Last update di 06 april 2021


The representation and processing of distributivity and collectivity: ambiguity vs. underspecification

co-authored with Adrian Brasoveanu
Glossa 6(1): 1-22

Sentences with plural expressions can receive at least two interpretations. For example, the sentence The boys hold a balloon could mean that the boys as a group jointly hold one balloon (the collective reading) or that each boy holds one balloon, which would imply that as many balloons were held as there are boys (the distributive reading). Building on Frazier et al. (1999), we show that the human processor favors collective readings. Crucially, the preference for collective readings is only observed when the distributive reading has to be established through the means of phrasal distributivity (e.g., triggered by distributive quantifiers), and the preference disappears in case of lexical distributivity (e.g., the distributive interpretation of win). The findings provide evidence for different mental representations of the two types of distributivity and shed light on why the processor exhibits a default preference for collective interpretations.

A dynamic semantics of single-wh and multiple-wh questions

co-authored with Floris Roelofsen
In Proceedings of SALT 30

We develop a uniform analysis of single-wh and multiple-wh questions couched in dynamic inquisitive semantics. The analysis captures the effects of number marking on which-phrases, and derives both mention-some and mention-all readings as well as an often neglected partial mention-some reading in multiple-wh questions.


Production-based cognitive models as a test suite for reinforcement learning algorithms

We introduce a framework in which production-rule based computational cognitive modeling and Reinforcement Learning can systematically interact and inform each other. We focus on linguistic applications because the sophisticated rule-based cognitive models needed to capture linguistic behavioral data promise to provide a stringent test suite for RL algorithms, connecting RL algorithms to both accuracy and reaction-time experimental data. Thus, we open a path towards assembling an experimentally rigorous and cognitively realistic benchmark for RL algorithms. We extend our previous work on lexical decision tasks and tabular RL algorithms with a discussion of neural-network based approaches, and a discussion of how parsing can be formalized as an RL problem.

Reinforcement learning for production-based cognitive models

co-authored with Adrian Brasoveanu
In Proceedings of ICCM 2020

We introduce a framework in which we can start exploring in a computationally explicit way how complex, mechanistically specified, and production-based cognitive models of linguistic skills, e.g., ACT-R based parsers, can be acquired. Cognitive models for linguistic skills pose this learnability problem much more starkly than models for other `high-level' cognitive processes, as they call for richly structured representations and complex rules that require a significant amount of hand-coding. In this paper, we focus on how Reinforcement Learning (RL) methods can be used as a way to solve the production selection problem in ACT-R. Production rules are treated as the actions of an RL agent, and the ACT-R model/mind as the environment in the RL sense. We focus on a basic learning algorithm (tabular Q-learning) and a very simple task, namely lexical decision (LD), framed as a sequential decision problem, with the goal of learning when a specific rule should be fired. Learning is faster and less noisy for shorter LD tasks (fewer stimuli), but the Q-learning agent manages to learn longer tasks fairly well. Realistically long LD tasks and more complex models, e.g., parsers, are left for future research.

Computational Cognitive Modeling and Linguistic Theory

co-authored with Adrian Brasoveanu
Springer, Language, Cognition and Mind (LCAM) Series.


This open access book introduces a general framework that allows natural language researchers to enhance existing competence theories with fully specified performance and processing components. Gradually developing increasingly complex and cognitively realistic competence-performance models, it provides running code for these models and shows how to fit them to real-time experimental data. This computational cognitive modeling approach opens up exciting new directions for research in formal semantics, and linguistics more generally, and offers new ways of (re)connecting semantics and the broader field of cognitive science.

The approach of this book is novel in more ways than one. Assuming the mental architecture and procedural modalities of Anderson’s ACT-R framework, it presents fine-grained computational models of human language processing tasks which make detailed quantitative predictions that can be checked against the results of self-paced reading and other psycho-linguistic experiments. All models are presented as computer programs that readers can run on their own computer and on inputs of their choice, thereby learning to design, program and run their own models. But even for readers who won't do all that, the book will show how such detailed, quantitatively predicting modeling of linguistic processes is possible. A methodological breakthrough and a must for anyone concerned about the future of linguistics! (Hans Kamp)

This book constitutes a major step forward in linguistics and psycholinguistics. It constitutes a unique synthesis of several different research traditions: computational models of psycholinguistic processes, and formal models of semantics and discourse processing. The work also introduces a sophisticated python-based software environment for modeling linguistic processes. This book has the potential to revolutionize not only formal models of linguistics, but also models of language processing more generally. (Shravan Vasishth)


The book integrates formal semantic theories, specifically, DRT, and mechanistic processing models formulated in the ACT-R cognitive architecture into a new computational framework that can simulate human participants in real-time psycholinguistic tasks, fit these cognitive models to experimental data by embedding them into Bayesian models, and quantitatively compare qualitative (symbolic) theories.

In more detail, we introduce an extensible framework for mechanistic processing models that integrate formal syntax and semantics theories, an independently motivated cognitive architecture (ACT-R; Anderson and Lebiere 1998, Anderson 2007) and Bayesian methods of data analysis, parameter estimation and model comparison. The book culminates with the development of computationally-explicit processing models for natural language semantic phenomena that integrate Discourse Representation Theory (DRT; Kamp 1981, Kamp & Reyle 1993), ACT-R and Bayesian models.

These mechanistic processing models implement end-to-end simulations of human participants in real-time psycholinguistic experiments. End-to-end means that we model visual and motor processes together with specifically linguistic representations and processes in complete models of the experimental tasks. The models embed theoretical hypotheses about linguistic representations and about parsing processes in ACT-R. The resulting ACT-R cognitive models are then embedded in Bayesian models to fit them to experimental data, estimate their parameters and perform quantitative model comparison for qualitative (symbolic) theories.

A novel Python3 reimplementation of ACT-R called pyactr (current version; book version, including all book code) provides the crucial software infrastructure for developing ACT-R models that can be easily embedded in Bayesian models.

A Multiple Cue Explanation of Collective Interpretations with each

co-authored with Anna de Koster, Jennifer Spenader and Petra Hendriks
Proceedings of the 44th Annual Boston University Conference on Language Development. Cascadilla Press


Dynamic inquisitive semantics: Anaphora and questions

co-authored with Floris Roelofsen
Proceedings of Sinn und Bedeutung 23

This paper develops a dynamic inquisitive semantics and illustrates its potential to capture interactions between anaphora and questions.


co-authored with Adrian Brasoveanu
The Oxford Handbook of Experimental Semantics and Pragmatics, ed. by Chris Cummins and Napoleon Katsos

Quantification is abundant in natural language and is one of the most studied topics in generative grammar. Sentences with multiple quantifiers are famously ambiguous with respect to their quantifier scope, representing a type of ambiguity related to, but not necessary the same as, structural ambiguity. Two key questions in the psycholinguistic study of quantification are: (i) how does the human processor assign quantifier scope?; and (ii) how and under what circumstances is this scope assignment reanalyzed? The investigation of these questions lies at the intersection of psycholinguistics and theoretical linguistics. We summarize both strands of research, and discuss experimental data that played an essential role in the (psycho)linguistic theorizing about the topic of processing quantification and quantifier scope.


Building an ACT-R reader for eye-tracking corpus data

Topics in Cognitive Science

Cognitive architectures have often been applied to data from individual experiments. In this paper, I develop an ACT-R reader that can model a much larger set of data, eye-tracking corpus data. It is shown that the resulting model has a good fit to the data for the considered low-level processes. Unlike previous related works (most prominently, Engelmann, Vasishth, Engbert and Kliegl, 2013 ), the model achieves the fit by estimating free parameters of ACT-R using Bayesian estimation and Markov-Chain Monte Carlo (MCMC) techniques, rather than by relying on the mix of manual selection + default values. The method used in the paper is generalizable beyond this particular model and data set and could be used on other ACT-R models.

An extensible framework for mechanistic processing models: From representational syntax-semantics theories to quantitative model comparison

co-authored with Adrian Brasoveanu
In Proceedings of ICCM 2018

We introduce a Python3 reimplementation of ACT-R (Anderson and Lebiere 1998, Anderson 2007) in which we build an end-to-end simulation of syntactic parsing in a typical self-paced reading experiment. The model uses an eager left-corner parsing strategy implemented as a skill in procedural memory (following Lewis and Vasishth 2005), makes use of independently motivated components of the ACT-R framework (procedural memory, content-addressable declarative memory – cf. Wagers et al. 2009), and explicitly models the motor and visual processes involved in self-paced reading. The ACT- R model can be embedded in a Bayesian statistical model to estimate its subsymbolic parameters and perform model comparison.

Cumulative comparison: experimental evidence for degree cumulation

co-authored with Rick Nouwen
In The semantics of gradability, vagueness, and scale structure: experimental perspectives

In this paper we address the question whether it makes sense to assume that the domain of degrees, as used in degree semantics, consists not just of atoms, but also of degree pluralities. A number of recent works have adopted that assumption, most explicitly Fitzgibbons et al (2008); Beck (2014); Dotlačil and Nouwen (2016). In this paper, we provide experimental evidence for degree pluralities by showing that comparatives may express cumulative relations between degrees.

Quantitative Comparison for Generative Theories: Embedding Competence Linguistic Theories in Cognitive Architectures and Bayesian Models

co-authored with Adrian Brasoveanu
In Proceedings of BLS

We introduce a new framework in which (i) qualitative generative grammar hypotheses can be embedded into performance/processing models, and (ii) these integrated competence-performance models are embedded into Bayesian models, which enables us to fit them to experimental data and do quantitative model comparison for qualitative theories. Building on Lewis and Vasishth (2005), step (i) takes generative theories and embeds them in processing theories formulated in the ACT-R cognitive architecture (Anderson and Lebiere 1998). In a new Python3 computational implementation of ACT-R, we do step (ii), namely use ACT-R models as the likelihood component of full Bayesian models. We demonstrate (i) and (ii) by comparing models for lexical decision tasks that differ qualitatively and/or quantitatively.


The scope of nominal quantifiers in comparative clauses

co-authored with Rick Nouwen
Semantics and Pragmatics 10

We identify a new scope puzzle for quantifiers in comparative clauses. In particular, we argue that nominal quantifiers take scope at a higher level in the degree clause than previously assumed. On the assumption that quantifier scope is clause-bounded, this entails that there must be more structure in the clause than standardly assumed.

Children's Understanding of Distributivity and Adjectives of Comparison

co-authored with Anna de Koster and Jennifer Spenader
Proceedings of the 41st Annual Boston University Conference on Language Development. Cascadilla Press

How children learn to interpret sentences with plural expressions is still not fully understood.

Adults prefer collective readings with “the”-DPs, e.g. The boys pushed a car means one car. Replace “the” with “each” and adults only get distributive readings. Children allow both initially, gradually rejecting collective readings with “each”-DPs. Study 1 found that correctly rejecting collective readings with “each” DPs correlated with rejecting distributive readings with “the” DPs.

Study 2 investigated children's understanding of the adjective of comparison (AOC) “different” related to distributivity. If The boys pushed a different car , are there multiple cars (a sentence internal reading, similar to distributive readings), or did they push a different car than e.g. the girls (a discourse reading, similar to collective readings.)? Study 2 found children were more adult-like. However, rejecting the distributive interpretation with “the” in Study 1 did not correlate with rejection of internal readings with “different” in Study 2, contrary our predictions.


Experimental evidence for neg-raising in Slavic

co-authored with Mojmir Dočekal

This article discusses new experimental data that provide evidence for the existence of neg-raising in Slavic languages (in particular, in Czech). The results of the experiment are interpreted in Romoli's scalar theory of neg-raising (Romoli, 2012, 2013).

The comparative and degree pluralities

co-authored with Rick Nouwen
Natural Language Semantics, 24(1), 45-78

Quantifiers in phrasal and clausal comparatives often seem to take distributive scope in the matrix clause: for instance, the sentence John is taller than every girl is is true iff for every girl it holds that John is taller than that girl. Broadly speaking, two approaches were developed that derive this reading without postulating the (problematic) wide scope of the quantifier: the negation analysis and the interval analysis of than clauses. We propose a modification of the interval analysis, in which than clauses are not treated as degree intervals, but degree pluralities. This small change has significant consequences: it yields a successful account of differentials in comparatives and it correctly predicts the existence of hitherto unnoticed readings, viz. cumulative readings of clausal comparatives. Finally, this paper also makes the case that using degree pluralities is conceptually appealing: it allows us to restrict the analysis of comparatives by mechanisms that are postulated independently in the semantics of pluralities.

The Influence of Linguistic and Cognitive Factors on the Time Course of Verb-based Implicit Causality

co-authored with Arnout Koornneef, Paul van den Broek and Ted Sanders
The Quarterly Journal of Experimental Psychology, 69(3), 455-481

In three eye-tracking experiments the influence of the Dutch causal connective `want' (because) and the working memory capacity of readers on the usage of verb-based implicit causality was examined. Experiments 1 and 2 showed that although a causal connective is not required to activate implicit causality information during reading, effects of implicit causality surfaced more rapidly and more pronounced when a connective was present in the discourse than when it was absent. In addition, Experiment 3 revealed that - in contrast to previous claims (Long and De Ley, 2000) - the activation of implicit causality is not a resource-consuming mental operation. Moreover, readers with higher and lower working memory capacities behaved differently in a dual task situation. Higher span readers were more likely to use implicit causality when they had all their working memory resources at their disposal. Lower span readers showed the opposite pattern as they were more likely to use the implicit causality cue in case of an additional working memory load. The results emphasize that both linguistic and cognitive factors mediate the impact of implicit causality on text comprehension. The implications of these results are discussed in terms of the ongoing controversies in the literature, i.e., the focusing-integration debate and the debates on the source of implicit causality.


Incremental and Predictive Interpretation: Experimental Evidence and Possible Accounts

co-authored with Adrian Brasoveanu
Proceedings of Semantics and Linguistic Theory 25, 57-81

The main question we investigate is whether meaning representations of the kind that are pervasive in formal semantics are built up incrementally and predictively when language is used in real time, in much the same way that the real-time construction of syntactic representations has been argued to be. The interaction of presupposition resolution with conjunctions vs. conditionals with a sentence-final antecedent promises to provide us with the right kind of evidence. Consider the following `cataphoric' example and the contrast between AND and IF : Tina will have coffee with Alex again AND / IF she had coffee with him at the local cafe. We expect the second clause to be more difficult after AND than after IF : the conjunction AND signals that an antecedent that could resolve the again presupposition is unlikely to come after this point (the second conjunct is interpreted relative to the context provided by the first conjunct), while the conditional IF leaves open the possibility that a suitable resolution for the again presupposition is forthcoming (the first clause is interpreted relative to the context provided by the second clause). We present experimental evidence supporting these predictions and discuss two approaches to analyze this kind of data.

Pragmatic inferences with numeral modifiers: novel experimental data

co-authored with Stavroula Alexandropoulou, Yaron McNabb and Rick Nouwen
Proceedings of Semantics and Linguistic Theory 25, 533-549

We present results from two experiments that investigate pragmatic inferences triggered by numeral modifiers ("at least" and "more than" under universal quantifiers). We found that while both modifiers give rise to implicatures, they differ in the extent to which they trigger such inferences, contra some current theoretical analyses. We also found that the precise nature of the inferences in question is one that favours a theory positing a rather modest set of alternatives for modified numerals.

Why is distributivity so hard? New evidence from distributive markers and licensors in Czech

Slavic Languages in the Perspective of Formal Grammar : Proceedings of FDSL 10.5

Sentences with pluralities can be interpreted in several ways. One such interpretation involves collective action, in which the plurality, as a whole, functions as an argument. In another, distributive, interpretation, each member of the plurality fulfills the same argument role. It has been noted that the distributive interpretation is often marked. This paper discusses an experiment on Czech providing evidence that distributivity is marked because (i) readers prefer simpler syntactic structures (Minimal Attachment, Frazier, 1978), and (ii) distributivity is syntactically more complex than the collective interpretation and the difference is marked on dependent arguments, most commonly, objects.

Strategies for scope taking

co-authored with Adrian Brasoveanu
Natural Language Semantics, 23(1), 1-19

This squib reports the results of two experimental studies, a binary choice and a self-paced reading study, that provide strong support for the hypothesis in Tunstall (1998) that the distinct scopal properties of each and every are at least to some extent the consequence of an event-differentiation requirement contributed by each (Experiment 3 in Tunstall 1998, Chapter 5 designed to provide such evidence does not reveal the predicted effect). However, we also show that the emerging picture is more complex than the one painted in Tunstall (1998): English speakers seem to fall into at least three groups regarding the scopal properties of each and every.

The manner and time course of updating quantifier scope representations in discourse

co-authored with Adrian Brasoveanu
Language, Cognition and Neuroscience (formerly Language and Cognitive Processes), 30(3), 305-323

We present the results of two experiments, an eye-tracking study and a follow-up self-paced reading study, investigating the interpretation of quantifier scope in sentences with three quantifiers: two indefinites in subject and object positions and a universal distributive quantifier in adjunct position. In addition to the fact that such three-way scope interactions have not been experimentally investigated before, they enable us to distinguish between different theories of quantifier scope interpretation in ways that are not possible when only simpler, two-way interactions are considered. The experiments show that contrary to underspecification theories of scope, a totally ordered scope-hierarchy representation is maintained and modified across sentences and this scope representation cannot be reduced to the truth-conditional/mental model representation of sentential meaning. The experiments also show that the processor uses scope-disambiguating information as early as possible to (re)analyze scope representation.

Sentence-internal same and its quantificational licensors: A new window into the processing of inverse scope

co-authored with Adrian Brasoveanu
Semantics and Pragmatics, 8, 1-52

This paper investigates the processing of sentence-internal same with four licensors (all, each, every and the) in two orders: licensor+same (surface scope) and same+licensor (inverse scope). Our two self-paced reading studies show that there is no general effect of surface vs. inverse scope, which we take as an argument for a model-oriented view of the processing cost of inverse scope: the inverse scope of quantifiers seems to be costly because of model structure reanalysis, not because of covert scope operations. The second result is methodological: the psycholinguistic investigation of semantic phenomena like quantifiers and the licensing of sentence-internal readings should always involve a context that prompts a deep enough processing of the target expressions. In one of our two studies, participants read the target sentences after reading a scenario introducing the two sets of entities the quantifier NP and the same NP referred to and they were asked to determine whether the sentence was true or false relative to the background scenario every time. In the other study, the participants read the same sentences without any context and there were fewer follow-up comprehension questions. The relevant effects observed in the study with contexts completely disappeared in the out-of-context study, although the participants in both studies were monitored for their level of attention to the experimental task.


Probabilistic semantic automata in the verification of quantified statements

co-authored with Jakub Szymanik and Marcin Zajenkowski
Proceedings of CogSci 2014

Strategies used by people to verify quantified sentences, like `Most cars are white', have been a popular research topic on the intersection of linguistics, computer science, philosophy, and psychology. A prominent computational model of the task, semantic automata, has been introduced by van Benthem in 1983. In this paper we present a probabilistic extension of the model. We show that the model explains counting errors in the verification process. Furthermore, we observe that the variation in quantifier verification data cannot be explained by probabilistic number estimations, in particular, Approximate Number Sense.


What a Rational Interpreter Would Do: Building, Ranking, and Updating Quantifier Scope Representations in Discourse

co-authored with Adrian Brasoveanu
Pre-proceedings of Amsterdam Colloquium 2013

We frame the general problem of rationally (in the sense of Anderson et al's ACT-R framework) integrating semantic theories and processing, and indicate how this integrated theory could be explicitly formalized; an explicit formalization enables us to empirically evaluate semantic and processing theories both qualitatively and quantitatively. We then introduce the problem of quantifier scope, the processing difficulty of inverse scope, and two types of theories of scope, and discuss the results of a self-paced reading experiment and its consequences for these two types of theories. Finally, we outline how probabilities for LF construction rules could be computed based on the experimental results.

Reciprocals distribute over information states

Journal of Semantics, 30(4), 423-477

Reciprocal sentences require distributivity for the correct interpretation. In the semantic literature on reciprocals, it is standard to assume that distributivity spans almost everything in reciprocal clauses, including not only reciprocals but also predicates and their arguments. In contrast to this tradition, I argue that nothing but reciprocals themselves should be interpreted distributively. Such a semantics of reciprocal clauses can be accommodated in dynamic frameworks which can store and retrieve dependencies between pluralities, modeled as plural information states, i.e., sets of variable assignments (van den Berg 1996, Nouwen 2003, Brasoveanu 2007, among others). I argue that a reciprocal distributes over nothing else but a plural information state, and I provide compositional interpretation of reciprocity using Plural Compositional Discourse Representation Theory (the extension of Compositional Discourse Representation Theory of Muskens 1996 by Brasoveanu 2007, 2008).

Peeling, structural Case, and Czech retroactive infinitives

co-authored with Radek Šimik
Formal Desription of Slavic Languages: The Ninth Conference

We characterize a newly observed type of embedded infinitive in Czech which we call the "retroactive infinitive". From a distributional, semantic, and to some extent syntactic point of view, retroactive infinitives correspond to English retroactive nominals and gerunds. We argue that the syntactic behavior of retroactive infinitives supports the view that Dative can be a structural Case, that structural Dative is licensed higher than Accusative, and that so called "peeling" is the default mechanism of structural deficiency.


The online interpretation of sentence internal same and distributivity

co-authored with Adrian Brasoveanu
Proceedings of Semantics and Linguistic Theory 22

This paper investigates how sentence-internal same is processed with three of its licensors (each, all and the) and two orders: licensor+same, i.e., surface scope (Each student saw the same movie) and same+licensor, i.e., inverse scope (The same student saw each movie). Our study shows that (i) there is no effect of surface vs inverse order, which we take as an argument for a model-oriented view of the processing cost of inverse scope, and (ii) all is processed faster than each and the, which we take as an argument for a particular semantics of distributive licensors.

Licensing sentence-internal readings in English: An experimental study

co-authored with Adrian Brasoveanu
Logic, Language and Meaning. 18th Amsterdam Colloquium. Springer

Adjectives of comparison (AOCs) like same, different and similar can compare two elements sentence-internally, i.e., without referring to any previously introduced element. This reading is licensed only if a semantically plural NP is present. We argue in this paper that it is incorrect to describe a particular NP as either licensing or not licensing the sentence-internal reading of a specific AOC: licensing is more fine-grained. We use experimental methods to establish which NPs license which AOCs and to what extent and we show how the results can be interpreted against the background of a formal semantic analysis of AOCs. Finally, we argue that using Bayesian methods to analyze this kind of data has an advantage over the more traditional, frequentist approach.

Binominal each as an anaphoric determiner: A compositional analysis

Proceedings of Sinn und Bedeutung 16. MITWPL

The quantifier each can express distributivity at distance, as in The men read one book each, which could be paraphrased as each of the men read one book. Distributivity at distance is problematic for many semantic accounts, which, in order to capture it, need to resort to a non-compositional analysis. Compositional accounts, as Zimmermann (2002) and Blaheta (2003) can deal with basic cases but run into troubles when each appears in adjuncts or is embedded in another noun phrase. I offer a novel analysis of distance distributivity in which each is treated as an anaphoric determiner. This analysis is compositional but it avoids pitfalls of previous accounts. Finally, I show how the analysis can be formalized using Plural Compositional DRT.

The acquisition of distributivity in pluralities

co-authored with Gaetano Fiorin and Elena Pagliarini
Proceedings of the 36th Annual Boston University Conference on Language Development. Cascadilla Press

Definite plural noun phrases (NPs) can be interpreted either collectively or distributively. For example, the sentence the boys lifted two boxes can mean that the boys acted as a group in lifting two boxes (collective reading) or that each boy lifted two boxes (distributive reading). However, it has been observed that adults fully accept collective readings of definite plural NPs, but find distributive readings marginal. We present the results of an experiment testing the hypothesis that the degraded status of the distributive reading of plural definite NPs is a case of conversational implicature: a definite plural NP such as the boys is not interpreted distributively because, if the speaker wanted to express a distributive interpretation, she would have used the NP each boy which unambiguously expresses a distributive reading. The following predictions were tested: (i) in language development, the rejection of distributive readings with non-quantificational noun phrases follows the acquisition of each; (ii) the level of acquisition of each predicts whether subjects will reject distributive readings or not. The results of the experiment show that children learn to reject the collective reading of each before they learn to reject the distributive reading of plural definite NPs. The results also show that children's acceptance of the collective reading of each is a significant predictor of children's acceptance of the distributive reading of plural definite NPs. Therefore, the results support the hypothesis.

Fastidious distributivity

Proceedings of Semantics and Linguistic Theory 21

I argue that nominal predicates cannot be interpreted distributively if their subject is not a distributive quantifier. This is surprising and problematic for the current theories of distributivity. I offer a novel analysis of distributivity, fastidious distributivity, in which only arguments can be interpreted distributively, and formalize fastidious distributivity using team logic.

A null theory of long-distance reciprocity in English

co-authored with Oystein Nilsen
NELS 39: Proceedings of the 39th Annual Meeting of the North East Linguistic Society

Higginbotham (1981) notices that the sentence John and Mary think they like each other has a reading paraphrasable as John thinks he likes Mary and Mary thinks she likes John. This reading has been subsequently named long-distance reciprocity and it has been noticed that such a reading poses a challenge to the theories in which reciprocals find their antecedent locally, i.e., within the same clause. We present an analysis which (i) builds on the semantics of attitude verbs of Cresswell and von Stechow (1982) and (ii) event semantics of Landman (2000). By allowing cumulation on events and de re attitudes being formed about such events we derive the correct reading without postulating long-distance movement or long-distance binding of reciprocals.


Anaphora and Distributivity. A study of same, different, reciprocals and others

LOT dissertation series


This thesis investigates two issues. It studies the interpretations of sentences with plural arguments and it analyzes anaphoricity triggered by the presence of semantically plural nouns. Regarding the first point, it is standardly assumed that sentences with plural arguments can have many interpretations. This thesis discusses various psycholinguistic experiments, as well as two new questionnaire studies, which show in detail how the preferences among these interpretations depend on the type of noun phrase, including preferences that are commonly not acknowledged in the semantic literature. One goal of this thesis is to offer a semantic and pragmatic account which captures all the interpretations but can also explain why some are preferred over others. The explanation is used for studying the anaphoric expressions 'same', 'different', 'others' and reciprocals. What these expressions share is their ability to be anaphoric sentence-internally if there is a semantically plural noun in the clause. However, the availability of this reading depends on the type of noun phrase used, in a similar way that preferences in the interpretations do. Based on these and other data, the thesis offers a novel formal semantic analysis of reciprocals and argues for a particular semantics of 'same', 'different' and 'others'. It is shown what consequences the analysis has for our understanding of anaphoricity and plurality.

This study is relevant for scholars working on binding, anaphora, and the semantics and pragmatics of pluralities. More generally, it is of interest to scholars in the field of semantics and pragmatics, as well as for linguists working on the syntax-semantics interface.

Distributivity in reciprocal sentences

Logic, Language and Meaning. 17th Amsterdam Colloquium. Springer.

In virtually every semantic account of reciprocity it is assumed that reciprocal sentences are distributive. However, it turns out that the distributivity must be of very local nature since it shows no effect on the predicate or other arguments in reciprocal sentences. I present a semantic analysis of reciprocals that treats reciprocal sentences as distributive but captures the local nature of distributivity.


'The others' compared to 'each other' -- Consequences for the theory of reciprocity

co-authored with Oystein Nilsen
Proceedings of Semantics and Linguistic Theory 18

Two markedly distinct analyses have been developed for reciprocal expression like each other. One very popular approach treats them as anaphoric noun phrases, i.e., noun phrases containing free variables that come with various constraints on their binding configurations (Heim et al., 1991a,b, Schwarzschild 1996, Sternefeld 1998, Beck 2001). The other approach treats them as polyadic quantifiers (Keenan 1992, Dalrymple et al. 1998, Sabato and Winter 2005: a.o.). We argue that both approaches are in fact neccessary but for different expressions. More in particular, we argue that the latter approach is correct for each other, while the former one leads to an elegant account for the properties of the others.


Across-the-Board extraction and clitic omission in Czech

Formal Approaches to Slavic Linguistics 16. Michigan Slavic Publications

I analyze clauses with deleted clitic clusters in Czech, showing that these cases of clitic omission are possible only in coordinations, and therefore, they should be treated as ATB extractions. In the second part of the paper, I focus on cases of ATB extractions in which clitics are phonetically realized inside the first conjunct, arguing that these examples support Nunes's (2005) analysis of sideward movement but are problematic for Citko's (2005) analysis of ATB extractions.


Syntax and semantics of spatial P

co-edited with Anna Asbury, Berit Gehrke and Rick Nouwen
John Benjamins


Quantifier modification

co-edited with Rick Nouwen
ESSLLI Proceedings

Clitic omission in Czech as across-the-Board extraction

Czech in Generative Grammar. LINCOM Studies in Slavic Linguistics

I argue that clauses with deleted clitic clusters present an argument for the application of Boskovic/Franks analysis of clitic placement to Czech data. In such an analysis clitic placement in Czech must be the result of an interplay of syntax and phonology. More concretely, the Czech data can be derived if we assume that several syntactic copies of clitics created by displacement are later filtered out by phonological requirement of clitics to follow the first host in a prosodic constituent.

Why clitics cannot climb out of CP: A discourse approach

Formal Approaches to Slavic Linguistics 15. Michigan Slavic Publications

It is well-known in Slavic linguistic literature that clitics can move out of their clause only if the clause is an infinitival one. I argue that this restriction follows from specific discourse requirements that elements moved out of a finite clause must fulfill. As I show, these requirements clash with the discourse properties of clitics.


Non-local binding in Slavic languages and restructuring

Proceedings of Console XIII

Non-local binding of anaphors is one of the most discussed issues in the theory of anaphora. The classical analysis, starting with Pica (1987), connects the variation in the binding possibilities with the type of the anaphor that enters the binding relation. I argue that Czech and Russian exhibit non-local binding of a different origin. Non-local binding in these languages arises because of variation in the architecture of infinitival clauses. More concretely, it is argued that infinitival clauses may lack PRO, and the absence of PRO makes non-local binding possible. This account is compared to the analysis of restructuring elaborated in Wurmbrand (2001).


Jedno syntakticke omezeni u neprojektivnich konstrukci s klitikami
Korpus jako zdroj dat o cestine. Masarykova Univerzita

The syntax of infinitives in Czech

Master's thesis. University of Tromso

The thesis studies properties of Czech infinitival clauses within the framework of Principles and Parameters. It focuses on the so-called restructuring clausal properties.

  • chapter 4 appeared in Non-local binding in Slavic languages and restructuring
  • parts of chapter 3 were expanded into Why clitics cannot climb out of CP: A discourse approach