Many readers have probably noticed that the manuals of historical method which deal with verbal societies are primarily concerned with the sources available and the application of a critique to them. This is true for McCall's or Gabel and Bennett's works on Africa. But what is to be done with the sources once they are ready for evaluation remains vague. How does one reconstruct the past? How does one explain, or eventually interpret, history? Of the two works mentioned, only the first pays some attention to the question of “historical synthesis.’ McCall lists three possibilities: (a) that the sources support each other; (b) that they contradict each other; and (c) that they have no common reference or meeting point. This last situation is the most common in African history and indicates merely that not enough is known and that eventually new data could lead to new interpretative situations-either (a) or (b). The manual stresses that sources should be classified by discipline so that comparison of sources yields either confirmation or contradiction, with obvious and known data reinforcing the validity of the result. Once this is achieved it would seem that the job is finished, except for the warning that historical reconstruction requires a certain type of mind: imaginative yet disciplined.

Yet the job is not finished. By comparing we have only established the degree of validity of reported events or situations. We have only verified how the observation, to borrow a term from the scientific experimental method, is correct. The impression remains that historical research is fairly mechanical: to find sources, subject them to a critique, assemble them. Reconstruction follows, with suitable use of imagination. That is the craft. Yet anyone who works with historical materials knows that that is not the practice of the craft. Josephine Tey's novel The Daughter of Time features a police sergeant who more nearly exemplifies historical practice–he guesses, ponders, backtracks, and finds sources almost by intuition. If he had made a few more mistakes he would have been a recognizable historian at work. A recent volume, The Historian's Workshop, though impressionistic, also yields a more realistic picture. In the real world historians start out with a hunch, an idea which leads them to an interest in documents or in oral traditions. Then the data suggest what Popper calls a historical interpretation – “untestable points of view.” The practitioner feels that the interpretation is not enough. It should be doubted and controlled by reference to more data until the point is reached at which no more control is possible. Then the historian feels satisfied with the result–even though it still remains an interpretation, because there remains the selective point of view implicit in the idea that initiated the research.

Alfons Vermeulen published a novel about his life in a Congo town as a trader in 1933 and also left drafts about the same subject for his memoirs when he died in 1965. This essay systematically compares both documents using the rules of evidence and concludes that although it does not appear to be so on the surface the novel distorts the evidence he adduces considerably more than the memoirs. At the same time though both documents are equally deeply affected by the spirit of the time of their composition. Hence the combined evidence of both documents together is not only richer but also more reliable than their separate use.

Despite their uninviting aspect, dictionaries make for engrossing historical reading. Seemingly dull compilations conceal an ample description of a society and its culture at a given moment in time. Each entry in the original language constitutes a bit of information. For African languages the amount of information varies between one thousand or fifteen hundred entries in the shorter vocabularies to as many as eighty thousand for the fullest dictionaries. Anything less than a thousand words may be considered to be only a wordlist, the longest of which rarely exceeds five hundred items. Most dictionaries seem to contain between six thousand and thirty thousand entries. The extent of information is astonishing because the form is usually so condensed that it consists only of a gloss as translation of the term given. The best dictionaries contain explanations of the gloss, references to relevant publications, and examples taken from everyday speech as well as from literary usage, and they begin with a short grammatical introduction. Unfortunately, this kind of dictionary has been and remains all too rare; usually the gloss is unnecessarily laconic. Even so, the most exciting as well as the most trivial information is to be found in these tomes. For example, we learn from the 1652 Kongo dictionary (7200 entries) collected or recopied by van Geel, both that the Kongo were matrilineal–a fact stated nowhere else for the old kingdom–and that playing cards was a popular pastime.

The special features of written vernacular history as a specific category of African historical documentation still await a general theoretical analysis. This article makes no attempt to remedy the deficiency, but considers two possible hypotheses from the relationship between Xhosa traditional historians and the Lovedale Press during the 1930s. First cf Vansina, that it is not only oral traditions which are affected by their mode of transmission. Second, cf Goody and Watt, that it is one thing to be literate, but quite another to find a publisher.

Perhaps the first printed work in Xhosa was that of a stoic-looking cow bestriding the legend “All cattle come from God,” which appeared in 1823. The writer was Rev. John Bennie of the Glasgow Missionary Society, and the printing was done at the Chumie mission station, shortly to be renamed Lovedale. From that time, Lovedale remained the focal point of the literate Christian culture which emerged among the Xhosa of South Africa's Eastern Cape. This primacy was reinforced in 1915 when the South African Native College (now Fort Hare) was established nearby under the chairmanship of the Principal of Lovedale. The Lovedale Press flourished along with its host institution. The only available estimates indicate that up to January 1939, 238 books were produced in Xhosa, more than in any African language except Swahili.

Interest in the question of Bantu expansion rose dramatically in the 1950s as historians, archeologists, and anthropologists all joined in the fray. This reflected both the rise of Africa in world affairs and the expansion of research in general. The scholars involved were typically a new breed of professionals, and as such more dependent than their predecessors on universities or research institutions. The School of Oriental and African Studies in London achieved overwhelming dominance from about 1950 until the late 1960s, so that opinions held by its staff found the widest audience. The new scholars also were, for the most part, anti-racist, sympathetic to African nationalisms, and of liberal or socialist persuasion. They tended to reject the notion of “conquest,” believing in gradual change rather than abrupt cataclysmic mutation, perhaps because they were repelled by their recent experiences during the war. As had happened earlier, these extraneous circumstances left a deep imprint on the speculations that were now proposed. Early in this period a new paradigm almost achieved consensus, but after 1968 this fell apart and during the last decade two new trends have appeared: the single-minded quest for a new paradigm and the search for better understanding through the study of analogous processes, coupled with a more radical skepticism.

In 1754 the newly-arrived governor of Angola, António Alvares da Cunha, announced to the court in Lisbon that, inspired by plans his uncle had made in 1725, he had decided to revive the old attempt to find a way overland to link Angola and Mozambique. He was initiating this enterprise by sending, at his own expense, an expedition to the fair of Cassange, the farthest known point eastwards of Luanda, to gather information about the lands between Angola and Mozambique. The governor's envoy was Manoel Correia Leitão— an experienced trader, but one who had never been to Cassange— accompanied by a pilot, António Roiz Grizante, whose task was to fix the latitude of Cassange.

A suspicious reader could conclude with the subsequent governor António de Vasconcelos, that this venture was a slave trading expedition by Leitão to recover some debts. But it was financed by the governor and one suspects that the whole enterprise was in fact a joint business venture, disguised as an exploration, because governors were forbidden to trade. Be that as it may, the reaction from Lisbon to this news was unexpected. The governor was ordered to abandon his plans and to recall his envoys because this enterprise was too “delicate.” Knowledge about it was to be shrouded in absolute secrecy and a report about this was to be treated as top-secret. Whereupon the governor recalled Leitão.

Despite its very visible career in anthropology and folklore, structuralism has been little used by historians of precolonial Africa. Only Ronald Atkinson has applied the method of Lévi-Strauss in the Edmund Leach variant, although a number of historians have attempted to elucidate symbolic meanings by other means. Rather surprisingly as well, given the two decades or so that have elapsed since Lévi-Strauss developed its axioms and analysis, no historian of Africa has ventured to discuss the validity of structuralism for coping with the interpretation of myths of origin or other oral traditions, except in passing. The topic has surfaced only here and there in the never-ending debate about traditions as expressions of the present, or of the past, or of both. Given the influence of structuralism elsewhere, though, it is due time that the approach be discussed for its own sake.

The reticence to do so became especially incongruous when a senior structuralist, Luc de Heusch, began to cover ground that historians had recently trod. In his Le roi ivre he discussed at length myths in the kingdoms of southeastern Zaire and adjacent areas. This did prompt publication of two articles about his Luba and Lunda interpretations, but no general assessment of this work in toto. Jeffrey Hoover faulted de Heusch's Lunda material but still praised his “provocative ideas” and the method, “which bore some good fruit,” while Thomas Reefe prefaced his critique of Luba material by calling the book “stimulating” and sidestepped the issue by noting that “no matter what the final assessment of this book will be by historians…” Others were equally bland in their references to this work, while still refuting de Heusch on specifics. Everyone felt, it seems, that a general assessment was beyond or outside their competence. Yet a general critique would have been of use for de Heusch is one of the oldest and most experienced structuralists in anthropology, perhaps the first disciple of Lévi-Strauss. Trained in Paris, he imbibed the approaches of the Griaule school, the protostructuralism of Georges Dumézil, and the early teaching of the master himself. Of all structuralists he remains the most faithful to the method of Lévi-Strauss.

Between 1590 and ca. 1610 the English sailor Andrew Battell lived in Central Africa, first in Angola until 1606/07 and then in Loango. His reports about these lands are a priceless source for the otherwise poorly documented history of Angola between ca. 1590-1606, especially since his is the only known eyewitness account about the way of life of the notorious Jaga. He actually lived with one of their bands supposedly for at least twenty months (26-27). In addition his account is also one of the very earliest about Loango. Hence modern historians of Angola and Loango have relied extensively on him. They all, myself included, have used the text edition by E.G. Ravenstein of The Strange Adventures of Andrew Battell of Leigh (London, 1901) and did so without referring back to the original documents. These are, first Battell's information in Samuel Purchas' Purchas His Pilgrimage, or Relations of the World and the Religions observed in all Ages and Places discovered from the Creation unto the Present (London, 1613), and later, the more detailed “The Strange Adventures of Andrew Battell” in Samuel Purchas His Pilgrimes (London, 1625), also known as Hakluytus posthumus after its frontispiece. Given the absolute reliance of modern scholars on Ravenstein, it is worthwhile to evaluate its reliability compared to the original publications.

Did Africans once independently invent the smelting of metals or did they obtain this technology from Europe or the Middle East? This continues to be an unresolved and hotly disputed issue, mainly because the dates for the earliest appearance of smelting in Africa south of the Sahara remain inconclusive. All the earliest sites in Western and West-Central Africa from Walalde in Senegal to the Tigidit cliffs and Termit in Niger, the firki plains south of lake Chad, Taruga, and perhaps Nsukka in Nigeria, Ghwa Kiva (Nigeria), and Doulo (Cameroon) in the Mandara mountains, Gbabiri (Ndio district) in the Central African Republic, and a few sites in Rwanda, Burundi, and Buhaya cannot be dated more closely than between 840 and 420 BCE. Greater precision is impossible because the C14 curve runs flat during these four centuries, hence all these sites yield the same date. (Alpern, Killick, Me Eachern, Holl, Jézégou/Clist, Kanimba Misago). If the earliest “real” dates fell before 800 BCE, they would support independent invention, while later dates strengthen the case for borrowing. Still, this information does tell us that ironworking was adopted in the northern parts of West and West -Central Africa and in the region of the Great Lakes within the span of a mere four centuries.

The emergence of ironworking must have left linguistic traces in the relevant terminology irrespective of whether it spread by borrowing or by independent invention—hence historical linguistics can contribute to this debate. That approach is best tested by an examination of the relevant vocabulary in Bantu languages because the historical study of those languages is further advanced than that of any other language family in Africa (Nurse/Phillipson). Moreover Bantu-speakers occupy a large portion of the continent.

In the initial paragraph of a set of book reviews dealing with a sixteenth-century relation and a map by Ptolemy, Michel Doortmont informs us casually that:

Quite a statement! How arrogant and how gratuitously insulting to all those who do study “ancient and early modern history.” If Doortmont were alone in his views, one might just as well ignore this as an example of regretable idiosyncrasy. But the cavalier way in which he delivers his opinion suggests that he merely voices a truism, i.e., an opinion that he thinks is shared by most of his colleagues. Moreover, such a statement will surely frighten budding scholars away, which in itself is a good reason not to let such a sweeping condemnation of the study of earlier African history pass without comment.

Four main claims are made here: (a) a dichotomy and a contrast exists between those who who study the more remote past and those who study the recent past; (b) the first group is small and out of the mainstream; (c) its research yields disappointing results; (d) its scholarship lacks seriousness and method—with (b) and (c) of course in implicit contrast to the second group.

Unesco's General History of Africa (henceforth GHA) is now nearing completion—seven out of the eight volumes having been published in English, at least—and the time is ripe to draw attention to its role in the historiography of African history, a role very different from that played by the Cambridge History of Africa with which it is often compared. In certain ways the Unesco project has been a unique venture, and not just in African history but, in general, because it broke with old established practice. The work was not guided by one or two editors but by a large committee. History-writing by committee seemed not only distasteful but impossible to achieve to many, both because of the practical difficulties involved and because it seemed incredible that so many editors could agree on a common text, without falling into sheer banality. Now that the volumes are out, readers can judge for themselves. Many among them actually wonder how exactly these volumes and chapters were created. Because I have been an active member of the committee since its creation in 1971 and of its bureau since 1983, I can provide a general answer to this question. But the time has also come to draw attention to the records generated by this project. Future researchers will find a huge mass of papers involving hundreds of historians of Africa that touch on practically all aspects of Africa's historiography between ca. 1965 and today.

Around 1850 the peoples of central Africa from Duala to the Kunene River and from the Atlantic to the Great Lakes shared a common view of the universe and a common political ideology. This included assumptions about roles, statuses, symbols, values, and indeed the very notion of legitimate authority. Among the plethora of symbols connected with these views were the leopard or the lion, the sun, the anvil, and the drum, symbolizing respectively the leader as predator, protector, forger of society, and the voice of all. Obviously, in each case the common political ideology was expressed in slightly different views, reflecting the impact of differential historical processes on different peoples. But the common core persisted. The gigantic extent of this phenomenon, encompassing an area equal to two-thirds of the continental United States, baffles the mind. How did it come about? Such a common tradition certainly did not arise independently in each of the hundreds of political communities that existed then. However absorbent and stable this mental political constellation was, it must have taken shape over a profound time depth. How and as a result of what did this happen? Is it even possible to answer such queries in a part of the world that did not generate written records until a few centuries ago or less?

This paper addresses this question: how can one trace the social construction of such a common constellation over great time depths and over great regional scale? All the peoples involved are agriculturalists and the political repertory with which we are concerned could not easily exist in its known form outside sedentary societies.

One wonders what Fernand Braudel and the school of the Annales have done to become a kind of Trojan Horse for the wholesale condemnation of the historical value of oral tradition. Yet they are the banner raised by W.G. Clarence-Smith in a recent article in his journal to preach jihad against its historical value. Clarence-Smith claims that the historiographical revolution effected by Annales has resulted in the definitive exclusion of oral traditions from the halls of Clio. Oral traditions are at best ambiguous “signs” about the past and are very much of the present. They lack absolute chronology and they are selective, so away with them. If they be worthy of attention at all, let anthropologists and sociologists be concerned, save in a few rare instances where a historian wants to check on some European printed source. And even then, caveat emptor. Significantly, the article is not just the expression of the views of one person; rather it is symptomatic of much of the criticism which has been leveled at oral tradition, mostly by fasionable anthropologists. And it brings this criticism to its logical conclusion.

But first a word about Braudel, the Annales, and oral tradition in general. The Annales School was founded by Lucien Febvre and Marc Bloch before World War II. Fernand Braudel is its most distinguished exponent. His major theoretical pronouncements can be found in his Ecrits sur l'histoire, a collection of articles reprinted and published in 1969. This and his two major historical works should be read by those who want to know more about his views and ways of dealing with history. The basic tenets that members of the Annales School hold is that the history of events is but the spray of past developments; other time depths tell us more about the waves of the past. There is the time of the conjoncture, the trend, and the even longer time periods -- sometimes many centuries long -- the longue durée or long term. Successful history writing does not liminate the study of events, but analyzes them against the movement of these longer and deeper-running trends.

A forged travel account reminds me of a raffia palm in central Africa, because there is a use for every part of such a palm: the wine (sap), the nuts (edible), the raffia (for textiles), the other leaves (for roofcovering), the branches (for furniture), its pith (for making various articles), and lastly the grubs inside the pith (also edible). Nothing is wasted. In the same way a forged travel account can be deconstructed until all its parts down to the very last sentence or proper name can be used as evidence for one or another kind of history. The considerable interest fraudulent travel accounts can have for the historian of Africa is usually far underrated because once they are exposed as forgeries they tend to be summarily dismissed and henceforth to be avoided like the plague. At most, it is conceded that sometimes part of a forged account rests on the author's observations and experiences at the time and in the place where his (the known forgers seem to have been all male) narrative placed them and may therefore actually be genuine.

The usefulness of forgeries as evidence goes well beyond this, however, and rests on two arguments. First, a narrative forgery is never totally the product of a person's imagination, if only because it strives to achieve the verisimilitude required to be passed off as genuine. A good part of any such forgery must therefore rest on valid observations made by someone, somewhere. If one can discover from where and when such elements stem, they add new evidence to the record about that where and when. Secondly, the very choice of topics and themes; raised in a forgery is historical evidence in its own right, for it tells us much about the expectations of both the social milieu in which the work was written and its intended audience at the time (not always the same social aggregates). To develop and illustrate these points, there may well be no better instance than the notorious book whose unmasking raised a great geographic furor in the earlier nineteenth century—the notorious Douville forgery.

The historian of pre-nineteenth century Africa…cannot get far without the aid of archaeology.

Nevertheless, historians have good reason to be cautious about historical generalisations by archaeologists and about their own use of archaeological material…: it would be a rash historian who totally accepted the conclusions of Garlake and Huffman with the same simple-minded trust as I myself accepted the conclusions of Summers and Robinson.

In the beginning, historians of Africa put great store by archeology. Was its great time depth not one of the distinctive features of the history of Africa, a condition that cannot be put aside without seriously distorting the flavor of all its history? Did not the relative scarcity and the foreign authorship of most precolonial written records render archeological sources all the more precious? Did not history and archeology both deal with the reconstruction of human societies in the past? Was the difference between them not merely the result of a division of labor based on sources, so that historical reconstruction follows in time and flows from archeological reconstruction? Such considerations explain why the Journal of African History has regularly published regional archeological surveys in order to keep historians up to date.

Historical tales are the most abundant and the sources most used to reconstruct the history of the kingdoms of the area between the Great Lakes. This is especially true for the history of the Nyiginya kingdom in Rwanda, where such tales, preserved at the court as well as by local people on the hills, are even more abundant than anywhere else. It is not surprising then that they form the bedrock on which authors have built their reconstructions on the history of that kingdom. Yet little attention has been paid to a general critical examination of these tales. Here, and elsewhere in the region, their contents have generally been accepted as credible, after the arbitrary erasure of all references to passages judged to refer to miracles, after the arbitrary dismissal of the bits and the variants that do not conform to one's preferred version, and after either the exclusion of local “provincial” narratives, or as happened after 1960, the exclusion of all those that stemmed from the court. Such practices will simply not do.

Because these tales form the bedrock of the history of the kingdom the Institute for Scientific Research in Central Africa (IRSAC) under my direction instituted a large collection of such sources between 1958 and 1962, and made them available in the original and in translation first in depots at Butare and Tervuren and later in 1973 on microfilm. Surprisingly enough, this collection as well as even major editions of other sources, have been completely ignored by most scholars working after 1960.

The later precolonial history of a vast area in west central Africa between the Kwango and the Lubilash rivers starts with—and is dated by—the tradition of exodus of Kinguri and his companions from the heartland of the Lunda commonwealth. For the last two decades, however, several scholars have claimed that this tradition is merely a later addition to the older body of the traditions told by a dozen or so different peoples in west central Africa. Yet so far no one has examined where and when and how the Kinguri exodus tradition could have grafted itself onto the traditions of so many peoples over such a vast area. If true, this claim also requires a radical revision of the accepted history of western Lunda expansion. To examine the claim and its consequences is the aim of this article, which begins with the earliest written report of the Kinguri's exodus story.

From time to time one encounters a situation where a wisp of a source yields evidence of momentous consequence for the understanding of a whole society in the past. Such a situation raises the question of how much evidentiary weight a slender source can credibly bear, for when the evidence is so slender or tenuous, the slightest misunderstanding or misinterpretation can lead to hugely magnified errors in historical reconstruction. At the same time, the more slender the evidence, the more tempted researchers will be to interpret it according to their own a priori vision of the past in question. This is a genuine problem which can only be overcome by finding more relevant evidence.

Such situations naturally crop up wherever sources (written or otherwise) are very rare, so that every scrap of evidence must be scrutinized and exploited with the greatest care and to its fullest extent, a common circumstance in African history in relation to written documents before ca. A.D. 1450 and documents at all times about topics such as, for instance, women's history or specific techniques of agriculture that were normally not recorded. The situation then, while unusual, is frequent enough to warrant a more detailed discussion of one such instance to illustrate what the problem is and how it can be tackled. This illustration will be the case of oratoi, a single word in the Periplus Maris Erythraei.

The Periplus is justly famous in African historiography. Known from a single manuscript now in Heidelberg and held to date from the early tenth century and a much later copy from that now in the British Museum, written in a mixture of classical and vulgar Greek by either a Greek living in Egypt or an Egyptian, and firmly dated to between AD 40-55, the Periplus contains the first, albeit succinct, information about the inhabitants of the East African coast, more than half a millennium before other comparable written mentions occur.

Any mention in a text to something demonstrably more recent than its attributed date of composition is an anachronism. The presence of an anachronism is absolute proof that the purported date given for the redaction of the final the text is wrong and that the real date occurred later than the date implied by the anachronism. In rare cases an anachronism can also occur in reverse, namely, when it can be shown that the anachronism implies the presence of something (an archaism, a practice) which had been replaced or had disappeared by the purported date for the manuscript. Anachronisms can be very useful both to date a document and in some cases to show us something about the dynamics involved in the creation of a text. The detection of anachronisms is common with regard to written documents, but they can also be used for the study of oral documents which have been memorized word by word, especially when only one version of such documents has survived. That is the case for the Rwandan esoteric code of kingship, which yields an excellent example of how a systematic search for anachronisms throws light on such a document and allows a historian to use its contents with much greater confidence than was the case otherwise.

Ubwiiru is the name given in Rwanda to a set of eighteen pieces in prose, called “roads” or “ways,” which vary in length between 74 and 1252 lines and were learned by heart since “immemorial times” by specialists called abiiru.

Thirty-four years ago David Henige launched History in Africa (hereafter HA) at a time when scholars often cut corners in their rush to construct a history of Africa, and disregarded rules of evidence, thereby running the risk that many of their reconstructions would prove to be unsound. The question was not that these scholars were wholly indifferent to methodology, but that the precolonial history of the continent was the cynosure of the field at the time, and hence that all eyes were turned towards the use of oral sources to overcome the perceived scarcity of written sources for that period and to provide voices from the continent. In their haste to fill huge voids in the story of Africa's past, scholars debated the rules of evidence in relation to such unconventional sources. They often disregarded almost every methodological canon when it came to written data. Crucial differences between primary and secondary sources were ignored, archival research was scanty, new editions of older publications were mere reprints accompanied or not by new introductions that were so uninformed as to be useless, while issues about authenticity, authorship, chronology, or translation were all brushed aside as quibbles. Thus, in the days before 1974, methodological concerns focused exclusively on oral tradition and oral history to the detriment of everything else. As its initial editorial made clear, HA was launched as a forum where scholars interested in method could publish articles about all the facets of the historical method—from epistemology to heuristics, rules of evidence, and historiography. The journal was founded and the contributors came.