sign in sign up
<<
Home , Historical linguistics

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Ghent University Academic Bibliography

K. Bostoen. Historical 257

HISTORICAL LINGUISTICS Koen Bostoen1

Linguists and archaeologists offer complementary view- ‘historical-’, starts from extant points on human behaviour and in African and tries to reconstruct their from communities. While historical-comparative linguistics ancestral stages through the study of current-day varia- commonly deals with the immaterial traces of the past tion. Such inter- can be phonological, in Africa’s present-day languages, unearths morphological, syntactic, semantic or lexical. the material vestiges of ancient . Even if both In the case of Bantu, the hypothetical common ances- sciences share similar core concepts, their methods, data tor language reconstructed on the basis of similarities and interpretive frameworks are profoundly different. observed between languages known mainly from the Explaining some basic principles of historical-compar- 19th century onwards is commonly called Proto-Bantu. ative linguistics as applied to the Bantu languages and This proto-language is assumed to be the best possible debunking some common misconceptions are the central reflection of the ancestor language that was supposedly aims of this contribution. Due to space constraints, no spoken some 4,000 to 5,000 years ago in the area from detailed bibliographic references are provided through- where Bantu languages started to spread through Central out the text (see my earlier publications for extensive Africa and beyond. Bantu linguists agree to situate this bibliographies).2 Some essential readings for non-spe- homeland in the so-called Grassfields region of - Cam cialists are listed at the end of this chapter. eroon, not far from the country’s border with Nigeria. This zone displays the highest linguistic diversity (which I. DIACHRONIC LINGUISTICS ON THE BASIS OF means that parent languages had sufficient time to- di SYNCHRONIC DATA verge locally) and is close to the area where the Benue- Ideally speaking, historical linguistics is the study of Congo relatives of Bantu languages are spoken. distinct historical stages in the evolution of one single language or . This is the case in Ro- II. REFERENTIAL VS. HISTORICAL OR GENEA- mance, for instance, where the development of Latin LOGICAL CLASSIFICATIONS into its multiple daughter languages can be empirically The best-known Bantu classification is no doubt- Mal reconstructed. In Africa, examining language variation colm Guthrie’s. In 1948, Guthrie subdivided the Bantu through time on the basis of diachronic language data languages in 16 different zones labelled A, B, C, D, E, F, is hardly ever possible, due to the lack of written docu- G, H, K, L, M, N, P, R, S and T, which he reduced to 15 ments. in 1971 by merging the last two to one zone. Each zone The case of Kikongo, whose historical record starts is further subdivided into language groups, indicated by in the early 17th century, is exceptional, and even not a decimal number, in which individual languages are equalled by Kiswahili whose oldest surviving texts do indicated by a unit. Lowercase letters following certain not date further back than the mid-18th century. For units refer to of a same language, e.g. Ciluba most other Central African languages, written docu- (L31a) and Lulua (L31b). In contrast to what is often ments become at best available from the late 19th cen- believed, Guthrie’s classification is strictly referential tury onwards. Even today, there are still many undocu- and was never meant to be historical: Guthrie did not mented languages, several of which are on the verge of rely on the ‘’ (which is the core ap- extinction. Historical linguistics in Africa thus usually proach of historical-comparative linguistics) or ‘shared consists in the comparative study of historically-related innovations’, its basic principle for historical subgroup- languages. This up-stream approach, also known as ing. Shared innovations are lexical, phonological or grammatical changes that took place only once in some 1 BantUGent - UGent Centre for Bantu Studies, Department of Languages ancestor language from which its daughter languages in- and Cultures, Ghent University, Belgium. herited it and which are therefore indicative of the closer 2 http://research.flw.ugent.be/en/koen.bostoen 258 Field Manual for African Archaeology. Chapter 6

relatedness between languages. By attributing a unique communities3 is a reflection of shared , the study alpha-numeric to each language, Guthrie wanted to of widespread cultural vocabulary usually provides in- facilitate comparison between the several hundred Bantu teresting insights on the lifestyle of past . This languages known at the time. sub-discipline is also known as the ‘-and-things Despite its limited historical or genealogical value, method’ (see Ricquier, this volume, pp. 261-263) or lin- Guthrie’s classification remains a useful reference tool. guistic palaeontology. To archaeologists, language data Each one of the nearly 900 documented Bantu language are particularly useful as a source of indirect historical varieties can be approximately situated in space thanks to evidence for those aspects of human culture which are its unique code. That is exactly why Jouni Maho updated either immaterial or whose material traces do not con- Guthrie’s list by adding new languages, but remained as serve well. Similar words with similar meanings shared faithful as possible to the original approach. Other schol- by numerous languages can be inherited from a common ars did propose rearrangements on historical grounds. ancestor language and spread through the dispersal of its Only one of these gained relatively wide acceptance daughter languages. They can also have been adopted amongst Bantu linguists, i.e. zone J proposed by the through contact and spread across languages as loan- former linguistics department of the Royal Museum for words. Central Africa in Tervuren. To distinguish between inherited and borrowed vocab- Bantu as a language family has been established ever ulary, linguists depend on the principle of regular sound since Bleek (1851). Its homeland is the region where correspondences. These are phonological similarities the ‘Narrow Bantu’ languages, i.e. those conventionally between languages, which cannot be the outcome of his- classified as Bantu by Guthrie, meet the ‘Wide Bantu’ torical accident, because they are recurrent, systematic languages, i.e. their closest Benue-Congo relatives aka and without unexplainable exceptions. While synchroni- ‘Bantoid’. The small ‘Mbam-Bubi’ subgroup, consisting cally widespread inherited terms can be reconstructed of several languages of the Mbam region of central Came- into a putative proto-language via these regular sound roon and Bubi spoken on Bioko Island, is the genealogical changes, loanwords cannot. Several Great Lakes Bantu junction between Narrow and Wide Bantu. The (Narrow) languages, for instance, have a lexical doublet to refer Bantu family further branches into five major subgroups: to calabashes and glass bottles. These are two words that ‘North-Western’, ‘Central-Western’ (aka ‘North Zaire’ or are historically related, but one of them was acquired ‘Congo’), ‘West-Western’ (aka ‘West-Coastal’), ‘South- through regular intergenerational transmission from an Western’ and ‘Eastern’. We mainly owe this robust under- ancestor language, while the other was obtained from ve- standing of Bantu genealogy to quantitative analyses of hicular Swahili through contact-induced diffusion. The so-called ‘basic vocabulary’, such as and inherited for calabash is phonologically much more phylogenetics. Qualitative approaches based on phono- heterogeneous, e.g. Sukuma cuβa, Nyamwezi nsòhá, logical and/or grammatical features fit less with the tree Ganda ɛ̀ nsúwà, Shi nshùhá. These words were model of language divergence, and emphasize that con- to the regular sound changes that their language under- vergence due to language contact also had a significant went since Proto-Bantu for which *-cʊ́ pà ‘calabash’ has impact on the speciation of Bantu languages. been reconstructed. Such is not the case for the term for glass bottle, which they recently borrowed from Swa- III. LANGUAGE AS AN HISTORICAL SOURCE hili resulting in much more similar loanwords: Sukuma Our knowledge of the environmental, social, cultural, cupá, Nyamwezi cupa, Ganda ccúpà, Shi ìcúpà. In Swa- and historical phenomena underlying hili itself, the word chupa refers to both calabashes and is often very limited in Africa. Its languages most often glass bottles. When the latter type of containers were need to ‘speak for themselves’. The study of language introduced along the East African coast, Swahili speak- has in itself become an important method of reconstruct- ers called them after their traditional containers using the ing history to which not only linguists, but also historians word for calabash which they inherited from Proto-Ban- and archaeologists dedicate themselves. Founded on the basic premise that vocabulary shared between 3 A is defined here as a group of people who consider themselves to speak a same language. K. Bostoen. Historical Linguistics 259

tu. Swahili speaking long-distance traders subsequently the most scattered one, which is preferably attested in the introduced this new specimen of material culture and more outlying areas, while the younger form occurs in its Swahili word in several East-African communities, a group of adjacent languages, which may be large, but many of them already having a regularly inherited Bantu not as scattered as the older form. A judicious historical word for ‘calabash’. interpretation of requires a basic insight into Unlike archaeologists, linguists do not have a stand- the internal classification of a language family. The rela- ard and universally accepted method for the absolute tive time depth does not depend so much on the number dating of language change. In the absence of diachronic of languages in which a feature occurs, but rather on its language data and without the tentative association of distribution over distinct historical subgroups. Hence, a language data to archaeological data, linguists need to term that is rare but scattered amongst the north-western limit themselves to . To do so, they rely and western Bantu languages is considered older than on a number of principles for which they are indebted a synonym that is densely spread among eastern Bantu to archaeology: stratigraphy, geographic distribution and languages only. seriation. A final basic archaeological concept also found in his- Linguists refer to the concept of stratigraphy to dis- torical linguistics is seriation. Linguists usually rely on it entangle the successive strata in the formation of a for the sequential ordering of sound changes. Each lan- language. The and of a language are guage is subject to sound changes, which can be called transmitted through time and transformed due to the loss regular to the extent that they affect all words sharing of old elements and the incorporation of new elements. a given phonological environment. The chronological They accumulate formative layers, which are never neat- sequencing of sound changes is primarily used for the ly superposed. Unlike archaeological strata, language historical classification of languages through the princi- layers are not subject to the law of superposition. There ple of shared innovations. If closely related languages is permanent stratigraphic contamination, so to speak. It share a historical change (whether lexical, phonological is the task of the historical linguist to order the present- or grammatical), there is a good chance that this innova- day data into successive strata. The words for ‘calabash’ tion only happened once, i.e. in their most common re- and ‘glass bottle’ in the example above clearly belong to cent ancestor – although independent convergent change two distinct strata of language history. can never be entirely excluded. Once one has an idea of Linguistic geography or geolinguistics can help with the internal classification of a language group and the the relative dating of language layers. This method deals relative chronology of sound changes, seriation is also a with the geographic distribution of linguistic features. It helpful dating device for loanwords. The earlier foreign is used for mapping loanword diffusion routes and for words are borrowed, the more sound changes they have determining their direction of borrowing and also as a in common with regularly inherited words and the better relative chronology device. ‘Linguistic isoglosses’ are they are phonologically integrated, making it difficult to the equivalent of stylistic horizons in archaeology. They identify them as borrowed vocabulary. mark the geographic distribution of a given linguistic feature shared by a number of languages. For example, CONCLUSIONS words for Kimbundu njila (‘bird’) are only found The interaction between African archaeology and lin- in a geographically restricted cluster of Bantu languages guistics has been severely criticized in the past, among spoken in the southwestern part of the domain, while other things due to a lack of critical evaluation of under- for Kikongo nuni ‘bird’ are found throughout lying concepts and methods. Although this appreciation the Bantu domain. Such spatial distribution is interpreted is certainly not undeserved, this should not refrain us as a function of time: the Kikongo word is a shared re- from interdisciplinary collaboration. No discipline is tention going back to Proto-Bantu while the Kimbundu capable of solving on its own the many complex rid- word is a more recent shared innovation. The relative dles of African history. Sound archaeological-linguistic chronological interpretation of isoglosses is done ac- teamwork requires in the first place a good understand- cording to certain areal norms which are not strict rules, ing of each other’s concepts, methods and evidence, to but rather hermeneutic principles, e.g. the oldest form is which I have tried to contribute in this chapter. A second 260 Field Manual for African Archaeology. Chapter 6

fundamental issue is the importance of direct collabora- Dimmendaal, G. 2011. Historical Linguistics and the tion between scholars of different disciplines who per- Comparative Study of African Languages. Amsterdam/Phila- fectly command their own body of evidence and are able delphia: John Benjamins. to make a judicious assessment of its historical signifi- Grollemund, R., S. Branford, K. Bostoen, A. Meade, cance instead of leaving this task to scholars who only C. Venditti and M. Pagel. 2015. ‘Bantu expansion shows that master one method or none at all. Finally, it is crucial habitat alters the route and pace of human dispersals’. Pro- that archaeologists and linguists mutually benefit from ceedings of the National Academy of Sciences of the United their specific advantages, e.g. absolute dating in the case States of America. http://dx.doi.org/10.1073/pnas.1503793112 of archaeology or the possibility to reconstruct vocabu- Heine, B. & Nurse, D. (eds). 2000. African languages: an lary referring to immaterial or poorly preserved material introduction. Cambridge/New York: Cambridge University aspects of human life in the case of historical linguistics. Press. Heine, B. and Nurse, D. (eds). 2004. Les Langues afric- REFERENCES aines. Paris: Karthala. Bastin, Y. & Piron, P. 1999. ‘Classifications lexi- Hombert, J.-M. & Hyman, L.M. (eds.). 1999. Bantu His- costatistiques: bantou, bantou et bantoïde. De l’intérêt torical Linguistics: Theoretical and empirical perspectives. des “groupes flottants”’. In J.M. Hombert & L.M. Hy- Stanford, CA: CSLI. man (eds.), Bantu historical linguistics: theoretical and Maho, J.F. 2009. NUGL Online: The online version of the empirical perspectives. Stanford: CSLI Publications, New Updated Guthrie List, a referential classification of the pp. 149-164. Bantu languages (4 June 2009). Online file: http://goto.glo- Bastin, Y., Coupez, A. & Mann, M. 1999. Continuity calnet.net/mahopapers/nuglonline.pdf and Divergence in the Bantu Languages : Perspectives Nurse, D. & Philippson, G. (eds). 2003. The Bantu lan- from a Lexicostatistic Study. Tervuren: RMCA (series « guages. (‘Routledge language family series’, no. 4). London/ Annales de Sciences humaines, série in-8°», no. 162). New York: Routledge. Bleek, W. 1851. De nominum generibus linguarum Africae Nurse, D. and G. Philippson. 2003. Towards a histori- Australia. Bonnae: Formis Caroli Georgii. cal classification of the Bantu languages’. In D. Nurse & G. Blench, R. 2006. Archaeology, Language and the African Philippson (eds.), The Bantu Languages. London/New York: Past. Lanham: Altamira Press. Routledge, pp. 164-181. Bostoen, K., Clist, B., Doumenge, C., Grollemund, R., Schadeberg, T.C. 2003. ‘Historical linguistics’. In D. Nurse Hombert, J.-M., Koni Muluwa, J. & Maley, J. 2015. ‘Middle & G. Philippson (eds.), The Bantu Languages. London/New to Late Holocene Paleoclimatic Change and the Early Bantu York: Routledge, pp. 143-163. Expansion in the Rain Forests of West Central-Africa’. Cur- Vansina, J. 1995. ‘New Linguistic Evidence and the Bantu rent 56: 354-384. Expansion’. Journal of African History 36: 173-195.