This thesis presents a novel pragmatic account of the meaning and use of numerically-quantified expressions. It can readily be seen that quantities can typically be described by many semantically truthful expressions - for instance, if 'more than 12' is true of a quantity, so is 'more than 11', 'more than 10', and so on. It is also intuitively clear that some of these expressions are more suitable than others in a given situation, a preference which is not captured by the semantics but appears to rely upon on wider-ranging considerations of communicative effectiveness. Motivated by these observations, I lay out a set of criteria that are demonstrably relevant to the speaker's choice of utterance in such cases. Observing further that it is typically impossible to satisfy all these criteria with a single utterance, I suggest that the speaker's choice of utterance can be construed as a problem of multiple constraint satisfaction. Using the formalism of Optimality Theory (Prince and Smolensky 1993), I proceed to specify a model of speaker behaviour for this domain of usage. The model I propose can be used to draw predictions both about the speaker's choice of utterance and the hearer's interpretation of utterances. I discuss the relation between these two aspects of the model, showing how constraints on the speaker's choice of utterance are predicted to make pragmatic enrichments available to the hearer. I then consider applications of this idea to specific issues that have been discussed in the literature. Firstly, with respect to superlative quantifiers, I show how this model provides an alternative account to that of Geurts and Nouwen (2007), building upon that offered by Cummins and Katsos (2010), and I present empirical evidence in its favour. Secondly, I show how this model yields the novel prediction that comparative quantifiers give rise to implicatures that are conditioned both by granularity and by prior mention of the numeral, and demonstrate these implicatures empirically. Finally I discuss the predictions that the model makes about the frequency of quantifiers in corpora, and investigate their validity. I conclude that the model presented here proves its worth as a source of hypotheses about speaker and hearer behaviour in the numerical domain. In particular, it serves as a way to integrate insights from distinct domains of enquiry including psycholinguistics, theoretical semantics and numerical cognition. I discuss the claim of this model to psychological plausibility, its relation to existing approaches, and its potential utility when applied to broader domains of language use.

In learning language, children have to acquire not only words and constructions, but also the ability to make inferences about a speaker’s intended meaning. For instance, if in answer to the question, ‘what did you put in the bag?’, the speaker says, ‘I put in a book’, then the hearer infers that the speaker put in only a book, by assuming that the speaker is informative. On a Gricean approach to pragmatics, this implicated meaning – a quantity implicature – involves reasoning about the speaker’s epistemic state. This thesis examines children’s development of implicature understanding. It seeks to address the question of what the relationship is in development between quantity, relevance and manner implicatures; whether word learning by exclusion is a pragmatic forerunner to implicature, or based on a lexical heuristic; and whether reasoning about the speaker’s epistemic state is part of children’s pragmatic competence. This thesis contributes to research in experimental and developmental pragmatics by broadening the focus of investigation to include different types of implicatures, the relationship between them, and the contribution of other aspects of children’s development, including structural language knowledge. It makes the novel comparison of word learning by exclusion with a clearly pragmatic skill – implicatures – and opens an investigation of manner implicatures in development. It also presents new findings suggesting that children’s early competence with quantity implicatures in simple communicative situations belies their ongoing development in more complex ones, particularly where the speaker’s epistemic state is at stake. I present a series of experiments based on a sentence-to-picture-matching task, with children aged 3 to 7 years. In the first study, I identify a developmental trajectory whereby word learning by exclusion inferences emerge first, followed by ad hoc quantity and relevance, and finally scalar quantity inferences, which reflects their increasing complexity in a Gricean model. Then, I explore cognitive and environmental factors that might be associated with children’s pragmatic skills, and show that structural language knowledge – and, associated with it, socioeconomic status – is a main predictor of their implicature understanding. In the second study, I lay out some predictions for the development of manner implicatures, find similar patterns of understanding in children and adults, and highlight the particular challenges of studying manner implicatures experimentally. Finally, I focus on children’s ability to take into account the speaker’s epistemic state in pragmatic inferencing. While adults do not derive a quantity implicature appropriately when the speaker is ignorant, children tend to persist in deriving implicatures regardless of speaker ignorance, suggesting a continuing challenge of integrating contextual with linguistic information in utterance interpretation.

No description available.

Cette thèse s’est proposé d’explorer le rôle de l’intentionnalité et de l’affiliation sociale dans la compréhension des inférences communicatives. Trois des quatre études principales portent sur l’inférence contrastive, qui consiste à déduire d’une expression référentielle telle que le chien sec, l’existence d’un autre objet de même nature (un autre chien) dans le contexte de la communication en cours. Les études que j’ai menées dans ce travail de thèse suivent une réflexion autour de cette inférence particulière, s’inscrivant à ses débuts dans le champ traditionnel de la pragmatique expérimentale et portant une attention croissante au rôle de l’affiliation sociale dans la communication inférentielle.L’Etude 1 propose un paradigme original qui capte de façon précise la capacité des sujets à faire une inférence contrastive, dans un contexte théorique qui fait généralement l’hypothèse d’un traitement par défaut lorsque les conditions contextuelles sont réunies. Une trajectoire développementale a été mise en évidence chez les enfants, et les performances des adultes mettent en évidence le caractère optionnel de cette inférence. L’Etude 2 s’intéresse à l’âge à partir duquel les enfants se montrent capables d’utiliser leurs attentes de rationalité sur les actes référentiels du locuteur pour se poser la question des raisons qui incite celui-ci à utiliser une description plus informative que nécessaire a priori. Les résultats de cette expérience montrent qu’à 5 ans, mais pas à 3 ans, les enfants sont ralentis dans le processus de compréhension d’une instruction lorsqu’elle est sur-informative, suggérant qu’ils sont sensibles à la pertinence d’un choix référentiel.Dans la continuité de cette réflexion sur le rôle fondamental de la prise en compte des intentions d’autrui dans les inférences pragmatiques, la question s’est posée de savoir si la relation entre le sujet et le locuteur pouvait influencer le processus inférentiel lui-même. Une première démarche a été de regarder si une induction de groupe minimal aurait un effet sur la propension à prendre la perspective visuelle de l’autre dans une tâche communicative (Etude exploratoire). Les études suivantes suivent une approche différente pour répondre à cette même question, à partir de l’idée que les inférences pragmatiques jouent un rôle dans l’établissement et le maintien des liens sociaux. L’Etude 3 est une étude en électromyographie et porte sur le cas particulier des blagues. L’affiliation sociale entre le locuteur et le sujet a été manipulée à partir d’une induction politique. Cette étude sur les blagues montre que la compréhension d’une inférence humoristique est mieux valorisée et génère davantage de réactions positives lorsque le locuteur a une pertinence sociale pour le sujet. L’Etude 4 teste avec un paradigme d’eyetracking l’hypothèse d’un effet de l’affiliation sociale sur l’inférence contrastive. Les résultats montrent qu’une induction politique influence la manière dont l’inférence d’un objet contraste est traitée. Précisément, lorsque le locuteur est un membre de l’outgroup, les individus vérifient davantage le contexte épistémique du locuteur pour s’assurer de la présence d’un objet contraste qui justifie l’utilisation d’une expression référentielle sur-informative du point de vue du sujet. L’ensemble de ces études apportent un éclairage nouveau sur l’inférence contrastive, à partir d’une démarche issue des récents développements de la pragmatique expérimentale. Elles contribuent également à caractériser le rôle des mécanismes sociaux impliqués dans la communication inférentielle. / This thesis aimed to explore the role of intentionality and social affiliation in the processing of communicative inferences. Three of the four main studies deal with contrastive inference, which consists in inferring from a referential expression such as the dry dog, the existence of a contrast object of the same kind (e.g. another dog) in the context at hand. The studies that I conducted in this work were developed within the traditional framework of experimental pragmatics, with a growing interest for the role of social affiliation in inferential communication.Study 1 uses an original paradigm that captures in a fine-grained manner subjects’ propensity to draw a contrastive inference, in a theoretical background that generally assumes that its computation is done by default when the context allows it. A developmental trajectory was observed among children, and adults’ performance suggests that drawing a contrastive inference is actually optional. Study 2 was interested in determining when children start using their expectations of optimal relevance when interpreting under- and over-informative instructions. It shows that at the age of five, but not three, children are slowed in responding to a modified instruction in a context where it is over-informative, compared to optimal, suggesting that they are sensitive to the relevance of a referential choice.Building on the idea that intentionality is crucial for pragmatic communication, I asked the question of whether features of the speaker-listener relationship could influence the inferential process itself. A first step was to test the effect of group affiliation on perspective taking abilities (Exploratory study). The next studies follow a different approach, by viewing inferential communication as a way to establish and maintain social affiliation.Study 3 is an electromyography study that deals with the particular case of jokes. Social affiliation between speakers and subjects were manipulated using a political induction. The results show that a humorous inference is better evaluated and triggers more positive reactions when the speaker’s identity is socially relevant to the subject.Study 4 uses an eyetracking procedure to test the hypothesis of an effect of social affiliation on contrastive inference. Manipulating political affiliation proved to affect the processing of a possible contrast object when responding to an over-informative instruction. Specifically, when the speaker was a member of the outgroup, subjects tended to check more on a hidden image that possibly contained a contrast object, in order to ascertain the relevance of the speaker’s referential choice.In sum, these studies shed a new light on contrastive inference, taking advantage of the recent developments in experimental pragmatics. They also allow for a better characterization of the social mechanisms involved in inferential communication.

Zwei Theorien über die Verarbeitung von Metaphern postulieren die Beteiligung unterschiedlicher kognitiven. Die erste, die „Implicit Comparison View“, behauptet, dass Metaphern durch einen Prozess des analogen Denkens verstanden werden (z. B. Gentner et al., 2001; Gentner & Bowdle, 2008). Eine zweite Ansicht, die „Category Inclusion View“, sieht das Verstehen einer Metapher als einen Prozess, bei dem die lexikalische Bedeutung des metaphorischen Vehikels spontan moduliert wird, um eine ad-hoc, zielorientierte Kategorie zu schaffen (z. B. Glücksberg, 2008; Sperber & Wilson, 2008). Obwohl es eine große Anzahl an Experiment gibt, die die Vorhersagen dieser beiden Theorien testen (z. B. Bowdle & Gentner, 2005; Gernsbacher et al., 2001; Jones & Estes, 2005; Jones & Estes, 2006; McGlone & Manfredi, 2001; Wolff & Gentner, 2011) ist es bis jetzt nicht möglich gewesen, das Problem der Metaphernverarbeitung zu lösen zugunsten einer der beiden Theorien. Diese Dissertation versucht genau das zu tun, indem die Verarbeitung von zwei Arten deutscher nicht-nominaler Metaphern untersucht werden: verbale Metaphern und Verb-Objekt-Metaphern. Dies wurde gemacht durch eine Untersuchung der Rolle des Kontexts während der Verarbeitung von nicht-nominalen Metaphern. Dabei wurde auf die Literatur zur situierten und inkrementellen Sprachverarbeitung zurückgegriffen (siehe Huettig et al., 2011; Huettig et al., 2012; Kamide, 2008; Knoeferle & Guerra, 2016). Insgesamt die Ergebnisse von 14 verschiedenen Experimeten als besser zu vereinbaren mit der „Category Inclusion View“ als mit der „Implicit Comparison View“. / Two main sets of theories of metaphor comprehension posit the involvement of different cognitive mechanisms. The first one, the Implicit Comparison View, claims that metaphors are understood through a process of analogical reasoning in which the elements of a metaphoric expression (in the example above my cat, which is known as the ‘topic’ and princess, which is known as the ‘vehicle’) are scanned for relational similarities (e.g. Gentner et al., 2001; Gentner & Bowdle, 2008). A second view, the Category Inclusion View, sees metaphor comprehension as a process in which the lexical meaning of the metaphoric vehicle is spontaneously changed to represent a newly created, goal-oriented category (e.g. Glucksberg, 2008; Sperber & Wilson, 2008). Despite there being a large body of experimental data testing the predictions made by these theories (e.g. Bowdle & Gentner, 2005; Gernsbacher et al., 2001; Jones & Estes, 2005; Jones & Estes, 2006; McGlone & Manfredi, 2001; Wolff & Gentner, 2011), it has not been possible to settle this debate and tip the scale in favor of one or the other view. This dissertation attempts to do just that by examining the processing of two types of German non-nominal metaphors: Verbal metaphors and verb-object metaphors. This was done by investigating the role of context during metaphor comprehension in order to further specify the available theories, and, more generally, by drawing on the literature on situated and incremental language processing (see Huettig et al., 2011; Huettig et al., 2012; Kamide, 2008; Knoeferle & Guerra, 2016, for reviews). Overall the results of 14 experiments are interpreted as being more consistent with the Category Inclusion View than with the Indirect Comparison View.

Devoir et pouvoir, comme les verbes modaux d’autres langues, ont déjà fait l’objet d’une littérature abondante. Le fait qu’ils puissent recevoir des sens différents selon leur contexte d’emploi a particulièrement suscité l’attention des sémanticiens et des pragmaticiens. Cette thèse revient sur la question de la nature des différentes interprétations de devoir et pouvoir, et tente d’établir si leur plurivocité relève de la polysémie ou de la sous-spécification. L’hypothèse polysémique, la plus répandue dans la littérature française, implique qu’au moins le sens radical et le sens épistémique soient inscrits dans la langue et donc représentés en mémoire. Selon l’hypothèse de la sous-spécification, la multiplicité de sens relève de l’enrichissement contextuel d’un unique sens sous-spécifié stocké dans le lexique mental. L’état actuel des recherches, l’examen des différentes interprétations de devoir et pouvoir et de la sous-détermination potentielle de leur sens en contexte, comme l’étude de leur évolution sémantique en diachronie, ne permettent pas de falsifier l’une ou l’autre hypothèse. Des méthodes expérimentales, développées en psycholinguistique et en pragmatique expérimentale, sont donc convoquées. Notamment, l’examen des temps de traitement en lecture, dans une expérience d’eye tracking manipulant le sens et le contexte, suggère une représentation effectivement polysémique pour devoir, mais une représentation monosémique sous-spécifiée pour pouvoir. Devoir et pouvoir sont souvent traités ensemble, les études considérant que l’un constitue, dans son domaine modal, le pendant de l’autre. Nous relevons que le parallèle effectué n’est pas aussi motivé qu’il n’y paraît. / Devoir and pouvoir, as modal verbs in other languages, have already been the subject of extensive literature. The fact that they can convey different meanings depending on the specific context in which they occur is of particular interest to semanticists andpragmaticians. This thesis focuses on the nature of the various interpretations of devoir and pouvoir and attempts to ascertain whether their meaning multiplicity is a result of their polysemy or of their underspecified semantics. The polysemy hypothesis, which is the prevalent view in the French literature, implies that at least both the root sense and the epistemic sense fully belong to the linguistic system and hence that both are represented in memory. On the contrary, according to the underspecification model, contextual enrichment of a unique underspecified meaning stored in the mental lexicon accounts for meaning multiplicity. The current state of research, the review of the several possible interpretations of devoir and pouvoir, the investigation of potential meaning underdetermination in context, as well as the study of their semantic evolution in diachrony, do not allow us to rule out any of the two hypotheses. Experimental methods, developed in psycholinguistics and in experimental pragmatics, are thus used. Notably, analysis of processing times in reading in an eye tracking experiment in which both meaning and context are manipulated favours a polysemic representation for devoir, but a monosemic and underspecified representation for pouvoir. The two modal verbs are traditionally examined together since it is assumed that one matches the other in its own modal domain. This thesis casts doubt on this assumption.