Introduction^{Footnote 1}

It is true. I never met Jan Vansina. One does not need to meet, let alone interact with someone to be inspired and influenced by their colossal contributions to scholarship. However, I first met Jan Vansina’s works as an undergraduate student at the University of Zimbabwe at the turn of the twentieth century. My undergraduate major combined history with economic history and archaeology. It is in history where I was exposed to sources used in the reconstruction of the African past. Despite reading so many fine works, Vansina’s Oral Tradition as History ^{Footnote 2} stood out, not just for myself but also for my fellow students of African history. I still vividly remember how nearly impossible it was to get a copy of the book which was always reserved on the library loaning system despite the existence of strong alternatives. No doubt, Vansina’s methodological rigor and very persuasive but carefully constructed arguments contributed to this popularity.

The long-term perspective and wide-angle approach adopted in historiography and historical methods provided a rich comparative gist for situating knowledge production about Africa, inside and outside the continent, in a world context. As a student of archaeology and anthropology, I developed a unique ability to combine techniques from different disciplines to study Africa’s past and “present” over the long term. And even more important, I grew up in a rural setting in southern Zimbabwe participating in, contributing to, and observing daily routine and constantly shifting situational practices amongst the southern Shona people. This exposed me to their daily decision making, subsistence activities, cultural and religious practices as well as to governance systems. Fortuitous as this was, such a disciplinary and cultural background impressed upon me the advantage of studying the African past from a combined source perspective.

Although I later followed the archaeology trajectory, and focused more on deep history, I never lost engagement with later periods of African history. Such an academic itinerary involves overlaps, ruptures, and tensions between method and interpretation, particularly when it comes to developing analogues from the recent past and applying them to interpret the archaeology to reconstruct deep histories. This is also the methodology and approach that Vansina tellingly adopted with very strong effect – it allowed him to make an exceptional inter-generational impact on studies of the African past.^{Footnote 3} Necessarily therefore, in appraising the contributions of Jan Vansina to studies of Africa’s deep histories, this discussion initially focuses on tensions between sources and “the rule of evidence,” before emphasizing the need to continuously review generally accepted historical narratives in the face of new evidence and new ideas. In the process, the contribution will where possible, highlight some limitations associated with select works by Vansina, particularly those that intersect with my own research interests and knowledge fields.^{Footnote 4}

When Ethnographies Become Histories, and Histories Become Ethnographies

Inevitably, generations of scholars on Africa’s deep history are always confronted with the problems and opportunities presented by the available sources.^{Footnote 5} Between the early to mid-twentieth century, the writing of the African past was compromised by the mistaken belief that history could only be written in contexts with documentary sources which excluded a large part of the continent from history.^{Footnote 6} And yet, “people without history,” particularly those in Africa whom others imagined bereft of a past had rich forms of oral literacy that include traditions that were transmitted over generations. In their multiple forms that also include myths and legends, oral sources contain valuable information on the societies that produced them just as the history of words and languages. For the last five hundred years, oral sources contain within constraints, insider information on governance systems, economic organization, and social relations of communities such as those belonging to very large groups such as the Bantu. However, oral sources like written records, are not impervious to challenges and limitations. While oral traditions may not have been created for the purposes to which historians now put them, they can be easily distorted and are prone to chronological telescoping. Vansina recognized this and developed a robust and credible methodology for applying oral sources to the study of African history.^{Footnote 7} In a continent once derided for not having a history, Vansina’s contribution to the use of oral sources as a tool for studying the African past, alongside that of others, was game changing for outsiders. And, given the political interests such outsiders often had in the continent at the time these debates swirled, their opinions about the status of Africans’ practice of history mattered very much as part of the ground for colonial subjecthoods/subjection.^{Footnote 8} It is only fairly recently that Africa-based Africans have taken some control over the research agenda, translating method to their own ends.

Being the astute scholar, Vansina quickly observed that while existing oral and documentary sources were a formidable source for studying Africa’s recent past, their short temporal coverage limited their utility in exploring the deeper past.^{Footnote 9} This brought into his cross hairs the discipline of archaeology whose staple are the material and non-material remains that bear traces of past human activities. Such material signatures include remnants of settlements, debris from technological practices such as copper smelting, residues of crops and animals, and much more. While archaeological evidence is available in a fragmentary form and can be studied in multiple ways including scientific analysis to explore issues such as resource extraction and distribution and population movements on the landscape and multi-scalar contact, interpretation remains a major problem.^{Footnote 10} How might people in the present imagine how settlements that were occupied over a thousand ago functioned? This requires not just controled guesswork but carefully constructed hypotheses that must be critically assessed against the data and other possibilities.

Often, this motivated the development of analogues from the ethnographies and histories to build ideas that were evaluated against the evidence from the past. This process of analogical reasoning is unavoidable in archaeology. And yet, the use of analogies drawn from recent periods to study communities in the distant past is one of the most problematic if not most controversial aspects of archaeology.^{Footnote 11} By drawing interpretive leads from the present, are researchers not exposing themselves to the risk of anachronism? In his study of political centralization in Central Africa, Vansina was well aware of this challenge.^{Footnote 12} He observed that while the use of analogies drawn from sources dating to the recent past to interpret the distant past was unavoidable, it was prudent to consider the context in which the embedded knowledge was produced. For example, colonialism wrought many material changes in African communities which makes an uncritical extrapolation and mapping of leads from recent times into the distant past highly problematic.^{Footnote 13}

A very powerful example that readily comes to mind is the concept of tribes that features heavily in twentieth century anthropological documents and colonial administrative records. Colonial administrators who were keen to establish governance systems, which in their own biased views were better suited for natives, often created tribes that were previously non-existent and assumed that such tribes were static, monolithic entities.^{Footnote 14} In southern Africa, Herbert Chimhundu makes it explicit that Sotho and Tswana, languages with related phonology were transcribed with different orthographies by early missionaries which created separate tribal identities for populations that spoke them.^{Footnote 15} Similarly, the colonial government of Southern Rhodesia supported the Southern Rhodesian Missionary Conference which divided the Shona of Zimbabwe into tribal groupings such as Karanga, Zezuru, Ndau, Kalanga, and Korekore, labels that previously did not exist.^{Footnote 16} However, these colonial language projects relied on the behind the scenes, often unacknowledged contributions of a group of literate Africans (e.g. catechists) who worked as sources of information.^{Footnote 17} While the motivation for their participation varies and was often undisclosed, missionaries and colonial experts invariably treated them as informants and not as producers of knowledge. It may be therefore doubtful that the “catechists” and other African “readers” either agreed with or validated the outcomes of these language projects which they contributed to voluntarily or involuntarily, given the power imbalance in the colony and its skewed knowledge production structure. This means that using such invented ethnic labels to study the deep past in the region bound by the Zambezi and Limpopo rivers where the people now known as Shona were home to is highly problematic. Christopher Ehret and Margaret Kinsman suggest that a comparison of the development of the Shona language through time dissipates any “language differences” engendered by colonial inventions and therefore partially solves the problem caused by colonially invented language labels.^{Footnote 18} However, as Chimhundu powerfully shows, most non-language experts loosely extrapolate colonial tribes onto decorations of pottery made a thousand years ago to identify ancestral Zezuru, Karanga, Manyika, and so on, as if unaware of these controversies.^{Footnote 19} For example, the archaeologist Thomas Huffman used such tribal labels to argue for an eleventh century south to north migration which brought to the Zimbabwe plateau, Shona speakers who made a distinctive pottery tradition that he labeled Kutama.^{Footnote 20} This Kutama tradition was subdivided into regional clusters such as the southwest based Leopard’s Kopje (Kalanga), the southern based Gumanye (Karanga), and the northern based Harare (Zezuru) and Musengezi (Korekore) traditions. The generally accepted narrative for the history of Shona people over the last thousand years rests on this rather tenuous framework. Consequently, there are many reasons why such an association between invented tribal labels in the colony and designs on pottery dating to a thousand years ago must be viewed as problematic but the most important is that holding constant colonial tribal classifications for nearly a millennium years sidelines the very complicated and dynamic histories of multi-scalar and recursive mobilities that characterized African societies of all periods.^{Footnote 21} David Beach shows that in the nineteenth century, the northern plateau based Rufura people moved southwards to settle in the south central, while the Njanja moved westwards from Sena (modern day Mozambique) into the east-central region, and so on.^{Footnote 22} To account for such layered complexities, Vansina strongly recommended a rigorous “adherence to the rules of the evidence,” which involved a considered source evaluation and criticism, accompanied by an understanding of the bigger sampling issues and nested contexts in which the knowledge was produced and transmitted.^{Footnote 23}

Vansina’s use and criticism of the application of ethnographies to the study of the deep history of political centralization in Central Africa raises the question: are histories different from ethnographies? If history is the study of the past, then with each second, hour and day of lived experience passing, history is made, and altered while for the living, life goes on. To Vansina this exposed the dubious nature of the concept of ethnographic present which was neither a fixed nor a static temporal entity where western anthropologists were separated from their subjects of study as far as the progression of their societies was concerned.^{Footnote 24} The ethnographic present represented a “moment” when the evidence was generated that lies beneath a particular time, place, description, event, etc. Elucidating such factors was the first step in applying the rules of evidence to “ethnographic” evidence.

Further insights may be obtained from considering Johannes Fabian’s concept of time and the other.^{Footnote 25} The central thread of Fabian’s thesis is that during the colony, the subjects of ethnographic observation and enquiry, were sometimes denied coevality with the colonial masters. This was exploited by the colonials whose minds were firmly etched with the simplistic and arrogant idea that historical change was progressive, transitioning from smaller, de-centralized political formations to the centralized arrangements which colonials assumed as the pinnacle of their narrative. With this supremacist perception of how societies developed through time, colonials defined and in some cases cast African cultural “traditions” as timeless, in a biased colonial setting, where Africans and their societies were inferior relics from bygone eras. While this was not the view of Africans themselves, who as subjects (of the colony and of study) did not fully participate in knowledge production, the thinking underwrote the colonial conceit of civilizational “development” and/or denial of equal rights.^{Footnote 26} Of course, nothing about this formulation was correct or without profound contradictions and wilful blindness, and everything about it was indispensable to the ideology of colonial rule. Ethnographic analogy based on this denial of coevalness, is unhelpful and inappropriate. However, it must be acknowledged that there is nothing in analogy that inherently prescribes a progressivist theory of historical change. Historians and anthropologists must stop time in order to tell their stories, but they need not deny the coevalness of the socio-cultural worlds nourishing the men and women who provide them with some of the information which they use to explain the past.

Given some of the challenges associated with unnuanced engagements with the concept of time, it is essential to dialogue different evidential sources and lenses to craft the dynamism of practice and society, temporally and spatially. This is important because the othering inherent in the concept of an ethnographic present (which played to colonial ideologies), sometimes froze societies along synchronic and diachronic scales. And yet, because of inherent dynamism punctuated by interfaces, innovation, adaptation, and recontextualization of ideas, as lived experiences, ethnographies are histories. Because history is made daily, with the passing of time ethnographies become dynamic histories that represent continuity and change in lived experience as societies responded to multiple influences, be they from within or from without. David Schoenbrun raises a very interesting dynamic that speaks to sampling issues, both in the past and in the present. The core of his idea is that the present is never totally experienced as it passes such that individuals and societies encounter parts of it.^{Footnote 27} Therefore, it is only snippets of the “present” that are often intentionally and unintentionally retained in sources such as oral sources, written documents and archaeological sites – each source has its own limitations! In this complicated labyrinth, scholars, wrapped up in their socio-political context and theoretical orientation and motivated by the academic questions at hand further throw additional accessories such as biases onto the already scanty “record” of the present, that with time is othered into a mix of experiences that are known as the past.

Although the standard definition of history mostly applies to retrospective orientation that undergirds any account of change and continuity, history is also about understanding lived experiences. However, because the use of analogies to explore deep histories is unavoidable, observations from the recent past may be used to generate ideas for understanding interfaces with the deep past on a multi-scalar level, as conditioned by an adherence to the rules of evidence.^{Footnote 28} Then, the use of analogies between multiple presents (the actual, ever-passing present and those other, “ethnographic” presents), to explore multiple pasts is undergirded by a set of assumptions about a recognizable continuity between the source and the subject to expose convergences and disjuncture that reveal what may have persisted and what may have changed through time.^{Footnote 29} Recalling Ann Stahl’s work on the direct historical approach, this calls for carefully controled comparative analogies that consider the geographical, chronological, and cultural relationships between the source and the subject.^{Footnote 30} Therefore, despite their pernicious use in the invention of tribalism on the Zimbabwe plateau’s early twentieth century, language projects, particularly those by careful scholars such as Chimhundu and others, contain evidence in favor of applying the direct-historical approach in that region.^{Footnote 31} However, this must be done in tandem with the acceptance of the fact that long histories of languages that show continued occupation by cultural groups in a certain area are not an indication of unchanging continuities from earlier periods. Rather, such similarity or relatedness nests complex histories of interaction. For example, Shona is closely related to Swahili; from the sixteenth century it adopted some loan words from Portuguese – it is also a Bantu language which is related to other Bantu languages. This however is no accident of history: the evolution of the language is a culmination of complex, often mixed layers of mobility, interaction, interchange of ideas and trade and exchange that existed as people went on with their lives. This underscores the fact that the human condition is about feedback processes: feedback between the past and present, feedback between experiences of coeval people that interact based in similar or different geographical locations, and feedback of knowledges, skills and ideas all of which calls for caution and care in the study of communities whether in the past or present.

As far as this issue of feedback is concerned, one of the most important points raised by the historian David Beach but forgotten by those who criticize Vansina’s application of ethnographies to study history, careful as he does in some cases, is the problem caused by feedback between sources.^{Footnote 32} This problem stems from the fact that early written histories and ethnographies in colonial Africa were based on oral sources and the interaction between the colonial and the local. However, the local-colonial binary masks the extent to which “local” practices had great regional extent and time-depth while “colonial” ones could be idiosyncratic.^{Footnote 33} For example, the establishment of colonies in southern Africa fragmented deeply entrenched regional connections with the consequence that, for instance, related people such as the Shona and Tsonga ended up in different territories of Southern Rhodesia (British), Portuguese East Africa (Mozambique) at colonization. Thanks to colonialism, such people began to require passports to cross the colonially invented national boundaries. According to Steven Feierman, while in some respects colonialism actively sought to annihilate and discredit entire genres of African knowledge and practice and often challenged them for being unscientific, unwittingly, it authored subaltern and less visible pasts.^{Footnote 34}

With the advent of literacy, most people became schooled in written sources and read written texts. Interestingly, those who read such sources between the 1930s and 1970s are now elders who are frequently interviewed as sources of oral historical information. Under this maze of entanglement, how might researchers distinguish between that which was originally oral, that which was learnt from written sources, and that which was acquired from experiences? This problem of feedback can be partly resolved not by essentializing sources but through a nuanced engagement with different sources, acknowledging dynamism of practice and experience through time. If ethnographies are history and histories are ethnographies so what? This means that if the goal of history is to understand societies through time retrospectively, paying attention to dynamism of practice then we must use all sources at our disposal but after unpacking their limitations. As Vansina’s lead has shown, it is not prudent for historians to avoid ethnographies because they are histories of the more recent periods.^{Footnote 35}

One must however, acknowledge that Vansina’s view of archaeological evidence has been challenged just as the manner in which he sometimes interweaves sources is problematic.^{Footnote 36} For instance, to some, Jan Vansina’s “Siblings” paper might represent a strangely narrow account of theory in archaeology as, essentially, stuck on the kind of progressivist, colonial-inflected narrative of Africa that was frozen in time.^{Footnote 37} Admittedly, and as pointed out by Pete Robertshaw and others, by the mid-1990s archaeology theory had grown rich and complex but there is so much variability in African contexts which makes some of Vansina’s caricature applicable to some regions but not others.^{Footnote 38} It is tempting to think that Vansina might have been directing his critique to particular archaeologists but then generalized to avoid being criticized for making ad hominem attacks on colleagues. While the real motive of this caricature will remain unknown, his calls for revisionism were resolute and inspiring. Furthermore, in regions such as southern Africa, debate still rages on the use and abuse of ethnographies and the static, neo-evolutionary and narrow manner in which state formation is sometimes considered.^{Footnote 39} While I acknowledge some of the limitations pointed out by Robertshaw, and accept that the caricature in some of Vansina’s work might appear to be provocative, if not uncalled for, the bigger contribution of his work is that in a cautious manner, he contributed to the growth of African history as a rigorous and integrative field of human endeavor that uses sources at its disposal in a controlled manner. It is possible that the caricaturing might have prevented some archaeologists from fully engaging with his contributions to the development of tools that empowered a disempowered continent to write its own past following the denialism of typical of the colonial period.

Generally Accepted Historical Narratives Still Remain Hypotheses: Jan Vansina and Revisioning State Formation in Southern Africa

Being the ever cautious and meticulous researcher that he was, Vansina was clear that the main aim of scholarship was to advance knowledge to the extent that some of our generally accepted narratives remain hypotheses that must be constantly evaluated, critiqued and revised – even abandoned – as new observations were made as new data became available. Having been one of the pioneer scholars on the topic of the Bantu migrations, Vansina was quick to throw in words of caution to challenge and critique some of the generally accepted narratives.^{Footnote 40} This flowed in part from his appreciation of the fact that some of the early historical narratives in African history were based on very small sample sizes with generalizations being extended to very huge areas through the drawing of unidirectional arrows. Furthermore, Vansina’s constant awareness of the need to make methodological improvements saw him advocating that specialists in different disciplines must consider narratives from cognate disciplines as hypotheses to be assessed following the rules of evidence in their fields.^{Footnote 41} This was essential for them to avoid circular reasoning engendered by uncritically accepting conclusions made in other disciplines as fact. For example, archaeologists often used conclusions from historical linguists to support their own conclusions, which were then taken as evidence to support linguistic observations. This resulted in “illusory motion” where there was an appearance of movement in scholarship when no new knowledge was produced.^{Footnote 42} Be that as it may, scholarly caution demands that different knowledge fields must critically engage with knowledge from other disciplines to create robust scholarship.

Leveraging experiential knowledge acquired growing up amongst the southern Shona, my research began to question some of the generally accepted narratives for state formation in southern Africa.^{Footnote 43} In this region, it was generally believed, at least from the early 1980s onwards, that state formation originated at a single source – K2 (AD 1000–1200) and Mapungubwe (AD 1220–1290) in the Shashi-Limpopo Valley before being exported to Great Zimbabwe (AD 1300–1450) which eventually passed the baton to Khami (AD 1450–1820). The main reason for attributing all state formation to the adjacent sites of K2 and Mapungubwe was based on cultural precedence. The abandonment of K2 and the occupation of Mapungubwe was assumed to have been associated with spatial shifts associated with the crystallization of an ideology of sacred leadership, rain making and class distinction which was topographically expressed on Mapungubwe Hill.^{Footnote 44} Yet, a re-assessment of the archaeological evidence in southern Africa identified multiple archaeological sites with cultural and topographic attributes similar to, prior to and coeval with those at Mapungubwe. These places such as Mapela, Mananzwe, Mtanye, and many others have stone walling, solid dhaka floors, Leopard’s Kopje pottery, evidence of imports and chronological evidence which overlaps with that of Mapungubwe. In particular, excavations at the massively stone terraced site of Mapela yielded stone terraces, K2 pottery, and other associated material culture on the basal levels of terrace.^{Footnote 45} These were stratigraphically succeeded by Mapungubwe ceramics and material culture. More work in southwestern Zimbabwe identified other sites such as Mananzve while archaeological research in the adjacent regions of northeastern Botswana found other sites.^{Footnote 46} This demonstrated that the attributes only attributed to Mapungubwe were expressed on widely separated points on the landscape. This suggested that state formation did not originate in one place but in multiple ones scattered on the landscape with the corollary that the trajectory of state formation was neither linear nor simple and straightforward.

In re-assessing generally accepted narratives, readers must bear in mind that archaeology is affected by sampling problems. For example, the sizes of excavated areas across sites are often not comparable: some sites had big open excavations (e.g. K2 and Mapungubwe) while others only had small trenches (e.g. Mtanye). Some sites had their gold looted by antiquarians in the late nineteenth century (e.g. Chiumnungwa, Khami, Danamombe and many others) while others had their gold kept (e.g. Mapungubwe). This complicated fact means that the material coming out of excavations was variable to the extent that sometimes, it is almost meaningless to base the importance of places in the past on the basis of frequency of objects. Often, this motivated the use of analogy in building deep histories. Whatever this limitation imposed by sampling concerns may be, while the archaeological evidence appears to indicate the presence of multiple but widely separated communities on the landscape, what does the historical and ethnographic evidence suggest about the nature of political formations in southern Africa from the sixteenth century onwards? To add on, are there any ideas contained therein that might reach convergence with the archaeological evidence?

A survey of the nature of political centralization in southern Zambezia from the sixteenth to the seventeenth centuries suggest the existence of multiple polities scattered on the landscape across the Zimbabwe plateau and adjacent regions.^{Footnote 47} Most of this information was collected from oral traditions collected by David Beach and others following the methodology pioneered by Vansina and others (Beach included). The major limitation of the oral traditions was that often they were collected long after the historical events and in some cases from informants not directly related to the political formations under study. It bears noting that the way in which historians used those traditions departed in some key ways from the uses to which their tellers and audiences put them, in the sixteenth and seventeenth centuries. Other than the oral traditions, Beach and others such as Mudenge also studied a huge corpus of Portuguese documents relating to northern Zimbabwe. The documents were partly based on information provided by informants at the time (oral traditions), and partly on observations, both actual and imaginary of the Portuguese themselves. Subsequently, a considered source criticism and verification was performed to produce generally accepted surveys of the nature of political formations that existed from the seventeenth century onwards. These include the well-known Mutapa state (AD 1400–1896) which was based in northern Zimbabwe and adjacent Mozambican lowlands, the Torwa-Changamire state (AD 1400–1896) spread out in southwestern Zimbabwe and northeastern Botswana, the Buhera polity (c. 1600–1896) in the south central and many others.^{Footnote 48} Apart from having capitals that shifted based on political succession, these polities were of different sizes, self-governing and interacted with their neighbors, sometimes as peer polities, as conditioned by situational realities. While these communities belonged to a group that is now collectively known as Shona, Portuguese documents from the sixteenth century onwards expose some very pertinent points relating to the fluid nature of ethnicity. For example, they explicitly state that the inhabitants of northern Zimbabwe were known as Karanga who occupied a territory known as Mukaranga.^{Footnote 49} Mobility was a key feature of various communities such that various groups moved in different directions to the extent that boundaries both political, social, economic and otherwise were permeable, always made and remade. And yet, the Southern Rhodesia Missionary Conference of 1929 which created and crystallized modern day ethnic classifications labeled those people Korekore.^{Footnote 50} The label Karanga was given to southerners. There is however no known record of massive migrations that resulted in populations in these areas thereby discrediting the view that these groups may have swapped territories. The conclusion from an examination of historical and ethnographic records is that from the sixteenth century onwards southern Zambezia was characterized by cultural dynamism and that at no point was there a single state or polity ruling over all the other territories. In fact, some oral traditions connect ruling lineages to multiple places that were capitals scattered on the landscape.^{Footnote 51} A comparison of this picture with that in the archaeology reveals some similarities. In the archaeology of the period AD 1000 to 1500, there are multiple places that would have functioned as capitals of different states thereby reaching convergence with the historical picture. What however remains unknown are the names of such states and their territorial boundaries. Nevertheless, this makes it unlikely that southern Africa at any given historical moment had a single state whose collapse engendered a neo-evolutionary succession of state systems. This suggests that the emergence and flourishing of sociopolitical complexity in southern Africa was polycentric and not linear. Indeed, comparative evidence from Central Africa, shows that states were expanding, contracting and collapsing with new ones emerging in place of the old.^{Footnote 52}

This southern African example shows that a generally accepted historical narrative may be contradicted by using new evidence. Vansina had recognized the limitations posed on the reconstruction of the past by available sources and amount of research performed to the extent that he persuaded researchers to constantly revise their narratives. For example, vast swathes of land in Central and Southern Africa were unexplored archaeologically, with gaps in knowledge plugged by guesswork, it is always necessary to revise the picture as new data is generated. This observation also applies to southern Africa and other regions where research has mostly been directed to a few sites on the landscape. As more and more research is performed, the expectation is that narratives must be modified and remodified. One of the reasons why Vansina is inspiring and an example to follow is that he was always keen to revise his own narratives and to accept critiques and revisions of his own narratives. That way his scholarship was always in motion in a careful, cautious and considered manner.