wagiel + at + phil + dot + muni + dot + cz