A bantu ghost in Uralistics?

It is probably not necessary to call attention to the debate that is presently going on in Uralistics, a debate centred around the question of how the origin of the Uralic languages is to be interpreted. One can easily, at least when analysing the arguments of one of the parties in the discussion, get the impression that there is a backward traditional school, still clinging to one or another branch of the old language family tree, and that these traditionalists and their views are challenged by a group of radicals to whom the very concept of a language family tree is wrong and totally inadequate in the description of the origin of the Uralic languages. Instead, these radicals or rebels, as they like to call themselves, propose a different explanation according to which (if I have understood it correctly) the Uralic languages originate in a lingua franca, i.e. an auxiliary language. The basic assumption seems to be that this lingua franca was used as a means of communication by different groups of people speaking different languages and that these languages mixed or at least converged with each other to form simplified and modified lingua francas which were later to become dominant languages through processes of language shift. It also seems as if the radicals want to make the claim that speakers of Uralic languages while shifting to other languages, e.g. Germanic languages, are responsible for, i.a., phonological and morphological peculiarities in these languages. The ideas are to a very large extent influenced by the theories of S. G. Thomason and T. Kaufman (1988) according to whom so-called "mixed languages" exist. But whereas S. G. Thomason and T. Kaufman (1988 : 3) apparently are of the opinion that "there are indeed mixed languages [--] but most languages are not mixed", one can not escape the eerie feeling that the radicals in Uralistics seem to want to propose an explanation according to which unmixed languages do occur, but most languages are mixed.

If one browses through the papers included in the volume "The Roots of Peoples and Languages of Northern Eurasia II and III" (2000) one will find that these views are expressed over and over again, in one paper after the other, and this is not the only volume of its kind. The concept of an original Uralic lingua franca is obviously very appealing and, as A. Kunnap (2000 : 142) puts it, "the idea is spreading and has gained more and more supporters". The supporters of this linguistic theory sometimes represent other scientific disciplines, i.a. archaeology, genetics and even political science. The notion of a group of radicals or revolutionaries challenging the establishment originates in the way these scholars describe themselves and their work. K. Wiik (2000 : 202) is straightforward enough: "We are approaching a decisive breakthrough in the ways we think about the emergence of new languages" and according to A. Kunnap (2000 : 27) even "the [scientific] paradigm [in Uralistics] is changing just now under the leadership of some Uralistic rebels".

In principle, this is all well and acceptable. Anyone is entitled to come up with new ideas and test these ideas against the old and accepted ones. The proposal of new views and the scholarly debate around them are in fact the essential ingredients in any scientific discipline. And, just as the old ideas should rest on carefully gathered and evaluated material, so should the new ones. It is also probably a necessary prerequisite for the radical or revolutionary to be a trained scholar in the linguistic subdiscipline for which he or she is proposing totally new ideas. This is where things can and do become difficult, especially when evidence and skills are scant. If we take the idea of the Uralic languages originating in an auxiliary language, a lingua franca, as an example it is obvious that very little has actually been presented to support this idea, rather it is the idea itself that is constantly being put forward, not the evidence. The difficulty in presenting solid evidence becomes evident in Taagepera 2000. R. Taagepera, when asking the question of how a lingua franca comes about and spreads, states the following: "We can not observe the prehistorical ones, but we can look at the presently expanding ones" (2000 : 384). It is of course very convenient for anyone proposing a lingua franca origin for the Uralic languages that the earlier processes are not observable: there is no evidence, but there is no counterevidence either. The discussion of an original Uralic lingua franca has other limitations too, and generally the argumentation seems to concentrate more on the shortcomings of the old (and inadequate, according to the rebels) comparative method than on pure linguistic evidence from within Uralistics.

But if Uralistics has little or nothing to offer in support of an original Uralic lingua franca, what can be done to achieve a decisive breakthrough? One option is to seek for evidence in other languages and when done carefully and with proper training, such comparisons could possibly bear fruit. In times of revolution, however, the brains one has easily go to one's head. In a paper titled "Uralic as a lingua franca with roots", R. Taagepera, whose main qualification, according to himself, "is a near-total lack of qualification as a linguist" (2000 : 382) does his best to find a presently expanding lingua franca to support the general idea and eventually comes up with Swahili. This is poor performance, to say the least.

R. Taagepera (2000 : 384) defines a lingua franca in the following way: "[--] sometimes a mix of disparate languages emerges, combining words from various sources and simplifying the grammar down to the lowest common denominator (and then possibly starting to build a new grammatical structure). This is what I mean here by a lingua franca in a narrow sense. Swahili is a present example."

The main characteristics of Swahili, according to R. Taagepera (2000 : 384), are the following: "Swahili combines simplified and modified Bantu grammar with a heavily Arabic vocabulary. While only about 3 million speak it as their home language, it serves as an almost daily public language for 30 million people in East Africa. At least as an occasional means of communication it is spoken by a total of 50 million. Swahili is the official state language in Tanzania and one of the several official ones in Kenya, Uganda and Congo (Zaire). On the East African coast, it dates back at least to 1700s and most likely to much earlier times, but its spread inland has accelerated during the 1900s. As it spreads, it also develops local variations that may become permanent dialects [--]"

Recently, I had an opportunity to spend some time in Kenya to study Swahili (1). I am sorry--or when it comes to the idea of an original Uralic lingua franca rather glad--to say that I in no way can share the view expressed, although not explicitely, by R. Taagepera that Swahili is a mix of disparate languages combining words from various sources and simplifying the grammar down to the lowest common determinator. And I might add that I am not alone. G. S. P. Freeman-Grenville (1974 : 70) puts it in the following way: "One finds in many writers of earlier times the concept that Swahili is a hybrid language, a mixture of Bantu speech with Arabic. This is an old ghost which has now been laid to rest" (see also Gerhardt 1995 : 25).

Apparently, this ghost is now back to haunt Uralistics, or at least to be used as evidence for an original Uralic lingua franca. R. Taagepera bases many of his conclusions about Swahili on A. M. Mazrui and I. N. Sharif (1994). They have proposed a pidgin origin of Swahili, an opinion that presently stands out in isolation. Even A. M. Mazrui and I. N. Sharif themselves remark that "linguistic evidence for this genesis of Swahili is not easy to find" and that "the process of decreolization tends to obliterate the peculiar features of pidgins and creoles" (1994 : 67). This idea makes it possible to propose a pidgin origin for any reasonably unknown language and then say that no evidence remains because of decreolization, or, for that matter, to propose a lingua franca origin for an entire group of languages, in this case the Uralic languages. A. M. Mazrui and I. N. Sharif (1994 : 67) claim that the hypothesis would explain Swahili's substantial portion of Arabic words, its lack of a tonal system characteristic of virtually all other neighboring African languages, and the fewer distinctions in its concord system.

However, if R. Taagepera had taken the time to consult other sources, he would have been able to take other opinions about the origin of Swahili into consideration. D. Nurse and T. Hinnebusch (1993) have demonstrated that Swahili is a language in the Sabaki group of the Bantu languages and that it has developed fairly regularly from Proto-Sabaki, not from the mixing of an unknown Bantu language with Arabic. It is true that Swahili has a large amount of lexical items originating in Arabic, but since 80% of the words of Arabic origin in Swahili are present also in Malay (see Pearson 2000 : 43 for further references) it is possible to speak about "a corpus of traveling Arabic words", mainly terms connected with mercantile or religious activities. The presence of a large amount of originally Arabic lexical items is, by the way, characteristic of all languages spoken by muslims, e.g. Persian and other Iranian languages, and Hindi/Urdu and other Indic languages. According to D. Nurse and T. Hinnebusch, "the majority of words in Swahili taken from Arabic are relatively recent and most likely derive from the recent Omani period of influence on the coast" (1993 : 315). The structural influence of Arabic on Swahili is very limited and little studied (see Lodhi 1994; 2000a; 2000b).

The notion that Swahili combines a simplified and modified "Bantu grammar" with Arabic lexical material is hard to grasp. There are hundreds of Bantu languages in the southern, southeastern and eastern parts of Africa and although the strong resemblances between the individual languages indicate a fairly recent and rapid spread across the continent from an original heartland in present-day Nigeria and Kamerun, the Bantu languages do differ from each other in several respects. The idea of a specific Bantu grammar is as valid as that of a typical Uralic grammar. And exactly how "simplified" and "modified" is to be interpreted as yet. Swahili has a number of noun classes, often but not always using different prefixes for the singular and plural categories. Swahili has two different past tenses and a number of morphological means to express causativity etc. According to D. Nurse and T. Hinnebusch, the tense-aspect system of Swahili did undergo a huge simplification historically, but "this is also true of all NEC [North East Coast Bantu; A.H.] languages and can be assumed for PNEC [Proto-North East Coast Bantu; A.H.] and therefore be assigned to the early 1st millenium A.D.; long before any putative influence from Arabic" (Nurse, Hinnebusch 1993 : 309). Simplification is thus not something caused solely by the haphazardous mixing of Bantu grammar with Arabic, but rather a process that can take place in any kind of language development and is not limited to the emergence of pidgins and creoles.

Swahili is fairly easy to learn and for this reason it can serve as a common means of communication in many parts of East Africa. It does, as R. Taagepera correctly assumes, indeed date back to much earlier times than the 1700s and as a matter of fact poetry in Swahili was published already in the preceeding century. Swahili was in all likelihood the language of the Swahili city states, African muslim societies dominating the East African coastal strip and the African side of the Indian Ocean trade until the arrival of the Portuguese in the late 15th century. Swahili was a prestige language and the ability to speak and write Swahili was held in high esteem. Since then, the socio-economic status of the Swahili has changed but their language remains the national language of Tanzania and one of the official languages in Kenya. It is, however, a language in the Sabaki group of the Bantu languages, lexically influenced by Arabic because of close religious and mercantile ties with the Arab world. That Swahili was restricted to the East African coast until the end of the last century is due to the fact that Swahili society was a coastal mercantile society, orientated more towards India and the Arabian peninsula rather than to the interior of Africa. The spread of Swahili inland during the 19th century has many causes among which are escalating slave trade, German language policies in Tanganyika and the choice of national languages for recently independent colonies.

Swahili is not to be taken as evidence for an original Uralic lingua franca in this guilt-by-association way, simply because it does not originate in any pidgin or creole language. Arabic can be said to play the same role in the development of Swahili as it does in many of the Turkic languages, e.g. Tatar, where many religious and economic terms of ultimately Arabic origin can be found. In many cases, the same terms exist also in Swahili, simply because of the fact that Arabic through Islam has had a vast impact on both Tatar and Swahili. The study of Swahili, therefore, could be of interest to anyone studying the impact of the Turkic languages on the Uralic languages. Originally Arabic lexical items and syntactic structures (such as the Mari conjunction ja ... ja ... 'either ... or ...') can and have spread to Mari as a result of its contacts with Tatar, and the same goes for many Bantu languages in East Africa today which are currently being influenced by Swahili. Studies of these current processes in Africa can shed light on previous processes in, for instance, the Volga-Kama area of Russia. It is in this way Swahili is of use to Uralistics: properly described by trained scholars with an up-to-date view on the language in question, cooperating with scholars in other linguistic subdisciplines. Old ghosts of the past have no place neither in Uralistics, nor in Bantuistics. If R. Taagepera, the ardent supporter of the lingua franca idea, had taken the time to consult other sources than Mazrui, Sharif 1994, he would probably not have fallen for the temptation to use Swahili as evidence for alleged processes in the Uralic languages. If the "decisive breakthrough" in Uralistics is not to turn into a decisive breakdown or fall-through, the self-appointed rebels will have to refrain from presenting old ideas about other languages as fresh evidence.

REFERENCES

Freeman-Grenville, G. S. P. 1974, Some Aspects of the External Relations of the East African Coast: before 1800. - Foreign Relations of African States, London (Colston Papers XXV), 69-83.

Gerhardt, L. 1995, Swahili-Sprachgeschichte.--Swahili-Handbuch, Koln (Afrikawissenschaftlicher Lehrbucher, Band 7), 25-39.

Kunnap, A . 2000, About Some Morphological Features of Proto-Uralic.--The Roots of Peoples and Languages of Northern Eurasia II and III, Tartu 2000 (FU 23. Historia Fenno-Ugrica), 27-32.

--2000, On Linguistic Relationships.--The Roots of Peoples and Languages of Northern Eurasia II and III, Tartu 2000 (FU 23. Historia Fenno-Ugrica), 137-144.

Lodhi, A . Y. 1994, Arabic Grammatical Loans in the Languages of Eastern Africa.--University of Trondheim. Working Papers in Linguistics 22, [Trondheim], 60-74.

--2000a, Arabic Loans in Swahili.--Orientalia Suecana XLIX, Uppsala, 74-81.

--2000b, Oriental Influences in Swahili. A Study in Language and Culture Contacts, Goteborg (Orientalia et Africana Gothoburgensia 15).

Mazrui, A . M., Sharif, I. N. 1994, The Swahili. Idiom and Identity of an African People, Trenton, New Jersey.

Nurse, D., Hinnebusch, T. 1993, Swahili and Sabaki. A Linguistic History, Berkeley.

Nurse, D., Spear, T. 1985, The Swahili. Reconstructing the History and Language of an African Society, 800-1500, Philadelphia.

Pearson, M. N. 2000, The Indian Ocean and the Red Sea.--The History of Islam in Africa, Oxford-Cape Town-Athens, 37-59.

Taagepera, R. 2000, Uralic as a Lingua Franca with Roots.--The Roots of Peoples and Languages of Northern Eurasia II and III, Tartu 2000 (FU 23. Historia Fenno-Ugrica), 381-395.

The Roots of Peoples and Languages of Northern Eurasia II and III, Tartu 2000 (FU 23. Historia Fenno-Ugrica).

Thomason, S. G., Kaufman, T. 1988, Language Contact, Creolization and Genetic Linguistics, Berkeley.

Wiik, K. 2000, European Lingua Francas.--The Roots of Peoples and Languages of Northern Eurasia II and III, Tartu 2000 (FU 23. Historia FennoUgrica), 202-236.

ANDRE HESSELBACK (Uppsala)

(1) I am grateful to the Wallenberg Foundation in Sweden for providing the means for field work in Kenya.