aa

Langues et civilisations
à tradition orale
(UMR7107)

          Le CNRS           Accueil SHS          Autres sites CNRS          
cnrs paris3 inalco paris3 paris3
Rechercher
Intranet Lacito English




  Accueil > La recherche au Lacito > Structures du lexique : Typologie et dynamiques

Opérations de recherche en cours

 

 

 

 

Structures du lexique : Typologie et dynamiques

 

 

 

 

Responsables : Lameen Souag et Antoinette Schapper

 

La lexicologie est l'étude du lexique, et de la manière dont il est structuré dans chaque langue. On peut vouloir l'analyser dans une seule langue en synchronie – par exemple, en rédigeant un dictionnaire, ou en décrivant un champ sémantique particulier dans cette langue. On peut aussi adopter une perspective typologique en comparant les lexiques de plusieurs langues dans un domaine donné : par exemple, comment les langues structurent-elles le domaine de la parenté ? de la magie ? des émotions ? de la parole et de la pensée ? Retrouve-t-on partout les mêmes concepts, les mêmes distinctions sémantiques ? La typologie lexicale peut permettre d'identifier des polysémies récurrentes, des zones de stabilité et de variation, des universaux du lexique.

On peut également interroger ces structures lexicales dans leur dynamique, dans le temps et dans l'espace. Certaines polysémies, certaines associations sémantiques, sont apparues historiquement, ou au contraire ont disparu — soit par évolution interne, soit sous l'effet du contact avec d'autres langues. La tendance, chez les individus bilingues, à aligner les structures sémantiques des langues qu'ils parlent, a permis la diffusion de certaines catégorisations lexicales à l'échelle de vastes aires linguistiques et culturelles : c'est ainsi que certains découpages sémantiques, certaines polysémies ou phraséologies, deviennent les symptômes d'une aire donnée. Parfois, il est possible d'expliquer ces phénomènes aréaux par des liens entre pratiques langagières et pratiques sociales répandues dans la région: certains modes d'organisation familiale, par exemple, pourront être corrélés à des structures lexicales spécifiques dans le domaine de la parenté, ou dans le vocabulaire du mariage et des relations interpersonnelles.

Au fil des prochaines années, l'idée de ce séminaire sera de mettre en valeur les données de première main que nous avons recueillies sur tous les continents, ressources considérables et souvent sous-exploitées. Les dictionnaires existants, livresques ou électroniques, pourront nous servir également; ainsi que nos corpus de textes, pour peu que nos méthodes impliquent d'y recourir. Par ailleurs, nous nourrirons nos réflexions théoriques et méthodologiques des publications diverses dans le domaine de la sémantique lexicale, de plus en plus nombreuses ces derniers temps – voyez la liste des références.

Diverses approches sont possibles, et le LaCiTO pourra choisir d'en aborder plus d'une. En fonction des souhaits des participants, nous pourrons choisir d'établir la typologie d'un domaine sémantique particulier, ou d'une sélection de différents domaines, à travers des études parallèles. Nous pourrons nous pencher sur la quête d'universaux, sur les cas de convergence aréale, sur les parcours étymologiques, sur la théorisation du changement sémantique. Nous pourrons confronter diverses approches méthodologiques: bases de données, création de questionnaires, élaboration et visualisation de statistiques, cartes sémantiques…

 

Références :
-- François, Alexandre. 2008. Semantic maps and the typology of colexification: Intertwining polysemous networks across languages. In Martine Vanhove (ed.), From Polysemy to Semantic Change, Studies in Language Companion Series, vol. 106, 163–215. Amsterdam: John Benjamins. [accès en ligne]
-- François, Alexandre. 2013. Shadows of bygone lives: The histories of spiritual words in northern Vanuatu. In Robert Mailhammer (ed.), Lexical and Structural Etymology. Beyond Word Histories. Studies in Language Change 11. Boston/Berlin: Walter de Gruyter, 185-244.
-- Juvonen, Päivi & Maria Koptjevskaja-Tamm (eds.). 2016. The Lexical Typlogy of Semantic Shifts. Cognitive Linguistics Research 58. Berlin: De Gruyter Mouton.
-- Koptjevskaja-Tamm, Maria, Ekaterina Rakhilina & Martine Vanhove. 2015. The semantics of lexical typology. In Nick Riemer (ed.), The Routledge Handbook of Semantics. London: Routledge.
-- Koptjevskaja-Tamm, Maria, Martine Vanhove, & Peter Koch. 2007. Typological approaches to lexical semantics. Linguistic Typology 11.1, 159-185.
-- Mahieu, Marc-Antoine & Nicole Tersis (eds.) 2016. Questions de sémantique inuit / Topics in Inuit Semantics. Amerindia 38. 274 pp.
-- Mailhammer, Robert (ed.). 2013. Lexical and structural etymology: Beyond word histories. Studies in Language Change, 11. Berlin: DeGruyter Mouton.
-- Moyse, Claire & Volker Gast & Ekkehard Koenig. 2014. Comparative lexicology and the typology of event descriptions : A programmatic study. In Doris Gerland, Christian Horn, Anja Latrouite, Albert Ortmann (eds), Meaning and Grammar of Nouns and Verbs. Studies in Language and Cognition, 1. Düsseldorf : Düsseldorf University Press, 145-183.
-- Tersis, Nicole & Pascal Boyeldieu (eds.). 2017. Le langage de l'émotion : Variations linguistiques et culturelles. (Société d'Études Linguistiques et Anthropologiques de France, 469). Paris–Leuven: Peeters.
-- Urban, Matthias. 2011. Asymmetries in overt marking and directionality in semantic change. Journal of Historical Linguistics 1(1). 3–47​.
-- Vanhove, Martine (ed.) 2008. From Polysemy to Semantic Change. Studies in Language Companion Series, vol. 106. Amsterdam: John Benjamins.

 

 

Exposés à venir

  • Mercredi 13 novembre 2019 16h00-18h00
    Salle Mezzanine, Maison de La Recherche, 4 rue des Irlandais, 75005 Paris
    Exposé de Arienne M. Dwyer (University of Kansas), sur Using network analysis to explore etymological clustering in Chaghatay medicine
    Résumé en anglais :
    This methodological evaluation of ongoing research combines lexicography with the digital humanities, using network analysis to discover semantic and etymological clusters in the domain of healing and medicine in Central Eurasia. The source data are late Chaghatay (ISO 639-3: chg) manuscripts copied in late-19th c. Chinese Turkestan (Xinjiang), largely based on earlier Greco-Islamic humoral medical handbooks transmitted across Eurasia. We use etymologies as a rough proxy for the origin of medical ideas, healing substances (such as "manna" and "Chinese" cinnamon), and human anatomy. Network analysis (the algorithmic visualization of interconnected elements) has largely been used to analyze social networks. In the current project (Dwyer et al. 2018-2020), our team is experimenting with network analysis to visualize the semantic clustering and Perso-Arabic/Turkic synonymy of medical terms. These methods could aid the study of colexification and other areas of lexical semantics by visualizing correlations beyond simple frequencies for large data sets.
    Référence :
    -- Dwyer, Arienne M. (PI), Jeff Rydberg-Cox & Sandra Kübler (co-Is). 2018-2020. Analyzing Turki Manuscripts from the Jarring Collection Online (ATMO-2). Sponsored by the Henry Luce Foundation. Online: https://uyghur.ittc.ku.edu/atmo.html.

 

Exposés passés

  • 3 octobre 2019 à 11h
    Exposé de Tatiana Nikitina (LLACAN), sur Amphichronic semantic maps for closely related languages: Singulatives in Southeastern Mande
    Résumé en anglais :
    This study relies on an extended Radial Category approach to explore a network of meanings associated with singulative markers in four closely related Mande languages: Dan, Tura, Mwan, and Wan. The proposed model captures similarities and variation observed across these languages without making assumptions about the synchronic status of semantic relations between different uses. The amphichronic approach to micro-scale variation is shown to shed light on the history of individual languages and the ways in which semantic networks develop over time. The model's strengths and weaknesses are discussed.
    The reconstructed semantic network is argued to have fallen apart in Wan, while it is still preserved in Dan and Tura, and only partially lost in Mwan. Comparison with earlier results based on a study of diminutive markers helps identify two different types of change involved in the disintegration of semantic networks: renewal of the expression associated with a particular use, and loss of the use without replacement. While the former type is at the center of classical studies in lexical typology, the latter raises issues for the universalist assumption that underlies the semantic map approach and is too often taken for granted in current typological research.

  • 23 juillet 2019
    Exposé de Rachel Hendery, sur Colexification of meat and fish: Diachronic dimensions of semantic maps
    Résumé en anglais :
    A method of ‘semantic mapping’ has been developed in recent years with some variants (Haspelmath 2003, François 2008, Van der Auwera 2013, 2017). An example of such a semantic map is shown in fig. 1 (from François 2008: figure 4). Most semantic maps do not distinguish between historical and synchronic processes. For example, François (2008: 178) explicitly states that an advantage is that colexification is ‘neutral’ about historical or semantic interpretation, although he proposes that future research could investigate incorporating a diachronic dimension and using three dimensions instead of two. Several other authors (e.g. Narrog 2011; Van der Auwera 2013) encode directionality explicitly by connecting terms with arrows, but this still does not represent changes in the semantic distribution of terms over time. In this paper I build on these ideas to explore how adding dimensionality to semantic maps could allow for encoding of diachronic information. I test some solutions with examples of changes in colexification in Australia and the surrounding region.
    One possibility is to use a third dimension. I propose that a semantic map could be displayed as layers or slices of time along a y axis, with connections between the planes showing the relationship of terms across time. This is similar to the standard practice of viewing high-dimensional data in scatterplot matrices instead of individual scatterplots. An example is shown in figure 2. The example I use for this and the following diagram is the change from ‘fish’ to ‘meat’ in parts of Australia (see fig. 3, from McConvell 1997), which I will discuss further during our paper.
    A disadvantage of such a representation is that it is relatively difficult to see what is happening at the “back” of the diagram. In a digital space, the map could of course be rotated, zoomed, or explored in a virtual walk-through, ameliorating this difficulty. Nevertheless it is worth considering whether three dimensions could be better represented in a flat structure. The key to doing so is in recognising that the information to be represented consists of two network graphs with a mapping between the two that is itself a third network. Network 1 is lexical forms, among which diachronic relationships can exist. Network 2 is semantic terms, among which relationships can also exist (such as ‘x is a kind of y’). Network 3 is the way in which the lexical items map onto the semantic terms in different languages and/or at different times. In fig 4 I show how these three networks could be represented using two arc diagrams for the lexical forms and the semantic concepts, and a matrix for the mapping between the two.
    Our presentation will discuss these and other approaches for representing the diachronic (and geographic) dimensions in semantic maps, their advantages and disadvantages, as well as other complexities we have found when trying to account for change in colexification patterns in Australia and the surrounding region.
    Références :
    -- François, A. (2008). Semantic maps and the typology of colexification. From polysemy to semantic change. In: Towards a typology of lexical semantic associations. Vanhove, M. (ed). John Benjamins. 163-235.
    -- Haspelmath, M. (2003). The geometry of grammatical meaning: Semantic maps and cross-linguistic comparison. In: The new psychology of language. Psychology Press. 217-248.
    -- McConvell, P. (1997). Semantic shifts between fish and meat and the prehistory of Pama-Nyungan. In: Boundary rider: essays in honour of Geoffrey O’Grady. Tryon & Walsh (eds), Pacific Linguistics. 303-25.
    -- Narrog, H. (2010). A diachronic dimension in maps of case functions. Linguistic Discovery, 8(1), 233-254.
    -- Van der Auwera, J. (2013). Semantic maps, for synchronic and diachronic typology. In Synchrony and Diachrony. A dynamic interface Ramat, A et al (eds). John Benjamins. 153-176.

  • 5 juin 2019, INALCO, PLC (65, rue des Grands-Moulins, Paris 13e), salle 5.08
    Exposé de David Gil (Max Planck Institute for the Science of Human History, Jena, Germany), sur Who is your name?
    Résumé en anglais :
    Why does Indonesian ask Nama siapa? (name who), rather than What is your name? as in English? English speakers are often puzzled by the Indonesian construction, believing that since names are inanimate, WHAT should be used. Conversely, Indonesian speakers are often surprised to learn that English uses WHAT, arguing that since names refer to people, it should be WHO. This paper poses the question whether the different choice of WH words in Indonesian and English reflects some more fundamental property distinguishing between the two languages, or whether it is a superficial feature without deeper typological ramifications. The answer that is provided is: a combination of both.
    The first part of this paper presents the results of an ongoing world-wide cross-linguistic survey examining the choice of WH word in "What is your name?" questions, covering close to 800 languages. The results show that the Indonesian WHO construction represents a cross-linguistically widespread option, spanning a wide rage of seemingly typologically diverse languages, including, among others, Zulu, Amharic, Tsez, Mongolian, Dani, Tahitian and Squamish. Moreover the presence of areal patterning evident in the map shows that the choice of WH word is a feature that is readily borrowed across languages of different genealogical and typological groups. Thus, in large part, Indonesian uses WHO because it is a typical Insular Southeast Asian language, while English uses "What is your name?" because it is a run-of-the-mill Western European language.
    However, the second part of this paper shows that in spite of such areal patterning, the choice of WH word does indeed also reflect deeper aspects of morphosyntactic organization. The choice between WHO and WHAT is shown to correlate with the results of an in-progress cross-linguistic experiment on over 60 languages world-wide measuring the extent to which the assignment of thematic roles are grammaticalized. Specifically WHAT languages tend to exhibit more grammaticalization of thematic role assignment than WHO languages. Thus while in English, what and your name are related via thematic role assignment, in Indonesian, nama and siapa are connected through a looser relationship of association.
    The choice of WH word in "What are your name?" questions is thus partly arbitrary, reflecting the outcome of diachronic processes of language contact and borrowing, and partly principled, reflecting the degree of grammaticalization of thematic role assignment in the grammar. This case study underscores the way in a single linguistic phenomenon may simultaneously reflect an ontologically heterogeneous potpourri of factors, some diachronic, others synchronic — there can be no one single story explaining everything.

  • 27 février 2019
    Exposé de Guillaume Segerer & Martine Vanhove (Llacan) sur Areal patterns of colour naming in the languages of Africa
    Résumé en anglais :
    In an article (Segerer & Vanhove in press) we investigated lexicalization patterns in colour naming among languages of Africa in a balanced sample of 350 languages based on lexical data (mainly from the RefLex database (Segerer & Flavier 2011-2018) concerning the eleven "basic" color terms of Berlin and Kay's (1969) hierarchy. Based on the same sample, this presentation will address methodological issues concerning the way to discover colexification patterns with the tools provided in this large searchable database (over 1,200,000 entries), to differentiate between genetic similarities and areal patterns, between universal and accidental ones, and between significant and trivial similarities. It will be demonstrated that
         (i) there are very few areal patterns for color naming in Africa;
         (ii) some of them are trivial in the sense that (a) widespread lexicalizations in one particular area are linked to naturally and culturally salient objects (e.g. the widespread lexicalization of 'yellow' from 'locust bean' in western and central Africa which corresponds to the geographical distribution of the plant itself), or (b) they don't pattern geographically but are widespread crosslinguistically (e.g. the colexification of 'red' and 'ripe');
         (iii) contact induced areal patterns concern a word class that has been long believed not to be borrowable, namely ideophones (e.g. borrowings from Fula in related and unrelated neighbouring languages for 'black' and 'white'), challenging established borrowability scales. The scarcity of areal or genetic patterns in the domain of the colexifications of colour terms brings some support to Wierzbicka's (2005: 239) claim (in reaction to Berlin and Kay) that "[t]here are no 'color universals'" but our data does not seem to support the reason for this claim, namely that "'color' itself is not a universal concept".
    Références :
    — Berlin, B. and P. Kay. 1969. Basic Color Terms. Berkeley: University of California Press.
    — Segerer, Guillaume and Sébastien Flavier. 2011-2018. RefLex: Reference Lexicon of Africa, Version 1.1. Paris, Lyon. http://reflex.cnrs.fr/.
    — Segerer, Guillaume and Martine Vanhove. In press. Color Naming in Africa. In Lexicalization Patterns in Colour Naming: A Cross-Linguistic Perspective, Ida Raffaelli, Daniela Katunar & Barbara Kerovec (eds.). Amsterdam, Philadelphia: Benjamins.
    — Wierzbicka, Anna. 2005. There are no "color universals" but there are universals of visual semantics. Anthropological Linguistics 47/2: 217-244.

  • 30 janvier 2019
    Exposé de Anne-Laure Dotte (ERALO, UNC) & Claire Moyse-Faurie (Lacito), sur Approche typologique de « manger » dans les langues kanak et polynésiennes
    Résumé :
    Cette présentation vise à explorer la diversité des expressions décrivant l'ingestion des aliments dans les langues océaniennes (kanak et polynésiennes principalement), et tente de dresser un portrait des convergences et des divergences attestées dans ces langues. Dans les langues kanak, par exemple, le champ lexical des expressions pour « manger » ne renvoie pas à un domaine homogène ou consistant : plusieurs sous-groupes de langues sont identifiables en fonction des catégories d'aliments qu'elles distinguent, à partir de critères physiques, nutritifs ou symboliques (culturels), à partir du mode de préparation, ou de la combinaison possible entre différents types d'aliments, et ces catégories se distinguent sur la base du verbe avec lequel elles se combinent.

  • 10 octobre 2018
    Exposé d'Antoinette Schapper sur Baring the bones: the metaphorical association of bone and strength in New Guinea and the limits of colexification
    Résumé en anglais :
    The talk will be an exploration of some issues around comparing lexical meaning across languages. Metaphor-driven lexical associations are often only present in languages in conventionalised expressions, but we may wonder whether such limited associations still count as colexification. In this talk I will present a range of data from the languages of New Guinea, showing there is a scale of lexical-to-conventional-association of bone and strength. Capturing the range of manifestations of this metaphorical association is unlikely to be possible using CLICS or similar databases. Warning: This talk is meant to be a discussion piece, rather than a worked out thing.

  • 25 mai 2018
    Exposé de Masha Koptjevskaja-Tamm (Stockholm University) sur Hot as fire
    Résumé en anglais :
    My currents research interests are in cross-linguistic variability in semantic systems and its limits, both in space and time, approached through empirical cross-linguistic research with a special focus on the interaction between language structures, culture and cognition. Is linguistic categorization universal or language/culture-specific, how do semantic systems emerge and develop, where do words for categories in one semantic domain come from and how can they be used for others?
         My talk will revolve around two very basic and important semantic domains – TEMPERATURE and FIRE.
         TEMPERATURE is something I have been engaged in for quite a few years, as partly witnessed by the (huge) edited volume The linguistics of temperature (Koptjevskaja-Tamm, ed. 2015, John Benjamins). I am right now working on its sequel, the book "Temperature in language: typology, evolution and extended uses", which is also the main reason for my present stay at Llacan/Lacito. In my talk I will first present a general overview of the ways in which languages of the world carve up the temperature domain among their expressions. I will then focus on two particular issues:
         -- How (dis-)similar can closely related languages be in their semantic systems, with lexicalization of the temperature domain across Slavic as case study, and
         -- Co-lexification patterns of temperature expressions, with the metaphor "AFFECTION IS WARMTH" (cf. warm words, warm heart) as a case study
    The other semantic domain, fire, is obviously related to temperature, among other things, because 'fire' is one of the cross-linguistically recurrent sources for temperature expressions. I present a few ideas and illustrations on how this domain can be approached cross-linguistically

  • 31 janvier 2018
    Exposé d'Alexandre François sur Typologie de la colexification : Une base de données
    Résumé :
    La ​​colexification est la propriété structurale par laquelle une langue peut utiliser la même forme lexicale pour exprimer deux significations S1 et S2 (François 2008). Ainsi, certaines langues colexifient "pays" avec "village", ou "tête" avec "chef", ou "devant" avec "avant"… Comme toute propriété typologique, la colexification peut être étudiée sous divers angles : description d'un domaine sémantique particulier; identification de tendances typologiques; reconstruction historique; études aréales, etc. Parmi les travaux que notre opération de lexique pourrait aborder en commun, figurerait CoLex, une base de données de colexifications – base générique (ouverte à tous les domaines du lexique) et collaborative.
          L'unité d'observation ne serait pas le sens unique, mais un couple de sens S1-S2. Chaque couple (en abscisse) serait croisé avec diverses langues (en ordonnée), afin d'observer dans quelles langues il est colexifié. Cet exposé présentera un échantillon de 165 couples de sens, croisés avec 5 langues de Mélanésie et d'ailleurs. Mais cette base accroîtra son intérêt à mesure qu'y seront ajoutés plus de couples lexicaux, et surtout plus de langues de familles diverses — en profitant de la variété des langues qui sont représentées au LaCiTO.
          Au travers de quelques exemples, l'exposé illustrera l'intérêt d'une telle base comparative, notamment dans la réflexion sur la dimension aréale des structures lexicales (cf. Juvonen & Koptjevskaja-Tamm 2016). Nous discuterons ensemble des choix présidant à la constitution d'une telle base, des écueils possibles, des décisions à prendre, et des perspectives de développement et d'exploitation de nos futures observations.
    Références
    -- François, Alexandre. 2008. Semantic maps and the typology of colexification: Intertwining polysemous networks across languages. In Martine Vanhove (ed.), From Polysemy to Semantic Change, Studies in Language Companion Series, vol. 106, 163–215. Amsterdam: John Benjamins.
    -- Juvonen, Päivi & Maria Koptjevskaja-Tamm (eds.). 2016. The Lexical Typology of Semantic Shifts. Cognitive Linguistics Research 58. Berlin: De Gruyter Mouton.

  • 29 novembre 2017
    Exposé de Lameen Souag sur Kinship terms in proto-Berber: Using colexification patterns in reconstruction
    Résumé en anglais :
    The reconstruction of kinship terminology in Berber presents a number of difficulties. Northern Berber kinship systems are profoundly influenced by Arabic at the level of lexicon as well as structure. Saharan ones paint a different picture, broadly reflecting an Iroquoian system rather than the Sudanese one characteristic of Arabic and Northern Berber; in this respect, they match several neighbouring languages of the Sahel. To determine the original situation, it is necessary to examine the history of colexification patterns. Colexification patterns shared across distantly related branches are likely to be original, while more localized ones are likely to be innovative. Innovative lexicalizations, expressing meanings not originally independently lexicalized, tend to be formed through loanwords, transparent compounding, or metaphorical extensions from other domains. A further aid to reconstruction is the fact that particular family residence patterns promote particular colexicalization patterns. Taking all of this data into account, it appears that proto-Berber kinship terminology reflected a bilateral Hawaiian system unlike any Berber group today, and that Tuareg kinship systems have been profoundly influenced by Sahelian ones just as Northern Berber ones have by Arab ones…

  • 1er juin 2017
    Exposé d'Alexandre François visant à présenter le domaine – avec une attention particulière portée aux phénomènes de colexification (François 2008), et la possibilité de créer des cartes sémantiques en typologie lexicale.
    ​ L'exposé sera suivi d'une discussion collective pour réfléchir à la manière dont nous voudrions mener ce groupe de recherche au cours des prochaines années.


Imprimer Contacter le webmestre Plan du site Crédits Accueil