Patrick Caudal

1 Panyjima aspectual classes: new perspectives on formal models for event structure PatrickCaudal (LLF, CNRS/Université Paris 7) pcaudal@linguist.jussieu.fr Alan Dench & Marie-Eve Ritz (The University of Western Australia) {alan.dench; marie-eve.ritz}@uwa.au.edu Page 2. ...

ABSTRACT This paper proposes a formal treatment of collective nouns (e.g. police, orchestra, and forest) drawing on the rich lexical typing of the Generative Lexicon. Collectives can be alternatively interpreted as collections (in the sense of Link, 1983 [9]) or as individuals, depending on contextual factors. They are therefore polysemous. Their semantics includes a meronymic (part/whole) relation which may or may not be accessible, depending on which interpretation they receive. The polysemy of collectives will be modelled using dot objects nominal types (dot objects being the Cartesian product of simple lexical types within a hierarchical lexicon) and qualia roles, capturing the constraints observed on their meronymic structure and complex lexical typing. 1. Introduction Ideally, any semantic theory or NLP system should rule out (1) while allowing (2) : (1) ?John finished arriving/leaving. (2) The regiment finished arriving/leaving. Intuitively, the contrast in (1)/(2) stems from the fa...

La presente these a pour objectif de formuler une theorie des effets de polysemie aspectuelle en adoptant une demarche contrastive francais/anglais, appliquee a un corpus de donnees bilingue. Elle soutient que les interpretations aspectuelles d'un enonce sont constituees a partir de deux grands types d'information aspectuelle, qui sont la structure situationnelle (portee par un verbe et sa complementation) et les points de vue aspectuels (portes par un temps verbal). Ces derniers expriment la perspective selon laquelle l'enonciateur nous presente une situation (c'est-a-dire un proces), qui est vue comme une figure se detachant plus ou moins et selon des modalites variables sur le fond du temps percu. Ceci suppose une operation de focalisation, qui consiste a valider une zone particuliere d'une structure situationnelle (i. E. Une phase), parfois en combinaison avec une operation de coercition, alterant une situation. L'apport principal de la these est une theorie de la focalisation et de la coercition basee sur une conception riche de la structure situationnelle. A chaque situation est associe un decoupage linguistiquement motive en phases, ainsi qu'une hierarchie de saillance de ces phases, expliquant la possibilite d'une focalisation sur une phase pour un temps et eventuellement des modifieurs de gv donnes, ceux-ci apportant des contraintes sur la focalisation. Le decoupage avance a ete particulierement affine selon un angle contrastif pour la phase resultante d'une situation (i. E. Son etat resultant). Il a ete egalement propose d'enrichir la description de la structure situationnelle au moyen d'echelles de degres, qui permettent d'offrir une representation complexe des changements d'etat, et de rendre compte des effets de sens aspectuels des modifieurs de degre tels que presque, a peine, completement, a peine, etc. Enfin, une formalisation de ces resultats theoriques a ete ebauchee dans le cadre de la drt (discourse representation theory).

this paper. It will be established that those two languages exhibit substantial differences in the way they address resultative meanings, respectively through the pass compos (PC, henceforth) and the perfect (PF, henceforth). A notion abstracting away from the ontological and purely mereological view on situation structures, phasal structure, will be introduced. The contrast between French and English resultativity will be shown to arise from a difference between the semantics of the PC and that of the PF, rather than in the way French and English encode phasal structures. Arguments for positing similar phasal structures in the two languages will be discussed. Special attention will be paid to so-called scalar verbs (e.g., cook).

Introduction v Currently dominant neo-davidsonian view on aspect calculus has driven many to study event structure in terms of mereological, part-of relationships ; it will be argued here that this approach is not appropriate for theoretical & empirical reasons (cf. Asher (1993) & Verkuyl (1993) for related positions, on philosophical grounds); v An alternative approach based on discourse-like relations will be proposed. Main features : Explicit relations between situations, sub-situations (or situation stages) and aspectual viewpoints ; Empirical evidence in a variety of languages suggesting indeed that event types / event structure (= situation structure / eventuality structure) should be regarded as a language-specific issue. 2 Preliminary definitions : some basic aspectual concepts 2.1 Defining situation structure and lexical aspect

This paper investigates the diachronic evolution of so-called aspectual coercion (de Swart 1998, Bary & Egg 2012) in French, with respect to two major tenses, namely the simple past (passé simple) (PS) and compound past (passé composé) (PC); it will more specifically bear on cases of inchoative readings. Throughout a study spanning several diachronic stages and capitalizing on earlier work (Caudal 2015a, Caudal 2015b, Caudal, Burnertt & Troberg 2016), it will be shown that the PC and the PS exhibit striking differences in their acquisition of inchoative coercions, with the PC consistently lagging behind the PS in some respects. Initially, at the Old French period, the PC was totally deprived of any coercive power w.r.t. states, whereas the PS already had a broader and better established inchoative coercive capability. But across subsequent stages of the language, the PS gradually increased its inchoative potential at a steady pace-although it seems to retain some difficulties with certain types of stative utterances, especially those denoting individual-level states, and locative/posture structures. While the PC has often been claimed to have largely replaced the PS, I will here show that even in Modern French, the PC seems to still have a noticeably lesser 'inchoativizing power' than the PS. In order words, in spite of nearly a thousand years of parallel evolution and semantic convergence, the initial semantic gap between the two forms still hasn't been bridged. I will suggest that these consistent differences should lead us to consider so-called inchoative coercion as a distinctly conventionalized type of meaning expansion mechanism-rather than a simple matter of overcoming the violation of some aspectual semantic restriction.

Le présent article s'efforcera d'esquisser un programme de recherche novateur sur le conditionnel, et d'avancer quelques arguments empiriques assez détaillés pour étayer l'hypothèse avancée. On peut la qualifier de constructionnelle (au sens très large du terme, et pas simplement de (Goldberg 1995)). L'idée principale qu'il défend est que la diversité des « emplois » du conditionnel ne saurait s'analyser ni seulement à l'aide d'un invariant sémantique (monosémie), ni uniquement par des interprétations en relation de polysémie productive vs. d'homonymie, et qu'elle implique parfois des appariements formes/sens conventionnalisés au niveau de constructions syntaxiquement complexes. Abstract This paper will put forth an innovative research program on the French conditionnel, backing it with some relatively detailed empirical arguments. The research hypothesis here advanced can be described as constructionalist in the broadest sense (and not simply in the sense of (Goldberg 1995)). The main idea defended here is that the diversity of the so-called 'uses' (emplois) of the conditional cannot be accounted for by means of a single semantic invariant (monosemy), nor purely through productive polysemous or homonymous relations between various readings. Rather, it sometimes implies conventionnalized forms / meanings pairings at the level of syntactically complex constructions.

This paper aims at revisiting a fairly well-studied issue, namely the so-called 'attenuative' or 'polite' uses of the French imparfait. Departing from a currently predominant line of analysis in the semantics & pragmatics community, it will argue against a fully productive approach to said uses. It will, on the contrary, establish its conventionnalised (yet not completely entrenched) nature, on two grounds : (i) on interpretative grounds, because the corresponding meanings cannot be straightforwardly accounted for on the basis of 'productive' semantic/ pragmatic processes ; (ii) on form-related grounds, by establishing that the so-called imparfait atténuatif in fact points to a constructional network rather than a bona fide productive form/meaning pairing, cf. (Traugott & Trousdale 2013; Timponi Torrent 2015). Finally, the paper will replace the imparfait atténuatif within the larger, and cross-linguistic context of other conventionnalised TAM constructions exhibiting composite tense-aspect/modality marking without a fully productive (e.g. compositional) interpretation, thereby offering a general plea against aprioristic productive approaches to such composite TAM structures.

To appear as : Caudal, Patrick. 2017. Les « usages atténuatifs » de l’imparfait : vers une analyse constructionnelle. In Eta Hrubaru & Estelle Moline (eds.), Nouveaux regards sur le sens et la référence. Hommages à Georges Kleiber,  46 p. Bern: Peter Lang.

Les abondants travaux consacrés à l'imparfait n'ont pas encore, à notre connaissance, conjoint une dimension à la fois diachronique et formelle ; dans cet article, nous allons justement entreprendre d'appuyer une démarche diachronique sur un traitement formel. Nous montrerons tout d'abord (section 1) quelle a été l'évolution suivie par l'imparfait depuis son usage en ancien français, et qui est principalement marquée par le passage d'un temps à valeur de pur « imperfectif » ou « sécant » (c'est-à-dire décrivant une situation en cours), à un temps admettant des valeurs dites « narratives » ou « perfectives ». On verra que cette évolution doit être corrélée à celle du passé simple, qui était en ancien français une manière de prétérit (un peu à la manière du simple past anglais et des temps équivalents dans les langues germaniques) oscillant entre des emplois de type « en cours » (ou sécants), et des emplois où la situation décrite est au contraire vue dans sa globalité (cf. Guillaume 1929 pour l'opposition global / sécant) – emploi global dans lequel il est aujourd'hui confiné. Nous proposerons ensuite (section 2) un traitement formel des différents emplois de l'imparfait, qui permettra d'expliquer son évolution diachronique grâce au modèle ainsi produit, et de montrer comment il a pu annexer des emplois autrefois réservés au passé simple tout en restant distinct de ce temps à ce jour – en bref, en quoi il n'est pas encore un prétérit 1. Nous soulignerons en particulier que les usages dits narratifs de l'imparfait sont moins le fait du contenu propre de ce temps, qu'un effet de sens provoqué par des contextes narratifs, et avec lequel l'imparfait a cessé d'être incompatible au fil du temps (ce qui le distingue de temps tels que le past progressive anglais). 1 Nous employons le terme « prétérit » dans le sens de « forme aspectuelle ambivalente (sous-spécifiée), ayant des emplois perfectifs et imperfectifs ».

In Chapter 9, “The passé composé in Old French and Modern French: evolution or revolution?,” Patrick Caudal has recourse to both sentence grammar and the discourse relations defined in the SDRT model in order to solve a mystery concerning the diachronic evolution of the French passé composé (PC) (present perfect). The mystery lies in discovering the stages by which a form which originally depicted a simple present result state, as in Latin habeo litteras scriptas (I have letters written) and Old French (OF) desarmé sont (they are disarmed)
turned into the dynamic perfective form of Modern French j’ai écrit une lettre hier (I wrote a letter yesterday).
Against the claim of earlier researchers that the perfective point of view was introduced as late as the seventeenth century, Caudal and Roussarie (2006) and Caudal (this volume) argue that perfectivity was, on the contrary, an early component of the PC. It was at first a weak kind of perfectivity, however, which did not
accept a past time adverb like hier. By examining two lengthy medieval texts, Caudal shows how the PC structure evolved from depicting a simple result state to denoting an inchoative result state within a Sequence of Events (SOE) context in correlative constructions as early as twelfth- and thirteenth-century Old French (cf. (36)).

(36) Tant li prièrent li meillur Sarrazin/ Qu’el’ faldesteol s’est Marsilies asis. (Song of Roland, 12th c.)
(The best Saracens begged him so much that upon his throne Marsile sat
down)

Caudal shows that in this early period, the SOE function of the PC was essentially identical to the SOE function of Present Tense in narrative texts: it could occur in Sequence of Event passages, but did not yet allow a punctual past time adverb or date, unlike the modern French PC (“Il s’est assis sur la chaise hier/à 15h”—he sat down on the chair yesterday/ at 3 pm).
While the inchoative resultative was only weakly perfective in an embedded
position in a correlative structure, it was more strongly perfective in matrix position, as in (37).

(37) Quant la reine voit le roi/ . . . Si s’est contre le roi dreciee. (Chevalier de la Charette, 13th c.)
(When the queen saw the king . . . she stood up against the king)
Still, it was not yet possible to add “hyer” (yesterday) or a precise date to such sentences.

Caudal’s solution with respect to the final step in the evolution of the French
present perfect towards strong perfectivity combines a “conservative semantics” with an “innovative pragmatics” set within the SDRT. The conservative semantics takes both the present auxiliary and the past participle as being active in the interpretation of the verbal form. The innovative pragmatics refers to discourse functions within an SDRT framework which pragmatically enrich the interpretation of the structure from weak resultative perfectivity to a strong perfective interpretation. In time, this pragmatic mechanism was instrumental in the semanticization of perfective interpretations—with semantic perfectivity rendering the passé composé compatible
with definite past time adverbs. Caudal proposes that given an unstable two-part verbal form, discourse functions may determine, and in the case of the French PC did determine, whether the matrix present tense or the result-denoting past participle is focused and consequently grammaticalized within a particular discourse context.

This paper proposes a formal treatment of collective nouns (eg police, orchestra, and forest) drawing on the rich lexical typing of the Generative Lexicon. Collectives can be alternatively interpreted as collections (in the sense of Link, 1983 [9]) or as individuals, depending on ...

Les abondants travaux consacrés à l'imparfait n'ont pas encore, à notre connaissance, conjoint une dimension à la fois diachronique et formelle ; dans cet article, nous allons justement entreprendre d'appuyer une démarche diachronique sur un traitement formel. Nous montrerons tout d'abord (section 1) quelle a été l'évolution suivie par l'imparfait depuis son usage en ancien français, et qui est principalement marquée par le passage d'un temps à valeur de pur « imperfectif » ou « sécant » (c'est-à-dire décrivant une situation en cours), à un temps admettant des valeurs dites « narratives » ou « perfectives ». On verra que cette évolution doit être corrélée à celle du passé simple, qui était en ancien français une manière de prétérit (un peu à la manière du simple past anglais et des temps équivalents dans les langues germaniques) oscillant entre des emplois de type « en cours » (ou sécants), et des emplois où la situation décrite est au contraire vue dans sa globalité (cf. Guillaume 1929 pour l'opposition global / sécant) – emploi global dans lequel il est aujourd'hui confiné. Nous proposerons ensuite (section 2) un traitement formel des différents emplois de l'imparfait, qui permettra d'expliquer son évolution diachronique grâce au modèle ainsi produit, et de montrer comment il a pu annexer des emplois autrefois réservés au passé simple tout en restant distinct de ce temps à ce jour – en bref, en quoi il n'est pas encore un prétérit 1. Nous soulignerons en particulier que les usages dits narratifs de l'imparfait sont moins le fait du contenu propre de ce temps, qu'un effet de sens provoqué par des contextes narratifs, et avec lequel l'imparfait a cessé d'être incompatible au fil du temps (ce qui le distingue de temps tels que le past progressive anglais). 1 Nous employons le terme « prétérit » dans le sens de « forme aspectuelle ambivalente (sous-spécifiée), ayant des emplois perfectifs et imperfectifs ».

Degree Scales and Aspect Patrick CAUDAL CNRS, UMR 7110 and Universite Paris 7 1. Introduction Treating situation'structure in terms of scalarity is an already old idea underlying many a work about telicity, cf. the notions of odometer in Verkuyl (1993), path in Jackendoff (1996) or ...

Achievements and accomplishments are argued in this paper to differ wrt atomicity (rather than punctuality), a notion strongly but not exclusively related to incrementality, ie, to event-object mapping functions ; the latter will be shown to be unsufficient to account for certain cases ...

Publication View. 42544764. Proceedings of EACL '99 Result Stages and the Lexicon: The Proper Treatment of Event Structure (2007). Patrick Caudal. Abstract. I will argue in this paper that the standard. Publication details. Download, ...

(i) the distribution of tenses with morpho-syntactic units in general, but more particularly with so-called aspectuo-temporal modifiers (eg, for phrases, past temporal expressions…); this corresponds to the compositional semantics of tenses, since semantics and morpho-syntax govern ...

The purpose of this paper is to establish that so-called 'uses' of verbal inflections such as the French imparfait can articulate a conventionalized, 'entrenched' dimension, alongside with a productive, context-sensitive dimension at the semantics/pragmatics interfacei.e., that one type of modelling mechanism (entrenchment vs. dynamic context-sensitivity) does not preclude the other. In order to reveal the complexity of the matter at stake, I will here focus on two relatively well-known uses of the imparfait, namely the so-called attenuative imparfait (or 'politeness' imparfait), which associates with utterances conveying polite, negotiable requests, and the so-called 'narrative imparfait', which associates with sequence-of-events narrative discourses. The term 'construction' highlights a fundamental hypothesis explored in these pages, namely that tense uses are often conventionalized in one way or another. I will use it here in two distinct senses: in the (strong) sense of a special set of lexicalized syntactic constructs in the imparfait paired up with a conventionalized semantic content non-compositionally derivable from the elements making up the construct, and in the (weaker) sense of a conventionalized meaning attached to the lexical entry of a verbal inflection. More specifically, I will analyze the so-called 'attenuative imparfait' as an instance of strong construction, and the 'narrative imparfait' as an instance of weak construction, i.e., as a matter of semantic homonymy-the main contribution of this paper being that in spite of their constructional nature, these 'tense uses' are nevertheless endowed with significant context-sensitivity.

The purpose of this paper is to establish that so-called 'uses' of verbal infections such as the French imparfait can articulate a constructional, 'entrenched' dimension, with a context-sensitive, semantic/pragmatic enrichment dimension-i.e., that one type of modelling mechanism (lexical entrenchment vs. dynamic, context-sensitive semantic/pragmatic enrichment). In order to reveal the complexity of the matter at stake, I will here focus on two relatively well-known uses of the imparfait, namely the so-called 'attenuative imparfait' (also known as imparfait de politesse), which associates with utterance conveying polite requests, and the so-called 'narrative imparfait', which associates with sequence-of-events narrative discourses. Until recently (cf. e.g. (Caudal 2017; Patard 2017; Caudal 2018a)), most existing accounts of so-called 'tense uses' put the stress on various kinds of productive, online, semantic and/or pragmatic strategies, to contextually adapt and/or enrich some underspecified 'core' meaning. This was achieved by various mechanisms (semantic composition, discourse structural parameters such as discourse relations or other contextual parameters in general, either purely pragmatic or at the semantics/pragmatics interface), the nature of which do not matter here. However, concerns were soon voiced that this might not a suitable explanation for at least some so-called tense uses. This was notably the case for both the 'attenuative' and 'narrative' uses of the imparfait: thus (Anscombre 2004; Abouda 2004) were the first to observe that the attenuative imparfait looked suspiciously close to a lexified, syntactically entrenched construction. In a similar vein, (Tasmowski-De Rijck 1985), highlighted the syntactic constraints licensing the appearance of so-called 'narrative imparfait' structures-thus suggesting they were conventionalized to some extent. This gave rise to a (still limited) number of novel constructional analyses of some tense uses, and resulted in de facto opposition between 'uses' modelled as being lexicalized constructions (and amenable to a static semantics, in formal terms; cf. e.g. (Patard 2017; Caudal 2018a)), whereas non-constructional uses remain treated as context-sensitive (i.e., non-amenable to a static semantics; they rather required a dynamic semantics/pragmatics approach, conceiving meaning in terms of context update). In this paper, I will question whether or not it is legitimate to view constructionalized 'tense uses' as falling squarely within the realm of static semantics-i.e., whether the above de facto dichotomy has a theoretical basis. Although the present analysis will argue that 'tense uses' are generally the byproduct of some kind of conventionalization process, I will try and demonstrate here that it does not necessarily require all connections to be severed with context-sensitivity qua dynamic semantic and/or pragmatic mechanisms contributing to the interpretation of said 'uses'. By studying the attenuative and narrative uses of the imparfait in turn, I will compare the manner in which context sensitivity plays a different part in each case. I will first (§2) show that although a lexified multidimensional semantics à la (Potts 2005; Gutzmann & McCready 2016) is required to model 'attenuative uses' of the imparfait, a dynamic pragmatic account of the notion of attenuation (qua attenuated directives) based on (Portner 2018)'s theory of commitment management in dialogue is also required to explain their contextual, 'polite request' effects. In contrast to this, discourse structural patterns (i.e. so-called rhetorical relations) will be shown in §3 to play a key role in the emergence of so-called 'narrative imparfait' sequences-even though these also require the presence of some manner of support syntactic markers (in effect, a construction network), as we will see. These different modes of contextualization, I will argue, suggests that the study of so-called tense uses calls for far more diverse and complex approaches at the morpho-syntax to semantics/pragmatics interface than hitherto assumed in the literature.