From Affiniam-Boutem to Dakar : migration from the Casamance, life in the urban environment of Dakar, and the resulting evolutionary changes in local Diola organizations

Material Information

From Affiniam-Boutem to Dakar : migration from the Casamance, life in the urban environment of Dakar, and the resulting evolutionary changes in local Diola organizations
Reboussin, Daniel A.
Place of Publication:
Gainesville, Fla.
University of Florida
Publication Date:
Physical Description:
xii, 214 leaves : ill. ; 29 cm.


Subjects / Keywords:
Agriculture ( jstor )
Cash ( jstor )
Emigration ( jstor )
Gender roles ( jstor )
Labor ( jstor )
Labor migration ( jstor )
Rice ( jstor )
Slavery ( jstor )
Villages ( jstor )
Women ( jstor )
Anthropology thesis, Ph. D.
Dissertations, Academic UF -- Anthropology
Spatial Coverage:
Africa -- Senegal


General Note:
General Note:
General Note:
Includes bibliographical references (leaves 191-212).

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
002070520 ( aleph )
AKQ8794 ( notis )
34381813 ( oclc )


This item has the following downloads:

Full Text







Copyright 1995


Daniel A. Reboussin


More people helped me complete this dissertation than reasonably

can be listed here. Several, though, provided help without which I

could not have finished at all. My advisor and committee chair through

most of this long process, Ronald Cohen, has been a true mentor in many

ways, from my first contacts with him until now. He has been a strong

supporter of all aspects of my work and a sensitive critic of my

attempts to convey complex concepts as clearly as possible given the

limitations of my writing. His persistence and success in recuperating

from a severe stroke since August 1994 continue to be an inspiration to

me and others. Since Dr. Cohen's illness and subsequent retirement, H.

Russell Bernard has been kind enough to serve as my committee chair. I

would like to acknowledge his valuable editorial advice and professional

efforts to bring this work into its present form. Each member of the

supervisory committee has been in the first instance an exceptional

teacher. I am grateful that each has dedicated himself to teaching, and

that I have had the opportunity to have studied with him. Beyond this

each has been an essential critic and advisor, greatly assisting me in

bringing specific aspects of this research report into its final form.

I wish to thank Arthur Hansen and Marvin Harris of anthropology, R. Hunt

Davis of history, and Goran Hyden of political science. Although it

need not be said, all of the errors and omissions that remain are

entirely my own.

Antoine Badji also deserves special acknowledgment. He was my

assistant and chief informant in Senegal, helping me to adjust to the

particular difficulties of life in Dakar. He thereby exposed me to some

of the difficulties faced by Diola immigrants to the capital. He also

introduced me to village life in Affiniam-Boutem, and kept me well.

Through him I would also like to thank all of his family and the people

of his village, who made my stay with them comfortable and informative.

I reserve special acknowledgement for the women of Boutem, who

constantly impressed both Antoine and me with their strength, wisdom,

and kindness toward a stranger. I am honored to have spent the time I

had with them.

Two sources of federal support have assisted me in this work. In

1988 I received a Foreign Language Area Studies summer fellowship, to

take an intensive course in Wolof at the University of Illinois, Urbana-

Champaign. The Fulbright-IIE U.S. scholar program, along with the U.S.

Embassy through its USIA American Cultural Center, also provided me with

funding. Special thanks go to Jerome Faye and El Hadji Sarr, along with

the rest of the Senegalese staff in Dakar. I will not forget their

genuine friendship and critical support.

I would like to also thank Peter Malanchuk, who approved several

leaves of absence from my position at the University of Florida

Libraries. He has been a generous and tolerant supervisor, taking on

quite a bit of extra work while I have been gone, and my appreciation is

heartfelt. Thanks also go to Jenny Konwinski, who created a superb

digital map of the village from the diverse materials I provided her.

Finally, I would like to express sincere thanks to my own family.

They constantly demonstrate to me the importance of kinship. Patricia,

my mother, has taken several months of her time to help with household

management and domestic tasks that I have neglected while writing. My

brother David and father Roland have provided me with good advice on how

to proceed with quandaries of research design and statistics. My wife

Ann Glowasky has provided constant emotional support throughout my

research, suffering along with me since before the actual research

began. She has also taken upon herself all the financial

responsibilities of income-earning during the more than six months of

leave I have cumulatively taken from my job to complete the writing

phase of this research. Most importantly she has been an unwavering

source of emotional support despite sometimes not being entirely sure of

why it is that I am doing this.









The Case of Diola Women . ... 1
Migration as a Factor in Cultural Evolution 2
A Typological Outline of Migration 4
Time Period . ... 4
Duration . .. 5
National Boundaries and Distance .. .5
Economic Issues . .. 6
Purpose . .. 6
Degrees of Compulsion . 7
Typological Sketch of the Present Case 8
Time period . ... 8
Duration . .. 8
National boundaries . 9
Economic issues . 10
Purpose . ... 10
Compulsion . ... 10
Causes of Migration from Various Theoretical Perspectives 11
Neoclassical Models . 11
Social Groups as Factors in Migration .. .12
Dependency Models . ... .14
Neo-Marxism . ... .15
Conclusions on the Causes of Migration .. .17
Consequences of Migration . ... .19
Economic and Social Scientific Contributions .. .19
Dependency Views .. .21
Neo-Marxist Approaches .. .22
Conclusions on the Effects of Migration from the Village 24
Implications and Conclusions ... ... .. .25
Households and Voluntary Associations .. .25
African Women as Active and Independent Migrants 27
Conclusions . ... .31


Introduction . ... .34
Periods of Diola History . .. .37
Early Sedentism and Early Circumscription .. .37
Early States in Casamance (Thirteenth to Sixteenth
Centuries) . .. 40
Early European Trade, Slavery, and "Legitimate Trade" .. 44
Twentieth Century Colonialism and the Independent State of
Senegal . ... .52
Characteristic Patterns of Diola Migration for Each Period .62
Period One: Early Sedentism .. .66
Period Two: Early States .. .67
Period Three: Early European Trade .. .67
Period Four: Twentieth Century Colonialism and
Independent Senegal .. .69


Introduction . ... .73
Methods . ........ 74
Interviews . ... .75
Census . .. ... 79
Women's Association and Analysis of Dues Paying Records 81
Oral History of Migration from Boutem .. .83
Earliest Migrants to Dakar .. .89
Early Women Migrants .. .91
Diola Associations . ... .93
The Village "Men's" Association: "We Had No Big Brothers
Here" . ... 96
History of the Boutem Women's Association in Dakar 99
Summary . ... 103
Conclusions . ... 104


Introduction . ... 106
The contemporary village of Boutem .. .107
Women Migrants to Dakar: "Work Is Not for Finding Happiness" 113
Career Histories of Women Migrants .. .114
Working Conditions ... 122
Why Migrate? . .. 127
Commercial Endeavors ... 129
Household Expenses . .. 133
Rent and food ... 134
Water and utilities .. 136
Association dues .. 138
Remittances . ... 141
Clothing . 144
Health care . ... 146
School fees ...... ........ 147
Transportation ... 148
Summary . ... 149
Census Results . ... 150
Representativeness of Interviewees .. .155
Dakar Women's Association: The Contemporary Situation .158
Description of Women's Association Membership ... .163
Discussion . ... 167

. 34


Migration Theories ... 171
Historical Patterns of Diola Migration ... .175
The Case of Boutem .. 177
Discussion .. 179









Table page

1 Historical periods and associated characteristic forms of
migration . ... .. 72

2 Comparison of migrants' years away from the village ... 153

3 Percent of residents with given family name and quarter of
origin for three residential categories, women's association
membership, and interviewee status ... 156

4 Mean ages of all women, Dakar emigrants, association members
and interviewees . ... 158

5 Dues-payments as a percentage of membership paying in three
categories of regularity ... 161



Figure page

1 Map showing the Casamance region and the research site in
relation to the rest of Senegal . 9

2 Village of Boutem, Senegal . ... 109

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



Daniel A. Reboussin

December 1995

Chairperson: H. Russell Bernard
Major Department: Anthropology

This is a case study of rural-urban migration among the women of

Affiniam-Boutem, a predominantly Catholic Diola village in the

southwestern Lower Casamance region of Senegal, West Africa. It

includes the results of research conducted in 1989 and 1990, employing

several sources of information. Thirty interviews were undertaken in

Dakar with female emigrants of the village, and oral histories of

emigration and associated urban voluntary associations were collected.

In the village, a census focusing on migration histories was also

conducted. Migration from Boutem is best understood in terms of a

modified, contemporary approach to classical social scientific migration

theory. There are few opportunities for Diola women to earn money in

the rural setting. Because they have cash responsibilities towards the

support of their families, they have left their villages since the

beginning of this century to work in wage labor. .There are few social

constraints on their free movement, but voluntary associations in the

urban setting do restrict and direct the behavior of all emigrants to

some extent. Fines are levied for members who do not return to the


village by a set date to encourage wet season agricultural activities,

especially the planting and harvesting of rice.

Created among Boutem's emigrants to Dakar as social clubs in the

nineteen-thirties and nineteen-forties, these associations were later

employed as a means of collecting money for an early crisis: the first

death of an emigrant villager. Once regular dues were collected the

organizations changed, becoming increasingly formal to keep and record

these dues. While they maintain some aspects of more traditional rural

voluntary associations, which have helped the Diola to adapt to other

forms of migration for centuries, these changes were unlike previous

adaptations. They were enabled by an increasing number of formally

educated emigrants. My research found that the women's association was

able to collect substantial amounts of money over a few year's time.

These funds are employed to conduct development projects in the village,

such as improvements to the school and maternity clinic, and therefore

they represent an important means of improving the quality of life



The Case of Diola Women

This is a case study of rural-urban migration among the Diola of

Senegal, primarily focusing on women from the village of Affiniam-

Boutem. This case is particularly interesting insofar as Diola women,

both single and married, frequently migrate in greater numbers to urban

Dakar than do either single or married men from the same village. Diola

women often travel and live in groups, without other family members.

They also have achieved, on the whole, far greater success in gaining

urban employment than men from their villages.

The case of Diola women's migration is apparently unique in

Africa. In this dissertation, I present the results of nine months of

field research, and attempt to explain the causes and consequences of

the facts they represent. In accounting for the adjustment of Diola

women to the urban service sector, I apply theories of economic

development, gender roles, and migration. I examine local Diola social

institutions and show how they have survived in the transfer from rural

Casamance to Dakar. These institutions deserve special study, because

they can provide insight into the nature of state-society linkages in

Africa more generally.

Migration as a Factor in Cultural Evolution

Migration has played a key role in the evolutionary processes that

established human populations throughout the world. We have long known,

for example, that agriculture spread to Europe from the Middle East

during the Neolithic. The routes of specific cultigens have been

traced. Recently, one study examined evidence from 26 genetic systems

collected in 3,373 locations, finding that "the spread of agriculture

through Europe was not simply a case of cultural diffusion, but involved

significant differential reproduction of the new farmers whose origins

can be traced to the Near East" (Sokal et al. 1991:143). The authors

conclude that migration played a central role in the spread of

agriculture to Europe.

The literature on African migration overflows with examples of how

the movement of people has affected nearly every aspect of life,

profoundly changing economies, politics, religions, and social

organizations across vast periods of time. Archaeological and

linguistic evidence documents the role of early population movements in

shaping cultural, ecological, and demographic relations among African

peoples (e.g., Haddon 1911; Greenberg 1963; Mabogunje 1976; Shaw 1976;

Phillipson 1985; Rouse 1986; Johnson and Earle 1987; Austen 1987).

The literature on early states and long distance trade further

associates population movement with critically important processes in

African history (see Leary 1970; Quinn 1972; Oloruntimehin 1972, 1974;

Levtzion 1973, 1976; Ajayi 1974). Migration played a central role in

African political and economic systems as part of regular, systematized

patterns of village fissioning (Murdock 1949; Wilson 1951; Cohen 1978;

Johnson and Earle 1987).

When European traders established themselves in coastal Africa,

continental patterns of migration followed the changing loci of trading

activities. The commercial centers near trans-Saharan routes in the

interior slowly suffered due to the exclusive growth of what often

became coastal enclaves (see Hoselitz 1960:189; also Thomas 1960;

Mitchell 1969a; Rodney 1970; Leary 1970; Hopkins 1973; Flint 1974;

Austen 1987; and Hart 1987).

Contemporary Africa experiences ever more rapid change as migra-

tion urbanizes the continent at an unprecedented pace (Clarke and

Kosinski 1982; Adepoju and Clarke 1985:6-7; Hart 1987). Contemporary

migration in Africa must be considered within the context of a rapidly

growing work force, low job growth, and a set of multidimensional crises

that threaten the quality of life on many levels (Adepoju 1991).

With rapid change in the economic environment, migration provides

a quick way for people to adjust. Historically, West Africans have

actively used migration as an efficient means of adapting to changing

economic conditions (Hill 1963; Berg 1965; Little 1965; Coquery-

Vidrovitch 1991). Migration continues to function as a cultural means

of adjusting to economic changes in the present. The relationship

between the environment and migration, however, is complex and systemic

rather than causal. Migration causes urbanization to some extent, but

also functions as a means of adaptation to an increasingly urbanized

social and economic environment. While historically migration was used

to respond primarily to changes in the ecological environment,

increasingly it is used as a response to changes in the social and

economic environment.

As this case study illustrates, social institutions also play an

important role in this model. As I will outline in Chapter 2, certain

social organizations among the Diola served one purpose in historical

times, but since have been adapted to serve another purpose under

contemporary conditions. In particular, voluntary associations have

both affected how migration occurs and have themselves adapted to

changes in migration. The association of women migrants from Affiniam-

Boutem will be one focus of Chapter 3.

A Typological Outline of Migration

The migration literature is truly enormous, extending across all

world regions, academic disciplines and theoretical orientations. A

review of the literature on African migration alone (including

literatures on prehistory, on resettlement, on refugees, on labor

migration, and for example, on migration's role in economic development)

would require a book in itself. Heberle (1955), Petersen (1958),

Mangalam (1968), Du Toit (1990) and especially Byerlee (1972), Pryor

(1982), and Eades (1987) offer explicit, theoretically derived

typologies of migration.' In order to focus on the issues most relevant

to my own research, I outline some categories that I find useful for

comparison between the case with which I am most familiar and other

migration flows in Africa.

Time Period

First, I am concerned here with contemporary migration, not with

migrations in general or with precolonial migrations. I mention several

aspects of migration theory in general elsewhere in this chapter, in

order to locate the discussion in the context of the literature.

Several aspects of migration's role in prehistorical Africa have been

noted above, and I will consider precolonial migration patterns of the

Lower Casamance region of Senegal in particular in Chapter 2.

'For a typological essay focusing specifically on the forms of female
migration, see U.N. Secretariat (1993).


Second, duration is a critical variable for understanding cases of

migration. One may be absent from home for less than a single day, for

many years--or one may never return. In some cases, as in pastoralist

or hunter-gatherer groups, movement from place to place within a given

territory is a permanent part of life (see Petersen 1955).

Most often, however, migration is defined as movement from a

sedentary residence and as either permanent or temporary depending on

the duration of absence.2 There is of course a certain arbitrariness in

defining the length of time that separates an absence from home from a

migration. Seasonal migration introduces another complication. The

duration of movement is relatively short, perhaps three months, but the

number of individuals involved and the routinized nature of the

phenomenon may make seasonal migrations critical for the economies of

certain social groups or areas. Despite this, seasonal migration often

remains undocumented, especially in less-developed countries.

National Boundaries and Distance

The distances involved and whether or not migrants cross borders

are also important variables for comparing cases of migration. Crossing

a border in Africa may not be different in any practical sense from

remaining within a single nation. Except during crises, African bound-

aries are often abstract concepts to all but national bureaucrats and

foreign observers.

2See Mangalam (1986:7) and Beijer (1969:13) for two useful collections
of various social scientific definitions of migration; see also the
well-known formulation by Lee (1966:49-51).

Economic Issues

Most contemporary writers on African economies now explicitly

recognize the web of connections among rural and urban populations. A

single family may send members to both smaller local towns and the

capital city. There are cases where migration breaks up a community,

but more often migration creates connections among localities. Among

the benefits of these connections for rural dwellers are cash remit-

tances sent from urban kin (Keely and Tran 1989; Russell 1992), although

some theorists question the value of their impact on sending communities

(e.g., Kamiar and Ismail 1991). These questions "relate mainly to pro-

cesses of socioeconomic development per se" (Appleyard 1989:487), and

are not fruitful points for comparison among cases.

Especially when poorer members of a society migrate, there is some

evidence of a positive effect, relieving poverty (Russell 1992:273).

Remittances, for example, can be critical to the survival of rural

families and are frequently used to finance the construction of schools

and health facilities (Conde et al. 1986:108). Such social investment

is an important outcome of migration in the present case. Most urban-

ites remain strongly attached to their rural homes (Gugler 1969:148-151;

Peil et al. 1988), often strengthening their ties to the home village by

sending regular cash remittances.3


From the perspective of economic development, changing one's

employment from the agricultural to a non-agricultural sector of the

economy may be more important than the fact of migration itself

3See also the literature on voluntary associations and their role in
maintaining rural/urban ties in Africa and elsewhere (Mangin 1959;
Meillassoux 1968; Reveyrand 1986/87; Peil 1981).

(Johnston 1986; Mellor 1989). Since one may be employed in either or

both economic sectors regardless of one's residence, migration research

instruments cannot assume rural-urban migration and sectoral shifts in

employment are one and the same (Byerlee 1972). It would be best to

collect complete employment records of migrants. However, often

researchers are able to gather only information on the work a migrant

leaves and the job he or she gets (or hopes to acquire) at the destina-

tion of the move (see Winchie and Carment 1989).

Degrees of Compulsion

Finally, an important issue related to the purpose of a move is

the degree to which it is voluntary. The most dramatic illustrations of

this issue are found in the tragic cases of contemporary refugees4 and

the historical Atlantic slave trade." Other examples of involuntary

migration include forced labor such as corv6e, and village relocations

or resettlement, for which there is also a large literature (e.g., see

Koenig 1986, 1987; Fall and Mbodj 1989; Echenberg 1991; McMillan 1993;

Cook 1994).

Actually, much migration is not easily categorized as voluntary or

involuntary. Many people migrate to get better health care and other

social services, for economic gain, for better access to civil services

and infrastructures, or to be near family or friends as is the case

especially for many elderly women in Africa (Gugler 1989; Peil et al.

1988). Generally, Marxists tend to consider migration for economic

4See Brokensha and Scudder (1968); Cernea (1988); Chambers (1979, 1982); Gor-
man (1987); Hansen and Oliver-Smith (1982); INADES (1986); Kibreab (1985);
Koenig (1986); Refugee Studies Programme (1988); Schultheis (1989); Spring
(1979); UNHCR (1981); and U.N. Secretariat (1985).

SSee Curtin (1969) and Inikori (1982). Also, Basil Davidson's The African
Slave Trade contains an excellent reading guide for this most severe case of
involuntary migration (1980:289-293).

benefit as more compelled, while non-Marxists tend to consider such

migrations to be more a matter of choice. Underlying such contradictory

interpretations is the neoclassic theorists' assumption of individual

incentive and choice versus an emphasis by the collectivist thinkers on

the coercive capacity of social institutions.

As with many polar differences in interpretation, identifying

whether a migration is forced or chosen is insufficient. We need scalar

measures, so that examples of migration may be considered as more or

less voluntary rather than as either voluntary or compelled. The key to

understanding degrees of compulsion in migration is in the disaggrega-

tion of the implied variables. For example, factors hindering or

contributing to a particular movement can be elaborated with greater

precision, and the outcomes of a move may be defined more clearly in

terms of destination, purpose, economic effects, and duration.

Typological Sketch of the Present Case

The case of Diola women migrants is presented here briefly, to

introduce it in terms of the typology suggested. We will return to this

case in more detail in Chapters 2 and 3.

Time period

Wage labor migration became an important economic phenomenon

regionally in Lower Casamance only during the 1930s, although migration

for trade was a precolonial phenomenon. Historical migration patterns

are discussed in Chapter 2.


Today, large majorities of young Diola migrate during the dry

season (roughly January through June) from many villages to urban areas,

primarily Ziguinchor and Dakar.

Figure 1: Map showing the Casamance region and the research site in
relation to the rest of Senegal. Source: Adapted from United States.
Defense Mapping Agency. 1992. Digital chart of the world. (The source
permits free reproduction and adaptation).

National boundaries

Migration to the regional capital of Ziguinchor, as well as to

provincial towns such as Bignona, is common. Because of the peculiar

nature of the national borders of The Gambia, which itself is located .

entirely within the borders of Senegal, overland migrants from the

Casamance to Dakar must cross two international borders (see Figure 1).

The Diola certainly consider this an international migration. They most

frequently say "I'm going to Senegal" rather than "to Dakar" when they

travel to the capital. Nevertheless, this case is most appropriately

considered an internal migration, since both points of origin and

destination are within the national borders of a single country.

Economic issues

The purpose of my research was to investigate economic issues.

Findings from the research are presented in Chapter 3. Briefly,

however, I found that cash remittances were reported to be sent home

when a migrant had close relatives there. Most immigrants also bring or

send cash to their families at planting and harvest time, to pay for

cooperative labor groups. Voluntary associations in Dakar also organize

more substantial collections in cash and in kind for projects to repair

or construct schools, health facilities, and other village improvements.

This latter function was particularly interesting to me.


Most immigrants in my study, including almost all of the thirty

members of the women's association interviewed in Dakar, left family

farms in Casamance to work as domestic maids in the informal sector of

urban Dakar (see Lubell and Zarour 1990).


The question of how freely one undertakes a move from, for

example, a village with very limited land for staple rice agriculture,

is certainly a legitimate issue for debate. However, the village of

Affiniam-Boutem, the focus of this case, is universally understood to

have more than adequate arable land for both groundnuts and rice, as

well as a diverse set of natural resources providing adequate and nearby

fishing areas, fruit trees, and construction materials such as thatch

and clay. For residents of Affiniam-Boutem, it is a cliche to insist

that "we have everything we need here, except money."

Causes of Migration from Various Theoretical Perspectives

Neoclassical Models

Neoclassical explanations of migration begin with the assumption

that migration is caused predominantly by economic motives. Ravenstein

(1885, 1889) was the first to systematically formulate theoretical

statements about migration from this perspective. Despite their venera-

ble age, Ravenstein's "papers have stood the test of time and remain the

starting point for work in migration theory" (Lee 1966:47).

Social scientists have now applied this theoretical orientation to

various aspects of migration for over a century, making it one of the

longest-studied social phenomena. With such a long tradition of work on

this problem, researchers have had ample time to test many elements of

various theories, comparing them against both competing hypotheses and

entirely different models. The neoclassical approach is by no means the

only perspective garnering a substantial following in contemporary

migration research. However, it retains its vitality through adaptation

under sustained critique from traditional social scientists as well as

from its more recent radical critics, the dependency theorists and neo-


Economistic models within this orientation have tended to use

somewhat coarse units of observation throughout its history. Some

influential writers (e.g., Ravenstein 1889; Lewis 1954, 1955; Harris and

Todaro 1970; Todaro 1976) began modeling by associating broad economic

sectors (e.g., modern versus traditional) with whole geographic regions

(e.g., urban versus rural).6 Such models appear simplistic in

6The overly simplistic concept of the fully modern city surrounded by
increasingly distant and traditional villages, like the very definitions
of rural and urban, has inspired long and contentious debate. For a
recent review see Coquery-Vidrovitch (1991:6-10).

retrospect because the variability within such broad categories has

since been demonstrated repeatedly throughout the literature. Scholars

working within this orientation have, however, reacted creatively to

valid criticism, improving the power of this approach to explain the

causes of migration. For example, criticism of the 'dualism' supposedly

inherent to the model has been incorporated over time as a more fine-

grained approach to the observation of certain key variables has


Thus, groups previously assumed to be homogeneous (i.e., peasants)

are now commonly defined as members of smaller units according to a wide

range of cultural and other factors such as land tenure patterns,

specific economic conditions, political characteristics, and specific

measures of households' labor availability in relation to cultivable

land holdings (Byerlee 1972; Byerlee and Eicher 1972:4; Gluckman 1943;

Gulliver 1957, 1960; Harris 1959; Hill 1970, 1986; Miracle and Berry

1970; Niddrie 1954; Skinner 1960). Similarly, rural economies have been

demonstrated to have many elements, both modern and traditional, as do

urban-based enterprises (Byerlee and Eicher 1972:6,16) to which they are

often linked.

Social Groups as Factors in Migration

Regardless of the issue of unit scale in data observation, some

economists have argued that migration can be sufficiently explained with

economic data alone (e.g., Fields 1982; Knight 1972; Todaro 1976, 1980).

Proponents of this view have had to answer the criticism from their

colleagues that analyses based exclusively on economic data are often

insufficient to fully explain important noneconomic social phenomena

(Yotopoulos and Nugent 1976:220).

A consideration of social organizational factors improves the

analysis of such complex phenomena as migration. Neoclassical theory

has been modified (especially in the "new household economics" school)

to incorporate measures of such diverse factors as social networks, risk

aversion, stages of the life cycle, dependency ratios, and other infor-

mation theoretically influencing migration decisions (see Bender 1967;

Caldwell 1970; Epstein 1969, 1975; Goldstein and Goldstein 1981; Hammel

and Laslett 1974; Leslie and Richardson 1961; Sandefur and Scott 1981;

Sanjek 1982; Speare et al. 1982; Stark and Levhari 1982; Stark 1984a,

1984b; Tuma et al. 1979; Uhlenburg 1973).

Another challenge to the neoclassical approach involves studies of

the experience of individuals rather than of larger-scale (economic or

cultural) processes. The focus is on changes in the attitudes or values

of individual migrants. Scholars in this tradition assert that non-

traditional attitudes and values cause the breakdown of traditional

authority in addition to increasing the incidence of migration. Such

studies essentially challenge neoclassical assumptions by adopting

alternate presuppositions. They do so, however, in an abstract fashion

without providing any empirical support for their choice of assumptions.

To conclude from variant rates of migration across social groups that

one must study only individual migrants lacks sufficient basis. The

causes of migration in a given context cannot be established through

intensive studies of individual migration experiences. Rather, research

should be directed toward controlled comparisons of migration among

different groups and contexts. Case study work may be a necessary step

toward such comparative research goals.

Gugler (1968, 1969) provides an excellent discussion of the

differences resulting from a focus upon the individual incidence of

migration as opposed to an emphasis on the rate of migration in a given

population. He credits Mitchell (1959) with the earliest elaboration of

this distinction. While Mitchell considered the variables determining

migration to be either economic or personal, there are in fact many non-

economic factors that can be considered as political, social, or

domestic, among other categories (e.g., see Winchie and Carment 1989).

Examples include bottlenecks in labor availability, tax collection

intervals, domestic or life cycles, and dependency ratios. Variant

rates of migration among groups are most likely to be explained by

multidimensional variables including economic, cultural, social network,

household or family, and personal (e.g., demographic) factors.

While an atomistic approach to migration studies was common in the

1960s, it was challenged in the 1970s by models that used macro-level

units of analysis. The most popular of these new models was dependency


Dependency Models

As it does for other issues, the dependency approach to migration

stresses that the bifurcated rural/urban division, like the categorical

division of modern versus traditional, is a misguided attempt to depict

contemporary Third World societies simply. Writers from this perspec-

tive emphasize the interrelatedness of rural and urban economies, with

migrants carrying labor value out of the 'periphery' to the 'core' of

the world economic system.7

The value of the dependency critique was that it popularized

important inadequacies of the 'dual sector' model (e.g., see Lewis 1954,

1955). The dependency critique also established the importance of

international market factors more generally, emphasizing the effects of

the inequitable 'distortions' of market transactions undertaken between

Third World enterprises and those of the developed world (Emmanuel

7Shoemaker (1976) was the first writer to apply the dependency perspec-
tive to migration theory (Kearney 1986:339).

1972). Nevertheless, the dependency critique was itself rather


A key theoretical oversight of dependency models was their repro-

duction of the dualism of which they were so critical. In retrospect,

this dualism was probably a result of utilizing the very coarse units

adopted by both dependency models and the neoclassical models that were

targeted for criticism by this popular academic school.

The core/periphery division is no more successful as an analytical

tool for social science than the labeling of modern and traditional

factors. Empirically, these tools divide societies in very similar

ways. Without a successful analytical method capable of innovative

analyses of existing data, social research might still produce new data

through fieldwork guided by innovative theoretical insight.

Unfortunately, however, the dependency approach provides no theoretical

guidance for empirical research.

There is little detail even at the abstract level regarding the

mechanisms underlying the flow of labor and value in ways assumed by the

dependency theory. As Booth explained, the theory "could not specify

the mechanisms by which what capital 'needed' was translated into

reality at the local level" (Booth 1985:768). The 'core' economies

cause migration from without by means of the same vague mechanisms that

subject 'peripheral' areas to unequal economic treatment on inter-

national markets.


Interestingly, some of the most thorough and convincing critiques

of the dependency school have come from within the same tradition that

spawned it (e.g., de Janvry 1981; Brenner 1977; Warren 1980; see also

Palma 1978). Essentially, other Marxists have challenged dependency

models for their attribution of external causes for change in


'peripheral' society. Economic growth (and consequently, migration) is

explained only in negative terms by a "law of underdevelopment" (Frank

1967; Sweezy 1972; Amin 1974b).

The neo-Marxist ('modes of production' or 'articulation') orien-

tation is more convincing and withstands more intense criticism,

focusing on 'class' divisions within societies outside the developed

world and how these have influenced changes from within these societies.

This perspective emphasizes the diversity and vitality of rural

production, inquiring how the tenacity of indigenous organization is

maintained in the face of political and economic encroachments by

capitalist firms (see Goodman and Redclift 1982; Binsbergen and

Geschiere 1985).

Much important empirical research was undertaken in response to

the neo-Marxist French anthropologists working in Africa (e.g., Rey

1973; and Meillassoux 1972, 1981), particularly regarding social

patterns observable within villages and even households. The wider

social and political influences of such patterns upon women's roles,

their utility in explaining certain other gender-related patterns

(migration among them), and the effects of dense rural-urban linkages

are major emphases within this literature. Because of the attention it

places on interrelations between rural and urban economic activities

this approach is sometimes referred to as 'articulationism' (see Kearney

1986), but this term is jargon and offers no additional clarity.

Neo-Marxist writing has theoretically identified the household as

the pivotal locale from which individuals can be observed working,

consuming, and engaging in their own 'reproduction' as well as that of

their social structure. Rural areas maintain the structures of states

in the Third World first of all by subsidizing urban industry with

artificially low producer prices (Bates 1981). Additionally, these

areas provide services to laborers who can migrate to live temporarily

in urban areas without placing expensive demands on state distributions

for social services and welfare (Meillassoux 1972; Wolpe 1975). These

services are instead performed in rural areas, often by women in unpaid

roles within the household or by others in the village and throughout

the depressed rural economy, which benefits little itself in terms of

development (Schmink 1984; Mamdani 1985).

The focus on domestic, social 'reproductive' contexts in relation

to local political structures and economic production has proven a

fertile ground for research, cultivating vigorous debate relative to

issues of migration (e.g., Burroway 1976; CNFNA 1983). The popularity

of this literature is due in part to the fact that dependency theory

failed to gain any significant empirical support through field research.

Another important alternative to dependency theory is the literature on

women and inter-household relations in the Third World.S The atomism of

neoclassical economic models and the macro-level approach of the dual-

sector modernization and dependency models obscured the local contextual

determinants of migration. The latter became more visible through the

adoption of middle-range units of observation, largely inspired by

research and critiques from the neo-Marxist perspective.

Conclusions on the Causes of Migration

A valid critique of classical theory notes that broad economic

conditions cannot provide an adequate explanation of migration patterns

on the local level. There are nonrandom, measurable differences in the

rates of migration across social groups at this level. Individuals do

not react to economic conditions as atomistic units, as the classical

8For some examples from a burgeoning literature on the household, see
Chayanov (1966); Meillassoux (1972); Guyer (1981); Wood (1981); Netting
et. al. (1984); Schmink (1984); Leacock and Safa (1986); Moock (1986);
Boyd (1989); Bullwinkle (1989); and Pedraza (1991).

libertarian economic theory assumes. Rather, theoretical explanations

of variability in the rates of migration among different social groups

and categories provide satisfactory means of understanding why general

economic conditions do not affect everyone similarly. This challenges

the applicability of the assumption of rational and knowledgeable

individual decision making units (e.g., Harris and Todaro 1970). Econo-

mists such as Todaro (1981) and Harris (1978:110) have recast the

problem of migration in terms of behavioral adaptations to inequalities

in the structure of markets, responding to an essentially valid critique

from the left (see Kearney 1986:335-336).

The categories adopted by dependency writers were as monolithic

and cumbersome as the dualistic concepts used by the modernization

writers. Moreover, the dependency school has failed to adapt

successfully in the face of theoretical and empirical challenges. So-

called conventional theory, in contrast, has adapted well to

intermediate units of analysis.

The decade of the 1980s saw the development of an increasingly

sophisticated view of social organization by researchers interested in

economic and migration variables, among other research areas. As noted

above, economists incorporated concepts of household, small consumer

group behavior, and empirical 'distortions' in the marketplace from

other social science fields. Investigations into local social organiza-

tion and behavior were successful at disaggregating coarse, monolithic

categories such as 'rural', 'urban', 'modern,' and 'traditional' into

more meaningful variables and observable units. The resulting

intermediate-level units often have complex interactions among them. In

some sense, these improvements also were due to pressure from Marxist-

derived critiques of classical economic theory.

Consequences of Migration

Economic and Social Scientific Contributions

From the point of view of the receiving community, the effects of

migration are depicted most often in terms of the labor market. High

rates of migration reduce the cost of (especially unskilled) labor to

urban industry. Related food policy issues, such as whether governments

should subsidize consumer staples, are also important from this

primarily urban perspective (see Bates 1981; Timmer et al. 1983).

The literature on African urbanization points to many other issues

relevant to the communities receiving large influxes of migrants. These

include problems related to rapidly growing needs in housing, urban

infrastructure, and public services such as health and education, as

well as the difficulties of incorporating formerly rural peoples into

multi-cultural urban settings (see Mangin 1959; Gluckman 1961; Kuper

1965; Little 1965; Plotnikov 1967; Mayer 1969; Hance 1970; Middleton

1979; Hannerz 1980; Peil 1981; Coquery-Vidrovitch 1991).

The effects of migration on sending communities were rarely

considered prior to studies conducted by the British social

anthropologists during and after World War II (see the exceptional early

work by Thomas and Znaniecki 1927; Sorokin et al. 1932; and Thomas

1938). In British colonial Africa, a seasonal or "circulatory" pattern

of migration was the dominant means by which labor was supplied to urban

enterprises in many of the white settler economies.

This "circulatory labor" phenomenon appeared to damage rural wel-

fare, as in some cases there were not enough men left in rural areas to

grow the amount of food needed by consumers there (Richards 1939; Wilson

1941). The survival of traditional authority and culture seemed to be

at risk (Schapera 1947). For the administrators of indirect colonial

rule, this was a threatening possibility (Eades 1987). Fortunately,

however, the improved worldwide economic climate after 1945 averted

rural disintegration and the crisis it would have created for rural

peoples as well as colonial governments (Eades 1987; see also Hart 1987;

Hopkins 1973).

In the years since World War II, most non-Marxist anthropology

research has supported a moderately positive view of emigration from

rural areas. In Africa, rural households often receive half their total

incomes from remittances returned by members employed in urban areas

(Schapera 1947:62; Hyden 1980; Keely and Tran 1989; Russell 1992).

Empirical studies undertaken throughout Africa have demonstrated that

migration may in some cases strengthen traditional forms of authority by

providing resources to senior members of rural households, supporting

economic growth and innovation in rural agriculture (Read 1942; Watson

1958; Van Velson 1961; Hill 1963). Such work did much to weaken the

atomistic assumptions of neoclassical economists regarding rural Africa.

Early models of African migration assumed that labor was abundant

to the point of surplus, though unproductive in rural areas (Harris and

Todaro 1970; Lewis 1954; also see Ravenstein 1889). Field studies of

rural labor undertaken during the 1960s demonstrated instead that most

African countries face both seasonal labor peaks and bottlenecks during

different points in the agricultural cycle (de Wilde et al. 1970; Cleave

1974). Thus, rural labor was not plentiful and unproductive as it

appeared in such models, but instead faced constraints associated with

severe seasonal fluctuations in demand.

By concentrating people into areas that can both more efficiently

use labor resources and provide goods and services to increasingly large

conglomerations of consumers, migration plays a key role in the

development process (Caldwell 1969:204; Todaro 1976, 1980, 1981; Knight

1972; Harris 1978; Southall 1979). West African history contains

numerous examples of migration serving as an efficient adaptation to

changing economic conditions in both urban and rural areas (Berg

1965:161; Hill 1963; Little 1965).

Dependency Views

Writers in the dependency school have been particularly uninspired

with regard to research on the effects of migration on rural areas. The

model assumes a net loss in value for sending areas in the 'periphery.'

It provides no theoretical support for empirical research in the context

of these points of origin. These facts together have resulted in a

distinct lack of dependency research on this topic. With little empiri-

cal research spawned by this orientation, there is a correspondingly

small need to discuss any contributions of dependency theory toward

understanding the effects of migration on rural sending communities.

The dependency literature has provoked research on the effects of

so-called 'free trade zones,' with an almost singular focus on 'runaway

shops.' A relatively recent phenomenon in some Third World areas, this

term refers to contracts with large multinational firms that sponsor

satellite assembly operations for the garment and electronic industries.

Because these firms once operated only in developed economies, they are

called 'runaways.' Relocations from developed areas are often encour-

aged by favorable government policies in the Third World as well as

international trade treaties such as NAFTA. Another factor drawing such

firms to less developed countries is a plentiful supply of cheap (often

primarily female) labor for such operations in parts of East Asia,

Africa, and the Caribbean (see Lim 1978; Frobel et al. 1980; Nash and

Fernandez-Kelly 1983; Sassen-Koob 1983; Safa 1986; and Pedraza 1991).

Notwithstanding important work on this phenomenon, statements by

dependency researches often over-generalize from such specific cases.

An example is the following assertion: "Modern labor migration is a


very highly organized State-controlled movement of workers" (Kamiar and

Ismail 1991:562). Only a very specific and overly restrictive

definition of "modern" could make this a reasonable statement in light

of the numerous and highly variant cases of migration to urban areas

throughout the Third World.

Neo-Marxist Approaches

The neo-Marxist literature, in contrast, has supported a great

deal of research on a variety of specific local effects of migration

from rural areas. It demonstrably is more open to debate on the issue

of the net effects of emigration from rural areas than the dependency

school has been (Kearney 1986; e.g., see de Jonge et al. 1978). One of

the important aspects of this debate has been whether or not return

migrants contribute positively to the local economy. Do such indi-

viduals invest in productive growth either directly or through remit-

tances to relatives in the rural area? Or are they more apt simply to

improve their own families' level of consumption, without spreading

benefits more generally within the local economy?

The issues of return migration and remittances are related, in

that an evaluation of the role each plays in the home village often

signals an evaluator's attitudes toward migration and economic

development in general. Evaluations ostensibly weighing the economic

outcomes of diverse case studies instead often appear simply to use such

cases as evidence cited in support of a prior position on development.

Kearney summarizes the literature, primarily Latin American

(1986:346), and suggests a lack of empirical support for the thesis that

local economic benefits are a likely outcome of return migration (see

Mines and Massey 1985; Reichert 1981; Rhoades 1978, 1980; Stuart and

Kearney 1981; Swanson 1979; Wiest 1979). Indeed, studies undertaken

from diverse theoretical orientations suggest that expenditures by

returnees are more likely to be made for the purchase of consumption

goods for the family, including housing, land, and the education of the

return migrants' children (Cornelius 1978; Dinerman 1982; Chilivumbo

1985). However, as Gmelch (1987) argues, returning students and

professionals may bring more benefits back to their rural home areas

than migrant laborers (see Miller 1984). Furthermore, levels of

productive investment appear more significant if expenditures for such

'necessary' costs as housing are excluded from the analysis (Gmelch

1987:137). Finally, expenditures for so-called consumptive purposes can

have important, positive effects in economically depressed rural


The debate is not focused on the issue of whether (or how much)

money returns to the home village either through remittances or returns.

An enormous sum, representing an international financial exchange second

only to the trade in crude oil, is estimated to return annually to

villages worldwide through remittances (Russell 1992:269). Rather, at

issue is whether or not these monies represent a positive contribution

to economic development, either locally or nationally. Responses to

this question often have more to do with how individual writers evaluate

economic development itself than to analysis of remittance data

(Appleyard 1989:487).

Many writers disregard the value of local consumptive expenditures

to economic development. Nevertheless, particularly in the Sahel, these

have been "crucial to financing expansion of educational facilities in

rural areas" (Russell 1992:275; see Cond6 and Diagne 1986; Gould

1988:4.1.49; and Bradshaw 1988). Such expenditures, including the

construction of health care facilities and relying heavily on

remittances, represent an important investment in human capital. The

migrants I studied in my own research actively invested in just these

kinds of facilities in their home village. More generally, it has been

noted that especially when poorer members of a society migrate, while

"there is no automatic mechanism by which. migration and remittances

result in development. .limited available evidence suggests a positive

effect on poverty" (Russell 1992:273).

Conclusions on the Effects of Migration from the Village

The question of whether rural areas experience net material losses

or in fact gain from migration is an important research topic, ripe for

further empirical inquiry. Relatively few data have been collected to

clearly indicate actual capital flows in and out of specific rural areas

(Eicher and Baker 1982:226).

Notwithstanding the important issue of net capital flows,

empirical research on the diverse effects of rural emigration has

documented such dependent outcomes as the growth and expansion of

markets serving small urban centers (Southall 1979, 1989; Middleton

1979; Nicolas and Gaye 1988), increased national integration (Paden

1980; Skinner 1985), improvements in rural family consumption and

education levels (Chilivumbo 1985; Cond6 and Diagne 1986; Russell 1992),

and changes in the gender composition of the rural labor force (Staudt

1975; Chaney and Lewis 1980; Palmer 1985).

Dependency writers often claim that modernization and classical

social science theories assume that change, such as an increased rate of

rural-urban migration, is good (see Lipton 1980; Swanson 1979). In

fact, this is an often-repeated (but false) depiction of much of the

social science literature of the first half of this century. As noted

before, some colonial British anthropologists in fact argued that

migration was destructive of traditional African culture (e.g. Richards

1939; Wilson 1941; see Eades 1987). Reacting to such conclusions,

others have countered that it is wrong to assume that migration is

harmful to rural welfare (Read 1942; Watson 1958; Van Velson 1961).

Dependency models themselves generally consider certain outcomes

as foregone conclusions. An example is the assumed loss of value in

rural areas due to urban migration (Kearney 1986:354-355; see also

Griffin 1976; cf. Amin 1974a; Bohning 1975; Swanson 1979). Empirical

evaluation of specific outcomes is not supported by the theoretical

model. The value of empirical social research is instead that the costs

and benefits of such a phenomenon can be measured and weighed against

one another (Miracle and Berry 1970).

Only in subsequent academic generations have Marxist collectivist

traditions begun to actively support empirical research. Field

researchers in the neo-Marxist school have since observed migration in

its local context and gathered data concerning the functions of diverse

household strategies, including migration. It is no surprise that

empirical field research challenges the validity of outcomes that were

assumed in the dependency literature. Empirical work from the neo-

Marxist school has recently demonstrated that migrants may bring

significant benefits to their rural communities of origin (Wood 1981,

1982; Schmink 1984; Hart 1987; Griffin 1976).

Implications and Conclusions

Households and Voluntary Associations

A focus on village and inter-household social organization has

proven an innovative and useful way to consider specific local causes of

migration. In my view, the most useful aspects of this perspective owe

more to 'conventional' anthropologists than to Marxists. Certainly,

though, the interaction of these traditions has heightened interest in

this fertile research area. The household focus successfully provides a

means to combine macro- and micro-level analyses, as well as both

structural and individual approaches to the study of the causes and

consequences of migration.

The concept of the household as a basic domestic unit of

production and consumption remains difficult to operationalize, since it

cannot be defined similarly for all places and times (see Yanagisako

1979). However, in any given context it can be a heuristic interme-

diate-level model for reconciling problems encountered with analysis at

the structural or individual level. The household provides a context in

which various migration situations or circumstances can be interpreted.

Processes operating at the highest levels of social analysis, such

as urbanization, international trade, economic development, and the

like, have important consequences that affect individual choices to

migrate. At the same time, individual variables, including demographic

characteristics and personal migration experience for example, have also

proven to be important determinants of migration. At an intermediate

level of analysis, meanwhile, "the control the productive unit in the

rural economy is able to exert over the timing and length of the mi-

grants' absence can be crucial" (Gugler 1969:476).

At another, also intermediate level, as we have discussed,

voluntary associations often have an important effect on members'

contributions in support of basic needs in the home village. To

understand the role of these intermediate level institutions requires

location- and context-specific research in the field. However, this

focus promises to elucidate, for any specific case, causes of

variability in migration rates left unexplained by either macro-level or

individual variables. Households and voluntary associations have

important influences on their members, as has been demonstrated in the

migration literature (Mangin 1959; Little 1965; Plotnikov 1967;

Meillassoux 1968; Mitchell 1969b; Peil 1981, 1988; Wood 1981; Speare et

al. 1982; Schmink 1984; Traeger 1984; Boyd 1989; Lambert 1994; Woods

1994). Thus, studying the household as well as voluntary associations

and their influences upon members can provide important insights into

the determinants of specific migration patterns.

African Women as Active and Independent Migrants

Empirical research on women's roles as migrants anywhere in the

world remains uncommon and certainly is not yet well represented in the

literature. Research in Africa frequently has focused on women left

behind in rural villages, managing homes and farms alone, while men

undertake urban migration (e.g., Richards 1939; Wilson 1941; Cooper

1979; Chaney and Lewis 1980; Wilkinson 1983; Hirschman and Vaughan 1984;

Palmer 1985). Much of the research that does address women as migrants

assumes they are associationall" movers, accompanying husbands or

families, rather than undertaking to move independently. This notion is

largely outdated (Bilsborrow and Zlotnik 1992) but persists in the

literature just the same.

Some research on female migration per se has been published in

recent years, although there are still few examples of any kind (see

Diner 1983; and Lee 1989 for two historical examples; and Pedraza 1991

for a general review). In the past those that considered women in the

migration stream itself often focused on normative issues, rather than

questions chosen to advance theoretical or contextual understanding.

For example, unmarried or independent migrant women were frequently

assumed to be prostitutes (e.g., Nadel 1942; see also Little 1965, 1973;

Plotnikov 1967; Gugler 1968, 1969; but cf. Cock 1980; Brydon 1987;

Sudarkasa 1977). While this role does exist for African women migrants,

the presumption is distorted and accentuated by cultural bias both

within some African cultures and by Western observers (Pittin 1984;

Brydon 1987:167).

There is still a need for more case studies focusing on women as

migrants (Byron 1994), and particularly as independent migrants seeking

legitimate work. The image of the woman migrant as either a prostitute

or the dependent of a migrating man, while not entirely baseless,

certainly is not indicative of the important role many women migrants

play in contemporary Africa. In West Africa, for example, commerce has

provided an important attraction for women to leave their rural homes:

"Most of the millions of women involved in internal migration within the

various countries [of West Africa] would fall under the category of

commercial migrants" (Sudarkasa 1977:183). While this may overstate the

case, it indicates that the role of women as commercial migrants is

important in this continental sub-region. Other important research on

African women as independent migrants and important economic actors in

the urban setting includes Little (1965), Schuster (1979), Hansen

(1985), Moran (1990), and Bozzoli (1991). Hansen (1985) and Bozzoli

(1991) are concerned most directly with female migrants as domestic

servants, the focus of my own research.

Stichter (1985) asserts that relatively high rates of female

migration in Asia and Latin America indicate the greater economic

subjugation of women in these regions, calling them "free laborers."

Meanwhile, the general lack of female mobility in Africa indicates a low

social status for women, where in precolonial times their "status was

not dissimilar to that of slaves or serfs" (Stichter 1985:148). A view

of African women as entirely dependent on men for their mobility is not

without precedent nor is it entirely false in certain cultural contexts

(see Nadel 1942; Thandi and Todaro 1979, 1984; Cock 1980; Shah 1983;

Brydon 1987; Boyd 1989). However, African women migrate in surprising


numbers9 where there are the combined conditions of independence at home

and opportunity abroad, while remaining quite distinct from the Marxist

notion of a landless 'free laborer' (for West Africa see Caldwell 1969;

Hamer 1981; Oppong 1983; Sanjek 1976; Sudarkasa 1977; Yacoob 1983;

Zachariah and Nair 1980; also see Ochollo-Ayayo n.d. for evidence of

migration for East African single women; and Wells 1982 for the same in

South Africa). My own research supports the view that under certain

circumstances, "gender differences in the division of labor may favor

migration of women more than men" (Boyd 1989:657). In this case, the

reasons have more to do with the gender division of labor than with

issues of women's social status.

As elsewhere in the undeveloped world, high quality migration data

for Africa are scarce (Adepoju and Clarke 1985:17). It is rare to find

any data set in which gender variables are associated with valid

indicators of migration. National level census data are even less

likely to be adequate for investigating women migrants. Male migration

in Africa has been relatively well documented over a substantial period

of time, but information on women migrants is almost totally lacking.

Among other things, because women are more likely to work in the

informal sector, data on them are especially difficult to collect

(Bilsborrow and Zlotnik 1992; see Lubell and Zarour 1990). In

particular, specific "evidence on the determinants of female migration

in Africa remains virtually nonexistent" (Brockerhoff and Eu 1993:561;

see Thadani and Todaro 1984).

African women increasingly are migrating, both internationally and

alone (Adepoju 1991). If historical data from elsewhere are indicative

of the future trends in Africa, women's migration rates there will

9Brydon's (1987) data indicate that in Ghana, Avatime women leave for
the same reasons as men, and for those women under the age of thirty,
leave in much the same numbers as men.

continue to increase over time (see Byerlee 1972; Byerlee and Eicher

1972; Caldwell 1969; Connell 1984; Easterlin 1980; Fawcett et al. 1984;

Khoo et al. 1984; Melville 1978; Orlansky and Dubrovsky n.d.; Singh

1978; Roy 1983; Thadani and Todaro 1979; Thomas 1970; Traeger 1984;

Whiteford 1978; Lee 1989). Therefore, documenting female migration and

(perhaps more importantly) discovering the relationships between male

and female migration in Africa will become increasingly important.

Empirical data collected in Africa on women indicate that they

generally migrate at a younger age than do men. They are constrained

from migrating by such things as high fertility and marriage (Bilsborrow

and Zlotnik 1982; Brokerhoff and Eu 1993; U.N. Secretariat 1993).

Interestingly, women also are less likely than men to return to their

home villages once they move to an urban setting. This leaves many

African cities with predominantly female populations, especially among

the older age groups. Among the issues deserving of further research

attention are the feminization of older urban populations, inter-house-

hold relations among migrant families, and the economic roles women

migrants play in rapidly changing African cities. Migration also can

have important effects on gender relations, fertility, and the division

of labor in rural areas. "Internal migration, and particularly its

rural-urban form, is inextricably linked with other demographic phenom-

ena, as in the case where regional fertility differentials essentially

reflect the age-sex selectivity of migration" (Pryor 1982:25; see Farber

and Lee 1984; Brokerhoff and Eu 1993).

Historical evidence on coastal areas of West Africa during the

early twentieth century indicates that women were demonstrating an

economic independence from men in both the commercial arena (Brooks

1976) and in wage employment. The latter case is supported by the fact

that at least some Diola women in Casamance were being employed as dock

workers (seasonally, after the groundnut harvest) in Bathurst (Banjul)

as early as the 1880s, and in Ziguinchor by 1910 (Mark 1985:74; Roche

1976:316; Snyder 1978:240). Such opportunities would not be sufficient

to cause migration from a given community. However, this would require

a social environment that both supported the participation of women in

the cash economy and allowed their movement independent of men.

Presumably, this also indicates a need for cash among Diola women at

that time. This particular aspect will be discussed in Chapter 2.

In my own research, I focused on contemporary Diola women

originally from the Casamance region of Senegal. Many of them were

seasonal migrants, earning wages in the urban service sector, working as

domestic maids. I was interested in whether or not their migration

functions as a means of attaining capital for agricultural and other

productive investments in their village of origin. Women are prominent

in the migration flow from Casamance, and are particularly successful at

gaining urban employment. In the village, the division of agricultural

and other labor has undergone extensive change through time,

particularly since the colonial era (Linares 1970, 1981, 1985). Subse-

quent expenditures in their home village, including a repair of the

school roof and the construction of a maternity clinic, were financed

through dues and other contributions to one of several voluntary organi-

zations. Some funds, either sent as remittances or brought with them on

their return to the village, were used to hire cooperative labor groups.


No general theory of migration exists to integrate multiple and

competing models successfully. Because they often operate at different

levels of analysis, they are not necessarily mutually exclusive (Massey

et al. 1993). Several have benefitted from critical interaction with

their academic competitors. The models that have benefitted most from

continual modifications are able to operate effectively at intermediate

levels of analysis. The household concept provides one means by which

both macro- and micro-level data can be integrated and considered within

a given cultural context. However, it remains a heuristic device,

without promise of integrating diverse theoretical models.

Perhaps there is no need for an integrated theory of migration,

except in terms of its role in economic development. Migration patterns

diverge greatly depending on a great number of contextual situations and

variables. Careful case study work remains to be done to adequately

describe the full range of contexts in which women migrate. But if

there is no need for a separate theory of migration, there certainly is

no more reason to develop a separate theory of the female migrant. In

the case of Diola women, as we shall see in Chapter 2, women began to

use migration as a means of acquiring cash in the newly transformed

economy soon after Casamance was integrated into the colonial state of

Senegal. They were affected by economic changes differently from men

because of their social position in the agricultural economy. Diola

women's migration from the Casamance does not indicate that they were

targeted for exploitation. In fact, their status in traditional society

was relatively strong. They owned land and could divorce their

husbands, for example (Pelissier 1966:687). However, becauseQf_-the.

changes introduced by the colonial administration, their role as rice

producers was inadvertently devalued. As cash became increasingly

necessary, they sought access to the cash economy and found no

opportunities for earning wages in the rural setting. Thus, they sought

work in town, first nearby in Ziguinchor, but eventually further afield

in Dakar.

If we are to understand the causes and effects of migration in the

truly complex context faced by African women, more data are needed on

historical as well as contemporary economic opportunities at home and

abroad. These data need to be understood within the specific social

context of the source community, with its network of connections to

individuals and groups at the destination. Nevertheless, this is a

neglected aspect of migration research overall, and the particular

context of female migration may vary a great deal from that of men in

the same cultural setting. This situation requires specialized research

agendas and a particularly focused attention by researchers if good,

valid data are to be collected for women migrants. Similarly, while

there is no need for an "African theory of migration" (Byerlee 1972:17),

migration research in Africa is a specialized task requiring preparation

in a diverse range of background material. Wage-earning opportunities

have been generally unavailable to African women, perhaps due in some

degree to the colonial legacy. Some exploration of the cases in which

these opportunities have existed over time is therefore warranted.



Any discussion of the history of Diola migration before the

twentieth century must acknowledge the constraints implicit in the

scarcity of relevant, valid data. The available data are insufficient

to support the construction of a complete history of Diola migration,

and my goal is decidedly not to predict trends. It is nevertheless both

possible and useful to synthesize what data are available, interpreting

them in the light of comparable cases. The goal of this chapter is to

identify particular periods in Diola history during which the rates of

change were sufficiently slowed and sustained to permit a generalized

description that is applicable during a relatively long, rather well-

defined span of time.

The result of any such exercise is necessarily limited in its

usefulness as history. However, it may provide a useful

characterization of long term historical trends for a chosen cultural

phenomenon, in this case the changing patterns of Diola migration. The

goal of describing such long term trends in the patterns of Diola

migration will be pursued systematically in this chapter by first

introducing the criteria used to identify particular historical periods.

The identification and characterization of these periods will be the

focus of the second section of the chapter. The third section will

outline and describe the general characteristics of Diola migration

during each of these periods. Finally, the characteristics of Diola


migration during each period will be summarized in a table at the end of

the chapter.

What we know about the history of Diola migration is determined,

largely although not completely, by the limited availability of

information on the Diola before the twentieth century. Epistemological

issues are fundamental to the evaluation of scientific work, as they are

for all claims to knowledge (Kaplan 1964; Lakatos 1970). Such issues

are merely highlighted in a case such as this, where sources of data are

particularly scarce. In light of this scarcity, the criteria used to

identify generalized periods of Diola history are provided explicitly

below. They form relatively (not absolutely) stable periods within the

long history of dynamic change characteristic of the area. These

periods are emphatically heuristic categories. That is, they are useful

for my purposes--the study of Diola migration--but of unknown utility

for particular historical research, whether within or between identified


Periods of Diola cultural history are defined here in several

ways. For example, archaeological data indicate that important changes

in Diola subsistence patterns were taking place during the earliest

period. Because such patterns are theoretically associated with certain

types of migration, the first period is defined to separate the time

before such a change from the time after it. Thus, the first criterion

for defining a period is evidence for a theoretically important subsis-

tence change. In this instance, local evidence is considered signifi-

cant because we can infer changes in the dependent variable, migration,

based on observed changes in the independent variable, subsistence.

The second way periods are defined is based on a change in broader

regional conditions that is hypothesized to act as a selective factor

influencing specific local outcomes. For example, historians of the

broader region have documented the influx of other ethnicities into the

local area. Evidence of such large-scale population movements is also

observable in local archaeological data. The establishment of political

organizations that defended particular trade interests and

transportation routes in the local area restricted Diola territorial

expansion. This restriction, in effect, would have selected for certain

subsistence strategies, specifically for economic intensification and

increased sedentism.

Finally, a period may be identified by historical events having a

known effect either directly on the group itself, or on a range of

groups that may be compared with the Diola and its effects inferred

based on known cultural similarities and differences among these groups.

A relevant example is the effect of the Atlantic slave trade on the

Diola. While little specific information on Diola groups in particular

is available (e.g., see Bowser 1974; and BUhnen 1993), historical

documentation does exist regarding the effects of slavery on neighboring

groups such as the Balanta and Manjaku. Particular cultural

characteristics of these groups may be compared and contrasted with the

Diola in order to infer how they were affected by this important

historical process.

To summarize the criteria used to define each period of Diola

history in this chapter, I have focused on three kinds of evidence.

First, archaeological evidence of change in local subsistence patterns

is the best means of defining the earliest period. Second, historical

evidence of change in broad regional conditions, such as the

establishment of new trading patterns or states, has been used to define

the second period. Third, indirect evidence of change inferred from

documented changes in similar, nearby cultural groups is used to define

the third historical period. Finally, direct historical evidence is

available for the most recent period. For the purposes of this chapter,

I consider evidence for rapid cultural change based on these sources as


sufficient for separating historical periods. In the second section of

this chapter I will identify and define four historical periods using

the above criteria. In the third and final section, I will characterize

the forms of migration that are associated with each period.

Periods of Diola History

Early Sedentism and Early Circumscription

Linguistic, archaeological, and oral history data indicate that

Diola peoples originated along the Upper Guinea Coast of the Atlantic,

somewhat further south than their present location. A long-term, large-

scale movement of Diola populations northward brought them to the

southwestern corner of the present day Lower Casamance area of Senegal

as early as A.D. 200 (Linares 1971:41-43; Mark 1985; Baum 1986). This

trend continued until the eighteenth century when Diola advances into

the Fogny district to the northeast (primarily at the expense of the

Banyun ethnic group) were reversed by the Mandinka, whose own large-

scale, state-reinforced migration from the interior succeeded in pushing

the Diola back south and west of the Songrougrou River (Lauer 1969:59;

Quinn 1972:25; Brooks 1993).

Archaeological data for the description of this period come

exclusively from the work of Olga Linares (1971) Linares' evidence

suggests that the Diola were coping with subsistence stresses due to

persistent population growth as early as the second century A.D. Her

analysis is based on the changing frequency distribution of shellfish

species in the shell midden strata of the part of Lower Casamance

longest occupied by the Diola. Linares interprets these changes as the

IThis is the only article I know of reporting Diola archaeological
evidence. The limitations of a view based on a single source apply,
although its findings are consistent with other data sources.

result of particular shellfish resources, important foods at the time,

becoming scarce due to over-exploitation. Thus, new resources--

different shellfish species--had to be located where familiar ones could

not be exploited more effectively. Smoked oysters, for example,

continue to be an important protein source for many Diola in the


The earliest evidence of Diola culture in the Lower Casamance

indicates a reliance upon mixed agriculture as well as these foraged

marine resources. The arrival of Diola peoples in the Lower Casamance

probably was the result of groups moving northward to exploit new lands

suitable for paddy production, and toward more abundant supplies of pre-

ferred marine resources. There is evidence, already deposited in the

archaeological record by A.D. 200, of rice cultivation2 and animal

husbandry in the area. The presence of cattle bones in the record

suggests an early trade in cattle.3 Other domesticates such as pigs and

dogs, common in contemporary Diola villages, only appear in the record

about the time of European contact (Linares 1971:43).

Mixed agriculture was probably becoming an increasingly important

means of subsistence throughout the period. A population that continued

to rely extensively upon gathering dwindling natural resources would

have faced increasing nutritional deficiencies and disease. The Diola,

however, were already familiar with the benefits of a subsistence

strategy that included agriculture. From the second through about the

twelfth century, various Diola groups spread throughout the tidal

ecological zone of southwestern Lower Casamance (see Adams 1993). As

2West Africa may represent an independent point of origin for irrigated
rice (see Dresch 1949; Porteres 1956, 1970; Johnny et al. 1981).

3Herds of trypanosome-resistant N'Dama cattle are maintained in the
Lower Casamance, but their rates of reproduction are very limited.
Trade is the only feasible means of expanding herds (see Starkey 1984).


this occupation became increasingly complete, eventually villages could

no longer simply fission to maintain a balance between people and

natural resources (see Cohen 1978:35,53). Instead, they had to rely

increasingly on subsistence resources for which production could be

intensified through management, primarily through the increased

application of manual labor in agriculture. Irrigated rice production

is particularly responsive to this strategy.

In summary, during this first period, the primordial process of

village fissioning or hivingg off" most familiar among foragers and

horticulturalists continued alongside the intensification of rice

agriculture and other relatively newly-introduced economic strategies,

such as cattle trade and husbandry. Increasingly then, the Diola

pursued sedentary strategies as populations expanded relative to marine

and forest resources, and as less territory was available for exploita-

tion in the coastal ecological zone. Archaeological evidence from this

period indicates that single Diola villages expanded in population and

area over the course of up to four hundred years in some cases. This

process required a substantial intensification of inhabitants' subsis-

tence activities, through such enterprises as land reclamation from the

saline mangrove marshes and the artificial irrigation of these new rice

paddies (Linares 1971:41-43; Vieillefon 1977; Loquay 1981; Pellissier

1966). While large scale population movements into new areas were

becoming a less important means of maintaining the balance between

people and resources throughout this period, groups within these now

sedentary villages continued to fission from them and to diffuse

throughout the Lower Casamance. Eventually, these groups established

new villages in some of the most remote delta plateaus to the north of

the Casamance River4 by the late seventeenth century (Roche 1976:24;

Baum 1986:74; Linares 1983).

Early States in Casamance (Thirteenth to Sixteenth Centuries)

Given the evidence of cattle bones in the archaeological record,

as well as the limitations of trypanosomiasis, the Diola were presumably

involved to some extent in cattle trade from the time of their earliest

occupation of Lower Casamance. Even if trade was not undertaken until

much later, it remains clear that relations with the indigenous states

of the western Sudan were well established by the Middle Ages. Mande

population expansion and migration5 from the interior westward facili-

tated contact between these states and the stateless peoples of the

Atlantic coast.6 This migration was therefore an event of great

regional significance.7 Archaeological data support historical sources,

indicating that Mandinka peoples expanded from the interior westward and

southward toward the Upper Guinea Coast during the Middle Ages through

about 1700 (Lauer 1969:59; Leary 1970:39-43; Rodney 1970; Quinn 1971:9-

10, 1972:25; Mark 1985:11; Baum 1986:80). They had organized trade

there sufficiently to be exporting kola nuts, "that eminently perishable

product," to North Africa by the twelfth century (Person 1984:304).

4Sapir (1965) provides a map of about fifteen Diola dialect differences,
illustrating the linguistic effect of fissioning and the subsequent
separation of groups into many remote locales.

5Brooks (1993:87) attributes this migration to an extended dry interval
in the interior climate from about 1100-1500, approximately concurrent
with this second period.

6The Diola, Balanta, and Manjaku ethnicities all speak languages in the
Bak group of the West Atlantic family, and therefore probably are of
similar origins (Lauer 1969:7-8; Sapir 1971:45-112; Fivaz and Scott
1977:17-18, 309).

7Person (1984:318) compares its importance with the nineteenth century Zulu
migrations in southern Africa.


The Mande peoples are particularly important to the early history

of the western Sudan, as they built the powerful Mali empire during the

1200s. Mali controlled trans-Saharan trade in gold, kola, salt, and

slaves for four centuries, surviving until the mid-1600s. Indigenous

populations of the Upper Guinea coastal region were contained or

circumscribed (Carneiro 1970) by the fragmented population expansion

from the interior and by the growing Mande states, established in the

area from about the thirteenth century. The arrival of small states in

the area surrounding the Lower Casamance marks the beginning of a second

historical period. Early 'colonists' from the interior were traders

with ties to North Africa as early as the eighth or ninth century,

leaving no archaeological evidence of subsistence activities adapted to

the coastal environment (Person 1984; Linares 1971:38). Among the

peoples moving westward from the interior were the Banyun,' who played a

particularly important role vis-a-vis Diola migration during the third

period. Thus, the four major ethnic groups occupying the area between

the Gambia and Geba Rivers today, as they did prior to 1500, are Diola,

Balanta, Manjaku and Banyun (Lauer 1969:3-7).

'The Banyun originated in the area that is presently northeastern
Guinea, as indicated by linguistic evidence, and were probably pushed
westward by the Mande expansions (Lauer 1969:7-8). While they call
themselves "Iagar" or "Ihadja", they are known by various names,
including Bainunk, Bagnun, and Banhun, all derived from Portuguese
Creole. The Banyun language is a member of the Eastern Senegal-Guinea
group of the West Atlantic family, along with Tenda and Biafada, among
others (Westermann and Bryan 1952; Greenberg 1963:6-41; Lauer 1969:6-8;
Sapir 1971:45-112; Fivaz and Scott 1977:18-19, 309). The Bassari of
southeastern Senegal also speak a language in this group (see Person
1984:306). By local tradition the Banyun are considered autochthonous
to the Lower Casamance (Niane 1989:9; Baum 1986:102-103), although the
ethnic origins of the area's original inhabitants are far from clearly
established (see Baum 1986:46-57).

The Mali empire reached the Atlantic coast from its core on the

Upper Niger by the thirteenth century.9 The trade of goods between

coastal Casamance and the interior consisted primarily of salt, but

included rice, slaves, dried or smoked fish, and even perhaps smoked

oysters in return for iron, horses, and small amounts of gold (Lauer

1969:26; Person 1984:313; Niane 1989:10). In order to control similar

trade, the Mandinka founded states all the way to the Atlantic on the

north bank of the Gambia River. Significantly, however, on the south

bank these states reached only to the Vintang Creek, the terminus of the

Banyun trade network (Brooks 1980:6). This network, dominated by the

primarily Banyun state of Kasa, flourished over the long term (Mark

1985:14-15), eventually linking all peoples of the Lower Casamance.

Over time, its traders forged communications and exchange ties from the

Lower Casamance to the south bank Gambia state of Geregia to the north,

the Mali and Kaabu empires to the east, and south to the Portuguese

commercial enclaves by means of the "most important commercial channel

to Cacheu until the nineteenth century" (Mark 1985:11-15; also see

Brooks 1980:6, 1993)'.

These Banyun trade routes were controlled by the small Kasa state,

oriented toward the interior from its location in eastern Lower Casa-

mance. This state, sometimes referred to as Cassanga, was itself

originally an outlying vassal province of the Mali empire (Lauer

1969:61; Baum 1986:80).1" Over time, however, this formerly peripheral

region began to assert political control as Mali's power waned in the

9Lauer (1969:25) estimates a Malinke arrival in the lower Gambia by the
early fourteenth century, while in Brooks' (1980:6) judgement Mandinka
trade routes were established there during the eleventh or twelfth
century (see also Person 1984:304).

"Note 81 on the cited page refers to the following historical sources:
Rodney (1970:109-113); Pereira (1971:88); Boul6gue (1972:6); Monod et al.
(1951:57-58); and LeBlanc (1649:28-31).


fifteenth century. An indication of this westward shift in the relative

center of political power was the incorporation of the Kasa state into

the Mandinka Kaabu empire (Brooks 1980:7; see also Quinn 1972:33; Person

1984:313; Forrest 1992:9; Girard 1992).

The existence of the Kasa state was first documented by European

writers between 1580 and 1669 (Mark 1985:25). Parenthetically, Mansa is

the Mandinka term meaning kingdom, thence the generally recognized

origin of the regional name of Casamance. The subjects of this kingdom

were known as Cassanga, although the term has also been used variously

to refer to a Banyun clan and a purportedly independent ethnicity (see

Baum 1986:46-57; Person 1984:314; Brooks 1980:13; Lauer 1969:25). The

Cassanga were apparently an ethnically diverse people, incorporating

Diola, Balanta, and Banyun peoples, as well as Luso-African langados

after European contact (Mark 1985:17; Brooks 1980).

To summarize the most influential features of this second period,

trade grew in importance from the time that Mande population expansion

brought the trans-Saharan trade network as far west as the Lower

Casamance. While this expansion brought with it new trading opportu-

nities, it also circumscribed the Diola, limiting opportunities for

continued territorial expansion as well as for political and economic

independence. Most importantly for the Diola, the state of Kasa con-

trolled a Banyun trade network that linked them with the other states of

the region, including the Mali empire until its decline during the

fifteenth century. Kasa's political organization, economic functioning,

and ethnic composition are all rather poorly documented in the

historical record. Diola were noted by early European observers to form

one part of its citizenry (Mark 1985:14-15). Most probably, however,

the state was composed primarily of Banyun, and organized along the

lines of other Mandinka states founded on the exploitation and defense

of valuable trade resources and transportation routes.'

Early European Trade, Slavery, and "Legitimate Trade"

More so than the second, this third period (dating from about the

fifteenth to nineteenth centuries) is defined by the activities of

political organizations far removed from the Diola themselves.

Specifically, the arrival of European merchants was a critical factor in

catalyzing and speeding changes already taking place. The Banyun, for

example, had already demonstrated the economic and political power of

Kasa by preventing powerful Mande states from encroaching on their

territory. Banyun economic strength was, however, increasingly linked

to the fortune of European merchants (especially the Portuguese) as the

importance of the Atlantic slave trade grew. This was to play a central

role in the eventual failure of the Banyun to predominate in their

territorial conflicts with the Diola in Lower Casamance.

The introduction of new trading opportunities with European

merchants had the effect of raising the stakes of competition among the

various political groups of the Lower Casamance. However, other factors

were as important as the increased economic value of the early European

trade. For example, trade with Europeans oriented African economic

activities toward their coastal enclaves, rather than toward the

overland routes controlled for centuries by the states of the interior

"The Banyun developed a strong reputation as traders among newly
arrived Europeans (Lauer 1969:7-8), and were noted as the sponsors of
large market fairs every eight days, for example (Mark 1985:12). Later,
the Banyun were most often hired for extensive periods as navigators of
trading craft, compradors, and the like (Brooks 1980:5). In contrast,
the Diola, Balanta and Manjaku groups were noted by Europeans during the
fifteenth century as generally avoiding extensive involvement in trade
relations (Lauer 1969:32-35). They "excluded Portuguese and Luso-
African traders from their territories and restricted commercial
exchanges to places and arrangements of their choosing" (Brooks 1980:5).

(Hopkins 1973:79). This economic and geographic "about face" had pro-

found effects on the relative political strengths of many groups.!

The nature of the slave trade in particular also had penetrating

effects on the societies among whom slaves were captured, and indeed

perhaps more so among those who participated in raids for the capture of

people for sale. There is evidence that those societies most intimately

tied to and benefitting from this trade in human chattel were also those

most devastated by it.

Finally, the vastly increased availability of iron due to the

European trade was a remarkable fact in itself, as it was a critical

resource in both warfare and agriculture. The Diola were "pre-adapted"

(see Cloak 1986) or best suited to take advantage of this profound

change in the environment. This was due in part to the fact that the

Diola, unlike the Banyun, used iron not only to trade or to fabricate

effective weapons, but also to make agricultural tools. Iron enabled

further intensification of their agricultural economy (Lauer 1969:62).

The increased availability of iron through trade allowed them eventually

to succeed in gaining a relative political advantage over the Banyun

that was never relinquished. Today the Banyun have largely been

incorporated into Diola communities (Mark 1985:19-20,31).

While the effects of the slave trade and the increased availabil-

ity of iron had an importance independent of their economic value, the

value of these trade activities is well documented and does provide an

indication of the general importance of European trade in the Lower

Casamance during the sixteenth and seventeenth centuries. For example,

by one estimate between fifteen and twenty-five metric tons of iron were

imported to the Lower Casamance annually during the late sixteenth

12The effect of such changes in geo-economic orientation will not be discussed
here, but see Austen (1987:81-108); and Hart (1982).

century (Mark 1985:29). This large amount of iron was exchanged for

slaves in addition to material goods. Slaves in particular were being

exported from the area in very large numbers during this period; in

1676, for example, documents indicate that 220,000 people were sold into

slavery from Lower Casamance (Baum 1986:154)."

The dramatic expansion of trade that accompanied the establishment

of European outposts in the Lower Casamance gradually drew Diola

participation. This is indicated by an increasing Diola population in

the south bank Gambian state of Geregia (Mark 1985:24). Nevertheless,

the Diola continued to maintain their noted distance from direct trade

with the Europeans. The export trade in goods such as beeswax, ivory,

hides, and eventually captives, although undertaken within Diola vil-

lages, was most likely conducted by Banyun traders, who traveled among

many Diola villages at the time (Mark 1985:24; see also Coelho 1953:30).

A combination of political, economic, and ecological factors

increasingly supported a Diola advantage over the Banyun. Early in the

seventeenth century, the Banyun were at the peak of their political and

economic power. They were preying on the Diola for captives to sell to

the Portuguese slave trade, and had recently gained their independence

from Kasa (Lauer 1969:55-56; Mark 1985:24-25). However, the Portuguese

trade itself was in decline by mid-century. The Diola, particularly

north of the Casamance River, were able to continue their access to

trade through British and French posts (Mark 1985:53). Furthermore,

they could use the trade in iron greatly to their advantage. As noted,

iron was a key factor supporting an intensification and expansion of

"This sort of precise information on trade reflects the substantially
improved historical sources available for the late seventeenth century
(Mark 1985:22): see Coelho (1953); Teixeira da Mota (1977). Cultru
(1913) describes the voyage of de la Courbe in 1685, an account later
plagiarized by Labat (1728); see also Froger (1698).

their staple rice agricultural system, as well as an important resource

for their military power (Lauer 1969:62).

While the slave trade along the Gambia River continued through to

the nineteenth century, it peaked there during the late seventeenth

century (Quinn 1972:8).11 Quinn attributes this relatively early

decline in comparison with elsewhere in West Africa to the high prices

that Senegambian suppliers began demanding during the eighteenth

century. The Diola appear to have begun the seventeenth century with a

defensive posture toward the slave trade (Quinn 1972:10). However, by

the eighteenth century at least some Diola groups participated heavily

in it. Raiding among Diola villages became commonplace (Mark 1985:25-

31; Baum 1986:157).

Other ethnic groups in the area apparently suffered more during

this period, perhaps due to an earlier participation in slave raiding.

Internal factors such as greater social stratification, as well as

external factors such as vassalage to politically dominant states, were

also important differences between the Diola and many neighboring

groups. Lauer (1969:32-33) notes that large numbers of Manjaku and

Biafada have been reported in studies of some New World slave popula-

tions. He attributes their relative over-representation in the Americas

to these factors.15

The Diola reaction to slave raids against their villages was

primarily defensive at first. Houses in the affected areas, in Fogny

for example, were often surrounded by pikes and thick walls to defend

against raids from the Mandinka and Banyun (Pelissier 1958; Thomas 1968;

Quinn 1972:10; Linares 1983; Mark 1985:26; Baum 1986:96, 184-185). By

"See also Curtin (1975).

1But see Baum's (1986:155) reference to Bowser's documentation of 387 Diola
slaves being taken to Peru in 1605 (Bowser 1974:40-42; also see BUhnen 1993).

the late eighteenth and early nineteenth centuries, however, the Diola

had begun trading with Europeans more fully (Lauer 1969:35). This in-

cluded participation in the Atlantic slave trade, albeit rather late in

its history, and often through African middlemen rather than directly

(Baum 1986:155-156). Eventually many Diola communities saw exhaustive

participation in the slave trade as both victims and aggressors,

including raids between Boulouf and Bandial (north and south shore)

Diola groups (Quinn 1972:26; Mark 1985:25-31; Baum 1986:159-163). By

way of confirming the extent of their participation Baum (1986:164-175)

reports a set of detailed rules among the Esulalu Diola regarding the

capture and sale of slaves. Baum reports an "increasing frequency of

raids for captives during the second half of the eighteenth century"


As the Atlantic slave trade declined and legitimate forms of trade

with Europeans increased, a new set of opportunities arose for young

Diola men in particular. Like the slave trade, the legitimate trade had

profound effects on the structure of Diola society. At the same time,

Europeans were experiencing great difficulties with trade in the highly

factionalized and competitive West African economic environment (Hopkins

1973; Austen 1987). Unfortunately for the historian, the decline in

Portuguese trade in Lower Casamance left fewer sources for the

eighteenth century than are available for the seventeenth century. The

sources that do exist, however, confirm earlier accounts regarding the

transition of political power in Fogny from the Banyun to Diola, and the

disruptive effects of the slave trade (Linares 1983; Mark 1985:53).

From the northernmost districts of Fogny and Combo the Diola had

access to British trading posts along the Gambia (Mark 1985:65-66).

There, the exchange of wild rubber and palm kernels became a popular

means of acquiring cash and consumer goods in the late nineteenth

century (Leary 1970:223; Mark 1985:70-74). Men could incorporate some


of this newly introduced form of trading activity with their traditional

dry season migrations for fishing or collecting palm wine (Thomas 1958-

1959:495-498). Further south, the Banyun maintained a better commercial

position relative to the Diola and Mandinka around Ziguinchor (Mark

1985:55). South of the Casamance River in Esulalu, warfare between the

Diola and Banyun, "the Koonjaen wars," continued until the early eigh-

teenth century (Baum 1986:101). A notable decline in commerce at

Ziguinchor slowed the establishment of Diola trade in forest products

and rice on the south bank until the French opened their first post in

the area, at Carabane in 1836 (Mark 1985:55).

The nineteenth century saw a general decline in trade as a result

of difficulties associated with the transition from the Atlantic slave

trade to economic colonialism, including a decline in the barter terms

of trade (Leary 1970:225; Hopkins 1973:135,142-155). As a result of the

decline in Portuguese commercial fortunes, the French were able to

pursue an aggressive and successful policy of expansion in the Lower

Casamance from 1800-1880. This expansion was marked by the establish-

ment of a trading post at Carabane Island in 1836, which was successful

at halting trade between the Portuguese and Diola and Banyun groups in

the area (Mark 1985:55-57). Sedhiou, further up river in Middle Casa-

mance, became an important comptoir (trading post) for the newly intro-

duced trade in groundnuts" by 1850 (Mark 1985:55-57; Baum 1986:203-

266). Groundnuts had only been introduced from Brazil during the 1840s

(Quinn 1972:9), but gained an immediate acceptance as a cash crop with

the Mandinka, who accounted for much of the trading activity at Sedhiou.

Interestingly, the increased demand for cash-cropping labor in the

"Known as peanuts in the U.S., this term is considered derogatory by
Anglophone West Africans. Therefore, I use groundnut here.


Middle Casamance spurred raiding for slaves in the Lower Casamance (Mark


During the 1870s through about 1900, the French as nominal

colonial authorities attempted to fund their local military operations"

through lower prices offered at their trading posts in Lower Casamance.

Diola traders north of the river simply responded by transporting their

goods to British posts in The Gambia (Mark 1985:65-66,94). This

experience led to an increasing emphasis by the French upon tax

collection as a means of supporting their colonial operations in the

area during the late nineteenth and early twentieth centuries (Mark

1985:93). Several means of establishing an institutionalized monetary

economy were available to the officials of Afrique Occidentale Frangaise

(A.O.F.). Corv6e (a form of taxation in kind through forced, unpaid

labor) and regular in-kind tax collections of rice proved the most

brutal and effective means of all (Geschiere 1985; Fall and Mbodj 1989).

In Lower Casamance between 1910 and 1916, while cash payments were being

required of individuals rather than the former village payments in kind,

the threat of military coercion had to be invoked directly against each

village in order to enforce its compliance with this new demand (Roche

1976:187,311). The expense of such an ad hoc enforcement of colonial

policy was too great to sustain for long.

In summary, indigenous states of the western Sudan were probably

trading with the Diola for salt, dried fish, and rice, as well as

raiding for domestic slave markets before the beginning of this third

period. These activities all continued throughout the period, but the

arrival of European merchants vastly increased the scale of the markets

"Where, for example, French troops were indirectly drawn into fighting
associated with the Marabout-Sonink6 wars and conflicts resulting from the
Islamic revolts led by the Fulani against Kaabu, as well as direct conflicts
associated with Diola "pacification" (Leary 1970:153-155; Roche 1976:91-


and qualitatively changed the nature of trade in many ways. Perhaps the

most important example of this is the slave trade, which became

incorporated into a plantation complex that spanned the Atlantic Ocean

(Curtin 1990). As a result of this enormous increase in the demand for

slaves, many African states established direct ties to European

merchants and began intensive slave raiding on a much larger scale.

Eventually the Diola were involved not only as victims, but as agents

and captors as well.

Clandestine slavery continued into the nineteenth century in the

Lower Casamance, well after the official abandonment and condemnation of

the trade. However, as the overseas demand dwindled and enforcement of

new anti-slavery laws became more effective, the relative profits

attainable in the legitimate trade increased. European merchants

reacted with a vigorous pursuit of the sources of beeswax, gum, rubber,

and other forest products, all of which the Diola would provide in

exchange for iron, guns, and cloth. A strong competition among buyers

of these natural products of the forest, combined with relatively

uncontrolled access across the nominal borders introduced by the

European states, worked in favor of Diola suppliers.

Unfortunately for many Diola, the French reaction was to invoke

military and police powers in order to force them to support the

imposition of a colonial state organization in Lower Casamance through

taxation, forced labor, and artificially low rice prices. At first,

threats of force were insufficiently certain to induce widespread

compliance. Actual military attacks were rare. The Diola also were

notoriously evasive, and tax collection from them was totally inadequate

(from the colonial point of view) until the 1920s, after the implementa-

tion of an integrated, systematic means of control was finally insti-


Twentieth Century Colonialism and the Independent State of Seneaal

The third period of Diola history, exemplifying quantitative in-

creases in European mercantile influence over the Lower Casamance, and

the fourth period, representing qualitative changes from that influence

to actual political control there by the French, are best divided about

1930. In my interpretation, the historical division between these two

periods can be established only after the A.O.F. administration

successfully implemented its Brunot plan in 1917 (see Roche 1976:339-

345). In fact, French control was not firmly in place until the late

1920s or early 1930s in many remote villages. The Lower Casamance

remains relatively isolated to this day (Linares 1992:211-212), and a

violent secessionist movement has caused serious trouble for the current

government in the 1990s (Cormier-Salem 1993; DaCosta 1993; Marut 1994).

The trend toward relative political control and the institution of a

cash market, however, began in earnest for much of the area about 1930.

Characterizing the sixty-five years from 1930 to present as a

single historical period may require some justification. However, it is

by far the shortest period in this schema. Because it is relatively

shorter than the other periods, one could reasonably expect greater

justification be offered to divide this relatively brief span of years

into shorter periods. Nevertheless, the convention for Africa has been

to consider the colonial and independence years as separate for most

purposes. Therefore it is appropriate that we briefly direct our

attention to this issue.

As is true for much of Africa, an important continuity exists for

Senegal from colonial times through the present. As with previous

periods, our characterization of the fourth period is directed at

identifying relatively consistent conditions. Among these are a

politically dominant, centralized state power, the use of its political

power to insulate strategic economic domains from market forces, a

growing civil service sector," and the general isolation and exclusion

of the Casamance region from the benefits of the political system.

Thus, the thirty years from the implementation of effective colonial

power in Casamance about 1930 until national independence in 1960, are

considered here together with the thirty-five years from independence

until the present.

The continuity of colonialism and independence throughout Africa

is not simply an academic issue. Of course it runs counter to

nationalist ideology and teaching by independent governments, as is the

case in Senegal. The Diola themselves express this continuity, however,

when they say "Inje bei Senegal"("I'm going to Senegal") instead of "I'm

going to Dakar."'1 They demonstrate a lack of identification with the

national culture too when they complain about the increasing use of the

Wolof language (and the power of Wolof traders) in Ziguinchor: "Igi on

est trop colonis6 par le Wolof" (Julliard 1991:48). The fact that a

secessionist movement exists in Casamance today, and that it is

perceived as a continuation of the anti-colonial struggle, lends further

credence to my categorization of colonialism and independence together

in this fourth period. Before elaborating the reasons for the

incomplete integration of the Lower Casamance region into the

independent state of Senegal, we first need to consider the history of

its incorporation into the colonial state.

"Rapid and sustained civil service sector growth has fostered increased
urbanization in Senegal from the time Dakar was the administrative
capitol of A.O.F., and is one result of a clientelistic governing
strategy (Diop 1981). Only recently has this growth been checked by
structural adjustment policies dictated by international donor agencies.

"Linares (1992:212) has published this example, but use of the phrase
is ubiquitous among the Diola.

By the time it initiated serious efforts to integrate the Lower

Casamance, the French colonial administration had been consolidating its

hold over the rest of Senegal for many years. During the 1800s, for

example, French administrators successfully established an economy based

upon the large-scale cash cropping of groundnuts, most famously in the

Sine-Saloum region, but elsewhere as well (see Klein 1968). By 1852,

for example, primarily Mandinka farmers in Middle Casamance were

producing one quarter of the national output of groundnuts (Roche

1976:87). This fact implies that an important economic change had

already occurred in a neighboring region by then. The Mandinka were

traders rather than farmers until the French essentially forced them to

accept the cash cropping of-groundnuts. They had been the Diola's

longstanding source for cattle, for which they exchanged their

indigenous varieties of rice."2

Since just after the turn of the century, the French colonial

authorities had used tax collection and artificially low prices at their

comptoirs in the Casamance as a means of supporting local military

operations. The Brunot plan of 1917 had as its goal the full

incorporation of Lower Casamance into the Colony of Senegal. At the

same time it would reduce the cost of establishing colonial authority.

From 1910 until 1916, administrators had to enforce their authority to

collect taxes there through annual military operations in each village

where they wanted to collect cash payments (Roche 1976:311). Such

direct and ad hoc coercion was too expensive to maintain; Brunot's plan

proposed to make it unnecessary. It was successful largely because it

2As noted earlier, some varieties of West African wet rice are probably
indigenous in the sense that they were not imported from Asia even prior
to European contact (see Dresch 1949; Porteres 1956, 1970; Johnny et al.

implemented an integrated strategy that incorporated many of the most

successful tactics already operating elsewhere in the colony.

The plan had as a primary objective pacification, or the

imposition of complete political-economic control throughout Casamance.

It would establish a cash economy in Lower Casamance and create a free

circulation of labor. Once individual men were wrested from their

traditional labor obligations, they could grow cash crops and use their

earnings to pay taxes. To meet these objectives, colonial authorities

employed the use of political power to prevent market forces from

operating freely in several economically strategic areas, notably the

cash cropping sector and in particular the groundnut market (see

Geschiere 1985). Until then, Diola traders often traded at British

posts in The Gambia (Mark 1985:65-66,94). Thus, borders had to be more

effectively controlled.

Other efforts to introduce a cash economy included the devaluation

of traditional exchange goods (primarily rice), the institution of

corv6e to initiate a free circulation of labor," and universal adult

male taxation. Among the tactics employed were an aggressive military

recruitment, the installation of non-Diola chiefs at the village

level,22 the suppression of some still-continuing inter-village slave

raids, and increased control over the power of Mandinka traders. All of

these together represented an effective, integrated effort to support

policies of total disarmament, universal tax collection, and broad price

controls over the sale of cash crops (see Roche 1976:339-345).

21Though its use was more limited in Senegal than in Guinea or Sudan,
corv6e labor was responsible for all road construction in Senegal up
until 1936. Defined as a demand on tax payers for a fixed number of
days labor in addition to taxes paid in cash, eight days were required
annually of adult men in 1926 (Fall and Mbodj 1989: 256-260).

'2Previously, there had been no political integration at the village
level (Linares 1992).

Colonial initiatives to establish control operated to undermine

traditional Diola political authority in several ways. First, rice was

imported from Indochina for exchange with groundnuts. The Mandinka were

more receptive to farming groundnuts than were the Diola, contributing

to the production in Casamance of one quarter the national output of

this crop in 1852 (Roche 1976:87). French imports were cheaper than

Diola rice, undermining the position of Diola seniors who had relied on

this trade with the Mandinkas as a primary means of controlling benefits

in their villages. By 1906, rice was the most valuable import to

Casamance, further eroding the Diola position in traditional exchange

relations (van der Klei 1986:85; Pelissier 1966:762; Roche 1976:317).

Second, male labor was consequently redirected away from its traditional

employment (especially during the dry season) in maintaining irrigation

dikes and in preparing the rice fields. Instead, an increasing number

of men engaged in commercial trade or produced cash crops, particularly

groundnuts (Linares 1981:568). By the 1920s, labor migration, a simple

way to earn cash wages in order to pay the newly instituted individual

cash taxes, had become a pervasive dry-season activity among the Diola

(Thomas 1958-1959; Mark 1985:49).

Diola senior men for centuries had relied upon Mandinka trade

networks to exchange locally-produced rice for cattle. These long-

established trade ties were critically weakened as the Mandinka began

favoring the purchase of cheaper French rice imports from Indochina with

the proceeds of their groundnut crop sales. Thus, the price of

indigenous Diola rice was undercut during the early part of this

century, its external market value essentially destroyed by subsidized

imports. The enforcement of cash tax collection (Roche 1976:341)

created political pressure for men to earn a regular cash income. This

began to force an acceptance of groundnut cultivation among most Diola

men during the mid-1930s. At the same time it encouraged many young men


to enter into more extensive dry season trade activities away from their

home villages." The exchange of rice had been the principal means by

which Diola seniors controlled access to the main prestige good,

cattle"4 (Pelissier 1966:760-762). Thus, the loss of the indigenous

rice market removed the basis for their legitimate authority and their

control over labor was rendered impotent (Pelissier 1966; Roche 1976;

van der Klei 1986).

Many Diola men gained their first experience farming groundnuts by

migrating to Mandinka farms as agricultural laborers (Thomas 1958-1959).

By the 1920s, Diola men were beginning to appreciate the benefits of the

new cash crop on their own fields. They were further encouraged to

adopt groundnuts after 1921, when the newly established Soci6tes de

Prevoyances (early marketing boards, later replaced by ONCAD) began to

provide seed on credit in the Casamance, to be reimbursed in kind upon

harvest (Robinson 1950; Mark 1985:105). This institutional

encouragement combined with the significantly lower labor requirements

of groundnuts relative to the arduous inputs necessary in the indigenous

rice farming system (Loquay 1981:98) did much to encourage the adoption

of groundnuts and, importantly, of the cash economy as a whole.

Again, these activities were outside of lineage elders' control.

Cash was earned individually through trade or wage labor, unlike

traditional economic activity, which was developed and maintained

communally under senior male control. Wealth was becoming,

increasingly, an individual characteristic. In a sense, the control of

23This economic conversion was, interestingly, synchronous with a widespread
religious conversion from the Diola indigenous religion, known as kawasen, to
Islam, primarily north of the Casamance River (see Leary 1970; Mark 1985;
Linares 1986).

24The religious importance of cattle beyond simple luxury should not be
overlooked. "A man without cattle is not just poor; he is without the
ability to protect himself spiritually against calamities and sudden
twists of fate" (Baum 1986:365; see also Mark 1988).

prestige itself was being wrested from its traditional source. The

influence of the French colonial state upon Diola political society was

perhaps unintentional, but nevertheless direct and pervasive. From the

point of view of young Diola men, this was a liberating experience. As

a result of this weakening in traditional authority, young men became

free to pursue economic activities on their own. These changes, of

course, had a powerful effect on Diola women as well.

As young men continued to migrate out of the village during the

dry season, either to trade or earn wages, and as those who remained in

the village put more effort into cultivating groundnuts, staple rice

agriculture was relatively neglected. Traditional late dry season

activities for men included the preparation, maintenance, and expansion

of irrigation dikes, and the preparation of nursery beds for rice

seedlings (Pelissier 1966; Linares 1970, 1981; Loquay 1981). A general

movement away from these activities both slowed the expansion of the

most productive form of staple rice agriculture, and over the long run

probably has reduced the productivity of those paddy lands that were

already actively in production. The Diola continue to cultivate rice,

but because men in particular pursue cash-earning opportunities during

the dry season, the former long-standing expansion and intensification

of rice paddies has been reversed.

These changes in men's labor practices have had an important

impact on the traditional division of household responsibilities for

Diola parents (Hamer 1983:75-78). Fathers were expected by tradition to

provide rice for their children during the wet season, while mothers did

so from their granaries during the dry season. As staple rice

production has fallen relative to population, women are often unable to

grow enough rice for their children's needs, requiring their own cash to

buy imported rice when their granary supplies are low. The most

important result of this introduced difference is that women began

migrating to the urban areas to find work for wages. In contrast, men

can still earn a cash income growing groundnuts or perhaps vegetables in

the transformed rural economy.

Groundnut cultivation not only excludes women from the production
process, it also alienates men from rice production; its influence
goes even further than the sexual division of labor. By eroding
rice production, growing groundnuts undermines the very rituals
that insure overproduction, reciprocity, and redistribution of
paddy at the village level. It encourages the abandonment of
paddy fields--especially the deep fields in the mangrove swamps
that required a great deal of work even before the 1970s drought--
and makes it increasingly more difficult for the Jola to shift
resources between the subsistence and money sectors of the economy
when conditions demand it. (Linares 1985:92)

While both men and women continue to work the rice fields to the

present day, productivity cannot be maintained at traditional levels

without extensive labor inputs in soil preparation and dike repairs.

Since the 1920s and 1930s, women's labor has been unavailable at suffi-

cient levels to replace the efforts formerly contributed by men. With

this trend toward lower yields due to insufficient labor availability,

women have increasingly left the rural areas of Lower Casamance for wage

labor during the dry season (Hamer 1981). As has been previously noted,

the precedent for female wage labor is old in Lower Casamance: women

were noted as exclusively comprising the labor force on the docks at

Ziguinchor in 1910 (Roche 1976:316; Journet 1976:197).25

These economic trends, the need for cash incomes among both men

and women, and the resulting popularity of wage labor migration, have

been sustained and reinforced consistently in the Lower Casamance since

the 1930s. Political circumstances have also remained relatively

consistent since the incorporation of Casamance into the colonial state

of Senegal. In order to strengthen my assertion that a political

2"Linares (1992:79) attributes their association with this work to the
heavy lifting activities Diola women perform in agriculture.

continuity exists from about 1930 through the present, I will now turn

my attention to issues of governance.

A loose style of control was typical of African colonial

governments, which often were required to rule under difficult

circumstances and with limited administrative budgets.26 In the absence

of broad political legitimacy, a patrimonial strategy of governing often

was pursued (see Foltz 1969; Lemarchand 1972; Flynn 1974; Eisenstadt and

Lemarchand 1981; Colvin 1986; Fatton 1987; Young and Kant6 1990).

Benefits were distributed to those few subjects who could demonstrate

relatively strong influence over civil groups. Often the only benefits

available for distribution by administrators were prebendal offices.

Thus, a tax collector or village chief might not be so closely

supervised that corruption and graft for his personal benefit would be

prevented. A chef de canton judging a dispute might rule in favor of

businesses that could reward him directly for his trouble. In this

manner, as they grew in power, local fiefdoms could develop into

regional political forces that could exert certain influences over the

central government. Civil service jobs might be exchanged for political

support, for example. Senegal's civil service, seated in the capital of

Dakar, was swollen under colonialism to administer the entire A.O.F.

territory until independence in 1960. After five years of independence,

the civil service budget amounted to a staggering 47.2 percent of total

government expenditures. It continued to grow into the 1980s, and

continues to be an especially burdensome and sensitive issue today (see

Cruise O'Brien 1971:271-272; Zecchini 1984).

The key to understanding the incomplete integration of the

Casamance region into the state of Senegal lies not-oly in its

2"See Migdal (1988) and Rothchild and Chazan (1988) for the development
of the concept of the weak state.

geography, but also in the nature of patrimonial politics. Under

patrimonialism, a delicate balance must be maintained between support

for regional patrons who can deliver votes to the central government's

leadership and suppression of broader regional movement that threaten

the government's control. Regional political movements and their

centrifugal potentials in particular are feared by the central

authorities who govern weak states; often, therefore, radical strategies

are employed against such unities. As independent Senegal's first

president from 1960-1971, Leopold Senghor was widely noted to be a

master of this style of rule.27 He was able to play regional and other

political coalitions against one another in a way that maintained just

enough instability in his opposition to maintain his own relative

political strength.

The Mourides and other locally-dominant Islamic sects have for

many years garnered the majority of political support throughout rural

Senegal north of the Casamance River. Through their religious

institutions, and through their control of many important economic

firms, they have proven themselves the most influential leaders in

Senegalese civil society. They have been the primary power to be

reckoned with for every government from colonial times to the present

regime of Abdou Diouf, a follower of Mouridism himself (see Cruise

O'Brien 1971; Diop 1981).

To a large extent, this is the source of contemporary conflicts

between the Lower Casamance and the Independent state of Senegal:

regional politics there, because they threaten the integrity of the

27The following studies draw a "remarkably consistent portrait" of post-
Independence politics in Senegal (Boone 1990:346): Behrman (1970);
Zuccarelli (1970); Adamolekun (1971); Barker (1973); Cruise O'.Brien
(1975); Schumacher (1975); Coulon (1981); and Jackson and Rosberg


state, are suppressed through policies that disfavor it economically as

a region:

Casamance particularism is explained by the geographic isolation
of the region, the poor quality of its infrastructure, and more
generally by the neglect of the region from Dakar. It also stems
from the growing presence of merchants and bureaucrats from the
'North' who tend to impose their language and their religion. .
as the only legitimate ones. Faced with this 'internal
colonialism', a strong sentiment of frustration produces the
search for a distinctive identity (Coulon and Cruise O'Brien
1989:159; see also Darbon 1984, 1985; Benoist 1991). (Linares

Characteristic Patterns of Diola Migration for Each Period

Migration is in constant flux because it changes dynamically with

change in the environment, although the relationship in not direct. It

is a cultural means by which people adapt to change in the environment.

Because migration is used by people, who perceive and react to different

aspects of their environment at different times, what the environment is

(in terms of any model of migration) must be flexible enough to reflect

such changes. Each particular characterization of a period here defines

what is meant by or interpreted as environment for that particular time.

Linares (1992) focuses on the physical environment for her model of

Diola migration. I find her model convincing, but she tends to exclude

the social environment and contemporary forms of rural-urban migration

among the Diola. Migration is a long-established aspect of Diola

culture, rather than a modern introduction. It cannot correctly be

labeled as a simple indicator of social disintegration in the face of

colonialism and the encroachment of Western culture (see Thomas 1960).

Certainly, however, the forms migration has taken have changed over

time. For example, under colonialism Diola men were subject to con-

scription into foreign military service. These new experiences of

forced migration (and, at times, the avoidance of them) led some to

undertake voluntary migration, at times to urban areas, in search of

wage labor. Women soon followed the village men to Dakar, eventually

overtaking them in terms of their predominance in the migration stream.

By 1961 there were 100 Diola women for every 60 men in Dakar.2-

Certain cultural institutions pre-existed this new form of migra-

tion, providing the basic structures that contemporary urban migrants

have manipulated to suit their needs in this relatively new setting.

Such institutions, in this case voluntary associations, may be consid-

ered as serendipitous "pre-adaptations" to cultural ecological changes

(see Cloak 1986). The pre-existence of these associative institutions

has allowed Diola culture to adjust more rapidly to a situation that,

while it has had important negative effects, has also provided villagers

with opportunities that might otherwise have remained unfulfilled.

Because migration has had a long history of full integration into

Diola culture, it exists as one aspect in a nexus of cultural adapta-

tions, an "adaptive tradition" common to many Diola groups. Some other

aspects of this tradition include: a diverse set of associations,

social institutions that mediate relations between various cultural

categories; an established means of incorporating foreigners or

strangers more or less completely into the life and functioning of the

village; an ethic of treating strangers as guests;2' and finally, a

flexible set of kinship and land tenure rules and social regulation

thereof that can accommodate and adapt to large-scale movements of indi-

viduals over time.30

28Hamer (1983:250) citing this figure, refers to Martin (1968:368).

2"See Baum (1986) on how this was broken down during the slave trade,
but also how shrines were established to protect people from the threat
of this tradition's dissolution.

"See Linares (1983); see also Girard (1969);.Snyder (1977, 1981); and
Hamer (1983).

Linares (1992) provides a model of the historical changes and geo-

graphical movements of Diola groups that illustrates the cultural

adaptations these groups have made to accommodate the conditions

challenging them in the historical past. Her model adopts three

villages in three different parts of Lower Casamance, illustrating the

cultural differences in each setting. The cultural differences are, she

argues convincingly, the result of each group's adaptation to the

different cultural ecology in each of these three areas.

As Diola groups migrated north across the Casamance River from

about the sixteenth century, they encountered a physical environment

that was quite different from the one that they had left further south

(e.g. savannah rather than forest, lower average annual rainfall, and

much less land on which rice could be cultivated). They also found

themselves in a new cultural environment. They were faced with a

majority of neighbors, ethnic Mandings, who followed Islam, a very

different religious tradition from their local kawasen religion.

Linares has chosen one village to study in each of the three zones to

represent "pre-change," "transitional," and "post-change" periods of

time. She presents a descriptive analysis of land tenure, kinship, and

labor practices, among other things, in each of the three villages.

These differences are the result, she asserts, of the specific cultural

ecological conditions present in each of the three settings. She makes

it clear, however, that all of these groups continue to change in

observable ways, even during the relatively short period of her study.

She also argues the importance that ideological changes have

played in mediating material aspects of the transitions. Traditional

associations functioned, among other roles, to mediate conflicts among

genders, generations, residential wards or quarters, families, and

other potentially fissive categories and groups via the kawasen spirit

shrines. These associations served a cohesive role, socially cementing

together people who might otherwise have a tendency to break away from

the group. Associations achieved this cohesive function by establishing

patterns of labor sharing, for example, among all women who have married

into the village (see Linares 1988, 1992:50).31

That these sorts of associations existed as traditional institu-

tions was a fortuitous circumstance for urban Diola migrants, who were

able to adapt traditional forms of associations quite rapidly into

institutions that could help to serve important new functions in this

new environment. Associations rapidly evolved into what became perhaps

the most important means of cementing group relations in the city.

These associations now act to assist new urban migrants in fulfilling

their needs, but eventually are often also successful in re-directing

their attention back to the village, at least during important periods

of the agricultural cycle.

In order to understand contemporary Diola migration, it is

important to consider how and why migration patterns have changed over

time. Linares is convincing in her model of the changes necessitated

with the move north across the Casamance River. My goal here, though,

is to take a broader view of history, presenting changes in migration

that have taken place over a longer span of time. Having elaborated

four periods of Diola history in the previous section, I will present

the forms of migration associated with each period. Then, in Chapter 3,

I will present the data I collected on contemporary Diola migrants to


31Hamer (1983) and Reveyrand (1986-87) elaborate some other functions of
rural women's groups in particular.

Period One: Early Sedentism

In cultural evolutionary terms the most important change in

migration accompanies the transformation from reliance on a nomadic or

hunting and gathering subsistence strategy to dependence on sedentary

agriculture. I call the pre-sedentary form "carrying capacity"

migration, because groups move, generally, toward higher concentrations

of natural resources upon which the group relies most for its

subsistence. The primary factor determining migration in this setting

is the relationship of population to subsistence resource availability.

Period one marks this division for the Diola. At the beginning of

this period the Diola began to rely less directly and exclusively on

foraging for marine and forest resources. Instead, they began to rely

more on rice production. As populations grew, the Diola developed a

mixed agriculture economy, relying less on gathered shellfish, for

example, and more upon cattle trade and husbandry as time progressed.

Throughout the period one would expect an increasing trend towards

agricultural intensification, although expansion into new areas

continues into the present (Linares 1992).

As the Diola began to settle permanently and as they relied more

substantially upon agriculture, their movements were increasingly

determined by the availability of cultivable paddy land. Migration was

determined by more complex circumstances through period one, including

the relationship of village populations to cultivable land, perhaps

political rifts within villages causing fissioning, the viability of

current technologies to intensify production on available paddy land, or

even the ability of labor managers to induce others to work harder. As

the amount of land suitable for rice cultivation became more scarce, an

increasing reliance had to have been placed on the intensification

option. Linares presents archaeological evidence that these processes

were in fact occurring in Casamance from about the second century A.D.

(Linares 1971).

Period Two: Early States

The large-scale population movement of Mande peoples from the

interior westward provided new trade opportunities for the Diola. As I

have discussed in previous sections, they eventually developed an

economic system that relied to an important extent on the exchange of

locally-grown rice for cattle. The presence of small states in the

region from the thirteenth century also had other consequences for the

Diola. While they provided opportunities for trade, these states also

circumscribed the Diola, limiting their ability to continue historical

patterns of territorial expansion. As populations grew in Diola

villages, agricultural intensification was much more a necessity than an

option during this period and into the next.

It is during the second period that early patterns of seasonal

trading were probably established. These fit well into older patterns

of dry-season migration directed toward the collection of forest

products and fishing for consumption. These were forms of migration and

not simply another off-farm economic activity, because they often

involved periods of several months away from the village. However, they

were migration and not simply a continuation of a hunting and gathering

subsistence strategy because individuals lived in sedentary villages,

cultivating crops for much of the year.

Period Three: Early European Trade

The third period, spanning the fifteenth through nineteenth

centuries, is the least stable of this schema. Nearby Mande states vied

during this time for control over lucrative trade routes, expanding

warfare and slave raiding activities. North of the Casamance River, the

Diola and Banyun were often at odds over the same territory and

consequently in a constant state of war. The Diola eventually gained an

advantage over the Banyun in the early seventeenth century, partly due

to the way they employed the iron received in trade with Europeans.

During the second half of the eighteenth century the slave trade gained

in importance, further extending this long period of conflict and

uncertainty. Many Diola responded by withdrawing from contact with

outsiders, defending their villages but retreating from trade

activities. During periods of conflict, capture into slavery (a form of

forced migration) was a constant threat to those who ventured away from

the confines of the village, so voluntary migrations were severely


While the Diola were generally noted by Europeans to be

uninterested in trade, they had for centuries exchanged a number of

forest and coastal products with other Africans. After the abolition of

slavery, the European legitimate trade emphasized exchanges for wild

rubber and palm kernels. In Casamance, this trade was greatly expanded

between the French and Diola after the establishment of the trading post

at Carabane in 1836. Earlier, some Diola traders had crossed into The

Gambia to get the better prices offered by the British posts there.

Despite a general decline in European (particularly Portuguese)

trade during nineteenth century, the French maintained an aggressive

trade expansion in the Casamance from 1800-1880. Groundnuts were

introduced during the 1840s. The post at Sedhiou in Middle Casamance

became important as a trading center by 1850, due to the extensive

adoption of this cash crop by the Mandinka there. This date therefore

marks the earliest possible beginnings of Diola male wage labor

migration, widely initiated to harvest groundnuts on Mandinka farms in

order to pay the cash taxes imposed by the French as early as 1910.

Period Four: Twentieth Century Colonialism and Independent Seneaal

Each successive period of this schema illustrates the addition of

one or more forms of migration overlying the continuing patterns that

existed in previous periods. Subsistence foraging, the basis of the

economy prior to period one, was incorporated during the first period as

an additional activity pursued along with sedentary farming. These

forest and marine resource collecting activities were very similar to,

and most likely continued alongside migration undertaken for trade in

the second period. During the third period, with the beginnings of the

slave trade in the Casamance, forced migrations into slavery were an

additional (albeit undesirable) possibility, although other forms of

trade continued to exist together with new forms introduced by


During the late nineteenth and early twentieth centuries, but most

commonly during period four, military conscription and corv6e labor were

imposed by the French colonial government on Diola men. A less clearly

forced form of migration, though still indirectly imposed by the

colonial administration, was male wage labor migration to Mandinka

groundnut farms. Further removed from forced migrations, but still

indirectly caused by colonial impositions was the female wage labor

migration noted at the docks at Ziguinchor early this century. These

forms all preceded the more contemporary form of rural urban migration

that has become increasingly popular since the 1950s.

Wage labor migration began among the Diola well into the present

century. It grew in popularity as a means of earning cash, which was

needed primarily to pay newly-imposed taxes. Migration to Dakar in

particular began with the first military conscriptions, but expanded

with the growth of the civil service sector there. Temporary dry-season

wage labor was available for men in Ziguinchor, too: examples of men


working as carpenters and masons were documented during the ethnographic

interviews I conducted in 1990. The growing A.O.F. bureaucracy in Dakar

also created a demand for technicians, soldiers, police, and other

salaried positions.

A large, seasonal "rural exodus" of young Diola to Dakar dates

from about the 1950s. From soon after the earliest migrations from

rural Casamance to urban Dakar, women have represented a high proportion

of these movers (see Martin 1968:368). The reasons for this unusual

situation are complex, but can be explained within the context I have

provided in the description of the fourth period. To summarize my

discussion of twentieth century events, French colonial policies

diminished the value of Diola rice with Indochinese imports. This was

part of a deliberate systematic effort to replace the indigenous economy

with one based on cash, which would allow taxation. Men were targeted

for taxation and cash earning opportunities in the rural areas,

particularly the promotion of groundnuts as a cash crop, were

effectively provided only to men.

These colonial policies were an important cause initiating Diola

women's migration. However, they would not have necessarily had this

particular effect if Diola social organization, specifically the social

division of labor, were different (see Hamer 1983:75-78). Women's roles

in rice agriculture are essential and preeminent but, as rice was

devalued, men diverted their efforts to the favored crop, groundnuts.

Women did not have the same opportunities to grow this cash crop as men.

This was due in part to colonial efforts to target men as workers and

taxpayers, but also to traditional Diola practices. Women own and work

lower fields suited to rice, but not rainfed fields in which groundnuts

are cultivated (Pelissier 1966:687). Without the former level of male

labor inputs into the intensive cultivation of rice, productivity in

this traditional staple crop dropped and could not be maintained by

women alone.

At the same time, women in the modern economy are expected to

provide for themselves and their families, and their monetary

obligations have expanded over time (Hamer 1983:76). Without access to

cash earning opportunities in the rural setting, women tended to migrate

to urban areas to find.wage employment. I noted the case of Diola women

working the docks at Ziguinchor early in the century as the earliest

example of rural-urban migration. This job in particular fits with

traditional women's agricultural work, which includes the transportation

of water and wood to the home, and cow manure to the rice nurseries.

These tasks all involve heavy lifting. Most importantly, however, this

case demonstrates that women were seeking cash earning opportunities

quite early in this century. As opportunities developed in other areas,

Diola women were willing to travel, and in fact since the nineteen-

fifties they have been in particular demand as maids in Dakar.

There are a number of reasons why Diola women in particular found

it to be relatively easy to find employment in Dakar. There was a

growing demand in the market of the colonial capital for maids, in part

due to the number of expatriate men (with or without their families)

working in the colonial government there. More local Wolof families

were also earning cash wages at the time, and they too were interested

in having domestic workers to cook, clean, and care for their children.

At the same time, the French, most of whom were Catholics, preferred

hiring Catholic maids. This is perhaps simply a matter of prejudice,

but is probably also attributable to a sense among expatriates of

alienation from the majority Muslim community of Africans in Dakar.

Furthermore, the Mandinka and Wolof societies were based upon a caste

system that discouraged women in these groups from seeking work in

domestic service. The majority of Diola from south of the Casamance


River are Catholic, and being from a relatively egalitarian society, see

no stigma attached to domestic labor. On the contrary, they view such

work as quite honorable.

In the following chapter I will present the data I collected in

thirty ethnographic interviews, conducted in Dakar in 1990. In

particular, Chapter 3 is focused on the history of the women's

association. However, as a whole the chapter provides a sense of what

migration from Casamance to Dakar entails for the women who undertake


Table 1: Historical periods and associated characteristic forms of

Data source


Inferred changes in
migration patterns


1. archaeological early sedentism population expansion
200-1100 evidence on and mixed through ecological zone,
subsistence agriculture then intensification of

2. regional encroachment, trade, dry season
1100- history of circumscription migration patterns
1400 Western Sudan by states of established
Mali and Kasa

3. history of early European expansion of trade
1400- related trade, wars and migration, but
1800 groups, some slavery, then withdrawal in times of
local history establishment of war; evasion and "exit"
1800- (least stable legitimate trade from control
1930 period)

4. direct cash markets wage labor migration
1930- historical firmly patterns established
present evidence developed,
colonialism and



In this chapter I begin to present the data I gathered during

field research conducted in Senegal, both in Dakar and the Casamance

region. I use a case study approach to illustrate this example of

twentieth century West African rural-urban migration, focusing on the

recent history of migration from the village of Affiniam-Boutem, known

simply as Boutem. While the historical importance of slavery and

warfare should not be disregarded as limiting factors, urban migration

represents the most important change in the pattern of migration among

the Diola since they first began to rely on agriculture for subsistence

(see Table 1). Because of the gender division of agricultural labor in

Diola society and because of the way that the Diola and the Casamance

region have been incorporated into the economy of Senegal, Diola women

migrate from the Casamance in particularly large numbers (see Hamer

1983:74-78). The case study approach of this chapter provides some

insight into the nature of this migration for the residents of and

emigrants from Boutem.

I first report briefly how I conducted various aspects of the

research, including interviews, a census, and an analysis of the

membership of the women's association. Then I describe the recent

history of migration from the village of Boutem, as it was told to me by

residents and emigrants. The final section describes Diola voluntary

associations more generally, and provides a short history of several

Diola associations in the urban setting of Dakar.


The data presented in this chapter and in Chapter 4 were collected

during field research in Senegal, conducted during the nine months from

December 1989 through August 1990. The core of the data collected

consists of two parts. First, thirty directed, open-ended interviews

with emigrant women from the village of Boutem were conducted at their

residences in Dakar. Second, I completed a census, including the

migration histories of all individuals in each household of four

quarters (wards) of Boutem.' Many of the data presented here were

gathered during these two activities.

I also gleaned much additional information through daily

conversations with my principal informant and research assistant,

Antoine Badji, himself an emigrant from Boutem. Together we attended

general meetings of the village and women's associations, steering

committee meetings with the officers of these associations, and met with

individuals he considered to be particularly knowledgeable about

specific aspects of the research. Several conversations with Emile

Djiba, the former president of Boutem's youth association in Dakar, were

particularly fruitful, and he was kind enough to lend me documents

relating to the historical boundaries of the village. The officers of

the women's association also eventually allowed me to copy information

from their official record book, and I was able to interview the

president. I am especially grateful for their trust in lending me their

'No census data were collected from either the Bougafou or Boutoupa
quarters. For the purposes of this aspect of the research, these
quarters are not considered as part of the village.


records, because this book included three years of dues payments records

and was closely guarded. It was a unique and invaluable source of

documentation for my work.

Later, during about ten weeks spent in Casamance, we met with the

village women's association, separately with its officers, and with the

members of its maternity clinic committee. We also gathered a focus

group of emigrants resident in Ziguinchor for a discussion of the costs

and benefits of migration. Regular visits with many members of

Antoines' family, friends, and relatives in the village, Dakar,

Ziguinchor, and elsewhere, as well as attendance at his family

association in the Ouakam neighborhood of Dakar, rounded out the diverse

set of information sources I was able to draw upon.


Interviews with emigrant women were arranged through the village

women's association in Dakar. I first arranged to meet with the Boutem

women's association soon after my arrival in Dakar in December 1989,

when I attended a village association meeting at a neighborhood Catholic

church, the Martyrs de Louanda. At the first meeting I attended, on

January twenty-first, after members completed old business, I was

invited to present my project to the group. I explained my research

purpose, goals, and desire to work with them. Their response was

favorable, and I was asked to attend their next meeting to present my

work in further detail to the rest of their membership. The following

week I met with the group's officers. Together we agreed that, in

exchange for members' cooperation and help, I would help the association

to pursue funding for their current project, the construction and

equipping of a maternity clinic in Boutem.

After these meetings with the membership of the group, and after

some disagreement over the nature of the exchange among the members,

this arrangement was finally approved. Members agreed to meet with me

and answer my questions, while I would help the group to plan and fund

the building of a maternity clinic in the village. I would donate what

money I could afford to the project upon completion of my research. I

insisted the amount of my donation would be quite limited, less than

fifty U.S. dollars. However, I promised to look for and indeed was able

to locate a funding source, a small-projects development fund at the

U.S. embassy. The application I helped Antoine to complete was

eventually approved after my return home, and I have since received news

that the clinic itself has been completed and inaugurated. Meanwhile,

no individual compensation for interviewees was requested or offered.

I met with the entire women's association three more times,

attending their monthly meetings in the Benn Tali neighborhood in March,

April, and May. These were held outside the small concession or group

of homes inhabited by two interviewees and their families. I missed two

of the monthly meetings scheduled during the time I was in Dakar. I

tried to find the meeting place alone for the second meeting and was

unable to locate it, and I was sick the day of one other meeting.

However, I considered meeting with the women's association important,

and did my best to attend each monthly meeting. I also met with the

officers at their monthly executive meetings twice, in January and

March, when I was invited to discuss specific issues with them.

I attended all of these meetings with my principal informant and

research assistant, Antoine Badji, who translated my presentations,

which I made in French, into Diola. He also would translate questions

and answers, at times with added input from younger women who were more

fluent in French. The treasurer of the association was particularly

helpful in this regard. Antoine's credibility with the group was an

essential part of the success of our arrangement to work with the women.

While most members understood French, often using it at work, they were

not generally comfortable with a presentation conducted in French.

Antoine was also able to clarify issues that came up, since he was

familiar with the research goals and procedures of the research.

Initially I believed we could interview every member of the

women's association. I was told that we could acquire a list of the

entire membership early on, but this took longer than expected. I also

thought that we could conduct interviews more quickly than we, in fact,

could. It often took several days to successfully meet with potential

interviewees, and we often had to make several attempts to meet with a

woman at her home, either after work or during her day off. Once it

became clear that we would have to limit ourselves to interviewing about

thirty of the 100 or so members, interviewees were selected for

questioning based largely on convenience. However, we planned

interviewee inclusion to provide as wide a range of representation from

as many families, ages, neighborhoods of Dakar, and quarters of the

village as was possible in the time we had. I targeted employed women

for inclusion, in part because several studies of migrant women have

already focused on the youngest and most vulnerable populations (e.g.,

see CNFNA 1983; Philpott 1986). We avoided interviews with more than a

few of Antoine's relatives, friends, or close neighbors from the

village. I often had other contacts with them, and used informal

conversations in family settings to ask questions of them as they

occurred to me.

Interviews themselves were conducted in Diola. Although most

interviewees used French to communicate at work, they generally were

uncomfortable with using it outside of that context. Because Diola is

not taught in the U.S., before undertaking this research I studied Wolof

in an intensive study program for eight weeks at the University of

Illinois at Urbana-Champaign. Wolof is Senegal's lingua franca, and I

developed some conversational ease with it in the field. However, the

Diola do not speak Wolof natively. I therefore studied Diola in

Senegal, but lost my tutor after only a few weeks of formal lessons. I

progressed enough on an informal basis to exchange greetings and make

simple statements about, for example, such common topics as eating, who

I was, where I was from, and what I was doing.

In order to conduct the interviews, I developed a schedule of

questions after much discussion with Antoine. I then translated it into

French and analyzed it with him point by point. We discussed the goals

of the work together, and modified the presentation somewhat before he

translated it into his native Diola. We worked on back-translating it

several times before beginning our first interview. Finally, we made a

few changes after certain questions required explanation to several of

the early interviewees. I kept longhand notes of each interview,

recording it on audio tape as Antoine posed questions in Diola according

to our schedule. Many interviews included some responses made directly

to me in French, often when a particular point interested the

interviewee, or during more informal conversation as we closed the


After each interview, Antoine would translate from the tape,

orally and (after the first five or six sessions, which were translated

more loosely in the third person) verbatim into French. I would

transcribe his translation in longhand, writing in English. This was a

cumbersome process: after pouring over my own translations from his

French, I would read them back to him in French to confirm or correct my

interpretations. We generally had to listen to audio taped interviews a

second time. It was often a very difficult and frustrating task for the

two of us to fill-in for my benefit much information that was implicit

in an exchange between two Diola speakers. I did, however, learn much

about village life in Boutem from these intense exchanges. Finally, I

would write out a corrected translation. This entire process often


required three or more hours of work around the kitchen table for every

recorded interview of perhaps forty-five minutes. In the end it took

seventeen weeks, averaging about one interview every four days from

February through May, to complete thirty interviews.


The village census was organized in conjunction with the officers

of the village youth association in Boutem. Forming six teams, we

contacted one member at each residence in four quarters in the village

during the first week or so in July. Each team filled out a simple one

page questionnaire for each household.2 It requested information on the

name, gender, age, relationship to the household head, and the migration

history of each household member. I recorded each response in my notes,

and then questioned team members if I found inconsistencies or any

missing information. After my return to the U.S., the data were entered

into a computer file for analysis.

The village-based phase of the research design was intended in

part to confirm interviewees' responses regarding plans to return to

Boutem during the current rainy season. I asked each respondent if she

planned to return for the 1990 season. Almost every interviewee said

she did. However, most often this response was qualified, with "God

willing," or "If I can possibly do so." Therefore, I decided that as

well as censusing the entire village, I would list all of the

interviewees who actually returned before the beginning of the

agricultural season. This seemed to be a particularly simple and

convenient procedure, because the village association sets a date after

2The census and interview instruments are translated and reproduced in
Appendix A, while the codebooks for each are included as Appendix B.

which, if a member has not returned, he or she is fined.3 I would

simply count returnees on that date, at the village association meeting

itself. I had been told that everyone attends, so we could set up a

table in the meeting hall and record returnees, perhaps tracking down

the few who remained home that day for one reason or another.

Unfortunately, the August fifteenth meeting of the village

association was a complete disaster. The officers were drunk well

before noon (the president literally fell off of his stool), attendance

was low, and one young woman, whom I had interviewed and visited with

socially several times in Dakar, was informed during the meeting that

her son, about eight years old, had suddenly died. The fact that

attendance was so low despite reports that everyone would be there was

the first disturbing event that day. It prevented me from completing an

important part of my work. Antoine was visibly upset at the poor

attendance and behavior at the meeting. This annual meeting was not

being taken as seriously now as it had been just five years ago, when he

last attended. In a relatively short time, the village young people

apparently had lost interest and involvement in their local government.

This change is another indication that their attention is increasingly

focused away from the home village, toward the cities and migration.

I never completed the list of returnees, but did witness a

dramatic set of events that day, including the stricken child's wake,

funeral and burial. He reportedly had not been ill before the meeting.

These events were punctuated with loud disagreements and witchcraft

accusations. One of the first storms of the season added to the human

drama. A thunderstorm produced threatening clouds during the wake,

poured cold rain on the procession to the village church, and pounded

3See Snyder (1978) for a description of what he calls a village police.
I also recorded a set of Boutem's village association laws.

its corrugated roof so loudly that the short funeral service was nearly

inaudible. The rain subsided for the burial itself. The mother of the

dead child left the grave side wailing, in tears, and accompanied by the

women who had attended with her the wake, funeral, and burial. After

they had left the small cemetery, a clearing in a small but dense stand

of forest near the church, several men engaged in a loud argument. As

they lowered the shrouded body down into the muddied earth, one man was

shouting and standing in the grave itself. I later learned that they

were arguing, among other things, over who had the right (normally

reserved for a close relative) to take the cloth used as a shroud.

The next day, as I was leaving Boutem for the last time, I passed

by a home behind which a teen-aged girl was screaming. A daughter of

the attendant of a spirit shrine4 at which we had several times paid our

respects, was fully entranced and writhing on the ground, in the midst

of voicing a witchcraft accusation regarding the child's sudden death

the previous day. This was an aspect of village life I had not sought

out, but events associated with witchcraft and its suppression

confronted me on several occasions. Witchcraft, its prevention and

related intrigues are important aspects of village life. I was shaken

up by the experience, but also felt gratified to have witnessed these

dramatic events. However, my list confirming those emigrants who

actually returned never was completed.

Women's Association and Analysis of Dues Paying Records

Construction of this data set was begun from a list of all women's

association members compiled from the association's record book, which

was lent to me to copy with the explicit approval of the association

4The particular shrine (chin) is devoted to kajumo, "the renowned." The
general term for such spirits is bokin (pl. inaati) (see Mark 1985:32-

officers. Sixty-nine members who could be identified in the village

census were included in the first phase of constructing the data set,

leaving sixty-four individuals for whom dues paying data was collected

but who could not be linked to census information. While these members

were included in an earlier, preliminary analysis of dues paying

behavior, they were dropped from the present analysis because their dues

payments could not be associated with any other characteristics. The

association of dues information with data collected in the village

census was of critical importance, because it enabled the comparison of

dues paying behavior with the member's age and, for example, her

migration history information collected in the village census.

To complete the construction of the data set, nine additional

cases were added from information gathered in my interviews with indi-

vidual members of the association, some of whom no longer had relatives

in the village to report on them for the census. While complete census

data for these individuals were not available, in all but a few cases

they did provide me with complete information on their income and

duration of residence in Dakar, to which I could add their dues paying

records from the association book. The other twenty-one interviewees

had already been identified in the census, and were included among the

original sixty-nine cases. About half of these women self-reported a

different number of years residence in Dakar than had been recorded for

them in the census by family members resident in the village. In these

twelve cases the length of residence in Dakar was corrected to match the

self-reported figures.

Seventy-eight cases were therefore made available for analysis in

this data set, although every case did not include all of the types of

information gathered. The representativeness of this sample can be best

assessed by comparing a description of its characteristics to that of

the other sub-groups and categories of the relatively more complete

population as censused in the village. Such a description is provided

in the discussion of the association members, based primarily on the

sixty-nine women for whom I have census data available.

Oral History of Migration from Boutem

It is difficult to gather oral histories among the Diola that

cover the time before contemporary adults' personal memories begin.

Unlike many Senegalese and other West African (especially Muslim)

groups, the Diola do not recite geneologies or exhibit much interest in

their family or cultural past. Elders generally do not discuss their

own lives with their children either, as I discovered during many of my

interviews with migants to Dakar: "My parents never spoke about their

life, so I don't know if they ever migrated" (Interview 12). Many other

respondents also told me they had no idea whether their parents had ever

migrated or not: "I wonder if [my mother] even knows her way around

Ziguinchor!" (Interview 9). "They never told me about that, and I was

never curious enough to ask them" (Interview 30).

When one elicits them successfully, however, local oral accounts

generally agree with the migration history of the area as presented in

Chapter 2. Everyone I spoke with on the subject agreed that Diola

migrants into Boulouf had originated south of the river. It is likely,

given this and other evidence, that a general migratory movement

northward brought individuals just north of the Casamance River to

villages such as Affiniam and (probably somewhat later) to Boutem. Oral

accounts also point to Affiniam as the parent village of Boutem: among

other indications, it was referred to me as the "old village" at one


point.5 From there, people likely moved further north to other Boulouf

villages such as Thionk-Essil.6 "Boutame" is one subward of Thionk-

Essil (Hamer 1983:289). This fact, in particular, is interpreted by the

villagers of Boutem as proof that Thionk-Essil was founded by emigrants

from there.7 Hamlets such as Bode and Djilapao, located nearby the

relatively larger village of Boutem and sitting on either side of it,

most likely were founded or expanded as a result of fissioning from the

larger villages, as Boutem probably split off from Affiniam. Violent

disputes over agricultural land continue to break out at times in the

area. Boutem was actively "at war" with Diatok, to its north, during

the mid-1970s; its residents do not consider the dispute settled to this

day. This dispute led to the dissolution of the village association

comprising Affiniam, Boutem and Diatok, as is elaborated in an oral

history of the women's association presented later in this chapter.

In Chapter 2, we considered various forms of migration that

predominated during each of a set of four defined historical periods

(see Table 1). Several forms of migration characteristic of the earlier

periods can be observed in contemporary Lower Casamance. For example,

men still leave their villages to fish during extensive dry season

expeditions (diapang). This form of traditional rural-rural migration,

which can extend through an entire six-month dry season, consists of

camping and trapping fish along the rivers and marigots. It was

5The name Affiniam originates from village residents who tried to sell
fruits in Ziguinchor. When they failed to negotiate successfully they
would complain "Attiham" (literally, "You ate me") or, "you tricked me".
They were so ill-suited to commerce, the story goes, that they became
known in town by a corruption of this phrase, Affiniam.

6Hamer (1983:230) dates the origin of Thionk-Essil at about 1720, based
on an oral history of named circumcision ceremonies, which among the
Diola occur at more or less regular intervals of about twenty years.

7Interestingly, the residents of Thionk interpret the historical
sequence precisely in the reverse, and this is the basis for a joking
relationship between the two groups.

mentioned in only one of my interviews when I asked whether a migant's

parents had ever migrated themselves:

My mother never went to the city, she always stayed in the village
until she became old. She never went to the city. If she
traveled, it was to other villages. She stayed in the village and
did agricultural work. My father only went fishing, he'd go
diapang. (Interview 28)

Alternatively, contemporary village residents, men in particular, may

undertake similar journeys to collect palm wine. These collecting trips

are often taken into Muslim areas, where the demand for this alcoholic

product is clandestine, if present at all. Selling the collected

product of such expeditions can represent a significant portion of a

rural resident's income. Furthermore, as Linares (1992) demonstrates

for another form of migration that was particularly important in the

past, the expansion of Diola communities northward into areas where they

did not reside previously, far from being only a historical process,

continues to this day.

In contrast, the forced migrations common only a few decades ago

are not directly observable in present day Lower Casamance, although as

elaborated above, older forms of rural-rural migration remain common.

Men who experienced corv6e labor, military conscription into the

colonial armies, or who migrated to The Gambia in avoidance of these

still live in the village, and therefore their experiences are

accessible to contemporary researchers. As discussed in the previous

chapter, corv6e labor was required by the French colonial administration

as a form of direct, in-kind taxation.8 The imposition of corv6e was

pervasive throughout Lower Casamance, as it was elsewhere in Senegal,

during and after the conquest of individual Diola villages. I was told

that "many" local men died as a result of the terrible conditions

8Fall and Mbodj (1989); see also Hamer (1983:240-241) for a
consideration of its effects in the nearby village of Thionk-Essil.


imposed upon them during their corv6e service. The French used whips to

force them to work beyond their normal capacities. Crews of Diola men,

using their long-handled traditional shovel, the kayendo, built many of

the roads that continue to be used today throughout the Lower Casa-

mance.' They also built many bridges along the river, using the trunks

of local ronier palms. They were required to construct these bridges

over deep water, which was particularly dangerous.

One of the themes of my interviews with migrants to Dakar was the

history of migration among their parents. As I noted above, many

respondents simply did not know if their parents had ever travelled to

the city. Furthermore, short term events were apparently not regarded

as "real migration" by the interviewees themselves. If a parent

travelled to urban areas in search of health care, or for example if a

woman's mother came to Dakar to stay with her daughter for a few weeks

while she sold lemon juice, interviewees tended to discount these events

as too short to be considered migration. This was so even when such

incidents were explicitly cited and successfully elicited in interviews.

These events, like military service or corv6e labor, were not generally

perceived as migration per se by contemporary emigrants from Boutem, and

therefore are likely to remain unreported, even when one asks

specifically about a parent's history of migration.

My mother said migration didn't begin in her generation, but this
work started with her younger sisters. She said that her younger
sisters went to work in Ziguinchor for a few months. My father
never migrated, except perhaps for. .military service. .
(Interview 20)

Interestingly, although this information never came to light by

other means, including the census, Antoine volunteered during one

conversation on this topic that his father had served time performing

'The first road constructed in Lower Casamance, from Bignona to Tobor,
was completed in 1921 with corvee labor (Mark 1985:106-107).


corv6e labor. At that moment, and later, when this man and his wife so

generously shared their home in the village with me, I was struck by the

immediate accessibility of this unfortunate but nevertheless important

aspect of migration history. In a very real sense, the men who served

in corv6e labor groups represent the first modern migrations out of

Boutem and many other Diola villages. Nevertheless, despite the real

hardships they represent, these migrations were quite temporary: they

were confined to a week or two during the dry season, and appear to have

had little acculturative affect on participants, as much victims as

pioneers. Perhaps it is because of this lack of permanency and

acculturative affect that current migrants tend to discount the

importance of this form of migration, undertaken only a generation or so


Corv6e caused some men to leave the village for the first time, so

on one hand (in a limited sense, because as I've explained above, the

acculturative effect was minimal) it may have increased the integration

of the village of Boutem into the rest of Senegal. On the other hand, I

was also told by villagers that as a result of corv6e they learned to

flee at the first sight of whites heading toward their borders.

Therefore, I was told, when missionaries first came into Boutem to try

and open its first school they were left alone, sitting in an empty vil-

lage "with nothing to do."10

While corv6e labor apparently had minimal acculturative effect on

the individuals forced to serve under this aspect of the colonial

Indigenat, others left Casamance to avoid the imposition of its hard-

ships. These men were among the earliest rather long-term modern

migrants from the village, most often traveling north into The Gambia.

'"Roche (1976) discusses this form of whole village desertion as a
generalized form of passive resistance to colonial rule that was
employed by Diola villagers throughout Lower Casamance.

One of my interviews also touched on an example of this form of

migration, although it was not explicitly related to corv6e itself.

My father used to go to The Gambia, when Jacques's father was
there. He'd stay until the rainy season, but he didn't work. .
He just stayed with relatives until the rainy season. He was
offered a job, but since he was an only son, and his father was
old, he had to return to cultivate. The only son can't stay away
from the village during the rains. He wanted to stay, because he
was asked to stay so that they could get him work, but that's what
stopped him from staying there. (Interview 27)

I was told specifically that the avoidance of corv6e was what "pushed"

one man, V. Manga, to go to The Gambia. Several alternate means were

employed quite skillfully by others in their attempts to avoid military

conscription and corv6e. For example, I was told of the example of L.

Djiba, a highly respected school teacher who now has retired from a long

career teaching in Ziguinchor. He first went to school (probably in

Ziguinchor) through the completion of his brevet." In a successful

effort to avoid corv6e service, he then continued his schooling at a

Catholic seminary. He left the seminary before completing his studies

there because, I was told, he had always intended to use it simply as a

"stepping stone."

The role of the World War II veterans (who served in the famous

Tirailleurs Sen4galais) is emphasized in local accounts of early

migrants from the village. This is an appropriate emphasis since, among

other things, these men often were the first individuals from the

village to learn French. Later, many young men became educated,

migrated for some time to Dakar, and returned to the village for the

rainy season. As part of the village association's rainy season

activities in the village, they organized comical skits that poked fun

at the poor or incomplete French language skills of the elder veterans,

"A diploma awarded in the French school system for the completion of
the first "cycle" of secondary school.

Full Text

xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EJH0MR5YH_18PLJS INGEST_TIME 2017-11-15T18:00:04Z PACKAGE UF00080558_00001