The Nena: Preliminary Notes On History, Ethnography And Language

THE NENA

Preliminary notes on language, history and ethnography Martin Walsh #1, May 1998

Introduction The Nena are an enigma. In the last quarter of the nineteenth century a group of people of this name were encountered in the north-east Livingstone Mountains by two different European travellers. The English explorer Joseph Thomson stumbled across them in 1879 (1881: 248-249), and they were met again some 18 years later, in 1897, by the German missionary-priest Alphonse Adams (1899: 106). Their accounts leave no doubt that they met the same group of people, who were reported to be remarkable for their distinctive physical appearance, unintelligible language, and comparatively primitive existence. Thereafter, however, the Nena virtually disappeared from trace in the ethnographic record. Friedrich Fülleborn, who travelled almost the same route (in reverse) as Adams, a few years later, did not hear of the Nena: it must be said, though, that he does not appear to have been looking for them at the time. When he later wrote his pioneering ethnographic survey of the region, Das Deutsche Njassa- und RuwumaGebeit, he admitted that he had been unable to find any other mention of the Nena in the literature besides the references by Thomson and Adams (1906: 438). This silence on the Nena continued through to the mid-1970s, when the priest-turned-anthropologist Hans Stirnimann reported that Nena were still to be found, albeit in a very different part of the Livingstone Mountains, in the north-west of Upangwa (1976: 41-42). Unfortunately, Stirnimann gave no further details beyond theorising why Nena had been found by earlier travellers in a quite different location. This paper represents an attempt to reopen research on the Nena. In writing this first draft I have had to rely heavily upon Fülleborn’s discussion of the original sources: unfortunately I have not had access to complete copies of the relevant sections of either Thomson’s To the Central African Lakes and Back or Adams’ Im Dienste des Kreuzes (‘In the Service of the Cross’). In addition to Stirnimann’s observations, however, I do have access to the results of more recent research in the region, including my own work on the Sangu and neighbouring peoples, and this provides additional evidence and interpretations which were not available to these writers. Location Our first problem is to locate Thomson and Adams’ Nena on the map. Fülleborn also grappled with this problem: “According to Adams, the Wanena live to the east of Pangulidala mountain, in the vicinity of Mwagwama village; Thomson also came upon his Wanena after he had arrived at a village

1

called Mwagwama; so nothing remains but to assume that the Wanena of Adams are the same as those of Thomson (see the strikingly similar descriptions of both authors): admittedly Thomson’s Mwagwama is shown about 20 km north of Pangulidala on older maps; however, following recent maps, it must be identical with that of Adams.” (1906: 438, my translation)

Thomson’s spelling of the village name was ‘Mwangwama’, not ‘Mwagwama’ (see below); nonetheless it seems reasonable to assume, as Fülleborn did, that both Thomson and Adams were both referring to the same place, or at least the same general area. Pangulidala mountain is not shown on large-scale maps of the region (a widelytravelled Pangwa speaker I asked suggested that the name might refer to a location which was, for ritual reasons, forbidden to women – avadala in Pangwa and other Southern Highlands languages). Fülleborn himself, however, marched over Pangulidala on his way from Tandala to Kidugala, and describes it as a striking tableshaped mountain which projects from the southern edge of the Elton (now Kitulo) Plateau (1906: 429). The name Mwagwama or Mwangwama does not appear on modern maps either. The rather similar name Wangama, however, does appear as a village name, in roughly the right place. There are in fact two villages called Wangama in the area of Kidugala, one to the east and one to south-west. The latter is clearly the most likely candidate for identification with the earlier Mwagwama / Mwangwama. Thomson, who was travelling westwards, crossed what he calls ‘the Ruaha gorge’ and climbed to the ‘mountain top’ on the other side before reaching his Mwangwama (see below). This ‘Ruaha’ was evidently the Mbarali river (one of the principal sources of the Great Ruaha) or one of its tributaries which rise in the eastern Livingstone Mountains. The main western tributary of the Mbarali cuts through the edge of these mountains between Kidugala and Wangama (the eastern Wangama lies on the Ubena Plateau), and it was probably this tributary which Thomson crossed. The modern Wangama village is the headquarters of an administrative ward of the same name, part of Imalinyi division in Njombe district. According to Cribb and Leedal, Wangama is 14.5 km from Kidugala: the Kipengerasi Escarpment, noted for its wild flowers, is reached by climbing up into the hills a further 8 km to the west of Wangama (1982: 29). Wangama is located on the eastern fringe of a large tract of virtually uninhabited mountainous terrain, a historical ‘no-man’s land’ between different ethnic groups. While we cannot be certain that the modern Wangama is precisely where Thomson and Adams encountered the Nena (given the propensity of village names to migrate and replicate), this would certainly appear to be a likely place in which to find members of a small, isolated, possibly ‘remnant’, ethnic group. Our second problem is to locate Stirnimann’s Nena. In his own discussion of earlier reports he simply notes that the Nena are “a group which is still today represented in north-western Upangwa” (1976: 41). This presumably refers to a location south of the Kilondo river, which forms the traditional boundary (as recognised by the British colonial administration) between Ukinga and Ubena in the north and Upangwa in the south, and the modern boundary between Makete and Njombe districts and Ludewa district. Stirnimann elsewhere refers to this boundary region as Utilili, noting that it harbours numbers of immigrants from Ukinga (1976: 22-23). Modern maps show a village called Utilili in this area, the last village on the road north from Mlangali into

2

Makete district (administratively Utilili being part of Lupanga ward in the Mlangali division of Ludewa district). Although we cannot be sure that this is precisely the area meant by Stirnimann when he referred to ‘north-western Upangwa’, it does seem to be the most likely candidate. Like the Wangama area, this would also appear to be a likely place in which to find people who have retained an ethnic affiliation from the precolonial past. Another early traveller, W. P. Johnson, did indeed find a marginal group of people in this boundary area. Johnson travelled beyond the southern boundary of Ukinga in 1883: “The next day [after camping ‘by a tall peak called Pekwawa’] we came to an end of these [Kinga] villages. All the Wa-kinga valleys are very beautiful and fairly populated; they said they have also lower lands, where their cattle are often sent. To the west the hills rise higher, rounded and peaked to the north-east. On the south they are bounded by the Lamkiu, which flows south-west into Nyassa. Between this and its small tributary the Nyelele we found the remnants of an oppressed people, and then ascending a mile or so we found the Wa-palameno, who gave us a wine distilled from the bamboos of the country. Crossing two more streams we came to many villages, and the next day ascended the eastern face of the Livingstone range again by a mount called Unsuti. Near the top was a village, and goats were there. From the top a lower terrace was visible to the west; on the east the hills rose as far as we could see to the north-east, but fell away to the south. We descended some five miles to a village; Nyassa was still hid by the dense fog. Even here marauders were dreaded, but whether Wa-bena or Angone [Ngoni] we could not tell – probably the latter.” (1884: 530)

It is tempting to suggest that Johnson’s ‘remnants of an oppressed people’ were the Nena, and the ‘Wa-palameno’ (who he also calls ‘Wa-pelememo’) the northern Pangwa. Unfortunately none of the topographic names mentioned by Johnson appear on large-scale maps of the area, and consequently it is difficult to trace his route. It may be that he was travelling to the east of the modern Utilili. This is a question, however, which cannot be settled at present, though it would certainly be worth further research. Population We do not know if the Nena still exist as a discrete community in the north-east Livingstone Mountains. Thomson described them as a ‘small tribe’ in 1879 (see below), and the fact that they were only otherwise encountered and reported on by Adams 18 years later suggests that they were a very small group, even at this time. We can only guess at their later population history, and whether these Nena ‘died out’ or were simply absorbed by their neighbours. The area in which the Nena were found at the end of the nineteenth century is still only very thinly populated. The whole of Wangama ward registered only 2,668 people in the 1978 census, and population was projected to grow at the low rate of only 1% per annum. Population density was, at less than 12 persons per km2, among the lowest in Njombe district. The northern part of Ludewa district in which Stirnimann’s Nena presumably live (see above) is rather more densely settled. In 1978 Lupanga ward had a total population of 4,825, a mean density of over 30 persons per km 2 (United Republic of Tanzania 1978: 4-5, 14-15, my calculations of population density).

3

Ethnographic classification Joseph Thomson had quite definite views about the racial status of the Nena he met in September 1879. His description of them is reproduced in full below: “After crossing the Ruaha gorge, and reaching the mountain top on the other side, we camped at the village of Mwangwama, inhabited by another small tribe called the Wanena. If the Wapangwa looked of an unusually low type the Wanena were infinitely lower. Here, indeed, were sufficient marks of degradation to arouse the enthusiasm and scientific delight of ethnologists and those who seek after connecting links between apes and men. Such miserable specimens of humanity I have nowhere seen in Africa. Even the Wakhutu [the Kutu of eastern Tanzania] looked intelligent and manly beside them. Their heads are very small, and their skulls are so narrow as to suggest the idea that they have been pressed so when young. The two upper incisors are remarkably large and prominent. Their language is different from any of the surrounding tribes, and we found it quite impossible to communicate with them except by signs. The men wear a small piece of hide over the shoulder, and the women’s dress is reduced to a bunch of grass. They live in the most wretched hovels, beside which a hole in the ground would be comfortable. These hovels are conical in form, eight feet in height, and seven feet in diameter. They are built of a few inclined sticks tied together at the top, and then rudely thatched with grass, leaving a hole eighteen inches high to serve as a doorway. My men would have been ashamed to have run up such a miserable structure in half an hour, and yet whole families huddle themselves like so many pigs in them, and must necessarily lie on each other. I observed no domestic utensils, though doubtless they have, at least, an earthenware cooking-pot. Their chief food is peas, at least that was the only kind of diet we could get. Of these they eagerly brought basketfuls to barter for a strip of cotton an inch or two inches broad. They are exceedingly timid.” (1881: 248-249)

Immediately before reaching the Nena, Thomson had travelled through the territory of the Bena, followed by that of a group of ‘Pangwa’. He described the Bena as a ‘subtribe’ of the Sangu, who ‘submit to whoever is master of the field’, and these Pangwa as another ‘tributary tribe’, remarkable in his eyes for their extremely dark complexion, general ugliness, and an affliction which left most men with a squint and many with no sight at all in their left (and only left) eyes (1881: 246-247). After leaving the Nena, he travelled on to the country of the Kinga, who, he was told, had lost their former unity and fallen under the rule of the Sangu, and then the Hehe (1881: 252, 272). Thomson’s description of regional politics was reasonably accurate and can be corroborated from other sources: in late 1877 the Hehe had decisively defeated the Sangu and supplanted them as the principal power in the region, if not as direct rulers over the Kinga and all of the Bena (Johnson 1884: 528-530). Fülleborn, however, questioned Thomson’s anthropological statements and suppositions about the ‘Pangwa’ and Nena, describing the people he had studied in the northern Livingstone Mountains (the Wanji, Kinga / Mahanzi, and Mawemba) as being far from physically degraded: “…on the contrary they are really strong, healthy people, who thrive excellently in their mountains when their plundering neighbours leave them in peace” (1906: 440, my translation). Stirnimann also roundly rejects Thomson’s ‘pseudo-Darwinian’ description as applied to the modern Pangwa and Nena: “As for this unjust characterisation of the Pangwa and Nena, it must be remarked from more recent field investigations that if Thomson’s statements have any real grounding, they can on

4

no account be taken to refer to the actual Pangwa and Nena living in the southern Livingstone Mountains. Neither the Pangwa nor the Nena are extraordinarily black. With good reason they even disapprove of being called black (ntitu), and really dark-coloured girls have difficulty in finding husbands. Apart from a few isolated exceptions of very dark colour, one meets people of all shades from earth-brown to copper-red. Like other hoe-cultivators, they are almost entirely nourished on carbohydrate-rich vegetables, and most Pangwa are on average of gracile, slight stature. Neither are they conspicuous for an inclination to squint, nor do they frequently suffer from left-side blindness…” (1976: 42, my translation)

Fülleborn pointed out that Thomson’s ‘Pangwa’ should not be confused, as some authors had done, with the Pangwa living in the southern half of the Livingstone Mountains; ‘Pangwa’ in this context simply being a general name given by the Bena to their mountain-dwelling neighbours (1906: 438, 440). Instead, these western neighbours of the northern Bena were identified by Fülleborn with the people more widely known as the Mawemba: “The people who live to the south of the Wabuanji [Wanji] of the Elton Plateau on the eastern side of the Livingstone Mountains were – at least to their western neighbours, the Kinga – known as “Wamawemba”. Personally I only know the northern part of the Wamawemba mountains, which I crossed on the Tandala-Kidugala road…; from the new map, however, it can be inferred that Wamawemba territory stretches southwards as far as Ukinga does and that its southern section belongs to the realm of Bejera [the Bena chief Mbeyela]; in general the Mawemba mountains seem to be only very thinly populated. With the Mawemba we must also count those referred to as “Wanena” by Thomson and Adams.” (1906: 438)

Fülleborn clearly believed that the people called ‘Pangwa’ and ‘Nena’ by Thomson and Adams were really Mawemba. This would explain why Fülleborn himself, not to mention others, had failed to come across them. It does not explain, however, all of Thomson’s observations about the Nena. While it is easy to discount Thomson’s physical anthropology as a product of personal and epistemic bias, and identify his ‘Pangwa’ with the Mawemba; it is more difficult to explain away other reported facts about his Nena, or even explain why Thomson should have distinguished between them and the ‘Pangwa’. Could these Nena have been Mawemba, as normally understood? We know very little about the Mawemba, who appear to share borders with the Kinga, Bena and Pangwa ‘proper’, and who, ever since the British applied their policy of ‘Indirect Rule’ have tended to be counted (and have counted themselves) together with one or other of their more prominent neighbours. As George Park notes, the Mawemba have not been mapped since early German times; indeed, Fülleborn himself was the last writer to have attempted this. Park suggests that the Mawemba are intermediate in language and culture – as well as location - between the Kinga and Bena, and can readily hide behind either identity (1988: 171-172). Whatever the case, it has never been suggested that Mawemba language and culture are anything other than a local variation upon regional themes; an observation which sits awkwardly with Thomson’s characterisation of the Nena and their language in particular. Stirnimann, while rejecting Thomson’s descriptions of the Pangwa and Nena, accepts that they were indeed Pangwa and Nena – and therefore not Mawemba, though he does not spell this out. He concludes instead that they must have been refugees from further south, in other words from the area of northern Upangwa in which Pangwa and Nena

5

are still found today (1976: 39). I will discuss this thesis in more detail below. It still does not resolve the question of who the Nena actually were, and how they might be classified ethnographically. Stirnimann is extremely scathing about Thomson’s descriptions of both Pangwa and Nena, which he only quotes from selectively and without indicating that he has done so. In particular he omits Thomson’s statement about the language of the Nena, which, as we shall see, is not as readily dismissed as his opinions about their position on the evolutionary ladder. Language We know nothing more about the language of the Nena – past or present – than the little that Thomson and Adams tell us and Stirnimann implies. This is hardly sufficient to confirm that the Nena once possessed their own language (as opposed to a dialect of one of the known languages of the region), let alone speculate on its genetic affiliation. Nonetheless, the nature of the early descriptions, together with a consideration of wider linguistic and anthropological circumstances, invites some speculation, and I have not hesitated to follow this course in the discussion which follows. Linguistic data It is particularly difficult to account for Thomson and Adams’ statements about the language of the Nena, unless we assume – as Stirnimann seems to have done in the case of Thomson – that they were grossly mistaken. Thomson’s statement about the speech of the Nena is particularly intriguing, and is worth quoting again: “Their language is different from any of the surrounding tribes, and we found it quite impossible to communicate with them except by signs.” (1976: 249)

Adams also noted that the language of the Nena was completely different from that of the Bena, their eastern neighbours (1899: 106; as cited by Fülleborn 1906: 443). This does not square with what we know about the speech of the Mawemba, nor indeed of any other group of people in this general region, the linguistic geography of which might be best characterised in terms of a series of interlinked dialect continua (if informants’ statements about the mutual intelligibility of neighbouring languages are taken at face value). Unfortunately Stirnimann says nothing about the language of the contemporary Nena of north-west Upangwa: we can assume that if he knew it to be significantly different from the speech of their Pangwa neighbours, he would have said so. Linguistic classification There are a number of different possible explanations for Thomson and Adams’ observations about the speech of the Nena they encountered in the north-east Livingstone Mountains: (1) They were mistaken, and they or their interpreters exaggerated the difference between Nena and Bena and other languages they encountered en route across the Southern Highlands.

6

(2) They were right in marking Nena as different from surrounding languages, but this may merely have been a function of the fact that the Nena, as a dislocated group, were simply speaking a language (a dialect of Pangwa?) which local interpreters were not familiar with. (3) The Nena did indeed speak a language which was markedly different from any other in the region. The last of these three possibilities is the most tantalising. All of the known indigenous languages of the Livingstone Mountains and surrounding region (northwards and eastwards) belong to a single genetic sub-group of Eastern Bantu, christened ‘Southern Highlands’ by Derek Nurse. The principal members of Southern Highlands in Nurse’s classification are Bena, Hehe, Sangu, Wanji, Kinga, Kisi, Pangwa, and Manda (1988: 40-43, 46-48, 92). The classification of Southern Highlands has not yet advanced to a stage where the different ‘dialects’ of these ‘languages’ have been taken into account – the present list of members merely reflects the gross ethnographic classification which ‘Indirect Rule’ produced, and will no doubt be modified in future when more detailed research into local ‘dialects’ and supposedly intermediate forms (like Mawemba) is undertaken. It is most probable that the language of the Nena, past and present, belongs somewhere within this genetic grouping. One interpretation of Thomson’s statement (and our third explanation of it, above) might be that the Nena he encountered spoke their own Southern Highlands language, sufficiently different from other known dialects to be accorded its own branch on the Southern Highlands tree. The other, more tantalising – albeit purely speculative – possibility, is that they did not speak a Southern Highlands language at all, but a geographical isolate of either Bantu or non-Bantu origin. The Livingstone Mountains certainly provide the right topographical conditions for the survival of one or more language isolates into recent centuries. Harry Johnston was so impressed by the conservatism and other peculiarities of Kinga – which include a large number of unusual lexical items – that he classified it in its own group, one of only two Bantu languages to which he accorded this honour (1922: 52-53). He was similarly impressed by the language of the Nyakyusa and Ngonde, on the western side of the Livingstone Mountains: “…one of the most peculiar and interesting of the Bantu languages, one which contains a considerable number of unrelated word-roots or roots which have far-away connexions: a speech, in fact, which has evidently long been isolated in its present head-quarters, the mountain region immediately north of the north end of Lake Nyasa” (1922: 61). Nurse has also noted the presence of unique words for several common items in Nyakyusa and its western relative, Ndali (1988: 73). The sources of these unusual lexical sets on both sides of this mountainous region remain to be investigated. The most obvious candidates are extinct languages in the Khoisan phylum, for which there is already some evidence in the Nyasa-Tanganyika corridor (Nurse 1988: 74). There is also some evidence, though somewhat anecdotal, for the former presence of genetically distinctive populations in the remote mountain valleys of this area. Anthropologists who have worked among both the Nyakyusa (and their linguistic kin) and the Kinga have commented on the markedly different physical appearance of particular individuals. Monica Wilson, influenced by the traditions of her Nyakyusa speaking informants, assumed that this provided evidence for a former population of

7

pygmies (1958: 1, 8); George Park, on the other hand, assumes a Khoisan origin for the features he has observed among individual Kinga (1988: 187). These observations recall Thomson’s earlier description of the Nena. Neither Wilson nor Park can be accused of the ‘pseudo-Darwinian’ bias which both Fülleborn and Stirnimann felt had determined Thomson’s statements about the physical appearance of the Pangwa and Nena he encountered in 1879. While nineteenth century views about race and human evolution certainly coloured Thomson’s interpretation of what he saw, it does not seem far-fetched to suggest that he did observe real physical differences in the Nena of the north-east Livingstone Mountains. However, while Thomson related these and other features of the Nena to their mountain location (see below), he did not explain them in terms of the now widely accepted thesis that prolonged isolation in such an environment often results in genetic and cultural (including linguistic) conservatism. Did this thesis – or does it still – apply to the Nena? In the absence of further evidence, we can only speculate on the truth of Thomson’s assertion that the Nena spoke a distinctive language. If they did, then we can only guess at its affiliation. As we shall see below, some of the available historical (including historical linguistic) evidence hints at the possibility that the Nena were (and are) surviving representatives of an early, possibly pre-Bantu, population in the Livingstone Mountains. Corroboration or otherwise of this possibility, however, must await further research – the most obvious place to start being an investigation of the language or dialect which contemporary Nena and/or their close neighbours speak. History We know two certain things about the history of the Nena: a group of people known by this name were settled in the north-east Livingstone Mountains at least between 1879 and 1897, and people using the same name were to found further to the south-west, in north-western Upangwa, more recently, during the period of Stirnimann’s research. Although this does not amount to much, it does suggest a number of possible historical scenarios: (1) The Nena of the north-east Livingstone Mountains and their namesakes to the south-west are (were) members of quite different ethnic groups which were not as closely related as their name implies. The only hard evidence for a historical connection between the two groups is their ethnonym, and we do not know if this was the self-applied name of the group in the north-east or simply a name by which their neighbours knew them. We may recall here Füllerborn’s thesis that this is how Thomson and Adams’ Pangwa got their name, leading to a confusion in the literature between two distinct groups of people: the ‘Pangwa’ (according to Fülleborn, really Mawemba) of the north-east Livingstone Mountains and the Pangwa proper in the southern half of the same range. (2) Both groups of Nena are (were) closely related historically and share (or shared) a common origin, as evinced by continuing use, either by themselves or others, of the same name. If this is the case, then the apparent geographical separation of the two known groups of Nena might be explained in one of a number of different ways:

8

(a) Both groups are the surviving remnants of a once more widespread group of people, who formerly occupied a continuous stretch of country which included the the northern part of the Livingstone Mountains. The isolation of two sub-groups of Nena at the edge of their former range, and/or in the thinly inhabited borderlands between the ethnic groups which now dominate this area, might then be explained as a result of the intrusion and expansion of the latter, absorbing and marginalising the Nena in their midst. (b) The two sub-groups are direct descendants of an original group of Nena which both dispersed, at some time in the past, from an earlier homeland. Such a homeland might have been located centrally, in a discrete part of the northern Livingstone Mountains between the areas subsequently occupied by the two subgroups. On the other hand, it might have been outside of this mountainous region. Their migration from this homeland might then be explained by factors similar to those already cited in (a) above (the same also applies to the two scenarios which follow). (c) The original homeland of the Nena was in the north-east Livingstone Mountains, where Thomson and Adams found them. The Nena described by Stirnimann, however, comprise a splinter group which migrated away from this area. Given that we have no certain evidence that Nena were present in north-west Upangwa in the last quarter of the nineteenth century, it is possible that they migrated there some time after 1900. Given that we have no evidence for the continuing presence of Nena in the north-east in the twentieth century, it is even possible that they all migrated to the south-west. (d) The homeland of the Nena is in the area of north-western Upangwa where a group of them still lives today (or at least were living in the mid-1970s). The Nena of the north-east Livingstone Mountains are (were) a splinter group which had migrated from the south-west sometime before Thomson encountered them in 1879. This, as we shall see below, is one explanation suggested by Stirnimann. These scenarios might be modified in detail and selected elements of different scenarios combined. Each of the four previous authors who has contributed to our knowledge of the Nena has also presented a theory of their origin, explicit or implicit. Of these, only Stirnimann had the advantage of knowing that there was more than one group of people in the Livingstone Mountains referred to by this name. In the following section I will examine each of their arguments in turn and the evidence they adduce in favour of them, adding my own observations and comments where relevant. Theories of origin Thomson followed his contentious description of the Nena with a passage in which he outlined a general theory to account for their apparent ‘degradation’, along with that of other ‘mountain tribes’ in Africa: “It is somewhat strange that while in almost every other part of the world the mountain tribes are the bravest and noblest, in Africa it is quite the reverse. Whenever inhabitants are found occupying mountains in Africa, they are almost in every case some miserable remnant of a tribe, few in numbers, cowardly in the extreme, degraded in mind and body, and living worse than brutes. This may be accounted for on the supposition that no African will take to the mountains till he is compelled to do so to save his life. In the lowlands nature supplies him abundantly with food in return for the minimum of work. The climate is warm and equable, and necessitates no clothing. On the mountains it is different. There he must work hard for

9

his living; he is subjected to the greatest extremes of heat and cold, and all his surroundings are wretched. Hence no tribe will ever betake itself thither till on the verge of annihilation; and then it retires, broken up and disorganized, with no unity and no spirit…” (1881: 250).

Although he does not say so directly, Thomson appears to have believed that the Nena were a group which had been forced up into the mountains, in effect refugees from conflict in the lowlands below. Did he have any direct evidence for this, or was he merely extrapolating from what he knew about the history of inter- and intra-ethnic conflict he had heard about and witnessed en route to the Nena? Whichever the case, Thomson’s theory falls in line with a version of our scenario (2b), implying that the Nena had originally migrated from a homeland outside of the Livingstone Mountains. This theory would probably attract little more attention but for the fact that Adams provided a more detailed version of it, supported by a tradition he apparently recorded from the Sangu, who live in the plains to the north of the Livingstone Mountains. Before his own journey across the Livingstone Mountains, Adams spent some months in Madibira, in the north-eastern corner of Usangu. On more than one occasion he met with the then Sangu chief, Mugandilwa Merere, whom the Germans had brought from exile in the west with a possible view to making him paramount ruler over the whole of Hehe territory as well as his own ancestral lands. Adams does not tell us the source of the following tradition, though it presumably came from one or more Sangu informants: “Under King Merere Mgogollo (Merere, the Old, who will be referred to as Merere I) who lived in the fifties and sixties, the Wasango built a town in the open savanna near to the confluence of the Nendembela and Great Ruaha, which they fortified with a palisade of hedges and called Utengula# (= capital, residence). Merere Mgogollo now attempted to extend his kingdom; he fought successful wars against the Wabena and drove the Wanena and Wapangwa out of their ancient home on the Mbarali river, forcing them to retreat with their flocks of goats and sheep into the high, almost inaccessible Livingstone mountains, where they have fought a battle for survival until the present day. The Warori or Wasango established their rule as far as the highlands of Ussafa, a fertile region north of Lake Nyassa.” (1899: 85; pp.227-228 of an unpublished translation in the possession of Alison Redmayne)

Although the chronology and other details might be corrected, most of the events related in this passage are corroborated by other sources on mid-nineteenth century Sangu history, both oral and documentary. Adams’ ‘Merere Mgogollo’ was the Sangu chief Mwahavanga, also called Munyigumba or variants thereof in the literature, though this name is no longer used by the Sangu themselves (it was adopted by and is more widely used to refer to the first great Hehe chief, the father of Mkwawa). Mwahavanga is the first Sangu chief about whom anything certain is known. In 1858 he was visited by traders from the coast, one of whom later gave Richard Burton an account of the event. Mwahavanga was reckoned to be 70 years old at the time, and it appears that he did not have much longer left to live. A bitter dynastic struggle followed his death, and he was eventually succeeded by Tovelamahamba Merere, his daughter’s son. Merere was certainly in power in 1866/67, when he was visited by Tippu Tip. Adams’ reference to Mwahavanga as ‘the old Merere’ (in Sangu uMelele umgugogolo) is an anachronism, probably fostered by informants wishing to gloss over the controversial circumstances of the first Merere’s (Tovelamahamba’s) accession (Adams was led to believe that he was Mwahavanga’s son).

10

It is not known when Mwahavanga’s rule began. It seems that he began to raid his neighbours (including the Bena in the east and the Safwa in the west) and expand his realm well before the first appearance of the Ngoni in the Nyasa-Tanganyika corridor at the end of the 1830s. Mwahavanga’s original capital is said to have been in the west of Usangu (near the modern-day Utengule-Usangu), but he was forced to flee eastwards and take refuge in western Uhehe when attacked by the Ngoni, possibly as late as the mid-1850s. Mwahavanga returned from Uhehe to build a new capital in the north-east of Usangu, close to the point where the Utengule Swamp (as it is now called) flows into the Great Ruaha river (north of the point where the Lyandembela river, Adam’s ‘Nendembela’, joins the swamp). This Utengule (Adams’ ‘Utengula’) is where Mwahavanga was residing in 1858. The traders who visited him were travelling back to the coast from Ubena, and the Sangu initially mistook their caravan for a Bena raiding party. It is evident that Mwahavanga was embroiled in conflict with one or more Bena chiefs at this time, and probably had been so for some years (Walsh 1984: 36-41 and references therein). Adams’ reference to ‘successful wars against the Wabena’ is probably accurate, and there is some evidence that Mwahavanga installed Sangu and pro-Sangu rulers in northwestern Ubena. It is even possible that he drove Bena speakers from the south-east of Usangu, an area which they have now colonised in large numbers, a process which may well have begun after the ejection of Tovelamahamba Merere from Usangu by the Hehe in 1877, if not before. However, there is virtually no evidence to support Adams’ assertion that Mwahavanga ‘drove the Wanena and Wapangwa out of their ancient home on the Mbarali river, forcing them to retreat with their flocks of goats and sheep into the high, almost inaccessible Livingstone mountains’, especially if this is taken to mean that people going by these names once lived on the lower course of the Mbarali in eastern Usangu. While conducting research in western Usangu in the early1980s I did not record any traditions about the Nena, or even hear their name mentioned (I should add, though, that I did not enquire about them by name). Nor are the Nena mentioned in other published and unpublished traditions which I have seen, including traditions recorded in eastern Usangu. Likewise, there is no other convincing evidence to suggest that a people called ‘Pangwa’ ever lived in Usangu. Stirnimann reports that a German map published in 1894 – before Adams was in Usangu – shows ‘Pangwa’ as settled on the upper course of the Great Ruaha in Usangu (1976: 39; referring to Ramsay’s map supplement in Mitteilungen aus den deutschen Schutzgebieten, which shows the route of Governor von Schele’s 1894 ‘Nyasa expedition’). However, he rejects this as evidence in support of Adams’ assertion on the grounds that the cartographer, Ramsay, was probably influenced by a reading of Thomson’s account. Stirnimann hints at the possibility that Adams might also have been influenced by Thomson’s description of the Pangwa and Nena (and, we might add, Thomson’s theory of their lowland origin), though subsequently he appears to grant Adams’ more detailed statement some truth (see below). Taking this sceptical line of argument, we might postulate three stages leading to Adams’ insertion of erroneous information about the Pangwa and Nena into an otherwise reasonably reliable account of mid-nineteenth century Sangu history:

11

(1) Thomson describes meeting people called Pangwa and Nena on different sides of the ‘Ruaha gorge’ in 1879. He speculates that these and other ‘mountain tribes’ in Africa have been driven there after being defeated by stronger groups in their original lowland homes. (2) Ramsay, having seen Thomson’s account (or perhaps a map derived from it), mistakenly marks Thomson’s Pangwa as settled lower down on the Ruaha, in Usangu, when drafting his 1894 map. (3) Adams, who is familiar with both Thomson’s account and Ramsay’s map, infers that both the Pangwa and Nena must have been driven from Usangu to their mountain refuges, and adds this statement to his own account of Sangu history, which is otherwise based on information he gathered in 1897. This inference fits best with what Adams has already learned about Mwahavanga’s ‘successful wars’ against the Bena, so he slots it in there. Being aware that Thomson’s ‘Ruaha gorge’ was a tributary of the Mbarali river (having encountered Nena himself in the vicinity), he also adds the detail that the Pangwa and Nena originally lived on the Mbarali – though he does not specify exactly where. The most suspect part of Adams’ history is the fact that it includes both Pangwa and Nena, precisely the same two groups that Thomson discussed. Given that other sources on Sangu history in the pre-colonial period have nothing at all to say about people of this name, it seems hard to escape the conclusion that Adams had invented the story of their flight from the Sangu. He may, of course, have believed it: we have no reason to suspect that he added it to his Sangu history knowing it to be false. If we are prepared to be less sceptical, however, we might grant the possibility of some truth to Adams’ account. A lot hinges on what he meant by the Mbarali river. Both Fülleborn and Stirnimann took it to mean the lower course of the Mbarali in the Usangu Plains. However, there is nothing in Adams’ reference to ‘their ancient home on the Mbarali river’ which pins the Pangwa and Nena down to a location in Usangu, though the context might seem to suggest that this is what he meant. The Mbarali, which is often taken to be the principal source of the Great Ruaha, is formed by a series of tributaries (one of which was Thomson’s ‘Ruaha gorge’) which rise in the north-eastern block of the Livingstone Mountains. The principal tributaries meet, to the north of Kidugala, after dropping down from these mountains, and the combined flow then winds its way down a long valley towards Usangu. If Adams’ was referring to a location in this valley, along the middle course of the Mbarali, then some of our objections have less force. We would be less surprised to find other Sangu traditions failing to mention small groups of people in this valley, which lies beyond the southern boundary of Usangu and is generally considered to be part of Ubena. We also know that Mwahavanga (and later Sangu chiefs) extended their authority into this area, and it is quite conceivable that valley dwellers took to the hills when attacked or otherwise looking to escape subjection to the Sangu. Indeed, Thomson found his Pangwa and Nena in the mountains just above the tributaries of this valley: a relatively short distance for them to have fled. At present we have no way of deciding between these two different interpretations of Adams’ account, the sceptical and the more charitable. The first possibility, that Adams created his story of Nena (and Pangwa) origins, leaves us none the wiser about Nena history. The second possibility, outlined in the preceding paragraph, provides some support for a more detailed version of scenario (2b), suggesting that the Nena

12

were originally settled to the east of the Livingstone Mountains. The evidence for the latter scenario, such as it is, only applies to the Nena encountered by Thomson and Adams. We would still have to explain how Nena also migrated to the other side of the Livingstone Mountains, or admit that there is no connection between the two groups (our scenario (1)). It should be repeated, though, that the evidence for this charitable interpretation of Adams’ account is weak, and we might be accused of simply reworking Adams’ original invention. Fülleborn also questioned Adams’ account of the origin of Thomson’s Pangwa and Nena. His main target, though, was Thomson’s description of their physical and mental state and the theory he put forward to account for this: “Thomson thinks that the Wapangwa, Wanena, and also the Wakinga, were perhaps the thickset remnants of an aboriginal people, who, by staying in the high mountains, had gradually degenerated physically and mentally. If Adams’ statement is correct (?), then the Wapangwa and Wanena were ejected from their home by the Sangu, and had been living in the Livingstone Mountains for barely 30 years prior to Thomson’s time!” (1906: 440)

Readers can compare Fülleborn’s characterisation of Thomson’s theory with what Thomson actually wrote (see above): Thomson did not simply ascribe the ‘degradation’ of Africa’s ‘mountain tribes’ to montane residence, but to their status as weakened and ill-adapted refugees from the lowlands below. In any event, Fülleborn was not prepared to grant much credibility to either Thomson or Adams’ reports about these people, and, as we have already seen, assumed that they had actually encountered groups of Mawemba. To Fülleborn, then, the Pangwa and Nena of the north-eastern Livingstone Mountains did not exist as ethnic groups in their own right, but were the products of Thomson’s racial prejudice and a mistaken identity which Adams also fell victim to. From this point of view there is no local Pangwa and Nena history to explain, only the history of their reification by misguided observers. Are Thomson and Adams’ Pangwa and Nena ethnographic fictions? Fülleborn suggests that ‘Pangwa’ is (or was) a term used loosely by the Bena to refer to all of their neighbours living high in the Livingstone Mountains, though only the Pangwa proper of the southern half of the range use this as their own name. Thomson, so the argument goes, climbed up from Ubena into the Livingstone’s, and must have simply recorded the ethnonym which his Bena speaking interpreter(s) provided (1906: 438). This is possible, though we have no other evidence to suggest that Bena have ever called the inhabitants of the north-eastern part of the Livingstone Mountains ‘Pangwa’. Today’s Bena certainly seem to be in no doubt that their immediate neighbours in the northern part of the Livingstone Mountains are Mawemba (and beyond them Kinga and Wanji). Moreover, Fülleborn fails to provide an explanation for Thomson and Adams’ use of the name Nena (for one possible explanation, however, see the section which follows). Although we now know that there are people in north-west Upangwa who call themselves Nena, we would have difficulty in extending Fülleborn’s argument to say that Bena generalised this name as well, given that these Nena are not their neighbours (and that we have already accepted that ‘Pangwa’ has this extended reference). Stated in this way, it is hard to sustain the hypothesis that the Pangwa and Nena of the northeast Livingstone Mountains were both products of terminological promiscuity on the part of the Bena. This does not exclude, however, other and more subtle interpretations. Perhaps, for example, the ‘Pangwa’ really were Pangwa, and it was

13

they who referred to their neighbours at higher elevation as ‘Nena’, generalising from the situation in north-west Upangwa? Stirnimann, like Fülleborn, is critical of Thomson’s account, one of his main concerns being to refute Thomson’s ‘unjust characterisation’ of the Pangwa and Nena (1976: 42).. Despite concluding that the people described by Thomson might not have been Pangwa and Nena at all, he does present one scenario that might explain their presence in the north-east Livingstone Mountains in the second half of the nineteenth century. Stirnimann suggests that they may have been splinter groups which had fled the Ngoni raids on northern Upangwa, settling first on lower course of the Mbarali river before taking to the nearby mountains (1976: 39). No evidence is presented in favour of this scenario, though elsewhere Stirnimann does discuss the impact of the Ngoni on the Pangwa in general (1976: 35-38). If the Pangwa and Nena of the north-east Livingstone Mountains were refugees from the Ngoni, then they must have been relatively recent arrivals. The first Ngoni to reach the eastern side of Lake Nyasa were the followers of Mputa Maseko, travelling from the south. Ebner estimates that they probably arrived in (what is now) the Songea area in the early 1840s (1987: 53-55). The period of greatest disruption in Upangwa, however, began some years later, when a second group – or rather two groups - of Ngoni, under Zulu Gama and Mbonane Tawete, rounded the northern end of Lake Nyasa and established their base at Mlangali in northern Upangwa. Ebner estimates that they probably set out from the western side of Lake Nyasa in 1854 and arrived in Mlangali in 1856 (1987: 61-62). It is conceivable that some Pangwa and Nena fled northwards at this time, or possibly as a consequence of later raids. This does not, however, give them much time to settle on the Mbarali river (whichever part of the river’s course this is taken to mean) and to be forced to flee a second time by the Sangu chief Mwahavanga. If Mwahavanga did indeed force these Pangwa and Nena up into the Livingstone Mountains, then this is most likely to have happened before Mwahavanga himself had been chased out of Usangu by the Ngoni, an event which probably occurred in the early or mid-1850s (see above). Stirnimann’s Ngoni hypothesis might be salvaged if we discount Adam’s report (including his statement that the ‘ancient home’ of the Pangwa and Nena was on the Mbarali river) and think of these splinter groups as moving straight to refuges in the north-east Livingstone Mountains. This is still, however, pure speculation, and Stirnimann was perhaps wise not to pursue his suggestion further. It has to be admitted that we possess far too little information about the Nena (as well as Thomson and Adam’s Pangwa) to enable us to choose between the different theories about them presented by our four authors. At the same time it should be noted that the suggestions of Thomson, Adams, Fülleborn, and Stirnimann, do not exhaust all of the possibilities which were listed at the beginning of this section. There is, moreover, one further line of investigation which neither of these authors considered. This, as we shall see, invokes linguistic evidence which is independent of their theories. Although this evidence is also open to different interpretations, it does widen the range of possibilities and provide us with possible insights into the earlier history of the Nena. The name of the Nena

14

The only certain linguistic fact that we know about the Nena is their name – or at least the name by which others know (or once knew) them. Nena does not, as far as I know, appear as an ethnonym anywhere else in eastern Africa, nor does it have an obvious etymology (like, for example, the widespread ethnonym ‘Nyika’ and its cognates, which ultimately derives from the proto-Bantu term for certain types of savannah vegetation). One possible derivation is from the proto-Bantu verb root *-néén-, ‘to speak, say’, the Sangu reflex of which means ‘to play, dance’. However, this is no more than guesswork (the Nena dubbed as ‘dancers’ or ‘musicians’?). We might be on safer ground in treating the ethnonym as opaque. There is no reason why Nena should be derived from a known lexical source: it is possible, for example, that the name originates in the local pronunciation of an original self-applied ethnonym in a quite different (possibly non-Bantu) language. A more suggestive line of enquiry is opened up, however, when we consider another set of possible cognates in neighbouring Southern Highlands languages. These are all directional terms, which fall into two groups according to their primary meaning: Sangu Hehe

kunena kunena

‘east, the direction in which the sun rises’ ‘east’ (Mwamudemu n.d.: ‘Vocabulary’)

Kinga Pangwa

[ku?]nena xunena

‘above, up, on top’ (Johnston 1922: 484) ‘up, upwards’ (Stirnimann 1983: 193, 200)

In each of these examples the root –nena is preceded by the class 17 locative prefix, ku/ xu-, with the general sense of ‘to, towards’. This is the normal form of directional terms in the Southern Highlands languages. Although we only have a partial set – information is lacking on other languages and dialects in the group – we can tentatively reconstruct on this basis an early or even proto-Southern Highlands term *kunena (or *ukunena, with preprefix included), meaning either ‘east’ or ‘upwards’, or perhaps (though this seems unlikely) possessing both of these senses. Pangwa /x/ derives regularly from proto-Southern Highlands */k/, and it is not difficult, as we shall see, to imagine circumstances in which the original meaning of the term changed in one group of languages. If kunena was a relatively recent loanword in one or even all of these languages, then it would be more difficult (though not impossible) to explain subsequent phonological and semantic divergence. The critical question is whether and how the directional term is related to the ethnonym. There are three possibilities: The two are not related, but it is an accident of linguistic history that they appear to be cognate. It should be pointed out here that available transcriptions of these lexical items are phonologically incomplete: they do not mark tone, vowel length, or even allow us to distinguish between /n/ and /nh/ (from underlying */nt/). If the two items prove to be phonologically dissimilar, then the case for cognation will disappear. Even if they are identical, then it is possible that this is no more than linguistic coincidence. (2) The ethnonym is derived from the directional term, the original meaning of Nena being either ‘people of the east’ or ‘people who live uphill’. In this case it would not matter that *kunena could be reconstructed for proto-Southern Highlands: the ethnonym might have been derived from the direction in which they were found at (1)

15

any time in the development of one or more of the Southern Highlands languages (and could then have been borrowed by the others as an ethnonym with no particular directional connotation). It may be, for example, that the Pangwa called the Nena this because they lived higher up in the Livingstone Mountains, and that this name was then adopted by the Nena themselves and/or by other neighbouring groups without reference to their relative location. A scenario like this would be supported if Southern Highlands *kunena is found to derive from an earlier inherited form or loanword with similar directional meaning. Johnston placed his Kinga ‘Nena’ in a series of presumed cognates which occur widely in eastern Africa, and would include, for example, Sangu kushanya and Hehe kuchanya, ‘above, up’ (1922: 484). Unfortunately his inclusion of the Kinga term (for which he also gives the alternative ‘-nina’) in this series is less than convincing, unless it is interpreted as a much skewed loanword. (3) The directional term is derived from the ethnonym, the original meaning of *kunena being ‘the direction in which the Nena live’, which may have been in the east, uphill, or both. It is not unusual for directions to be given in this way. An elderly Sangu informant described all four points of the compass to me in terms of which neighbouring territory these were in the direction of, telling me that kunena was in the direction of Ubena - this is, indeed, where the sun appears to rise from in western Usangu. This scenario implies that people speaking proto- or early Southern Highlands were familiar with the Nena, perhaps before 500 A.D., according to Nurse’s estimate of the antiquity of the Southern Highlands group of languages (1988: 62, 79). This would support the theory that the modern Nena are the remnants of an ancient ethnic group, possibly non-Bantu (Khoisan?) in origin. The use of a directional term to name them might even give us some clues as to where they lived in relation to early Southern Highlands speakers. In Sangu kunena is paired with kusixa, which means ‘west, the direction in which the sun sets’. In Pangwa xunena is paired with xusixa, which in this case means ‘down, downwards’ (Stirnimann 1983: 193). Sangu kusixa and Pangwa xusixa are cognates, deriving from a proto-Bantu root whose East African reflexes generally refer to the principal rainy season. The original meaning of a reconstructed protoSouthern Highlands *kusika would therefore be ‘the direction from which the rains come’, in approximate compass terms ‘west’. This implies that Southern Highlands speakers first encountered the Nena to their east, possibly (though not necessarily) at a higher elevation. This would place the speakers of proto-Southern Highlands either on the western side of the Livingstone Mountains (where the Kinga and Wanji now live) or possibly in Usangu (if the Nena were then located on the Ubena Plateau). If early Southern Highlands speakers lived to the west of the Nena, then it is difficult to explain how they might have named the opposite direction after them. It is much easier to explain how one group of Southern Highlands speakers continued to use *kunena and *kusika to mean east and west, while another group innovated by shifting the primary reference of these terms to a vertical axis. Again, we have insufficient information to enable us to choose between these different possibilities. Our knowledge of the languages and dialects of the Southern Highlands group is extremely uneven, and properly transcribed and analysed linguistic data are in short supply. We know very little about the genetic development of Southern Highlands and its external and internal relationships, and it is to early for us to do more than guess at the location of a proto-language. Only when the relevant information has

16

been collected and analysis undertaken will it be possible to evaluate hypotheses about the past of the Nena based upon linguistic data. Ethnography Our knowledge of Nena ethnography is almost entirely limited to the few observations made by Thomson and Adams about the people of this name they encountered in the north-east Livingstone Mountains. Stirnimann has little to say specifically about the Nena of north-western Upangwa, and his account gives the impression that there is little to distinguish them from their Pangwa neighbours. Mode of Subsistence It appears that the Nena encountered by Thomson and Adams were mixed farmers, cultivating crops and herding small stock. Thomson reported that “Their chief food is peas, at least that was the only kind of diet we could get. Of these they eagerly brought basketfuls to barter for a strip of cotton an inch or two inches broad” (1881: 249). ‘Peas’ can refer to different kinds of pulses, and we can only speculate about the identification of the particular cultigen (cow-peas?) which Thomson’s party obtained. However, the relative abundance of these ‘peas’ suggests strongly that they had been cultivated by the Nena themselves. Adams’ statement that the Nena had retreated into the high mountains “with their flocks of goats and sheep” (1899: 85) indicates that they also herded small stock, or at least had done so at some time in the past. Thomson does not mention seeing any domestic animals, though the men “wore a small piece of hide over the shoulder” (1881: 249), which may or may not have been goat- or sheep-skin (see below). It seems quite likely that Thomson’s Nena were living in reduced circumstances. His description of their domestic life is worth quoting again in full: “They live in the most wretched hovels, beside which a hole in the ground would be comfortable. These hovels are conical in form, eight feet in height, and seven feet in diameter. They are built of a few inclined sticks tied together at the top, and then rudely thatched with grass, leaving a hole eighteen inches high to serve as a doorway. My men would have been ashamed to have run up such a miserable structure in half an hour, and yet whole families huddle themselves like so many pigs in them, and must necessarily lie on each other. I observed no domestic utensils, though doubtless they have, at least, an earthenware cooking-pot.” (1881: 279)

It is easier to explain the impoverished state of these Nena in 1879 as a consequence of the inter-ethnic conflicts which were raging at the time than to imagine that this description reflected normal, peaceful conditions. The apparently temporary nature of their dwellings, and their evident lack of grain stores, domestic utensils and domestic stock, all suggest that they may have been recently attacked or taken refuge from conflict. If Thomson had not mentioned their possession of quantities of pulses then we might suspect that he had stumbled across an isolated group of hunters and foragers, but this seems unlikely.

17

Unfortunately Stirnimann gives no specific information about the mode of subsistence of the Nena living in north-western Upangwa, beyond describing them together with the Pangwa as hoe cultivators (1976: 42). Dress and physical adornment The only other aspect of past Nena practice which we know anything about is their dress and appearance. Our only primary source is Thomson: “The two upper incisors are remarkably large and prominent… The men wear a small piece of hide over the shoulder, and the women’s dress is reduced to a bunch of grass.” (1881: 279)

As Fülleborn noted, Thomson’s comments about the size and prominence of the Nena’s upper incisors, indicates that they did not follow what was then common practice among the Kinga. Kinga boys usually had their four lower incisors knocked out with a wooden chisel, and sometimes also the lower half of their front upper teeth. The Mahanzi and Wanji, however, were reported not to follow this practice, though isolated individuals among them and other eastern neighbours of the Kinga (including the Mawemba?) apparently did (Fülleborn 1906: 463). Stirnimann reports that dental modification is also rare among the Pangwa. A few individuals have had their incisors filed and canines chipped off with a sharp knife by specialists from Ukisi. Pangwa informants told Stirnimann that this practice had only become fashionable among women and girls because some of them had married Kisi and Kinga (1976: 42). It differs, however, from Kinga practice as described by Fülleborn. It is possible that Johnson’s otherwise unidentified ‘Wa-palameno’ (1884: 530; see above) did normally practise some form of dental modification: their name can be roughly translated as ‘people with teeth like claws’, perhaps alluding to the treatment described by Stirnimann. Nena dress appears to have been similar to that of their mountain-dwelling neighbours. Fülleborn described men’s clothing as “peculiar and the same throughout the northern Livingstone Mountains”. They covered their genitals with a piece of rag, a small animal skin, or a bunch of leaves hanging from a string belt; while at times men also wore mantles of animal skin or – if they could afford it – calico (1906: 459). The eagerness of the Nena to barter their ‘peas’ for strips of cotton in 1897 presumably stemmed from a desire to replace the small pieces of hide which Thomson saw them wearing over their shoulders (he does not say how man covered their genitals). Although Thomson commented on the scantiness of women’s dress, this also appears to have been little different from that of their neighbours. Fülleborn later reported that “grass skirts” were worn by Kinga, Mahanzi and Wanji women (1906: 460). Conclusion These ethnographic observations suggest that in many respects the Nena may have been little different from other peoples living in the Livingstone Mountains: hence Fülleborn’s view that they were indeed Mawemba. This still leaves many features of Thomson and Adams’ accounts unexplained, including observations about their physical anthropology and language. As we have seen, we lack sufficient information

18

to decide between competing hypotheses about the origin and identity of the Nena. Stirnimann’s report that Nena can (or at least could) be found in north-west Upangwa raises the hope that it is still possible to fill in many of the gaps in our knowledge about them. It is also possible that more careful ethnographic work in and around the northeast Livingstone Mountains will add to our knowledge of the Nena. Until further research is undertaken, however, they will remain an enigma. Acknowledgements This paper is based primarily upon research undertaken by the author while studying the Sangu and other peoples of south-west Tanzania (1979-85, 1995-ongoing). I am grateful to Alison Redmayne for first drawing my attention to a number of the sources referred to, and in particular for providing me with part of the English translation of Adams (1899) which she has been in the process of preparing for publication.

References Adams, A. (1899) Im Dienste des Kreuzes, Errinerungen aus meinen Missionsleben in Deutsch-Ostafrika. Augsburg: St. Ottilien. Cribb, P. J. and G. P. Leedal (1982) The Mountain Flowers of Southern Tanzania: A Field Guide to the Common Flowers. Rotterdam: A.A.Balkema. Ebner, Elzear (1987) The History of the Wangoni and their Origin in the South African Bantu Tribes. Peramiho: Benedictine Publications Ndanda – Peramiho. Fülleborn, Friedrich (1906) Das Deutsche Njassa- und Ruwuma-Gebeit, Land und Leute nebst Bemerkungen über du Schire-Länder. Berlin: Dietrich Reimer. Johnson, W. P. (1884) ‘Seven Years’ Travels in the Region East of Lake Nyassa’, Proceedings of the Royal Geographical Society, 6 (9), 512-536. Johnston, Harry H. (1922) A Comparative Study of the Bantu and Semi-Bantu Languages (Vol.1). Oxford: Clarendon Press. Mwamudemu, Emmanuel (n.d.) Teach Yourself Kihehe, unpublished manuscript (commissioned by Geoff Fox, Mufindi) [unpaginated]. Nurse, Derek (1988) ‘The Diachronic Background to the Language Communities of Southwestern Tanzania’, Sprache und Geschichte in Afrika, 9, 15-115. Park, George (1988) ‘Evolution of a Regional Culture in East Africa’, Sprache und Geschichte in Afrika, 9, 117-204. Stirnimann, Hans (1976) Existenzgrundlagen und Traditionelles Handwerk der Pangwa von SW.-Tansania. Freiburg: Universitätsverlag Freiburg Schweiz. Stirnimann, Hans (1983) Praktische Grammatik der Pangwa-sprache (SWTansania). Freiburg: Universitätsverlag Freiburg Schweiz.

19

Thomson, Joseph (1881) To the Central African Lakes and Back: The Narrative of the Royal Geographical Society’s East Central African Expedition, 1878-80 (Vol.I). London: Sampson Low, Marston, Searle & Rivington. United Republic of Tanzania / European Development Fund (1978) ‘Demographic Data’, Appendix 1 in Annex 2, ‘Land Use Planning’, of the Regional Agricultural Development Plan, Iringa Region, Final Report. Essen: Agrar- und Hydrotechnik Gmbh. Walsh, Martin T. (1984) The Misinterpretation of Chiefly Power in Usangu, SouthWest Tanzania, unpublished Ph.D. thesis, University of Cambridge. Wilson, Monica (1958) The Peoples of the Nyasa-Tanganyika Corridor, Communications from the School of African Studies, New Series No.29, University of Cape Town.

20