Beyond States and Empires: The Search for Order in Medieval West Africa

Beyond States and Empires: The Search for Order in Medieval West Africa

Wolfram Latsch

Henry M. Jackson School of International Studies University of Washington

400 Thomson Hall Seattle, WA 98195-3650

USA

Tel. +1-206-543-6001 Email: [email protected]

Abstract There was extensive specialization and long-distance trade in the medieval West African savanna, or Western Sudan. But almost nothing is known about the institutions that made this possible. It is often assumed that large states or empires provided political and economic order but there is little evidence for these. We propose an alternative account in which economic integration and political fragmentation coexisted. We propose that the “translocal order” of the savanna consisted of three distinct but complementary institutional complexes. Each of these was formed by groups of occupational and ritual specialists who developed reputational mechanisms that lowered the cost of long-distance exchange.

Keywords: Mali Empire, translocal order, enforcement, institutional complex, reputational mechanisms

JEL classification: N47, P48, Z11, Z12, Z13,

Draft - Please do not quote without permission

1

1. Introduction

We investigate the nature of economic and political organization in medieval West Africa’s savanna zone, a belt lying between the Sahara desert and the rainforest (also known as the Western Sudan). The medieval period is widely associated with the emergence of large-scale centralized and territorial political structures such as the ‘Mali empire’ whose legendary rulers are said to have amassed power and wealth by controlling the trans-Saharan trade. We draw on the anthropological and historical literature in reconstructing a possible economic history, and a plausible political economy, of the medieval savanna. We want to understand the nature of key institutions and the role these institutions played in fostering economic and political specialization and in supporting extensive regional and trans-Saharan exchanges.

We propose an explanation that incorporates three widely agreed-upon facts about medieval West Africa in the 13th to 16th centuries CE, the apogee of the ‘Mali empire’:

(1) There is considerable historical and archaeological evidence for

specialization and long-distance exchanges within West Africa as well as between West Africa and the Mediterranean world. This evidence dates back more than two thousand years. The best-known evidence of extensive specialization and exchange includes dense urban settlements on the Niger river and in the Sahel. Other evidence is provided by finds in trading entrepôts such as Gao and Timbuktu and by the West African goods that reached North Africa and Europe via the trans-Saharan trade;

(2) There is no direct archaeological evidence that supports the

existence of a ‘Mali empire’ or much evidence of any large-scale territorial states or empires in the Western Sudan. The scale or extent of archeological finds have so far disappointed those seeking archaeological confirmation of large political structures in the form of either monumental architecture or other conspicuous infrastructure. The whereabouts of most places mentioned in Arab sources and oral traditions are either unknown or strongly contested.

(3) There are persistent and widespread accounts of a ‘Mali empire’ in

the region’s oral traditions, in the few surviving contemporary Arab accounts, and in later written sources that date from the 16th and 17th centuries. The use of oral traditions and Arabic written sources in reconstructing the history of the ‘Mali empire’ is highly contentious but they continue to shape popular and scholarly perceptions of West Africa’s past.

These three facts cannot easily be reconciled. In the face of this

realization many accounts have simply continued to assume the existence of a large and centralized territorial state or empire to explain extensive

2

economic specialization and exchange. We call such theories global order explanations. They posit, implicitly or explicitly, that economic and political integration are highly correlated.

In contrast we attempt to reconcile the three facts by hypothesizing

that the political economy of the Western Sudan was a decentralized economic and political order characterized by considerable institutional pluralism. This order did not require any large-scale political structures like states or empires. It would therefore not have produced the monumental archeological evidence that is typical or expected of states or empires. It consisted instead of at least three distinct but complementary institutional complexes. These complexes were composed of translocal groups of occupational and ritual specialists. The worldly and supernatural enforcement mechanisms deployed by these groups provided the building blocks of an emergent but durable phenomenon we call a translocal order.

This was at once politically fragmented and economically integrated.

It provided an effective enforcement environment and allowed for the creation and protection of a wide variety of rents by occupational and ritual specialists with distinct translocal identities. Rents were derived from specializations in production, protection and exchange and enabled by far-reaching reputational mechanisms that lowered the cost of long-distance exchange. We argue that some of these reputational mechanisms enabled the creation of focal narratives in the form of heroic myths (such as the epic of Sunjata) which claimed that the ‘Mali empire’ was the prime mover behind the economic and political order of the savanna. These myths were productive in that they helped to reduce political transaction costs between independent but increasingly interconnected specialists in violence. This enabled Smithian growth which, in turn, is now attributed to the global order created by the Mali empire extolled in the myths.

2. Beyond Multilocal Order: The Medieval Savanna Economy

West Africa has a long history of complex economic organization involving extensive specialization and exchange. The economic integration of the region is evidenced by long-distance trading networks and by the local, regional and interregional exchanges they enabled. These patterns display considerable stability over many centuries (Malowist 1966). The earliest exchanges were encouraged by the expected gains of trading across the different ecological zones – desert, Sahel, savanna, woodland and rainforest – that are stretched across the region laterally in relatively narrow bands. Trade was boosted by the spread of iron-smelting after about 1500 BCE and the consequent expansion of iron use in agriculture, forest clearing and mining. Opportunities for specialization were associated with the emergence

3

of networks of traders who used porterage, donkeys and boats in assembling caravans of goods that linked sellers and buyers across the region and connected to far-flung inter-regional networks (Hopkins 1973, Sundström 1974, Austen 1987, Newman 1995, Webb 1995, Ehret 2002).

The appearance and growth of regular markets, especially entrepôt

towns located at environmental junctures, contributed to the emergence of an increasingly integrated regional economy. Cowries, copper, gold, cloth and iron served as currencies and standards of account and credit was used from an early date. From about 300 CE the West African economy was connected to North Africa and the Mediterranean by the trans-Saharan trade. This provided channels for the export of gold, kola, salt, slaves and textiles and was further spurred by the development and expansion of the camel-based caravan trade. The trans-Saharan trade was dominated by Islamic Arab and Berber groups of the Sahara and North Africa. The regional trade of the savanna was controlled by West African specialists speaking languages of the Mande sub-group of the Niger-Congo language family (which includes Bambara, Mandinka and Soninke).

The gold trade illustrates the extent of specialization in medieval West

Africa. Gold mining in West Africa probably originated in Bambuk at the upper tributaries of the Senegal river and dates back to at least the 4th or 5th century CE. Gold production later shifted eastwards towards the goldfields of Bure on the upper Niger River and its tributaries around the end of the 11th century, to the Lobi goldfields on the Black Volta River after the early 14th century, and to the Akan forests of modern-day Ghana after the mid-15th century. The eastward shift in gold production was reflected in an eastward shift of the key entrepôts of the trans-Saharan trade routes between the 10th and the 16th centuries. Gold mining in the savanna appears to have been geographically dispersed and carried out on a small scale using seasonal labor during the dry season when labor demand in agriculture was low and labor was relatively abundant (Austin 2008). Gold was traded across the region and channeled into the trans-Saharan trade routes, which was flourishing as early as the 6th century CE (Garrard 1982). Demand for gold in the Maghreb and in Europe increased in the 13th and 14th centuries. Gold deposits were not found in Europe until around 1320: before that time around two thirds of Europe’s gold came from West Africa (Mauny 1961, Curtin 1983, Austen 1987, Devisse 1988, Spufford 1988, Blanchard 2005).

Long-distance trade in gold required three broad types of

occupational specialization and entrepreneurship: production, protection, and exchange. Production required solutions to problems of labor mobilization and to the technological challenges of iron making and mining; protection required solutions to problems of violence and enforcement; exchange required solutions to problems of information and logistics. The growing demand for local products in regional and inter-regional economies

4

provided incentives for the emergence of occupational specializations and entrepreneurial solutions to emerge. Groups of specialists constructed effective institutions that created and protected the rents deriving from an increase in the scale and scope of transactions. A wide variety of these institutions had to be aligned in ways that reduced rent-dissipating conflicts between different enforcement mechanisms.

The extensive economic integration of medieval West Africa clearly

points to something more than a multilocal order. A multilocal order is defined here as an equilibrium pattern of spatial social organization in which transactions between local groups (for example villages) are quite limited and based mostly on ad hoc and spot transactions. In a multilocal order enforcement is based mostly on personal relationships rooted in kinship, lineage or close familiarity. Such mechanisms do not scale up easily to incorporate outsiders or strangers. They are therefore not conducive to long-distance transactions.

We consider two alternatives to multilocal order: a global order

involving large-scale and integrated political organization under a common enforcer (what is typically called a state or an empire), and a translocal order supported by an enforcement mosaic of autonomous groups in the absence of any overarching political organization or common enforcer. We argue that a model of translocal order provides a better way of reconciling what we know about the medieval savanna.

3. The Search for a Global Order: Empires and States of the Savanna

The term global order as used here indicates that there is at least some degree of global or overarching governance within a geographically well-defined area. It implies the existence of some set of global rules; that is, there are at least some rules which apply to all individuals in the relevant group. It is assumed that these global rules are effectively enforced on subjects or citizens by a global enforcer. Any reasonable definition of a state or empire has to include some delineation of the extent to which, and the ways in which, a global enforcer enforces global rules, however extensive or narrow these might be. It must also define who can (or who must) draw on this ultimate enforcement power (Barzel 2002). If a group of individuals has no common ultimate enforcer then these individuals are not part of any global order. Therefore they cannot be viewed as subjects of the same state or empire.

A “global order” explanation would suggest that the medieval savanna

was characterized by a relatively strong centralization of political power –

5

what we call here a global order. This view is influenced by accounts left by medieval Arab historians and travelers (Levtzion and Hopkins 1981). Global order views assume that the rulers of these states, of which Mali is the best known and the one we consider here, projected their hegemony over large territories, although it is widely noted how little is actually known about Mali’s economic and political structures (Delafosse 1912, Fage 1957, Mauny 1961, Trimingham 1962, Monteil 1968, Ly-Tall 1972, Levtzion 1973, 1977, Mair 1977, Niane 1984, McEvedy 1995, Oliver and Atmore 2001, Ehret 2002). Most histories of Africa provide maps showing the territorial extent of Mali and other empires or states.

These confident depictions of West African polities as territorial and

consolidated are partly conditioned by progressive narratives of African pre-colonial history popularized in the era of decolonization, although they also served a variety of pre-colonial and colonial interests (Masonen 1994, Parker and Rathbone 2007). These progressive narratives emphasized postulate a long precolonial tradition of state formation, military expansion, administration and charismatic rule in Africa, in part to counter widespread perceptions that Africa was a continent without a sophisticated political history (Miller 1999, Wright 1999). They imply a strong degree of central direction and coherent intent by charismatic political leaders in the economic and political spheres. These are also tropes in the region’s oral traditions, especially in the famous epic of Sunjata, of which more below.

It is widely agreed, going back to medieval Arabic sources, that

neither long-distance trade nor gold production were ever under the direct control of any centralized government or state in the West African savanna (Mauny 1961, Levtzion 1973, 1977, Hopkins 1973, Levtzion and Hopkins 1981). The fact that key economic activities like gold production and long-distance trade appear to have been controlled by autonomous groups weakens the global order interpretation. Many other empires placed their key economic activities, and certainly the extraction of precious metals, under state control. Also, archaeological research has so far provided only limited evidence of any large-scale or centralized political structures that might be linked to a Mali empire (McIntosh and McIntosh 1996, Insoll 2003, Mitchell 2005).

Recent research has instead emphasized the great diversity of West

African institutional arrangements, exploring the possibility of non-hierarchical political organization and highlighting the fact that, in all periods of West African history, markets did not require states to function (Northrup 1978, McIntosh 1999, Lovejoy and Richardson 1999, 2001, Austin 2004), even if states might improve on more decentralized outcomes in some settings (Lovejoy and Richardson 2004). Many contributions point out that the relative abundance of land in much of West Africa resulted in political fragmentation that limited the feasibility of centralized governance (Kopytoff

6

1987, Iliffe 1995, Newman 1995, Wright 1999, Herbst 2000, Austin 2008); some highlight the often fragmentary, tenuous or contradictory nature of information provided by different sources (Meillassoux 1991, Lange 1992, Jansen 1996a, 1996b). While there does appear to have been increasing political consolidation over time in the Western Sudan, global order explanations do not tell the whole story. The more centralized states were constructed on and around previously evolved institutional mechanisms that often outlasted the ebb and flow of empire. These mechanisms constituted the building blocks of what we call a “translocal order”, and are the focus of this article.

4. Behavioral and Institutional Foundations of Translocal Order

A “translocal order” differs from a global order in that it enables economic integration without an overarching, integrated or global political structure. A translocal order will typically not exhibit a well-defined territorial structure or geographical extent. Instead it is characterized by the specific institutions (translocal groups of specialists) that promote exchange and specialization by linking, through their own members, the members of various localized communities (local groups). Members of local groups (such as villages) remain subject to their own local enforcement mechanisms which are often lineage-based and headed by chiefs.

Translocal groups specialize in particular occupational and ritual

activities which define and limit the scale and scope of governance they will provide. Different translocal groups might cooperate or collude in ways that strengthen their respective abilities to create and protect rents derived from translocal exchange. In doing so they form institutional complexes. Following Greif (2006) an institutional complex is defined as a cluster of institutions (formal organizations or informal mechanisms) that are complementary in their enforcement characteristics; that is, they do not crowd each other out and they recognize each other’s domains in ways that limit enforcement gaming. Translocal groups do not try to replace or supersede the enforcement mechanisms provided by local groups but to complement them.

4.1. Relations between Local and Translocal Groups

The creation of feasible contractual interfaces with local groups is one of the main challenges faced by translocal groups. Good within-group cooperation in local groups typically demands fewer contacts across groups and wariness of outsiders as communities seek to maintain their coherence and the exclusivity of their own local enforcement domains. This parochialism might be maintained through high entry barriers, restricted

7

mobility and the shunning of those perceived as defectors from local agreements (Dixit 2004, 2009, Kranton 1996a, Kranton 1996b, Bowles and Gintis 1998, 2004, Pagel and Mace 2004).The governance provided by translocal groups may encourage opportunistic behavior as members of local groups attempt to arbitrage their enforcement environments. This might involve defecting from the commitments and agreements that are subject to local enforcement and attempting to substitute translocal for local enforcement. Enforcement gaming of this kind creates negative externalities between different enforcers and creates potential rent-dissipating conflicts at the interface between local and translocal groups. Similar considerations apply to the interfaces between different translocal groups. To minimize this risk translocal groups have to conclude bargains with local enforcers and other translocal enforcers. These bargains have to provide some mutual recognition of enforcement domains. If successful these bargains limit the negative externalities that are created by overlapping governance structures and instead lead to a complementarity of translocal and local groups. As part of this bargain translocal groups will credibly limit the scale and scope of the transactions and agreements they will enforce on members of local groups or other translocal groups. The transactions conducted across these contractual interfaces between local and translocal groups substitute more standardized and more impersonal contracts for the more personal or ad hoc agreements between local groups that are characteristic of a multilocal order.

4.2. Reputations and Rents

Translocal groups are created and maintained by translocal entrepreneurs. Translocal groups are clubs or coalitions formed by networks of occupational and ritual specialists with distinct translocal identities or reputations. Translocal identities are recognizable and meaningful across a number of local groups; they transcend purely local political and economic institutions. These distinct identities are governed by the reputational mechanisms that operate in translocal groups. Reputation is defined as a bundle of individual characteristics that determines a person’s perceived desirability as a transaction partner in a world of imperfect information. Reputations play an important role in enabling cooperation within larger groups, especially within groups whose members are dispersed as is the case in translocal communities.

Cooperation in these circumstances is typically sustained through a

system of indirect reciprocity in which exchange partners do not meet very frequently (in contrast with direct reciprocity in which repeated dyadic exchange dominates). An increase in the number of trading partners requires the ability to transmit a measure of one’s reputation to others who do not have direct knowledge of it. A reduction in transaction costs is achieved if this reputational score or standing can be transmitted speedily, reliably and

8

cheaply. Communicating an accurate reputational score or standing requires some standardized measure that can be evaluated reliably by a large group of potential transaction partners. Therefore the standardization of reputation creates the rudiments of impersonal exchange by lowering transaction costs between strangers. It increases the portability of reputations, increasing the size of the area in which – and the number of groups and people for which – particular reputations are meaningful.

In a system of indirect reciprocity effective enforcement (that is, the

punishment of cheating or shirking) has to be distributed and multilateral since highly coordinated or centralized punishment is typically not feasible. But multilateral enforcement requires punishment that is costly to the punisher: this creates a second-order free-riding problem that threatens cooperation in large groups (Boyd, Gintis, Bowles and Richerson 2003). The reputation-based multilateral enforcement mechanisms that undergird indirect reciprocity typically involve shunning and the withdrawal of help instead of requiring costly direct punishment. This makes them more immune to the second-order free-riding problem. Indirect reciprocity based on reputational mechanisms therefore scales up more easily than other forms of large-scale cooperation. It may have a long evolutionary history in enabling the development of complex societies (Panchanathan and Boyd 2004, Fehr 2004, Nowak and Sigmund 2005, Henrich and Henrich 2006, Rockenbach and Milinski 2009).

Mechanisms of indirect reciprocity define and protect the reputations

of the members of translocal groups. They involve processes of monitoring and enforcement that are maintained by the translocal group as a collective good. The translocal group therefore forms a reputational community. The occupational and ritual activities of a reputational community create a translocal reputational field. Controlling access to this field is a source of rents for the members of translocal groups. Rents are derived from the lowering of transaction costs which the reputational field enables, for its members as well as for their clients or customers. The reputational field achieves this by validating, certifying or assuring the quality of the people, the products and the promises that pass through it. The reputational field effectively standardizes these, thereby increasing the portability of reputations. Whereas the reputational fields of local groups tend to be parochial (limiting access through high entry barriers for strangers) those of translocal groups are more cosmopolitan (enabling selective access and transforming strangers into customers or clients).

A reputational field effectively transforms complete strangers into

near-strangers. Near-strangers have some governance in common, and this might be provided by a translocal group. The attendant lowering of transaction costs induces an increase in economic and political specialization and an expansion of markets within a process of Smithian growth.

9

4.3. Ritual Specialists and Supernatural Enforcement

The emphasis in economics is mostly on enforcement mechanisms that impose tangible material costs through direct punishment and shunning. But the ubiquity and durability of beliefs in supernatural agents means that effective enforcement might involve sanctions that are neither immediate nor tangible. Drawing on recent evolutionary explanations of religion (Boyer 2000, Atran 2002) we will label this supernatural enforcement. Supernatural sanctions can be linked to more tangible or material sanctions. For example, common beliefs create communities of believers who follow common rituals. Adherence to these rituals can be used to screen for cooperativeness and trustworthiness; that is, it can be used to create and gauge standardized measures of reputation. The threat of exclusion from such a community of believers, for example by shunning, can provide an additional enforcement tool (Iannaccone 1992). Supernatural agents are often believed to have full access to strategic information that might be concealed from potential transaction partners: supernatural agents can detect cheating or detect the intent to cheat. An internalized belief in such agents might therefore strengthen a person’s internal enforcement mechanisms, such as self-control, self-sanctions or guilt, thereby increasing that person’s reputational capital (Norenzayan and Shariff 2008). Belief in an afterlife can induce greater cooperation by lowering a person’s discount rate, especially if it is believed that supernatural agents use cooperativeness as a screen for rewards or punishments in a potentially infinitely-lasting afterlife (see Ekelund et al. 2006 for a discussion of Christianity, Kuran 2004 for a discussion of Islam). There are also advantages and beneficial network effects in adhering to a belief system that is common to a large number of people including those who have little or no direct knowledge of each other. A common belief system strengthens the portability of reputations by increasing the reach of supernatural enforcement and protection mechanisms across space and time.

The brokering of supernatural protection and enforcement provides a

source of rents for ritual specialists. Ritual specialists are people who are perceived to be, and recognized as, gatekeepers or mediators between humans and the supernatural agents that feature in their belief systems. Ritual specialists can capture some of the rents created by the increased cooperation that their activities help to induce. Extending the reach of belief systems and their supernatural enforcement mechanisms, especially when tied to an effective reputational system, appears to be an important component of translocal entrepreneurship. For translocal groups supernatural enforcement can complement and reinforce more tangible enforcement. The reach of supernatural enforcement is increased the more standardized and widespread are the relevant belief systems and rituals. And rituals of all kind, being highly schematic and repetitive, are essential to

10

memorization in non-literate societies and cultures. All reputational mechanisms require the standardization, collection, storage and distribution of information regarding the types and behaviors of relevant individuals. Rituals of storytelling, singing, or dancing can effectively form a kind of distributed database by recording and retelling behavior and actions, by setting norms and instructing the young. Oral traditions can help encode the rights and obligations defined by enforcement mechanisms of formal and informal institutions.

4.4. Institutional Complexes

All the translocal entrepreneurial groups discussed below appear to have deployed reputational mechanisms to create and maintain the distinct identities of their members and to manage transactions between their members and local groups. They did so in distinctive ways and in support of a wide variety of specialized activities and occupations. Worldly and supernatural enforcement mechanisms were effectively bundled and deployed to create and protect rents deriving from the increased protection, production and exchange activities that were enabled by these mechanisms. A typical pattern observed in the precolonial West African savanna displays the pairing of occupational and ritual specialists within an occupational-ritual complex (or ORC). Occupational and ritual specialization was typically incomplete with each ORC encompassing three broad occupations (exchange, production, and protection) in variable mixes; each ORC appears to have been assembled around one of these occupations. Each ORC comprised two types of specialists that were organized as translocal groups: (1) ritual specialists, defined as people perceived to be gatekeepers between material and supernatural domains, and (2) occupational specialists, defined as people engaged in economic and political activities in the material world. In some cases occupational specialists were also ritual specialists. In other cases groups of ritual specialists were affiliated with or allied with groups of occupational specialists.

The pattern of specialization and exchange that emerges from the

operation of these translocal institutional complexes constitutes the translocal order of the medieval savanna. It is characterized by the absence of any overarching or common third-party enforcement of agreements or contracts. The costs of long-distance transactions are lowered as they pass through the reputational fields of translocal groups and the institutional complexes they form. These complexes are broadly complementary. Their functionally differentiated enforcement domains or jurisdictions overlap within particular geographical areas. A translocal order is therefore characterized by both political fragmentation and economic integration. It is politically fragmented in that there is no common enforcer for all the individuals and groups within its penumbra. It is economically integrated in

11

that the enforcement mosaic created by the institutional mechanisms of translocal groups contributes to the domestication of violence and induces specialization through a reduction in political and economic transaction costs.

5. Three Occupational-Ritual Complexes In the following sections we introduce three occupational-ritual complexes: respectively, the exchange complex, the production complex and the protection complex. We hypothesize that these three ORCs supported the translocal order of the medieval West African savanna. Given the very limited evidence for the medieval period the information assembled here is inevitably partial and the portrayal of how the constituent translocal groups operated is inevitably incomplete. We draw on contributions from history, anthropology, and archaeology. Some of the evidence presented is based on later periods, including the recent past and present, for which there is better and more reliable information. Although projecting more recent accounts back in time involves an obvious element of speculation it is probable that the groups of occupational and ritual specialists discussed below are of considerable antiquity. We argue that, given prevailing economic, environmental and demographic constraints during the medieval period, these groups are plausible institutional mechanisms supporting complex economic and political order – and more plausible than large states or empires with strong political centers. 5.1. The Exchange Complex: Long-Distance Traders, Trading Networks, and Islam

Trade within the West African region was rooted in ecological specialization. The proximity of different ecological zones running parallel from East to West allowed for gains from trade over relatively short distances, especially on a North-South axis. Traders and trade networks played an important role throughout the medieval period (Posnansky 1973, Mauny 1961, Sundström 1974). Traders in the West African savanna were mostly, speakers of the Mande subgroup of the Niger-Congo language family, which includes Malinke and Soninke. These traders were known by names such as Juula, Jakhanke, Marka and Yarse. The Mande heartland was the area between the upper Niger and upper Senegal rivers and the early specialization of Mande traders is likely related to the discovery and development of gold in that area (Lovejoy 1978, Perinbam 1980). From there Mande-speaking traders spread throughout the savanna region (McNaughton 1988, Brooks 1993). Through migration, adaptation and assimilation they contributed to the emergence of a large Mande culture area with common

12

symbols, myths and cultural references (Miller 1990, Conrad and Condé 2004).

West Africa’s trade was carried out by networks of traders linked

through kinship, religion, ethnicity or affinity. This often led, over time, to the emergence of distinct, if flexible, ethnoprofessional identities (Brooks 1993). These trading networks traversed a politically highly fragmented landscape (Iliffe 1995, Thornton 1996, Herbst 2000). Regional trade in West Africa appears to have been enabled by a variety of distinct trading networks, or trading diasporas. These partly overlapped and intersected, creating a mosaic of translocal groups that displayed considerable stability over time.

Traders in West African trading networks adopted Islam from the early second millennium CE. Islam was the religion of Arab trans-Saharan traders and the brokers in the main Sahelian entrepôts of Awdaghust, Walata, Jenne, Gao and Timbuktu. The adoption of Islam by West African trading groups permitted an expansion and consolidation of trading networks with Islamic law and Arabic providing the standardized templates for many types of business organizations – a common and standardized commercial code or law merchant (Cohen 1971, Austen 1987, Ensminger 1997). The spread of Islam allowed for a more effective articulation of the different trading networks south and north of the Sahara. Islam also had standardized religious rituals – which could be used as a costly signal of cooperativeness and commitment – and a set of supernatural inducements and sanctions. This allowed supernatural enforcement to be deployed alongside worldly sanctions to induce cooperation and deter cheating. It could also be used to maintain reputational mechanisms that helped create a distinct translocal identity for long-distance traders.

Long-distance trade was organized into caravans that were effectively

large moving markets in which trade was conducted continuously as the caravans progressed between larger settlements and market towns (Curtin 1971, Hopkins 1973). Caravans were organized and owned by clans or extended families of traders and they organized their own internal enforcement. They also arranged their own protection by hiring armed groups and by allying themselves with local warlords (Wright 1977). Trading groups forged relationships between (mostly non-Islamic) local groups along the caravan routes (Brooks 1993).

Trading groups often founded separate towns or separate quarters in

market towns. In many cases they were Muslim enclaves in non-Muslim areas. These autonomous settlements were effectively market centers or caravanserais that consisted of traders and of allied craft specialists such as weavers, smiths and clerics (Arhin 1987). These trader towns functioned as entrepôts that were under the independent jurisdiction of trading groups, indicating a clear delineation of enforcement domains (Perinbam 1981,

13

Curtin 1975, Cohen 1971, Brooks 1993). Where trading networks connected to existing towns or settlements the institutional arrangements enabling cooperation would include formalized relationships between traders and local enforcers (landlords or chiefs), or between itinerant traders and resident brokers from the merchants’ own trading clans or networks. Brokers would host itinerant traders and provide a marketplace for their goods, often in the brokers’ own houses. They would extend credit to buyers and sellers. Brokers would also mediate between traders and local groups, effectively guaranteeing the identity, probity and credit of their clients (Hopkins 1973, Launay 1979, Curtin 1984, Cohen 1971, Sundström 1974).

Long-distance traders maintained distinct translocal identities and

reputations, often to the point of being perceived as separate ethnic groups. As discussed above, these reputations constituted a collective good and all members of a trading group had to contribute to the group’s multilateral enforcement mechanisms. These include the punishment, through exclusion or shunning, of those whose actions threaten the integrity of the reputational field maintained by the group.

In the case of traders or trading networks, punishment in the form of

exclusion from group imposed immediate material penalties. Islam provided traders with a standardized commercial code that could be coupled to the information flows that are essential to reputational mechanisms and therefore to enforcement both within and between trading groups. This type of material enforcement was supplemented by spiritual sanctions involving the perceived intervention of supernatural agents. Islam provided an extensive moral code which drew on supernatural beliefs and the enforcement powers of supernatural agents for its strengths; it has a concept of the afterlife in which standing is affected by lifetime actions and choices. Adherence to costly rituals could serve as a screen for religious commitment and devotion.

Traders also made use of unorthodox or mystical aspects of Islamic

learning in constructing the supernatural enforcement mechanisms for their translocal groups. These provided protection from nature and witchcraft to clients, endowing traders with a supernatural aura. Traders often made use of this power by combining the role of trader with that of ritual specialists such as Muslim holy men (marabouts) or by associating their activities with the marabouts that often accompanied the caravans and settled in the towns or quarters settled by traders. Marabouts performed healings and provided supernatural protection through the production and sale of medicines, amulets and talismans, creating a translocal economy in standardized religious artifacts (Last 1988, Soares 1996). The religious and mystical standing of Muslim traders and clerics was reinforced by their literacy: verses of the Koran were considered to provide powerful magical protection with a large geographical reach to the buyers of their ritual and religious

14

services. (Curtin 1971, 1975, 1984, Brooks 1993, Austen 1987, Wright 1977, Levtzion 1986, Owusu-Ansah 2000, Thornton 1996, Cohen 1971, Wilks 2000)

To summarize, the professional-ritual complex represented by

members of long-distance trading networks combined professional specialization with material and supernatural enforcement mechanisms that could be directed effectively both inward (to limit within-group opportunism), and outward (aimed at the enforcement of agreements between trading groups, between traders and outsiders, or between outsiders). Outsiders included buyers of traded commodities or goods as well as buyers of the ritual and enforcement services provided by traders or their allied ritual specialists – Muslim clerics representing Islamic orthodoxy, and marabouts representing a more esoteric and mystical Islam. Jointly, these enforcement mechanisms created an effective reputational field. Controlling access to this field was a source of rents to the occupational and ritual specialists who provided, for profit, quality assurances for commodities, products and credit that were supplemented by physical and supernatural protection. These helped lower the costs of long-distance transactions on the savanna and at the Sahelian entrepôts of the trans-Saharan trade.

5.2. The Production Complex: Blacksmiths, Castes, and Secret Societies

Mande-speaking blacksmiths migrated from the Mande heartland over large areas of West Africa, often jointly with Mande trading groups, and settled among local groups after about 700 CE. Gold extraction appears to have been under the control of Mande blacksmiths and Mande traders by the early 2nd millennium CE (Brooks 1993).

Blacksmiths were producers of physical inputs and final products.

They smelted and smithed iron and made iron tools and weapons for use in agriculture, hunting and fighting. Iron products were highly desirable items in the regional long-distance trade - which was also supported through the use of the standardized iron currencies produced by blacksmiths. The blacksmiths’ effective ownership and guild-like control of their technology, their production process and their apprenticeship process was an important source of rents.

Blacksmiths were also ritual specialists who produced supernatural

inputs and participated in the governance of the local groups they encountered during their migrations. This ritual role of blacksmiths was linked to their perceived ability to manage the ubiquitous and essential but potentially destructive life-force or energy known among the Mande as nyama which was believed to play a key role in the extraction of smelting of ores (McNaughton 1988, Conrad and Frank 1995, McIntosh 2000).

15

Blacksmiths deployed their ritual activities in support of their role as producers; this helped create and maintain high entry barriers to their profession helped maintain their markets power. Supernatural enforcement was organized through lodges or franchises of secret societies that blacksmiths founded all over the region. These were usually centered on a local shrine or grove and played an important role in the initiation ceremonies of local groups. In West Africa’s lineage-based communities blacksmiths mediated and resolved disputes between lineages, often wearing masks to convey impartiality (Butt-Thompson 1929, Little 1960). Secret societies were also used to organize the provision of collective goods such as roads and to organize collective action such as mobilization for fighting, mining or forest-clearing (Little 1965, 1966, Brooks 1993). Blacksmiths and secret societies also provided ritual services related that were believed to defend against witchcraft – defined as the harnessing of supernatural forces by one person to harm another (Austen 1993). Blacksmiths produce ritual items such as amulets, medicine bundles or statues. These were perceived to be effective in managing interactions with supernatural agents or in fending off harmful supernatural forces; in other parts of West Africa these ritual items, as ‘fetishes’, were used in oath-taking to validate contracts between strangers in cross-cultural trade (Pietz 1987). They were ritual gatekeepers who mediated between the material and the supernatural worlds, managed transactions between natural and supernatural agents, policed ritual boundaries, and controlled the supply and allocation of supernatural inputs used in governance and production. All these ritual activities were sources of rents for blacksmiths.

Maintaining their market power and standardizing their ritual

products and services provided quality assurance. This supported the distinct reputations of blacksmiths in their occupational and ritual roles. The institutional mechanisms of secret societies appear to have scaled up well. This allowed blacksmiths to transform a local system of production and ritual into a translocal network and an extensive reputational field. Local secret societies were networked into an expanding regional system of shrines and lodges that enabled long-distance communication and collective action to support long-distance transactions (Röschenthaler 1999). From this emerged, over time, a common framework of governance and supernatural protection that supported the emerging translocal order of the savanna economy (Kuba and Lentz 2002, Allman and Parker 2005, Insoll 2006, Obeng 2006, Lentz 2009). The effectiveness and importance of secret societies was enhanced by network effects as local cults and practices evolved into belief systems with considerable geographical coverage, with standardized meanings and rituals that helped confer assurances to travelers and transactors, resulting in lower transaction costs.

To this day many promises or oaths in the region are considered

binding when made in the presence of a blacksmith (McNaughton 1988).

16

These effects enabled blacksmiths extend the coverage of their enforcement and protection activities as well as extending the market for their products. But the effective reach of shrines and secret societies intersected with that of other enforcement mechanisms, such as those provided by local and translocal groups. For example, blacksmiths often shared the governance of local groups with chiefs or lineage heads. This created the need to negotiate and delineate their respective enforcement domains and to crowd-in each others institutions – that is, to provide for the mutual recognition of rents and to prevent the potentially destabilizing effects of enforcement gaming.

The reputations and identities of blacksmiths were further defined by

their caste-like status as nyamakalaw. Nyamakalaw are the endogamous and hereditary craft specialists or ethno-professional groups of the Mande culture area. They are viewed as ritual specialists and as producers of goods and services; other examples include leatherworkers, potters and bards. The Mande caste system exhibited a tripartite structure, consisting of free people – that is, farmers, hunters and warriors – slaves, and craft specialists. It likely diffused through the region during the medieval Mande expansions (McIntosh 1981, McNaughton 1988, Tamari 1991, Brooks 1993).

Economic studies have emphasized the complementarity of services

and products provided by an occupational caste system, inducing specialization by providing a stable framework for exchange particularly in land-abundant environments. Caste societies have also been viewed as distributed systems of political and economic control with decentralized governance based on reputational mechanisms (Lal 1998, Freitas 2007). In that case caste systems can contribute to a translocal order by inducing specialization and exchange, lowering transaction costs through the standardization of the status, identities and reputations of their members.

As casted craft specialists blacksmiths in much of the West African savanna were considered to be of low status and often viewed with distrust or disdain (Levtzion 1977, McNaughton 1988). They were nevertheless perceived as powerful ritual specialists, creating an apparent disjunction between ritual or economic power and social status (McNaughton 1988, 1995). Subordination in status and the acceptance of an often marginalized existence in the community might be interpreted as a costly signal. Self-stigmatization and subordination within a powerful if decentralized system of enforcement such as a caste system may in fact increase the blacksmiths’ reputations as producers and ritual specialists. It can do by providing quality assurances for ritual credence goods produced by blacksmiths (for Christian parallels, such as the poverty vows of ascetic orders, see Ekelund et al. 2006). The quality assurance provided by the ritual-occupational reputations of blacksmiths would have been particularly important wherever they were involved in the creation of the iron currencies that supported trade and

17

stored value in parts of the region, currencies that had to be at once standardized and credible.

To summarize, the occupational-ritual complex and the reputational

field constructed by blacksmiths around their secret societies and distinct identities as casted craft specialists contributed to the translocal order of the savanna. It complemented the enforcement mechanisms provided by local groups and by other translocal groups. The ritual activities of blacksmiths added elements of impersonal enforcement to what was an often fragile lineage-based system of local governance. The institutional mechanisms of secret societies and caste could be scaled up and their geographical reach extended. Standardization of the ritual or supernatural dimensions of enforcement provided the basis for the expansion of secret society governance that is evident in extensive networks of shrines. The reputational field constructed by blacksmiths enabled translocal communication and coordination. It also extended the reach of supernatural protection and enforcement which supported long-distance trade. The standing of blacksmiths as ritual specialists was likely fostered by their caste-like status. The linking of productive activities (like smelting and ironworking) with ritual and governance-related activities allowed blacksmiths to control the markets for both their physical and their ritual products and generate rents from both productive and protective activities.

5.3. The Protection Complex: Warlords, War Houses, and Bards

The migrations of blacksmiths and long-distance traders constituted the first wave of the Mande expansions that are believed to have continued in different forms between 700 CE and 1860. Mande horse warriors, here referred to as warlords, constituted the second wave, probing and expanding the commercial and political frontiers of the West African savanna (Brooks 1993). Horses were introduced to West Africa after the 13th or 14th centuries, the period popularly associated with the rise of the Mali ‘empire’ (Law 1980). Demand for horses was likely driven by Mande warlords whose primary occupation was the use of, or threat of, violence in search of rents from predation and protection linked to the growth of regional trade in goods and people.

Projecting back from later periods for which there is better evidence,

specialists in violence founded their own settlements or settled among small-scale local groups who made protection arrangements with them (Şaul 1998, Parker and Allman 2005, Werthmann 2007). The savanna warlords created systems of safe conduct for traders and travelers based on passes. They also built political support and clients by redistributing the spoils of violence which included confiscated goods, raided livestock, and enslaved peoples (Alber 1999). To the extent that these rents depended on revenues derived

18

from trade and exchange warlords had to maintain some order and peace in their areas of influence; they had to construct and maintain mechanisms of governance and enforcement. And they had to bargain with other warlords over the mutual recognition of enforcement domains and rents. In short, they had to create some form of political order that would support specialization and exchange.

The overall outcome of these complex interactions between warlords

and local groups, and between warlords, was likely an intricate web of alliances, agreements, trade, diplomacy and war, a pattern that is well-documented for later centuries (Werthmann 1997). This web of political and economic transactions constituted an emergent order that supported the movement of goods and people along regional and interregional trade routes by controlling endemic violence (Şaul 1998). This long-range and long-term political equilibrium has been characterized as pax Maliana (Belcher 1999), or pax Mande (Bird 1999). We argue that this equilibrium was supported by a reputational field maintained by warlords and allied ritual specialists. This field enabled credible agreements between different specialists in protection as well as between specialists in protection and specialists in production and exchange, limiting predation and promoting production and exchange.

One of the key institutions created and developed by enterprising

warlords was the war house (Şaul 1998). It provided an organizational form for their economic and political activities and for the creation and protection of their rents. The war house was effectively a bundle of contractual arrangements between protection specialists and a set of kin and non-kin followers engaged in production, trade, and the raiding of goods and slaves. They were extended households or estates constructed from personal duties and obligations based on patronage and lineage. War houses organized opportunistic raiding expeditions that were ad hoc and for-profit ventures in which other warriors could invest and participate in exchange for a share of the takings (Şaul 1998).

Warlords could provide violence-backed enforcement services to

other local groups, such as local villages, or translocal groups, such as traders, in return for protection and enforcement rents. Warlords could supplement these worldly types of enforcement with supernatural enforcement mechanisms in assembling an effective field of governance based on their distinct identities and reputations. Supernatural enforcement was provided by ritual specialists and their organizations: warlords might seek the ritual and political support of secret societies and blacksmiths (Brooks 1993); families of Islamic ritual specialists were sometimes affiliated with war houses and were also believed to provide supernatural protection (Şaul 1998).

19

A group of ritual specialists that was unique to the protection complex were bards, praise-singers, or griots (Mande jeliw). Bards were the guardians of oral tradition and history; they were the custodians of the unwritten collective memories that are essential to the functioning of non-literate communities and groups. Like the blacksmiths they were one of the casted groups or nyamakalaw within the Mande caste system. They were entertainers, historians, genealogists, advisers, mediators, musicians, and propagandists. Their artistic and ritual activities were central to the creation and maintenance of shared identities, shared cultural references, and shared symbols and shared norms in the wider Mande culture area (Miller 1990, Belcher 1999). The famous epic poem Sunjata was (and is) memorized and performed by Mande bards. It claims to recount the events surrounding the founding of the Mali empire, extolling the exploits of its supposed founder and hero, Sunjata, and the reputation of his war-house, the Keita clan (Niane 1965, Austen 1999, Conrad and Condé 2004). Families of bards were associated with particular houses. They praised their warlord patrons and heightened their prestige, defining and broadcasting the reputations of Mande warlords (Şaul 1998, Frank 2000). This created a reputational field which helped define and standardize their identities as warriors and rulers. It also helped create something like a regional field of governance that regulated the use of violence by warlords without creating the concentration of violence characteristic of a unified state or empire.

As knowledge specialists and ritual entrepreneurs bards were

monopolists who controlled access to their profession and to the market for praise and legitimate genealogies. In doing so they created the symbolic capital that constituted an important source of rents for themselves and for their patrons. This symbolic capital consisted of the reputations and the legitimacy of individual warlords and their houses; its value was increased as it was networked into a Mande-wide political and ideological construct, the reputational field that defined warlords and bards as translocal groups. Bards enabled this networking by creating shared meanings and a standardized idiom of reputation, praise, status and relationships. Standardization involving the narrow repertoire of tropes, stereotypes and clichés typical of oral tradition allowed the genealogies of different groups in the Mande area to be interwoven, memorized, and retold. These semi-mythological genealogies linked warlords to the exalted ancestors who featured prominently in oral tradition, particularly the Sunjata epic. These genealogies involved fictitious kinship relations that nevertheless generated real claims based on perceived obligations and reciprocities between distant individuals or groups (Jansen 1996a, 1996b, 2000).

Bards provided a kind of political charter, a form of minimal and

unwritten constitution. They were the impresarios of a common Mande political culture. For example, in some versions of the eponymous epic the triumph of Sunjata concludes with an assembly of the Mande warlords at

20

Kurukanfuwa in which each of them receive governorships or kingdoms (Niane 1965, Conrad 2004). Such a story may provide a basis for the mutual recognition of domains, rights and rents by groups of warlords, contributing to a lowering of political transaction costs – that is, the costs of making and enforcing agreements about the use of coercion and power. This might hold even if Sunjata, the assembly at Kurukanfuwa or, for that matter, much of the Mali empire, are entirely imagined. The story does not have to be true to have real effects. The Sunjata epic created focal points for the political specialists of the Mande culture area, evoking an overarching social and political order created by a supernatural culture hero. Sunjata became the personification of that order, and the Mali empire became its cognitively highly salient and enduring expression.

To summarize, bards were ritual specialists and the monopolistic

guardians of the narratives that created a common savanna political culture. They grew to be a translocal group as the reach of their narratives spread. The bards’ standardized narratives provided the warlords of Mande with status, reputations, and with distinct identities. This effective networking in turn made warlords an increasingly translocal group by helping to lower their political transaction costs. That is, bards helped to provide legitimacy and helped to strengthen the claims of warlords to rents from protection and predation. Political and economic transaction costs were likely lowered by the widespread perception – and the emergent reality – of a common savanna political culture. This common culture was given a cognitively very salient expression in the form of a quasi-supernatural and overarching political structure – a state or empire – founded on a charismatic and quasi-supernatural culture hero. A common political culture, though, is not the same as a common polity. But invoking a state or an empire makes easier the task of explaining a complex economic order an easier one – for bards and for historians, ancient and modern. This explanation suggests that the widespread belief in centralized and territorial west African empires may rest on a rather too literal interpretation of the narratives created and perpetuated by Mande bards and their warlord patrons. But, even if they were figments, these narratives served tangible and material interests. The stability provided by the translocal order these narratives helped to maintain was supportive of the savanna economy as a whole – which, in turn, helped to generate rents for the specialists who helped to maintain it.

6. Conclusion

We argued that it is not necessary to invoke large-scale or territorial political organization to account for complex patterns of economic integration. Savanna states or savanna empires may not have been necessary to support the savanna economy. There is very little robust evidence for territorial states or empires in the archaeological or historical record.

21

We sought to provide an integrated explanation by suggesting a more

nuanced institutional view of the medieval West African savanna. In this view economic integration and political fragmentation coexist alongside an appreciation of the savanna’s durable oral traditions. We call the resulting political economy a translocal order.

We argued that there were at least three interlocking institutional

complexes in medieval West Africa. Each of these complexes consisted of various translocal groups, institutions formed by occupational and ritual specialists who discovered feasible and sustainable institutional solutions to problems of production, protection and exchange. The occupational and ritual activities within each complex were complementary in their enforcement. Access to these activities was controlled, allowing translocal groups to create reputational fields which validated and assured the quality of people, products and promises in long-distance transactions. Translocal groups and the institutions they formed enabled entrepreneurial rents and protected them from dissipation by predation and opportunism. The three institutional complexes discussed here were interlocking with the various translocal groups bargaining over a mutual recognition of enforcement domains and memberships. This translocal order enabled a certain amount of Smithian growth. It might profitably be understood as an unintended consequence of the institutional entrepreneurship of translocal specialists who were searching for enforcement mechanisms that would allow them to create and protect rents. They were searching for economic and political order, and created it as they did so. Bibliography Alber, Erdmute. 1999. “Violent Conflicts in West African Borgou on the Eve of

Colonization.” In Dynamics of Violence. Processes of Escalation and De-Escalation in Violent Group Conflicts, edited by Georg Elwert. Sociologus (Beiheft 1): 119-132.

Allman, Jean Marie, and John Parker. 2005. Tongnaab: The History of a West

African God. Bloomington: Indiana University Press. Atran, Scott. 2002. In Gods We Trust: the Evolutionary Landscape of Religion,

Oxford: Oxford University Press. Austin, Gareth. 2004. “Markets with, without, and in spite of states: west

Africa in the pre-colonial nineteenth century”. Working Papers of the Global Economic History Network 03/04. London School of Economics and Political Science: Department of Economic History.

22

Austin, Gareth. 2008. "Resources, Techniques, and Strategies South of the Sahara: Revising the Factor Endowments Perspective on African Economic Development, 1500-2000". Economic History Review. 61 (3): 587-624.

Austen, Ralph. 1987. African Economic History, London: James Currey. Austen, Ralph. 1993. “The Moral Economy of Witchcraft: An Essay in

Comparative History”, in Jean Comaroff and John Comaroff (eds.), Modernity and Its Malcontents: Ritual and Power in Postcolonial Africa, Chicago: University of Chicago Press.

Austen, Ralph, ed. 1999. In Search of Sunjata: The Mande Oral Epic As History,

Literature and Performance. Bloomington: Indiana University Press. Barzel, Yoram. 2002. A Theory of the State: Economic Rights, Legal Rights and

the Scope of the State. Cambridge: Cambridge University Press

Belcher, Stephen Paterson. 1999. Epic Traditions of Africa. Bloomington: Indiana University Press.

Bird, Charles S. 1999. “The Production and Reproduction of Sunjata”. In Austen 1999.

Blanchard, Ian. 2005. Mining, Metallurgy and Minting in the Middle Ages. Vol.3.

Continuing Afro-European Supremacy, 1250-1450. Stuttgart: Steiner.

Bowles, Samuel, and Herbert Gintis. 1998. “The Moral Economy of Communities: Structured Populations and the Evolution of Pro-Social Norms”. Evolution and Human Behavior 19 (1): 3-26.

Bowles, Samuel, and Herbert Gintis. 2004. “Persistent Parochialism: Trust and Exclusion in Ethnic Networks”. Journal of Economic Behavior & Organization 55 (1): 1-23.

Boyd R, H Gintis, S Bowles, and PJ Richerson. 2003. "The Evolution of Altruistic Punishment". Proceedings of the National Academy of Sciences 100 (6): 3531-5.

Boyer, Pascal. 2000. Religion Explained: The Evolutionary Origins of Religious Thought. New York: Basic Books.

Brooks, George.1993. Landlords and Strangers: Ecology, Society and Trade in

Western Africa 1000-1630. Boulder: Westview.

23

Butt-Thompson, Frederick W. 1929. West African Secret Societies: Their Organisations, Officials and Teaching, London: H. F. and G. Whitney.

Cohen, Abner. 1971. “Cultural Strategies in the Organization of Trading

Diasporas”. In The Development of Indigenous Trade and Markets in West Africa, edited by Claude Meillassoux. London/IAI: Oxford University Press.

Conrad, David, and Barbara Frank, eds. 1995. Status and Identity in West

Africa: Nyamakalaw of Mande. Bloomington: Indiana University Press.

Conrad, David C., and Djanka Tassey Condé. 2004. Sunjata : A West African Epic of the Mande Peoples. Indianapolis: Hackett.

Curtin, Philip. 1971. “Pre-Colonial Trading Networks and Traders: The Diakhanké.” In The Development of Indigenous Trade and Markets in West Africa, edited by Claude Meillassoux. London/IAI: Oxford University Press.

Curtin, Philip. 1975. Economic Change in Precolonial Africa. Madison:

University of Wisconsin Press. Curtin, Philip. 1983. “Africa and the wider monetary world, 1250-1850.” In

Precious Metals in the Later Medieval and Early Modern Worlds, edited by J.F. Richards. Durham NC: Carolina Academic Press.

Curtin, Philip D. 1984. Cross-Cultural Trade in World History. Cambridge:

Cambridge University Press.

Delafosse, Maurice. 1912. Haut-Sénégal-Niger (Soudan franc̜ais). Paris: E. Larose.

Devisse, Jean. 1988. “Trade and trade routes in West Africa.” In UNESCO General History of Africa, Vol. III, edited by M. Elfasi. Berkeley CA: University of California Press.

Dixit, Avinash. 2004. Lawlessness and Economics: Alternative Forms of

Governance. Princeton University Press.

Dixit, Avinash. 2009. “Governance Institutions and Economic Activity”. The American Economic Review 99 (1): 5.

Ehret, Christopher. 2002. The Civilizations of Africa: A History to 1800. Oxford: James Currey.

24

Ekelund, Robert B., Robert F. Hebert, and Robert D. Tollison. 2006. The Marketplace of Christianity. Cambridge MA: MIT Press.

Ensminger, Jean. 1997. “Transactions Costs and Islam: Explaining Conversion

in Africa.” Journal of Institutional and Theoretical Economics 153 (1): 4-29.

Evans, E.W., and David Richardson. 1995. “Hunting for Rents: The Economics

of Slaving in Pre-Colonial Africa.” Economic History Review 48 (4): 665-686.

Fage, John D. 1957. An Introduction to the History of West Africa. Cambridge:

Cambridge University Press.

Fehr, Ernst. 2004. "Human Behaviour: Don't Lose Your Reputation". Nature 7016: 432- 449.

Frank, Barbara. 2000. Griots at War: Conflict, Caste and Reconciliation in Mande. Bloomington: Indiana University Press.

Freitas, Kripa. 2007. “The Indian Caste System as a Means of Contract

Enforcement.” Northwestern University, Dept of Economics, mimeo. Garrard, Timothy. 1982. “Myth and Metrology: The Early trans-Saharan Gold

Trade.” Journal of African History 23 (4): 443-461. Greif, Avner. 2006. Institutions and the Path to the Modern Economy: Lessons

from Medieval Trade. Cambridge: Cambridge University Press.

Henrich, Joseph, and Natalie Henrich. 2006. Culture, evolution and the puzzle of human cooperation. Cognitive Systems Research 7(1): 220.

Herbst, Jeffrey. 2000. States and Power in Africa. Princeton: Princeton University Press.

Hopkins, Anthony J. 1973. An Economic History of West Africa. London:

Longman. Iannaccone, Lawrence R. 1992. “Introduction to the Economics of Religion.”

Journal of Economic Literature 36 (3): 1465-1496. Iliffe, John. 1995. Africans: History of a Continent. Cambridge: Cambridge

University Press. Insoll, Timothy. 2003. The Archaeology of Islam in Sub-Saharan Africa.

Cambridge: Cambridge University Press.

25

Insoll, Timothy. 2006. "Shrine Franchising and the Neolithic in the British

Isles: Some Observations based upon the Tallensi, Northern Ghana.” Cambridge Archaeological Journal 16 (2): 223-238.

Jansen, Jan. 1996 a. “Polities and Political Discourse: Was Mande Already a

Segmentary Society in the Middle Ages?” History in Africa 23: 121-128. Jansen, Jan. 1996 b. “The Representation of Status in Mande: Did the Mali

Empire Still Exist in the Nineteenth Century?” History in Africa 23: 87-109.

Jansen, Jan. 2000. “The Griot’s Craft: An Essay on Oral Tradition and

Diplomacy.” Research on African Languages and Cultures 8. Hamburg: LIT Verlag and Piscataway, N.J.: Transaction Publishers.

Kopytoff, Igor, ed. 1987. The African Frontier: The Reproduction of Traditional

African Societies. Bloomington: Indiana University Press.

Kranton, Rachel E. 1996 a. “The Formation of Cooperative Relationships.” Journal of Law, Economics & Organization 12 (1): 214-233.

Kranton, Rachel E. 1996 b. “Reciprocal Exchange: A Self-Sustaining System. The American Economic Review 86 (4): 830-851.

Kuba, Richard, and Carola Lentz. 2002. “Arrows and Earth Shrines: Towards a History of Dagara Expansion in Southern Burkina Faso.” Journal of African History 43 (4): 377-406.

Kuran, Timur. 2004. Islam and Mammon: The Economic Predicaments of

Islamism. Princeton University Press. Lal, Deepak. 1988. Unintended Consequences: The Impact of Factor

Endowments, Culture, and Politics on Long-Run Economic Performance. Cambridge MA: MIT Press.

Lange, Dierk. 1992. “Das Alte Mali und Ghana: Der Beitrag der Oraltraditionen zur Kritik einer historiographischen Fiktion“. Historische Zeitschrift 255 (3): 587-623.

Last, Murray. 1988. “Charisma and Medicine in Northern Nigeria.” In: Charisma and Brotherhood in African Islam, edited by D. B. Cruise-O’Brien and C. Coulon. Oxford: Clarendon Press.

Law, Robin. 1980. The Horse in West African History. London: Oxford

University Press

26

Lentz, Carola. 2009. “Constructing Ritual Protection on an Expanding

Settlement Frontier: Earth Shrines in the Black Volta Region.” In Shrines in Africa: History, Politics and Society, edited by Allan Dawson. East Lansing MI: Michigan State University Press.

Levtzion, Nehemiah. 1973. Ancient Ghana and Mali. New York: Africana

Publishing. Levtzion, Nehemiah. 1977. “The Western Maghrib and Sudan.” In The

Cambridge History of Africa, Vol. 3, edited by Roland Oliver. Cambridge: Cambridge University Press.

Levtzion, Nehemiah, and John F.P Hopkins, eds. 1981. Corpus of Early Arabic

Sources on West African History. Princeton NJ: Markus Wiener. Levtzion, Nehemiah. 1986. “Merchants vs. Scholars and Clerics in West

Africa: Differential and Complementary Roles.” In Islam in West Africa: Religion, Society and Politics to 1800, edited by Nehemiah Levtzion. Aldershot: Variorum.

Little, Kenneth. 1960. “The Role of the Secret Society in Cultural Specialization.” In Cultures and Societies of Africa, edited by Simon Ottenberg. New York: Random House.

Little, Kenneth. 1965. “The Political Function of the Poro, Part I.” Africa: Journal of the International African Institute 35 (4): 349-365.

Little, Kenneth. 1966. “The Political Function of the Poro, Part II.” Africa: Journal of the International African Institute 36 (1): 62-72.

Lovejoy, Paul. 1978. “The Role of the Wangara in the Economic Transformation of the Central Sudan.” Journal of African History 19 (2): 179-193.

Ly-Tall, Madina. 1972. Contribution à l’histoire de l’empire du Mali (XIIIe-XVIe

siecles). Dakar: Les Nouvelles Editions Africaines. Malowist, Marian. 1966. “The Social and Economic Stability of the Western

Sudan in the Middle Ages.” Past and Present 33 (1): 3-15. Mair, Lucy. 1977. African Kingdoms. Oxford: Clarendon Press.

Masonen, Pekka. 2000. The Negroland Revisited: Discovery and Invention of the Sudanese Middle Ages. Helsinki: Finnish Academy of Science and Letters.

27

Miller, Joseph C. 1999. “History and Africa/Africa and History.” American Historical Review. 104 (1): 1-32.

Mitchell, Peter. 2005. African Connections: An Archaeological Perspective on

Africa and the Wider World. Lanham, MD: AltaMira Press. Mauny, Raymond. 1961. Tableau géographique de l’ouest Africain au Moyen

Age. Dakar: Institut Francais d’Afrique Noire. McEvedy, Colin. 1995. The Penguin Atlas of African History. New Edition.

London: Penguin. McIntosh, Roderick. 2000. “Social Memory in Mande.” In The Way the Wind

Blows: Climate, History and Human Action, edited by Roderick J. McIntosh. New York: Columbia University Press.

McIntosh, Susan K. 1981. “A Reconsideration of Wangara/Palolus, Island of

Gold.” Journal of African History 22 (2): 145-158. McIntosh, Susan K., ed. 1999. Beyond Chiefdoms: Pathways to Complexity in

Africa. Cambridge: Cambridge University Press. McNaughton, Patrick. 1988. The Mande Blacksmiths: Knowledge, Power, and

Art in West Africa. Bloomington: Indiana University Press. Meillassoux, Claude. 1991. The Anthropology of Slavery: The Womb of Iron

and Gold. Chicago: University of Chicago Press. Miller, Christopher L. 1990. Theories of Africans: Francophone Literature and

Anthropology in Africa. Chicago: University of Chicago Press.

Monteil, Charles. 1968. Les empires du Mali, étude d'histoire et de sociologie soudanaises. Paris: G.-P. Maisonneuve et Larose.

Newman, James L. 1995. The Peopling of Africa: A Geographic Interpretation. New Haven CT: Yale University Press.

Niane, Djibril Tamsir. 1965. Sundiata: An Epic of Old Mali. Harlow: Longman. Niane, Djibril Tamsir. 1984. “Mali and the Second Mandingo Expansion.” In

UNESCO General History of Africa, Vol. 4, edited by D.T. Niane. Paris: UNESCO.

Norenzayan A, and Shariff AF. 2008. “The Origin and Evolution of Religious Prosociality.” Science 322 (5898): 58-62.

28

Northrup, David. 1978. Trade without Rulers: Pre-Colonial Economic Development in South-Eastern Nigeria. Oxford: Clarendon Press.

Nowak, Martin A, and Karl Sigmund. 2005. "Evolution of Indirect Reciprocity.” Nature 437 (7063): 1291-1298.

Obeng, Pashington. 2006. “Religious Interactions in Pre-Twentieth Century West Africa.” In Themes in West Africa’s History, edited by Emmanuel K. Akyeampong. Athens OH: Ohio University Press.

Oliver, Roland, and Anthony Atmore. 2001. The African Middle Ages.

Cambridge: Cambridge University Press. Owusu-Ansah, David. 2000. “Prayers, Amulets and Healing.” In The History of

Islam in Africa, edited by Nehemiah Levtzion and Randall L. Pouwels. Athens: Ohio University Press.

Pagel, Mark and Ruth Mace. 2004. The Cultural Wealth of Nations. Nature 428 (6980): 275-278.

Panchanathan, Karthik, and Robert Boyd. 2004. "Indirect Reciprocity Can Stabilize Cooperation Without the Second-Order Free Rider Problem.” Nature 432 (7016): 499-502.

Parker, John, and Richard Rathbone. 2007. African History: A Very Short Introduction. Oxford: Oxford University Press.

Perinbam, Marie. 1980. “The Julas in Western Sudanese History: Long-Distance Traders and Developers of Resources.” In West African Culture Dynamics: Archaeological and Historical Perspectives, edited by B.K. Swartz and Raymond E. Dumett. New York: Mouton Press

Pietz, William. 1987. "The Problem of the Fetish, II." Res 13: 23-45.

Posnansky, Merrick. 1973. “Aspects of Early West African Trade.” World Archaeology 5 (2): 149-162.

Rockenbach, Bettina, and Milinski, Manfred. 2006. "The Efficient Interaction of Indirect Reciprocity and Costly Punishment". Nature 444 (7120): 718-723.

Röschenthaler, Ute. 1999. “Fetische aus Bulo N’Guti.“ Tribus: Jahrbuch des Linden-Museums. Stuttgart: Museum für Länder- und Völkerkunde.

Şaul, Mahir. 1998. “The War Houses of the Watara in West Africa.” International Journal of African Historical Studies 31 (3): 537-570.

29

Soares, Benjamin F. 1996. “The Prayer Economy in a Malian Town.” Cahiers d'études africaines 36 (144): 739-753.

Spufford, Peter. 1988. Money and its Use in Medieval Europe. Cambridge:

Cambridge University Press. Sundström, Lars. 1974. The Exchange Economy of Pre-Colonial Tropical Africa.

London: C. Hurst. Tamari, Tal. 1991. “The Development of Caste Systems in West Africa.”

Journal of African History 32 (2): 221-250. Thornton, John. 1996. “The African Background to American Colonization.”

The Cambridge Economic History of the United States, Volume 1, edited by Stanley L. Engerman and Robert E. Gallman. Cambridge: Cambridge University Press.

Trimingham, J. Spencer. 1962. A History of Islam in West Africa. London and New York: Oxford University Press.

Webb, James L. 1995. Desert Frontier: Ecological and Economic Change Along the Western Sahel, 1600-1850. Madison: University of Wisconsin Press.

Wilks, Ivor. 2000. “The Juula and the Expansion of Islam into the Forest.” In

The History of Islam in Africa, edited by Nehemiah Levtzion and Randall L. Pouwels. Athens: Ohio University Press.

Werthmann, Katia. 2007. “Gold Mining and Jula Influence in Precolonial

Southern Burkina Faso.” Journal of African History 48 (3): 395-414. Wright, Donald R. 1977. “Darbo Jula: The Role of a Mandinka Jula Clan in the

Long-Distance Trade of the Gambia River and Its Hinterland.” African Economic History 3: 33-45.

Wright, Donald R. 1999. “’What Do You Mean There Were No Tribes in

Africa?’ Thoughts on Boundaries – and Related Matters – in Precolonial Africa.” History in Africa 26: 409-426.

30