Search

The place of sociopragmatics in construction grammars : review and prospects. Under the influence of cognitive linguistics, construction grammars make no principled distinction between semantics and pragmatics. Most construction grammars claim that meaning does not derive from abstract principles and cannot be apprehended outside a specific context. In fact, meaning is part of a network of representations that are both culturally situated and structured by experience. In construction grammars, semantics is thus very close to traditional pragmatics. Even if construction-grammar supporters have good reasons to reject the semantics/pragmatics distinction, they have done so at the expense of pragmatics. This is problematic because pragmatics has a lot to offer when it comes to explaining the use of constructions in terms of context, interaction, and communities of practice. Most construction grammars are only too willing to explain sociopragmatic phenomena by means of semantics, with the result that key constructional features are lost. This may be because in many construction-grammar architectures, the pragmatic level is too narrow., En vertu d’un postulat hérité de la linguistique cognitive, les grammaires de constructions n’opèrent pas de distinction de principe entre sémantique et pragmatique. En théorie, cette non-séparation est justifiée. En effet, dans les grammaires de constructions, le sens ne dérive pas de principes abstraits et n’a pas vraiment de pertinence lorsqu’il est appréhendé hors-contexte. Il s’inscrit en fait dans un réseau de représentations culturellement situées et structurées par l’expérience. La sémantique constructionnelle est en cela très proche de la pragmatique traditionnelle. Mais dans les faits, la non-distinction entre sémantique et pragmatique s’est faite au détriment de la seconde. En regard du sens, l’information pragmatique a ceci de plus qu’elle est en prise avec des facteurs contextuels, liés notamment à l’interaction des locuteurs au sein de communautés de pratiques. Trop rapidement recouverte par des considérations sémantiques, la pragmatique dépasse du cadre finalement étroit que les grammaires de constructions lui ont trop rapidement assigné., Desagulier Guillaume. Le programme socio-pragmatique des grammaires de constructions : bilan et perspectives. In: Intellectica. Revue de l'Association pour la Recherche Cognitive, n°56, 2011/2. Linguistique cognitive : une exploration critique. pp. 99-123.

Cet article fait le point sur le statut de la frequence dans les grammaires de constructions sous l’angle de l’ancrage cognitif et des collocations. Je montre que le traitement intuitif des phenomenes de frequences au detriment de l’empirie mene souvent a une vision faussee de l’usage. L’article est structure comme suit. La premiere partie est epistemologique. J’y aborde le statut de la frequence dans le contexte du tournant quantitatif dans les approches centrees sur l’usage. La deuxieme partie est consacree a l’etude de A as GN dans un corpus d’anglais americain. J’utilise trois outils : l’analyse collexemique covariante, la classification ascendante hierarchique et ΔP, une mesure d’association permettant de reperer les collocations asymetriques. La troisieme partie propose une discussion critique des resultats et des methodes.

International audience; In this paper, I develop a composite methodology for handling and visualizing measures of significant attraction between lexical items in a set of near-synonyms. I aim to uncover revealing aspects of the conceptual structure of four English moderators (rather, quite, fairly, and pretty) and shed new light on previous studies on degree modifiers. This methodology combines univariate and multivariate statistics. Collexemes are used as input for hierarchical agglomerative cluster analysis, and multiple distinctive collexemes are used as input for correspondence analysis.

International audience; In this paper, I develop a composite methodology for handling and visualizing measures of significant attraction between lexical items in a set of near-synonyms. I aim to uncover revealing aspects of the conceptual structure of four English moderators (rather, quite, fairly, and pretty) and shed new light on previous studies on degree modifiers. This methodology combines univariate and multivariate statistics. Collexemes are used as input for hierarchical agglomerative cluster analysis, and multiple distinctive collexemes are used as input for correspondence analysis.

article à paraître dans M. Bouveret et D. Legallois (eds), Constructions in French. John Benjamins Publishing.; In this paper, I propose a Construction Grammar approach to the count/mass distinction in French. Rather than confine my analysis to NPs, I examine the effects of massive conversion in the broader context of two partially-filled idiomatic constructions: the CCDN construction (ça c'est de la voiture ! 'that's some car') and the CDN construction (cette voiture, c'est de la bombe ! 'that car rocks!'). Both inherit properties from the Copular Subject Predicate construction (c'est une voiture 'that's a car"), except their nominal predicates undergo count-to-mass conversion. Generally, count-to-mass conversion has a quantitative function: it turns NPs whose referents are numerically quantifiable into an NP whose referents cannot be quantified as separate entities. In the CCDN and the CDN, count- to-mass conversion has a qualitative function: it predicates a quality of the subject by identifying this subject with the prototype of the category denoted by the nominal predicate. I show that the CCDN and the CDN belong to the same constructional network, even if they differ as to the kinds of identification that they realize.

I investigate the use of two English intensifiers: quite and rather. When quite and rather modify attributive adjectives, they can occur in pre-determiner position, an idiosyncratic behavior that other intensifiers do not show: (1) I know it's a fairly / *fairly a difficult question. (2) That's proved to be a quite / quite a difficult question to answer. (3) That is a rather difficult / rather a difficult question to answer. Allerton (1987) observes that, depending on whether the adjective that quite modifies is scalar or absolutive, some restrictions apply, a sign that pre-determiner position is more than just a matter of style or formality: (4) I mean this is quite a good idea / ??a quite good idea actually. (4') This is ??quite an excellent idea / a quite excellent idea. The question that naturally arises is whether there is any difference in meaning between the pre-determiner and pre-adjectival positions of quite and rather. Another question is whether these two intensifiers are synonyms. Previous works on English intensifiers rely almost exclusively on the concept of gradability to explain the relationship between intensifiers and adjectives (Quirk et al. 1985, Allerton 1987, and Paradis 1997). However, several studies conclude that context dependency is not always decisive to decide how an adverb modifies an adjective (Paradis 1997, Nevalainen & Rissanen 2002, Athanasiadou 2007). Corpus-based approaches address this problem (Altenberg 1991, Kennedy 2003, Lorenz 2002, Simon-Vandenbergen 2008) but run up against methodological issues. Indeed, they tend to use measures such as percentages, counts per n-thousand words, or pointwise MI which are inadequate to reveal two-way interactions between collocants (Altenberg 1991, Paradis, 1997, Kennedy 2003). My working hypothesis is that quite and rather have a semantic component paired with a syntactic component over and above their specification of degree. To test this hypothesis, I propose an original method that combines analytical and multivariate statistics. First, I extract all combinations from the 100M-word British National Corpus (World Edition). Then, I implement a technique known as multiple distinctive collexeme analysis (Gries & Stefanowitsch 2004) to determine which adjectives are most distinctively attracted to each intensifier depending on the syntactic construction. Finally, I use the frequencies of distinctive adjectives as input for correspondence analysis (Benzécri 1984, Greenacre 2007), a multifactorial approach that provides a low-dimensional map of the data by calculating matrices between the rows and the columns of a contingency table using the χ2 test. My results show that: i. adjectives cluster differently depending on (a) the intensifier that modifies them, (b) the syntax of the intensifying construction where they occur; ii. quite constructions and rather constructions cluster differently depending on their syntactic profiles (pre-determiner position vs. pre-adjectival position; intensifier + attributive adjective vs. intensifier + predicative adjective); iii. quite and rather attract semantically distinct adjective classes and are not, as expected, exact synonyms. Allerton, D.J. 1987. English Intensifiers and their idiosyncrasies. In Steele, R. & T. Threadgold (eds), Language Topics: Essays in Honour of Michael Halliday, vol. 2. Amsterdam: John Benjamins, 15-31. Altenberg, B. 1991. Amplifier collocations in spoken English. In S. Johansson, & A. Stenström (eds.), English Computer Corpora: Selected papers and research guide, 127-149. Berlin: Mouton de Gruyter. Athanasiadou, A. 2007. On the subjectivity of intensifiers. Language Sciences 29, 554-565. Benzécri, J-P. 1984. Analyse des Correspondances, Exposé Elémentaire (2nd ed.). Paris: Dunod. Greenacre, M.J. 2007. Correspondence Analysis in Practice (2nd ed.). Boca Raton: Chapman & Hall/CRC. Gries, S. Th. & A. Stefanowitsch 2004. Extending collostructional analysis: a corpus-based perspective on 'alternations'. International Journal of Corpus Linguistics 9, 97-129. Kennedy, G. 2003. Amplifier collocations in the British National Corpus: Implications for English Language Teaching. TESOL Quarterly 37, 467-477. Lorenz, G. 2002. Really worthwile or not really significant? A corpus-based approach to the delexicalization and grammaticalization of intensifiers in Modern English. In: Wischer, I., Diewald, G. (Eds), Speech, Place, and Action: Studies in Deixis and Related Topics. New York: Wiley, pp. 143-161. Nevalainen, T., & M. Rissanen 2002. Fairly pretty or pretty fair? On the development and grammaticalization of English downtoners. Language Sciences 24, 359-380. Paradis, C. 1997. Degree Modifiers of Adjectives in Spoken British English. Lund University Press. Quirk, R., Greenbaum, S., Leech, G., Svartvik, J. 1985. A Comprehensive Grammar of the English Language. Longman. Simon-Vandenbergen, A.M. 2008. Almost certainly and most definitely: Degree modifiers and epistemic stance. Journal of Pragmatics 40, 1521-1542.

I investigate the use of two English intensifiers: quite and rather. When quite and rather modify attributive adjectives, they can occur in pre-determiner position, an idiosyncratic behavior that other intensifiers do not show: (1) I know it's a fairly / *fairly a difficult question. (2) That's proved to be a quite / quite a difficult question to answer. (3) That is a rather difficult / rather a difficult question to answer. Allerton (1987) observes that, depending on whether the adjective that quite modifies is scalar or absolutive, some restrictions apply, a sign that pre-determiner position is more than just a matter of style or formality: (4) I mean this is quite a good idea / ??a quite good idea actually. (4') This is ??quite an excellent idea / a quite excellent idea. The question that naturally arises is whether there is any difference in meaning between the pre-determiner and pre-adjectival positions of quite and rather. Another question is whether these two intensifiers are synonyms. Previous works on English intensifiers rely almost exclusively on the concept of gradability to explain the relationship between intensifiers and adjectives (Quirk et al. 1985, Allerton 1987, and Paradis 1997). However, several studies conclude that context dependency is not always decisive to decide how an adverb modifies an adjective (Paradis 1997, Nevalainen & Rissanen 2002, Athanasiadou 2007). Corpus-based approaches address this problem (Altenberg 1991, Kennedy 2003, Lorenz 2002, Simon-Vandenbergen 2008) but run up against methodological issues. Indeed, they tend to use measures such as percentages, counts per n-thousand words, or pointwise MI which are inadequate to reveal two-way interactions between collocants (Altenberg 1991, Paradis, 1997, Kennedy 2003). My working hypothesis is that quite and rather have a semantic component paired with a syntactic component over and above their specification of degree. To test this hypothesis, I propose an original method that combines analytical and multivariate statistics. First, I extract all combinations from the 100M-word British National Corpus (World Edition). Then, I implement a technique known as multiple distinctive collexeme analysis (Gries & Stefanowitsch 2004) to determine which adjectives are most distinctively attracted to each intensifier depending on the syntactic construction. Finally, I use the frequencies of distinctive adjectives as input for correspondence analysis (Benzécri 1984, Greenacre 2007), a multifactorial approach that provides a low-dimensional map of the data by calculating matrices between the rows and the columns of a contingency table using the χ2 test. My results show that: i. adjectives cluster differently depending on (a) the intensifier that modifies them, (b) the syntax of the intensifying construction where they occur; ii. quite constructions and rather constructions cluster differently depending on their syntactic profiles (pre-determiner position vs. pre-adjectival position; intensifier + attributive adjective vs. intensifier + predicative adjective); iii. quite and rather attract semantically distinct adjective classes and are not, as expected, exact synonyms. Allerton, D.J. 1987. English Intensifiers and their idiosyncrasies. In Steele, R. & T. Threadgold (eds), Language Topics: Essays in Honour of Michael Halliday, vol. 2. Amsterdam: John Benjamins, 15-31. Altenberg, B. 1991. Amplifier collocations in spoken English. In S. Johansson, & A. Stenström (eds.), English Computer Corpora: Selected papers and research guide, 127-149. Berlin: Mouton de Gruyter. Athanasiadou, A. 2007. On the subjectivity of intensifiers. Language Sciences 29, 554-565. Benzécri, J-P. 1984. Analyse des Correspondances, Exposé Elémentaire (2nd ed.). Paris: Dunod. Greenacre, M.J. 2007. Correspondence Analysis in Practice (2nd ed.). Boca Raton: Chapman & Hall/CRC. Gries, S. Th. & A. Stefanowitsch 2004. Extending collostructional analysis: a corpus-based perspective on 'alternations'. International Journal of Corpus Linguistics 9, 97-129. Kennedy, G. 2003. Amplifier collocations in the British National Corpus: Implications for English Language Teaching. TESOL Quarterly 37, 467-477. Lorenz, G. 2002. Really worthwile or not really significant? A corpus-based approach to the delexicalization and grammaticalization of intensifiers in Modern English. In: Wischer, I., Diewald, G. (Eds), Speech, Place, and Action: Studies in Deixis and Related Topics. New York: Wiley, pp. 143-161. Nevalainen, T., & M. Rissanen 2002. Fairly pretty or pretty fair? On the development and grammaticalization of English downtoners. Language Sciences 24, 359-380. Paradis, C. 1997. Degree Modifiers of Adjectives in Spoken British English. Lund University Press. Quirk, R., Greenbaum, S., Leech, G., Svartvik, J. 1985. A Comprehensive Grammar of the English Language. Longman. Simon-Vandenbergen, A.M. 2008. Almost certainly and most definitely: Degree modifiers and epistemic stance. Journal of Pragmatics 40, 1521-1542.

The development of nominal constructions including a categorical liaison in prereaders has been studied using usage-based theories (Chevrot et al., 2009). In this article, we investigate the development of the variable liaison, which is a sociolinguistic marker in French (for example, its frequency in adult usage depends on the speaker’s socio-economic status). One hundred and eighty-five children aged between 2 and 6, belonging to two distinct SES groups (higher- and lower-SES) participated in a picture naming task eliciting the production of variable liaisons in "adjective + noun" context.Our results show that social differences appear progressively during development. For 5-6 year olds, the social stratification of variable liaison use, observed in adults, appears in child productions.In line with the principles of usage-based theories, we extend the developmental model of the acquisition of nominal constructions in categorical liaisons to the context of variable liaisons., Le développement des constructions nominales en contexte de liaison catégorique chez l’enfant pré-lecteur a fait l’objet d’une modélisation dans le cadre des théories basées sur l’usage (Chevrot et al., 2009). Dans cet article, nous examinerons le développement de la liaison variable, marqueur sociolinguistique du français, dont la fréquence de réalisation, chez l’adulte, est notamment liée à l’origine sociale.Cent quatre-vingt cinq enfants, issus de milieux sociaux contrastés (des enfants de cadres et des enfants d’ouvriers) et âgés de 2 à 6 ans, ont passé une tâche expérimentale de dénomination d’images induisant la production de liaisons variables dans le contexte "adjectif + nom".Nos résultats montrent que les différences sociales se créent progressivement au cours du développement et qu’à 5-6 ans, la stratification sociale des usages de la liaison variable, relevée chez l’adulte, est en place dans les productions enfantines.En accord avec les principes défendus par les théories basées sur l’usage, nous généraliserons le modèle développemental de l’acquisition des constructions nominales en contexte de liaison catégorique à celui de la liaison variable.

Site du séminaire: http://www.utc.fr/utc-evenements/page00014e94.php

Much attention has been paid to a characterization of indirect requests in terms of general, rigid speech acts or broadly defined cultural patterns. Conversely, the cognitive roots of indirect requests and their realizations as linguistic constructions have received little attention. This is all the more problematic as linguists have thus left aside frequent situations in which the subtle form/meaning interplay made possible by some directive constructions echoes the ad hoc construction/negotiation of interactants' identities. This is particularly conspicuous when the specific role of each interactant is unclear or unconventional, and therefore subject to momentary negotiation. Pseudo-directive constructions provide a good case in point insofar as they occur when Speaker attempt to get someone to do something without giving the impression of impinging upon Addressee's self-assessment that s/he has a freedom of choice. One efficient (i.e. face-saving) strategy consists in blending a covert coding of Speaker's authority and an overt coding of Addressee's freedom of non- compliance into the same construction. This can be successfully achieved by using the

Powerpoint: http://www.univ-paris8.fr/desagulier/home/AFLiCo2_Desag.pdf; Grammatical blending and the indirect framing of directive constructions Speakers resort to pseudo-directive constructions when attempting to get someone to do something without giving the impression of restricting the addressee's freedom of choice of action or without impinging upon the addressee's self-assessment that s/he has that freedom of choice. One efficient (i.e. face-saving) strategy consists in blending a covert coding of the speaker's authority and an overt coding of the addressee's freedom of non-compliance into the same construction. This is generally done compressively, both at the formal and conceptual levels. In XXXX (2005), we showed that the you (don't) want to/ wanna X construction (e.g. you want to make a left turn, you don't wanna miss that) displayed the main properties of a grammatical blend (Fauconnier and Turner 1996), among which the dual inheritance principle and, above all, compression (morphosyntactic and conceptual). In this case, grammatical blending successfully frames the directive as collaborative work (the speaker appears as willing to help the addressee) and thus mitigates the inherent coerciveness of the act. In this paper I will show that grammatical blending may serve the same function in expansive (pseudo-) directive strategies: (1) A- I'm going to have to ask you to leave. B- You're kicking me out? A- No, I'm not saying that. B- Ok, then I can stay? (...) A- I'm asking you to leave. B- Are you kicking me out? A- No. (2) I told you again not to take any photographs. I'm going to have to ask you to stay behind. (3) I'm afraid I'm going to have to ask you to check with the Defense Department. (4) Mister, you can say anything you want about me, but I'm gonna have to ask you not to talk about my horse that way. In this construction (I am going to have to ask you to X) the driving force of the illocutionary intent (Speaker wants Addressee to X > structural type: 'Speaker asks Addressee to X') is disguised as something impending (going to) and extrinsic from the speaker's own will (have to). The compositional combination of directive strategies is blended into a composite construction. However, despite its internal complexity, the expanded construction assumes some morphosyntactic and prosodic unity, insofar as each of its formal components displays a high degree of collostruction strength to the composite structure (Stefanowitsch and Gries 2003, 2005). This paper will show that compression - a defining property of conceptual integration (Fauconnier and Turner 2000) - does not (always) correspond in a one-to-one fashion with morphosyntactic chunking in a grammatical blend. This observation will be corroborated by empirical research on innovative directive strategies in Japanese (Smith 1992, Sunaoshi 1995, Takano 2005). Those case studies, combined with a critical examination of recent works on the topic (Fauconnier and Turner 2002, Mandelblit 2000, etc.), will form the basis for a renewed (and expanded) definition of grammatical blending. Finally, these newly defined constructional integration networks will be described in light of the Fuzzy Construction Grammar framework (Desagulier 2005). REFERENCES Blum-Kulka, S. 1989. Playing it safe: The role of conventionality in indirectness. Cross- cultural pragmatics: Requests and apologies. Edited by Shoshana Blum-Kulka, Juliane House, and Gabriele Kasper. Norwood: Ablex Publishing, 37-70 Fauconnier, G. & M. Turner. 1996. Blending as a central process of grammar. In Conceptual Structure, Discourse, and Language. Edited by Adele Goldberg. Stanford: CSLI, 113-129. ----------------. 2000. Compression and global insight. Cognitive Linguistics 11(3/4), 283-304. ----------------. 2002. The Way We Think: Conceptual Blending and the Mind's Hidden Complexities. New York: Basic Books. Mandelblit, N. 2000. The grammatical marking of conceptual integration: From syntax to morphology. Cognitive Linguistics 11 (3/4), 197-251. Smith, J.S. 1992. Politeness and directives in Japanese women's speech. Language in Society 21(1), 59-82. Stefanowitsch, A. & S. T. Gries. 2003. Collostructions: On the interaction between words and constructions. International Journal of Corpus Linguistics 8(2), 209-43. ----------------. 2005. Co-varying collexemes. Corpus Linguistics and Linguistic Theory 1(1), 1-43. Sunaoshi, Y., 1995. Your boss is your "mother": Japanese women's construction of an authoritative position in the workplace. Texas Linguistic Forum, vol. 34, Department of Linguistics, University of Texas at Austin, 175-188. Takano, S. 2005. Re-examining linguistic power: strategic uses of directives by professional Japanese women in positions of authority and leadership. Journal of Pragmatics 37(5), 633-666. Desagulier, G. 2005. Grammatical blending and the conceptualization of complex cases of interpretational overlap: the case of want to/wanna. Annual Review of Cognitive Linguistics 3, 22-40.

Il s'agit d'un document non officiellement publié. Merci de ne pas citer sans autorisation de l'auteur.; Ce texte a pour but de présenter de manière simplifiée le questionnement qui a conditionné l'apparition des Grammaires de Constructions. Il s'agit d'un développement de l'argumentation de l'auteur à l'occasion du séminaire du PRI "Anthropologie et Linguistique" (EHESS) qui s'est tenu le 13 décembre 2006.

Les concepts de figure et de force sont d’autant plus pertinents en Linguistique Cognitive qu’ils font écho à l’éternelle tension entre principe d’inertie et principe de changement dont se nourrit la dynamique langagière. En effet, loin de n’être qu’un répertoire stable de représentations figuratives ou abstraites, la grammaire interne des locuteurs telle que nous la concevons est un équilibre négocié de forces motivées par l’usage et de représentations stabilisées.

Les concepts de figure et de force sont d’autant plus pertinents en Linguistique Cognitive qu’ils font écho à l’éternelle tension entre principe d’inertie et principe de changement dont se nourrit la dynamique langagière. En effet, loin de n’être qu’un répertoire stable de représentations figuratives ou abstraites, la grammaire interne des locuteurs telle que nous la concevons est un équilibre négocié de forces motivées par l’usage et de représentations stabilisées.

This dissertation explores the mechanisms of variation and language change within the cognitive framework of Construction Grammars. From a usage-based perspective, a construction is both conservative and innovative. On the one hand, the symbolic form/meaning pairing is relatively stabilized insofar as it is the conventional product of an abstraction from linguistic experience on a collective scale. On the other hand, speakers do not reproduce exactly the same pairings, which makes reanalysis possible. That differential defines a zone of potential development for each construction, which is precisely what enables grammar to keep pace with language flexibility. Our working assumption is that it is by studying intermediate forms that we can gain a better understanding of creativity and innovation, from both a linguistic and a cognitive perspective. That is why we argue in favor of a Fuzzy Construction Grammar. Our case studies tend to show that form/function reshuffling is best understood as a grammatical blend, for which we offer a new definition, based on a critical examination of the works by Fauconnier and Turner (1996). We see those constructional integration networks as the keystone of our model. They hinge on the following principle: a construction that is cognitively salient provides solid ground for the structure of speakers’ mental grammar. This stable symbolic unit can thus (i) be retrieved wholly or partially to provide a template for the composition of new constructions (ii) help speakers/hearers gain access to more complex pairings.; Cette thèse aborde les mécanismes de la variation et du changement linguistiques sous l’angle cognitif des Grammaires de Constructions. Dans une perspective centrée sur l’usage, une construction est à la fois stable et potentiellement innovante. D’un côté, l’assemblage symbolique de forme et de sens qui la constitue est relativement figé, dans la mesure où il est le produit conventionnel d’une abstraction collective réalisée à partir de l’expérience linguistique. De l’autre, les locuteurs ne reproduisent pas exactement les mêmes assemblages, ce qui fait que des réanalyses sont possibles. Ce différentiel crée, pour chaque construction, une zone de développement potentiel, grâce à laquelle la grammaire s’adapte à la flexibilité linguistique.L’hypothèse de travail développée ici est que l’analyse des constructions intermédiaires fournit un point d’accès privilégié dans la compréhension de la créativité et de l’innovation, d’un point de vue linguistique et cognitif. C’est la raison pour laquelle nous plaidons en faveur d’une Grammaire de Constructions Floue. Nos études de cas tendent à montrer que le réagencement forme/sens prend la forme de blends grammaticaux, dont nous proposons une redéfinition, suite à un examen critique des travaux de Fauconnier et Turner (1996). Ces schémas d’intégration constructionnelle forment la clé de voûte de notre modèle explicatif. Ils s’appuient sur le fait qu’une construction cognitivement saillante joue le rôle de jalon dans la grammaire mentale des locuteurs, ce qui la prédispose (i) à exporter tout ou partie de sa trame dans l’élaboration de nouvelles unités symboliques, (ii) à faciliter l’accès à des assemblages plus complexes.

This dissertation explores the mechanisms of variation and language change within the cognitive framework of Construction Grammars. From a usage-based perspective, a construction is both conservative and innovative. On the one hand, the symbolic form/meaning pairing is relatively stabilized insofar as it is the conventional product of an abstraction from linguistic experience on a collective scale. On the other hand, speakers do not reproduce exactly the same pairings, which makes reanalysis possible. That differential defines a zone of potential development for each construction, which is precisely what enables grammar to keep pace with language flexibility. Our working assumption is that it is by studying intermediate forms that we can gain a better understanding of creativity and innovation, from both a linguistic and a cognitive perspective. That is why we argue in favor of a Fuzzy Construction Grammar. Our case studies tend to show that form/function reshuffling is best understood as a grammatical blend, for which we offer a new definition, based on a critical examination of the works by Fauconnier and Turner (1996). We see those constructional integration networks as the keystone of our model. They hinge on the following principle: a construction that is cognitively salient provides solid ground for the structure of speakers’ mental grammar. This stable symbolic unit can thus (i) be retrieved wholly or partially to provide a template for the composition of new constructions (ii) help speakers/hearers gain access to more complex pairings.; Cette thèse aborde les mécanismes de la variation et du changement linguistiques sous l’angle cognitif des Grammaires de Constructions. Dans une perspective centrée sur l’usage, une construction est à la fois stable et potentiellement innovante. D’un côté, l’assemblage symbolique de forme et de sens qui la constitue est relativement figé, dans la mesure où il est le produit conventionnel d’une abstraction collective réalisée à partir de l’expérience linguistique. De l’autre, les locuteurs ne reproduisent pas exactement les mêmes assemblages, ce qui fait que des réanalyses sont possibles. Ce différentiel crée, pour chaque construction, une zone de développement potentiel, grâce à laquelle la grammaire s’adapte à la flexibilité linguistique.L’hypothèse de travail développée ici est que l’analyse des constructions intermédiaires fournit un point d’accès privilégié dans la compréhension de la créativité et de l’innovation, d’un point de vue linguistique et cognitif. C’est la raison pour laquelle nous plaidons en faveur d’une Grammaire de Constructions Floue. Nos études de cas tendent à montrer que le réagencement forme/sens prend la forme de blends grammaticaux, dont nous proposons une redéfinition, suite à un examen critique des travaux de Fauconnier et Turner (1996). Ces schémas d’intégration constructionnelle forment la clé de voûte de notre modèle explicatif. Ils s’appuient sur le fait qu’une construction cognitivement saillante joue le rôle de jalon dans la grammaire mentale des locuteurs, ce qui la prédispose (i) à exporter tout ou partie de sa trame dans l’élaboration de nouvelles unités symboliques, (ii) à faciliter l’accès à des assemblages plus complexes.

In this paper, grammatical blending is presented as an alternative to the conventional, linear overlap models of grammaticalization when it comes to conceptualizing complex cases of overlaps. The choice of the ‘emerging modal’ want to/wanna as a case study is motivated precisely by its interpretational complexity. For it appears that the grammaticalization of want to/wanna has been shaped by the compositional interaction of form and meaning. In this configuration, the linear model has to be modified for two reasons: being exclusively semantic, it does not take the form of the expression into account; being linear, it does not lend itself to a treatment of constructional compositionality. Grammatical blending, understood as the blend of constructions as defined in the Construction Grammar framework, is one way of altering the linear model so as to enable the representation of non-linear (i.e. compositional) constructional overlaps.

This dissertation explores the mechanisms of variation and language change within the cognitive framework of Construction Grammars. From a usage-based perspective, a construction is both conservative and innovative. On the one hand, the symbolic form/meaning pairing is relatively stabilized insofar as it is the conventional product of an abstraction from linguistic experience on a collective scale. On the other hand, speakers do not reproduce exactly the same pairings, which makes reanalysis possible. That differential defines a zone of potential development for each construction, which is precisely what enables grammar to keep pace with language flexibility. Our working assumption is that it is by studying intermediate forms that we can gain a better understanding of creativity and innovation, from both a linguistic and a cognitive perspective. That is why we argue in favor of a Fuzzy Construction Grammar. Our case studies tend to show that form/function reshuffling is best understood as a grammatical blend, for which we offer a new definition, based on a critical examination of the works by Fauconnier and Turner (1996). We see those constructional integration networks as the keystone of our model. They hinge on the following principle: a construction that is cognitively salient provides solid ground for the structure of speakers’ mental grammar. This stable symbolic unit can thus (i) be retrieved wholly or partially to provide a template for the composition of new constructions (ii) help speakers/hearers gain access to more complex pairings., Cette thèse aborde les mécanismes de la variation et du changement linguistiques sous l’angle cognitif des Grammaires de Constructions. Dans une perspective centrée sur l’usage, une construction est à la fois stable et potentiellement innovante. D’un côté, l’assemblage symbolique de forme et de sens qui la constitue est relativement figé, dans la mesure où il est le produit conventionnel d’une abstraction collective réalisée à partir de l’expérience linguistique. De l’autre, les locuteurs ne reproduisent pas exactement les mêmes assemblages, ce qui fait que des réanalyses sont possibles. Ce différentiel crée, pour chaque construction, une zone de développement potentiel, grâce à laquelle la grammaire s’adapte à la flexibilité linguistique.L’hypothèse de travail développée ici est que l’analyse des constructions intermédiaires fournit un point d’accès privilégié dans la compréhension de la créativité et de l’innovation, d’un point de vue linguistique et cognitif. C’est la raison pour laquelle nous plaidons en faveur d’une Grammaire de Constructions Floue. Nos études de cas tendent à montrer que le réagencement forme/sens prend la forme de blends grammaticaux, dont nous proposons une redéfinition, suite à un examen critique des travaux de Fauconnier et Turner (1996). Ces schémas d’intégration constructionnelle forment la clé de voûte de notre modèle explicatif. Ils s’appuient sur le fait qu’une construction cognitivement saillante joue le rôle de jalon dans la grammaire mentale des locuteurs, ce qui la prédispose (i) à exporter tout ou partie de sa trame dans l’élaboration de nouvelles unités symboliques, (ii) à faciliter l’accès à des assemblages plus complexes.

Under the influence of cognitive linguistics, construction grammars make no principled distinction between semantics and pragmatics. Most construction grammars, claim that meaning does not derive from abstract principles and cannot be apprehended outside a specific context. In fact, meaning is part of a network of representations that are both culturally situated and structured by experience. Construction-grammar semantics is thus very close to traditional pragmatics. Even if linguists who work in construction grammars have good reasons to reject the semantics/pragmatics distinction, they have done so at the expense of pragmatics. This is problematic because pragmatics has a lot to offer when it comes to explaining the use of constructions in terms of context, interaction, and communities of practice. Most construction grammars are only too willing to explain sociopragmatic phenomena by means of semantics, with the result that key constructional features are lost. This may be because in many construction-grammar architectures, the pragmatic level is too narrow., En vertu d'un postulat hérité de la linguistique cognitive, les grammaires de constructions n'opèrent pas de distinction de principe entre sémantique et pragmatique. En théorie, cette non-séparation est justifiée. En effet, dans les grammaires de constructions, le sens ne dérive pas de principes abstraits et n'a pas vraiment de pertinence lorsqu'il est appréhendé hors-contexte. Il s'inscrit en fait dans un réseau de représentations culturellement situées et structurées par l'expérience. La sémantique constructionnelle est en cela très proche de la pragmatique traditionnelle. Mais dans les faits, la non- distinction entre sémantique et pragmatique s'est faite au détriment de la seconde. En regard du sens, l'information pragmatique a ceci de plus qu'elle est en prise avec des facteurs contextuels, liés notamment à l'interaction des locuteurs au sein de communautés de pratiques. Trop rapidement recouverte par des considérations sémantiques, la pragmatique dépasse du cadre finalement trop étroit que les grammaires de constructions lui ont trop rapidement assigné.