logo labex efl

RT1 Généalogie des langues (Niger-Congo, Austronésien)

Reconstruction, classification interne et description grammaticale des deux principaux phylums du monde: Niger-Congo et Austronésien


coordinateurs:  V. Vydrin;  A. François;  I. Bril

LLACAN, LACITO

Ce projet vise à mieux appréhender l'histoire et la classification interne des deux principaux phylums généalogiques du monde: Niger-Congo et Austronésien. En dépit des progrès récents réalisés dans la reconstruction de leur histoire, chacune de ces deux grandes familles présente encore certains problèmes qui méritent d'être explorés.

[le texte ci-dessous est en anglais]

New perspectives in Niger-Congo and Austronesian subgrouping

This project aims at a better understanding of the history and internal classification of the two biggest language phylums of the world, viz. Niger-Congo and Austronesian. While recent scholarship has helped better understand the history of these two families, each of them still presents a number of problematic issues that remain to be addressed.

First, one important issue is whether Niger-Congo forms a genetic unit at all. Sands (2009) claims that due to the “lumping” habits of Africanists as opposed to the “splitting” habits of, e.g., Americanists, African languages are systematically underrepresented in typological samples that are supposed to be balanced. As for population history, Greenberg’s (1963) exhaustive classification of the languages of Africa into four phylums leads to the paradoxical situation that the continent that is identified as the cradle of humanity is also the one with the least genetic diversity from a linguistic point of view. A critical reappraisal of Niger-Congo will bring us much closer to knowing whether the current linguistic situation of Africa is due to contact between genetically unrelated languages or due to relatively recent migrations that wiped out the previously existing genetic diversity.

While Austronesian languages appear to have their history at times better understood, this comes with exceptions (e.g. the Temotu subgroup of Oceanic, or the precise paths of language change in languages of New Caledonia). But the main problem faced by Austronesianists is perhaps the excessive weight which has been given to the Tree model, at the expense of alternative representations. Several historical studies (Geraghty 1983; Ross 1988; Pawley 2012; François 2011) have shown that tree-like phylogenetic representations, based on divergence and population splits, are much less adapted to the Austronesian family than had been previously assumed. The manner in which subgroups arise historically is through entangled processes of divergence (migration) as well as convergence (spread of innovations) – a problem also identified, for instance, in Sinitic languages (Hashimoto 1992) or in Indo-European (Garrett 2000). New historical models need to be defined, that can take into account the diffusional component of language genealogy; promising has been the concept of linkage developed by Ross (1988), and formalised further by proponents of “Historical Glottometry” (Kalyan & François f/c; François f/c).

Bringing together some of the world’s leading experts in these two domains, we will strengthen the empirical basis of comparative work in these phylums by describing and documenting individual languages and we will propose better classifications and reconstructions. ‘Better’ essentially means ‘more reliable’ in the sense of the Comparative method, but also more faithful to what we can reconstruct of the actual paths of language change over the last millennia. In every domain of our research project – description, classification and reconstruction – specialists of Austronesian and Niger-Congo will exchange methods and insight from their own field.

Historical linguistics and typology

A further reason for reappraising the history of Niger-Congo languages has to do with the importance taken by language phylogenetics in proposing typological generalisations about language. Indeed, the rising prominence of typological approaches to the study of language universals (if these exist at all, see Evans & Levinson (2009) and ensuing discussions in Lingua & BBS) as well as evolutionary approaches to explanation in linguistics (Blevins 2004, Heine & Kuteva 2005, Dunn 2008) entails a renewed interest in the classification of the languages of the world. Not only are reliable genetic classifications indispensable for the creation of balanced language samples; but the reconstruction of proto-languages also contributes to our understanding of paths of language change.

Some of the standing challenges in this field are, first, the need to falsify classifications that are based on similarities between languages (Greenberg 1963). We intend to do so by rigorously applying the comparative method as well as new quantitative methods on more and better descriptive data of African languages – particularly in order to prove (or not) the genetic unity of Central Sudanic on the one hand and Niger-Congo and its alleged lower branches on the other hand. Freely accessible databases built by participant laboratories have already revitalized academic discussions on regions and phyla in Africa (Greenberg 1971 ; Güldemann 2008). Second, pushing reliable classifications further back. Third, distinguishing genetic relatedness from long term contact phenomena. Fourth, broadening the empirical basis for diachronic approaches to lesser studied language families, which includes identifying and describing language isolates. And finally, studying morphosyntactic change by comparing genetically related but typologically distinct languages. The three research projects in this theme will tackle these issues from different angles.

The empirical basis of our research will be strengthened by fundamental documentary and descriptive work on previously undescribed languages of Africa or Austronesia. Methodological innovation in classification, reconstruction, typology and description will be a major goal and a privileged domain of collaboration between specialists of different families and areas within our research theme.