Michael Dougherty

Abstract: Enset (Ensete ventricosum) is a banana-like plant grown throughout the Southern Highlands of Ethiopia as the major staple food crop by many cultural groups. The issue of soil fertility among enset-growing farmers of Sidama, located in the Southern Region of Ethiopia, is embedded within the larger process of how a household makes a living. The traditional Sidama enset production and processing script presented in this paper describes how enset production and processing fit into the larger household livelihood process. Enset growing households of Southern Ethiopia have undergone a gradual process of impoverishment over the past three decades. This erosion of household assets has tested the ability of the enset script to continue to meet culturally established and emerging household consumption objectives. While socioeconomic production conditions and household objectives have dramatically changed, the traditional enset production and processing rules have not kept pace. The impact of insufficient enset script adaptation on female-headed households is examined. The argument is made that the enset production script must be modified through farmer planning to be able to meet existing (and anticipated future) household consumption objectives under new socioeconomic conditions. It is argued that new soil improvement technologies, which will be an important part of this new, modified enset script, must be evaluated in terms of how they fit into the larger household livelihood system. It is concluded that participatory farmer planning is necessary to help households adapt the existing enset script to address changes in socioeconomic conditions and to meet changing household objectives.

INTRODUCTION

The main purpose of this paper is to highlight the need for analyzing soil fertility management and the adoption of soil improvement technologies among small farmers as only one aspect in the larger process of how households make a living. To illustrate how soil fertility management practices and household livelihood strategies in Sidama are intertwined, the increasing inability of the traditional Sidama enset production and processing script to meet existing household objectives under changing socioeconomic conditions will be analyzed. Particular attention will be given to the impact on female-headed households (FHHs). Studies conducted throughout Africa and the rest of the world show that FHHs are typically poorer, have less adult labor, grow relatively more subsistence crops than cash crops and in general have less access to financial and physical capital.1 The final section of the article will examine some of the lessons learned from the analysis of scripts and suggest how these lessons can help future soil fertility improvement programs better achieve their goals of increasing food security.

To analyze how soil fertility management is embedded within the household livelihood system, this paper employs the sustainable livelihoods analysis framework in the description of Sidama enset production and processing and the impact on FHHs.2 The sustainable livelihoods framework highlights the need to examine the complex ways that people make a living given their assets and cultural, political, economic, and environmental contexts when designing development policy. The sustainable livelihoods framework emphasizes examination in five categories of inquiry:

· contexts, conditions and trends

· livelihood resources

· institutional processes and organizational structures

· livelihood strategies

· sustainable livelihood outcomes.3

This paper will cover only four of these five categories. Category three, institutional processes & organizational structures, will not be discussed here. Contexts, conditions and trends and livelihood resources will be briefly discussed to familiarize the reader with the Sidama enset production system. Livelihood strategies will be discussed using the example of the Sidama enset production and processing script. Livelihood outcomes will be examined using a brief case study of a FHH in Sidama.

Sidama soil fertility management practices are embedded within a complex set of gendered cultural rules, guidelines, standard operating procedures, or what Schank and Ableson describe as scripts.4 It is argued here that these scripts are detailed representations of specific household livelihood strategies. As will be seen below, scripts provide a representation of household livelihood strategies in vivid detail, yielding important descriptive cultural information about how activities are completed, who is involved, and highlight the complex contingencies contained in household livelihood strategies. As will be seen below, scripts can be nested hierarchically with embedded decision points. The gendered relationships of household livelihood strategies that scripts represent provide an important tool for examining adoption and adaptation of soil fertility improvement technologies.

METHODOLOGY

The data on scripts and the gender division of labor used in this study is drawn from primary data from case studies of ten Sidama households in two communities conducted by the author and Degife Shibru in 2001. Due to the detail of the production data required for this study a case study research design was used. A snowball sampling procedure was used to select five households within each of the two study communities. Due to this research design, the conclusions drawn in this paper are based on the processes observed working in ten households in two communities and may be suggestive of processes operating on a wider scale. Determining the generalizability of this paper’s conclusions is left to other studies with generalizability as their express concern.

Descriptive secondary data is drawn with permission from an unpublished region-wide food security survey of Sidama (n=270 households) conducted by Degife Shibru in 2000 for the Sidama Zone Bureau of Agriculture.5 A stratified random sampling design was used to represent households within the three major agroecological zones (lowland, midland, highland) of the Sidama Zone.6

GENDER ISSUES WITHIN THIS FRAMEWORK

This paper will discuss men and women’s production roles, the gender division of labor, the gender division of skills and cultural knowledge, and gendered access to capital within the enset system. Discussion of these issues is necessary to understand the relationship between the process of soil fertility decline and households’ choice of livelihood strategy, specifically enset production and processing activities. Understanding the connection between the process of soil fertility decline and choice of livelihood strategy is necessary for the design of effective policy to address food security. As will be seen by the case study of a FHH, the food security of this household, through the livelihood strategies it chose, is integrated in a community-wide process of enset land soil fertility decline.

WHAT ARE SCRIPTS?

People need cognitive tools to assist them in figuring out how to make a living in their complex and uncertain worlds. To help in this process, people create scripts to simplify and codify complex cultural information. Schank and Abelson define a script as "a predetermined, stereotyped sequence of actions that defines a well-known situation . . . a structure that describes appropriate sequences of events in a particular context."7 People are usually unaware that these scripts even exist. They simply use them to complete everyday tasks. Complex localized agricultural production knowledge gradually becomes transformed into scripts as particular combinations of techniques are proven to be successful over time. Chayanov points out that to make a living, households must decide how to apply available resources to existing activities to meet their objectives, however these objectives are defined.8 Scripts represent the standard operating procedures that people use to guide how they will make these important livelihood decisions.

Script formation helps to simplify the complicated web of interrelated factors into an easy to follow sequence of actions. Scripts are the result of a gradual distillation of the process down to its essential steps by generations of users. The script user is freed from having to think about the myriad factor dependencies involved in getting a particular job done. One need only follow the steps in the script. A script is practical and results oriented, hiding most of the logic from the user. The script has proven successful to others in the past and therefore the user need not spend time contemplating the complex underlying procedural details. This freeing of cognitive resources is one of the primary functional purposes of scripts.

An important aspect of scripts is that they embody generations worth of successful indigenous knowledge. Scripts have been developed over time, being incrementally adapted and handed down through the generations as a proven set of successful standard operating procedures that have met the socioeconomic and environmental conditions experienced in the past. Scripts of important activities are diligently taught by parents and learned by children. These scripts are important pieces of technological and cultural information. However, individuals are not inextricably bound by culture to follow these scripts. Often, pioneering individuals deviate from these scripts for a broad range of reasons.9

The scripts that people use everyday are usually preconscious or preattentive and are typically not written down on paper (writing being done almost exclusively by researchers).1 0 The term “script” as used in this paper refers both to the preconscious, preattentive set of instructions that people use to get a job done (emic) and to the written form usually created only by researchers to represent peoples’ preconscious instruction set (etic). Because of the tremendous amount of cultural information contained in these preconscious scripts, the creation a written form of these scripts by researchers can be an important tool for documenting household livelihood strategies.

A TYPOLOGY OF LIVELIHOOD STRATEGIES

Livelihoods are constructed as a portfolio of activities.1 1 Livelihood strategies represent a web of choices individuals and groups make about how to make a living and are based on people’s perceptions of how a mix of available activities can best meet their objectives with existing assets, in a particular context.1 2 Steven Devereux has proposed a series of categories that form a continuum of livelihood strategies.1 3

“Poor households everywhere survive by pursuing a mix of livelihood strategies that seek to increase their income flows and stocks of assets (accumulation strategies), to spread risk through livelihood adjustments or income diversification (adaptive strategies), to minimize the impacts of livelihood shocks (coping strategies) and, in extremis, to prevent destitution and death (survival strategies).”1 4

For the purpose of this paper, one useful way of viewing this series of categories is as a continuum of asset accumulation.1 5 At the ‘positive’ end (in terms of livelihood sustainability) of the continuum accumulation strategies gain assets, adaptive strategies may gain or lose assets, while both coping and survival strategies lose assets. Devereux makes the observation that non-erosive dis-accumulation often takes place in the coping category while erosive disaccumulation takes place in the survival category.1 6

Hussein and Nelson organize livelihood strategies differently, creating three categories: intensification, diversification, and migration.1 7 However, this author proposes that intensification, diversification, and migration should not be seen as categories of livelihood strategies but as dimensions of particular strategies. When making livelihood decisions, individuals and households face a choice about whether to intensify or dis-intensify (extensify in the case of using more land in agriculture) production of a particular activity, regardless of whether it is an accumulation strategy or a survival strategy. Likewise in the case of diversification, when individuals and households consider their portfolio of activities they are also faced with the decision to diversify or specialize (which is independent of the decision to intensify/dis-intensify a particular activity). Finally, individuals and households must make a choice about the location of each activity.1 8

Livelihood strategies can also be discussed in terms of the institutional scales of a particular strategy. Five scales are commonly identified: intra-household, inter-household, community, market, and state. Table 1 groups examples of Sidama coping and survival strategies to food shortage into these five scale levels. Each of the livelihood strategy examples categorized into each of the five scale levels can be classified as either accumulative, adaptive, coping, or a survival strategy depending on the particular context of the household. As will be explained further in the case study, processing enset for other households could at one time be considered an adaptive strategy while at another time be considered a coping or survival strategy.

Table 1. Institutional scales of livelihood strategies. (adapted from Degife, 2001, p. 93)

Intra-household	diversify crop and livestock activities reduce consumption go without food eat wild crops process immature enset for consumption use own financial resources migrate (temporarily or permanently)
Inter-household	seek support from relatives share food, land, labor, equipment, or animals process enset for others
Community	participate in mutual- assistance organizations (idir, ayde, seera)
Market	do petty trading produce and sell rural crafts purchase food in market sell animals sell fuel wood and livestock grass sell immature coffee on the tree borrow money
State	receive famine relief participate in rehabilitation and development projects

Many coping and survival strategies are forms of informal safety nets.1 9 Coping and survival strategies often involve simply the intensification of existing activities rather than engagement in new activities.2 0The strategies chosen as stress increases follow a predictable sequence based on the cost and reversibility of the action.2 1 Informal safety nets are often based on patron-client relationships. The redistribution of wealth that these relationships provide are important coping and survival strategies for vulnerable groups such as FHHs.

STUDY AREA

The households interviewed for this study are located in the west-central part of the Sidama Zone of the Southern Region of Ethiopia. Sidama is both the name of the administrative unit and the Sidama ethnic group. The Sidama Zone is 7,672 square kilometers and the population is approximately 2.8 million people of which 94% of kebeles (smallest Ethiopian administrative district) are classified as rural.2 2 The study area is commonly divided into three agroecological zones.2 3 The semiarid lowland (Amharic: qola) of the Rift Valley comprises 30% of Sidama (1200-1500 meters above sea level; 400-800 mm average annual rainfall; 20.0-24.9E C average annual temperature range). The moist mid-altitude (Amharic: woinadega) comprises 54% of Sidama (1500-2300 masl, 1200-1600 mm average annual rainfall, 15.0-19.9E C average annual temperature range). The cool moist highland (Amharic: dega) comprises 16% of Sidama (2300-3500 masl, 1600-2000 mm average annual rainfall, 15.0-19.5E C average annual temperature range). The study sites are located in the woinadegaareas of Shebedino and Dale Woredas (roughly a county-sized administrative district), approximately 30 miles south of the capital city of the Southern Region, Awasa. Rainfall tends to be bimodal (main rains: June-September, short rains: mid February-March) with rainfall becoming more continuous as elevation increases. However, the short rains are highly variable and since they often fail, farmers claim they are relying on them for grain production less and less.

Figure 1. Map of Ethiopia.

Figure 2. Map of Southern Region.

DESCRIPTION OF ENSET

Enset (Ensete ventricosum), is a long-lived, banana-like perennial plant used for food, fodder and fiber throughout the Southern Highlands of Ethiopia.2 4 The part of the plant that is used for human consumption is not the fruit, but the enlarged pseudostem and underground corm that swell over time with carbohydrates. The leaves are mainly used for fodder and the fibrous pseudostem can be processed for fiber. Enset products are used for everything from food wrapping to medicine. What makes the enset system such an intriguing agricultural system is that enset plants are transplanted several times (2 to 4 times in highland Ethiopia depending on the cultural group) during their 3 to 12 year lifecycle.2 5 In Sidama, enset plants are transplanted once and sometimes twice as will be explained in more detail below. Time to maturity varies widely depending on variety, management, and climate.2 6 However, most of the variation is due to climatic factors that vary with elevation (time to maturity is positively correlated with elevation). Enset typically must be processed before it can be consumed as food by humans.2 7 An elaborate process is required to extract the starchy pulp from the pseudostem and corm of the plant. After extraction, an involved fermentation process is completed allowing the resulting food products to be stored for long periods of time, lasting months to as long as years.2 8

Figure 3. Diagram of the enset plant. [EnsetDiagram.gif]

From Brandt et al., 1997

STAGES OF ENSET GROWTH IN SIDAMA

The Sidama system of classifying stages of enset growth presented here is specific to Sidama (what Werner and Schoepfle call “native definitions”).2 9 Each enset-growing ethnic group has a unique system of enset production and classification.3 0Like the other enset-growing ethnic groups, the Sidama categorize enset into various stages of growth based on the age and size of each plant (Box 1). Moving from one stage to another occurs either due to transplanting or based on the size of a plant.3 1 Enset plants in Sidama are transplanted only once, or if they are suppressed during the transplanted-sucker stage (awulo), they will be transplanted twice (dukalo). Individual enset plants are referred to by their stage name.3 2

Box 1. Stages of enset growth

Unsprouted-corm stage (sima) – The corm typically used for the propagation of enset suckers is ideally taken from a plant from the two-years-after-transplanting stage (simancho), however plants from other stages can be used producing suckers producing suckers with less vigor. The pseudostem of the enset plant is severed from the corm (see Figure 3), the apical meristem is removed, and the corm is buried with manure. “Sima” is the Sidamic term for the corm before the corm has sprouted new enset suckers. This un-sprouted corm stage (sima) lasts about 5 to 6 months.

Sprouted-corm stage (funta) – Once the un-sprouted corm (sima) has sprouted suckers, the suckers, still attached to the corm are referred to as in Sidamic as “funta”. No transplanting is done between the un-sprouted stage (sima) and the sprouted stage (funta). The sprouted enset plants (funta) grow, still attached to the corm (sima). The sprouted corm stage (funta) lasts about one year after sprouting. When the suckers (funta) are big enough to go to the next stage, the corm with the suckers (funta) still attached is uprooted and the suckers (funta) are divided from the corm.

Transplanted-sucker stage (awulo/kasho/kora) – Once the suckers (funta) are divided from the sprouted corm they are ready for transplanting. Suckers (funta) grown on the farm, purchased, or received from friends are transplanted to a new area and the plants are then referred to in Sidamic as “awulo,” “kasho”, or “kora”. During the transplanted sucker stage (awulo), the spacing of plants is regular. The transplanted sucker stage (awulo) lasts one year. The transplanted sucker stage (awulo) is usually weeded/manured 4 times.

One-year-after-transplanting stage (katalo) – The main purpose of the one-year-after-transplanting stage (katalo) is to thin out plants suppressed during the transplanted-sucker stage (awulo), allowing the vigorously growing plants to remain in place for further growth. After the thinning process is complete, the remaining plants are referred to in Sidamic as “katalo”. Plants will hereafter remain in their existing location until harvest; no further transplanting will take place. The spacing between plants will hereafter be irregular for the rest of each plant’s life. The one-year-after-transplanting stage (katalo) lasts one year.

Suppressed-enset-plant stage (dukalo) – Plants that are suppressed during the transplanted-sucker stage and are thinned-out at the beginning of the one-year-after-transplanting stage in Sidamic are called “dukalo.” Plants from the suppressed-enset-plant stage (dukalo) can either be used as livestock fodder and/or transplanted to a new area for further growth depending on the relative need for fodder and/or enset growing enset plants. The suppressed-enset-plant stage (dukalo) represents a branch in the primary progression of enset stages (sima, funta, awulo, katalo, simancho, malancho, etancho, bujancho, kalimo). If plants from this stage (dukalo) overcome their stunting, they will reenter the standard progression of enset stages (at the katalo stage) and become indistinguishable from those plants that traveled the through the primary progression of enset stages.

Two-years-after-transplanting stage (simancho) – From the two-years-after-transplanting stage (simancho) onward, the plants receive no further management until harvest. Typically this is the stage used for the propagation of enset suckers (funta). However, earlier stages are currently being used due to a shortage of two-years-after-transplanting stage (simancho). This stage lasts one year.

Three-years-after-transplanting stage (malancho) – Traditionally, harvest for human food begins at this stage. However, harvesting at stages as early as the one-year-after-transplanting stage (katalo) has become common due to the severe reduction in numbers of older plants in most households. This stage lasts one year.

Four-years-after-transplanting stage (etancho) – This stage lasts one year.

Five-years-after-transplanting stage (bujancho) – This stage lasts one year.

Mature-enset stage (kalimo) – This stage is when the enset plant reaches physiological maturity (flowering).

TYPES OF ENSET PROCESSING IN SIDAMA

There are several types of enset processing that are done in Sidama (hassa, shaqisha, howowicho, udee, ulaame). When a household determines that enset processing is necessary, harvesting of enset begins on an as-needed basis in one, linked harvesting/processing operation. Only two types of enset processing will be described here, primary harvest (hassa) and rainy season harvest (ulaame). For enset processing, several women from the community (since many of the operations require several people) are hired by a household and come together to process enset.

Box 2. Types of enset processing

Primary processing (hassa) - This is the main type of enset processing where the bulk of a household’s enset products will be produced. Primary processing (hassa) takes place during the dry season (October-February). Enset plants from the three-years-after-transplanting stage (malancho) to the mature-enset stage (etancho, bujancho, kalimo) can be processed during primary processing (hassa). Primary processing (hassa) is conducted in three sizes of based on the number of enset plants to be processed. A large primary processing (big hassa), using 200 enset plants or more, is conducted by wealthy households with large enset plantations and provides even the largest households with sufficient food for more than one year. A medium primary processing (medium hassa), approximately 150 enset plants, and a small primary processing (small hassa), approximately 100 enset plants or less, are conducted by middle and low wealth households. Most households that conduct medium and small primary processings (hassa) typically do not have sufficient food to last until the following year. Farmers interviewed during case studies report that households engaging in large or even medium primary harvests (big and medium hassa) are rare or non-existent do to the low number of enset plants in most households’ plantations. Most households engage in small primary processing (small hassa) using even fewer enset plants than have traditionally been used (indicated above).

Rainy season processing (ulaamme) - This is a secondary type of processing taking place during the early part of the rainy season (May-June) whose purpose is to bridge the food gap that exists for households with insufficient food to last the year (households conducting medium and small primary processings). Rainy season processing (ulaamme) is usually used to provide food before maize can be harvested (green or mature). Choice of rainy season harvest processing (ulaamme) is an important indication of food insecurity since it indicates that households are running out of processed enset products before the grain harvest (September-October) and forced to process enset again during the rainy season. Rainy season processing (ulaamme) is less favorable since it has lower labor productivity (1.5 person-days per quintal of processed enset compared to an average of 0.7 for big, medium, and hassa) and is lower yielding (2.1 plants per quintal of processed enset compared to an average of 1.6 for big, medium, and small hassa) than primary processing (hassa) during the dry season based on farmer estimates. Rainy season processing (ulaamme) is a particularly sensitive measure of food security. Since households with small enset plantations typically have no plants larger than the three-years-after-transplanting stage (malancho), they prefer not to conduct rainy season processing with plants from the three-years-after-transplanting stage (malancho) because of the resulting low quality products, but instead consume these plants without processing. Households conducting rainy season processing (ulaamme) can be classified as moderately at-risk, while households opting not to conduct rainy season processing but consume plants un-processed can be classified as severely at-risk.

STRENGTHS AND WEAKNESSES OF THE ENSET SYSTEM

The traditional enset system of the highland regions of Southern Ethiopia is an indigenous, famine-avoiding agricultural system unique to Ethiopia.3 3 The primary strategic importance of enset in food security is that enset helps prevent famine by surviving during droughts when other food crops fail. Although enset is often said to be drought tolerant, it is not drought proof. Enset cannot be grown in semiarid areas. Enset must receive a minimum of about 1100 mm of well distributed rainfall annually for vigorous growth with less than 4-5 contiguous dry months since enset plants must rely on stored soil water to continue growing during dry seasons.3 4 However, once enset plants are established in areas of sufficient rainfall they are able to tolerate occasional years of very low total rainfall or a short rainy season.3 5

Other strengths of enset-based livelihood systems include: storage longevity, multiple uses, and high energy productivity per unit area. The ability to store processed enset products for long periods of time with little storage loss provides households with a mechanism to smooth consumption during food shortage periods. Enset plants provide multiple products that serve many different purposes providing the opportunity to flexibly diversify production of different enset by-products (various fiber products, wrapping materials). Kefale and Sandford estimate that enset yields 1.3 to 3.5 times as much food energy per hectare per year as maize grown under similar management conditions.3 6 For households facing a shortage of land, the higher energy productivity (based on area and time) of enset relative to cereals makes enset an important food security crop.3 7

However, enset-based livelihood systems do face some fundamental structural weaknesses including low protein content, bacterial wilt, continual harvesting, and the need for manure to maintain vigorous growth. The low protein content of enset products (12 g protein per kg of dry processed enset) compared to cereals (100g protein per kg of dry maize) leaves individuals vulnerable to protein deficiency as they come to rely more heavily on enset during crisis periods.3 8 Whereas disease in annual crops threaten only the current year’s harvest, diseases such as bacterial wilt (Xanthomonas campestris) in a perennial crop like enset threatens the harvest for several years into the future.3 9 The enset system of production is capable of providing households with food security during periodic annual crop food production failures and other crises provided that these crises are separated by several non-crisis years when no enset harvesting takes place and the number of enset plants is allowed to increase. However, if crisis years are spaced too closely and reliance on enset during these crises requires harvesting large numbers of enset plants, the future capability of the enset plantation to provide food security is severely reduced.4 0 Continual heavy harvesting of enset reduces the long term resilience of the enset system to provide food security. This reduction in enset system resilience is increased when the supply of livestock manure for enset fertilization is reduced due to the reduction of household landholdings and communal grazing areas and attendant reduction in livestock numbers that growing population causes.4 1 Farmers interviewed for this study claim that without sufficient manure application, enset growth is not sufficiently vigorous to sustain the high harvesting rates caused by the continual state of crisis that many households now face.

Rapid population growth during the Twentieth Century has dramatically reduced the amount of land available for each household, reducing the number of livestock, and thus manure available for enset.4 2 In addition, the assets of many households have been eroded away as the result of a constant chain of low-level crises in the post-revolution period (1974 to present) due to a combination of factors such as erratic coffee prices, rainfall shortages, endless government restructuring, debilitating and inconsistent macro policy, periodic civil war, and official neglect.4 3

It is estimated that since prehistoric times the enset system has helped prevent famine in the region and 15 – 20 million people currently depend on enset either as a staple food crop or as a famine crop.4 4 Historically, farmers throughout highland Ethiopia have incrementally intensified the enset system as they have been faced with gradually increasing population density. However, rapid contemporary population growth, and the social, political, cultural and economic changes it brings, now threatens further adaptation and the continued success of this once food secure agricultural system.4 5 For many in high population density areas, continually shrinking household landholdings are pushing the limits of traditional strategies to provide households with a food secure livelihood. In the face of such rapid contemporary socioeconomic change, it is unclear how these households will be able to adapt their livelihoods to achieve food security.

SIDAMA LIVELIHOOD STRATEGIES

Sidama households have traditionally engaged in various combinations of livelihood activities including: cereals (maize, sorghum, barley, wheat, tef), legumes (beans, peas), root crops (enset, taro, potatoes, sweet potatoes), fruit trees (banana, avocado, citrus, mango), livestock (cows, oxen, sheep, goats, chickens, pack animals, bees), stimulants (coffee, chat), timber (eucalyptus), off-farm work (shopkeeping, civil service, trading, enset processing, laborer, priest) and trades (pottery, black smithing, weaving, basketry, building).4 6 This is a list of the wide range of activities available in Sidama. However, no households would be engaged in all of these activities simultaneously. The combination of activities individuals and households choose depends on household resources, agroecological conditions, and the local and regional socioeconomic context. The overwhelming majority of rural Sidama are engaged in an integrated crop-livestock livelihood system.4 7 One of the keys to success of traditional enset-based livelihood systems is maintaining the proper balance between livestock and access to manure as a source of soil fertility and the size and vitality of enset plantations.4 8 Enset is certainly not the only household livelihood activity, nor is it the most important. However enset has historically played a critical role in household food security. Off-farm work has increased in prominence with the increase in population and the resultant shrinking of household landholdings for crop and livestock activities. Households engaging in trades are more commonly found in urbanizing areas and are often stigmatized for working in trades.

While discussing these livelihood activities, people initially describe a rigid division of labor between men and women. Upon further discussion one discovers a great many exceptions to the general cultural rules governing men’s and women’s work and even more flexibility exists between adults’ and children’s work. Men are typically responsible for food crop planting and harvesting, cash crop land preparation and marketing, livestock production and marketing, off-farm work, and various trades. Women are typically responsible for child rearing, food preparation, housekeeping, food and cash crop weeding and processing, food crop marketing, and some trades. However these rough guidelines are obscured by a mass of conditionality. If children are not attending school, boys are expected to assist their father and girls their mother. If boys are in school, they may have no interest in farm work and do everything possible to escape it. Girls in school have little choice about their taste for work and tend to work almost as long as if they were not in school.

According to Degife, the mean household size in Sidama is 9.1 members.4 9 Sidama contains a mix of multi-generational and nuclear households. Nuclear households formed by a son being given a portion of his father’s land at marriage have been the ideal in Sidama (8 of 10 case studies conducted for this study). However, multi-generational households (2 of 10 case studies conducted for this study), with married children living with parents until their death, have become more common as household land holdings have decreased. Over 60% of sampled households have less than 1.0 hectare of land; the mean landholding is 0.84 ha. Approximately 98% of sampled households relied on the division of their parents’ farm to obtain land. Degife reports that 6.7% of sampled households have insufficient land to divide a portion off for their children and still be left with a viable amount of land remaining. Fifty-two percent of sampled Sidama households report having 3 to 6 successors to the parents’ land and 56.7% of households claim that the dividing of land for married children is the biggest constraint to agricultural production. These statistics suggest that land shortage pressure may be influencing some households to adopt a multi-generational structure to maintain farm-size viability.

Traditionally the Sidama were polygamous, however with the expansion of Protestantism (currently 78% of Sidama households) beginning in the 1930s polygamy has gradually shrunk.5 0 Currently, 26% of households were polygamous, roughly proportional to the percentage of households belonging to religions condoning polygamy (Ethiopian Orthodox, Muslim, and Animist). Serial monogamy is another common practice (6 of 10 case studies conducted for this study). Although death of a spouse occurs (only 1 of 10 case studies conducted for this study), divorce is a more common cause of the serial monogamy phenomenon (5 of 10 case studies conducted for this study). Marital disputes often result in divorce and subsequent remarriage. Young wives often leave when their husband attempts to marry a second wife.

Due to the purposive sampling structure of Degife’s Sidama regional food security survey, 10% are FHHs and 90% are male headed households (MHH).5 1 Less than 1% of household heads are single and not widowed. The statistics for the number of FHHs and widowed household heads are identical (10%). This implies that almost all FHHs are widowed and that most divorced women get remarried relatively quickly. This interpretation of these figures is supported by the case studies conducted for this study. All of the case studies were male headed at one point, including both households that were female headed in 2000/2001 (both widowers). Most men reported preferring to marry young wives regardless of their own age and tended not to marry widowed women (0 of 8 MHHs married widowed women, 2 of 2 widowed FHHs remain unmarried at the time of the interviews and consider themselves very unlikely to remarry). Until the most recent regime, women had no legal property rights (officially there is no private land ownership, all land is owned by the state and long term leases are granted). Despite the recent national law that gave women the right to lease land, women in rural areas throughout most of Ethiopia are rarely able to gain access to land independent of a male relative.

However, FHHs tend not to remain female headed very long. After the death of a male household head, the search for a successor begins immediately. Sons that are interested in farming will become the new household head when they are old enough to marry, until that time the household remains female headed. In Sidama, FHHs are ideally only a temporary state. Why? Without a male household head, a female household head is particularly vulnerable to land claims that can potentially be made by her husband’s male relatives.5 2 Due to widespread land shortage, there is intense pressure for Sidama households to name a male successor. Therefore, FHHs, especially those with prime land, must often struggle with the delicate question of identifying a successor and must defend their children from rival male relatives.5 3 Those FHHs that remain female headed for an extended amount of time are likely to be virtually landless (see case study below). Sidama property rights regimes and the cultural norms governing FHH formation and dissolution conditions FHHs choice of livelihood strategy as shown in the case study below.

THE SIDAMA ENSET PRODUCTION AND PROCESSING SCRIPTS

The enset production and processing script presented here developed over centuries and has remained relatively unchanged despite the dramatic socioeconomic changes experienced during the Twentieth Century. The Sidama today produce and process enset much the same as they have for centuries. The only notable exception is the option of using inorganic fertilizer that emerged with increasing coffee extension that occurred in the later decades of the Twentieth Century. However, inorganic fertilizer is a very recent occurrence and its routine use has not been widely integrated by most farmers into the enset production script (Box 3). Although the inorganic fertilizer decisions are included in this script, most farmers quickly rule out its use, primarily due to cost and other constraints discussed briefly below.

During the case study interviews, when people are asked about who does what in Sidama, everyone claims that men are responsible for enset production and women are responsible for processing. However, this does not mean that only men are involved in production and only women are involved on processing. Upon further observation and questioning it becomes clear that both men and women are both involved in a range of activities ideally deemed appropriate only for the other sex. This crossing of gender lines is especially marked in FHHs and in their case little of what follows in terms of division of labor applies. These gender dynamics make more sense when traditional Sidama men and women's production roles are further examined.

Sidama believe that men's role is to produce enset and that women's role is to process enset. Men and women ideally have separate production domains, where one may enter the other’s domain, but strictly on the other’s terms. For example, although men may participate in specific aspects of enset processing (digging pits, mixing, squeezing, transporting enset products around processing areas), women are clearly in charge of the whole procedure, doing the skilled operations, and making all decisions with men largely following orders (Boxes 4 and 6). The same is true of enset production, with women providing much of the manuring and weeding labor, with men in charge of the operation, making most of the critical decisions, and doing the skilled operations (Boxes 3 and 5). There are two possible exceptions to this concept of gender domains that seem to prove the rule, one for production and one for processing (these decisions will be marked with an asterisk in the scripts for distinction). First, although men are responsible for enset production, men and women (often heatedly) discuss what mix of enset varieties (“clones” since enset is vegetatively propagated) will be propagated or bought based on the different qualities valued by men and women.5 4 The second exception is that men and women debate which clones and how many enset plants from each stage will be harvested for processing due to the different objectives of men and women. One explanation for these domain exceptions (women “meddling” in enset production and men “meddling” in enset processing) is that these are two decisions where men and women’s domains overlap. The production decision about what clones to propagate and/or buy and plant can cause conflict between men and women because different enset clones have different uses.5 5 Another traditional domain of Sidama women (as elsewhere) is providing food for the household, and another traditional domain of Sidama men is livestock production. Since enset can be used both for human food and animal fodder (as well as for other purposes, adding to its politicization), men and women often compete for enset products to meet the various needs within their domains. This competition for enset resources precipitates a complex process of negotiation between men and women within the household to fill their roles. As one might expect, this process plays out in a range of ways for different households and is beyond the scope of this paper.

Box 3. Enset production and processing script.
October-November, year 1 – Sub-script: Prepare land for propagation (sima).

Female labor: 1 hour/corm

Male labor (hired or household): 2 hour/corm

Four steps are required for bed preparation. First a small amount of land (about 4 sq. m per corm) is selected, usually near the house. Second, manure is applied prior to tilling. Transportation of manure from the house to the enset field has traditionally been the sole responsibility of women and girls. This is onerous work and everyone complains about it. One basket holding about 4 to 6 kg of manure (gimbola) is placed between the spots where the corms will be planted. The third step is done with a chest-height, double-poled hoe (wenencho) used to turn the grass under. The fourth step is done with a short handled hoe with a wide blade (jamba, safuya). The hoe is used to cut-up the turned clods of soil and mix the manure into the soil. Men typically do tilling.

November-December, year 1 – Sub-script: Propagate suckers (sima).

Female labor: none

Male labor: 1 hour/2 corms

There are three steps in propagation. First, typically corms from the two-years-after-transplanting stage (simancho) are used for propagation, however other stages can be used. The enset plant is prepared for propagation by removing the pseudostem of the plant just above the corm and the pseudostem is processed along with a sufficient quantity of other plants (for processing details, see Box 4). Second, the central shoot of the corm is cut out to induce the budding of sprouts from auxiliary buds. Third, holes are dug at 1.0 m spacing and the corm is buried.

February-March, year 2 – Sub-script: Weed unsprouted corm (sima).

Female labor: none

Male labor: 1 hour/2 corms

The unsprouted corm stage (sima) is weeded once approximately two months after planting at the end of the January-February rains (belg). No manuring is done at this weeding since manure was applied during land preparation. Men typically do weeding. There are two steps to this weeding. First, large weeds are pulled-up by hand, the dirt is knocked-off, and the weeds are left as mulch. Second, small weeds are simply hoed under using a short handled hoe with a single pointed head (helako or tike).

September-October, year 2 – Sub-script: Weed sprouted corm (funta).

Female labor: none

Male labor: 1 hour/2 corms

The sprouted corm (funta) is weeded once, typically after the June-September rains (meher). No manuring is typically done at this weeding since manure was applied during land preparation. The weeding process is exactly the same as the weeding of the unsprouted corm.

January-February, year 3 – Sub-script: Prepare land for sucker transplanting (awulo).

Female labor: 0.5 hours/2 sq. m plot

Male labor: 2 hours/2 sq. m plot

Four steps are required for bed preparation. First, a plot (about 2 sq. m) must be selected to plant the approximately 100 suckers produced from two average corms. Second, women prior to tilling apply manure. Typically, 1, 4 to 6 kg baskets of manure (gimbola) are applied to a 2 sq. m plot at this stage. The third step is done with a chest-height, double-poled hoe (wenencho) used to turn the grass under. The fourth step is done with a short handled hoe with a wide blade (jamba, safuya). The hoe is used to cut-up the turned clods of soil and mix the manure into the soil. Men typically do tilling.

March-April, year 3 – Sub-script: Transplant suckers (awulo).

Female labor: none

Male labor: 1 hour/2 sq. m plot

There are three steps to transplanting suckers. First, the sprouted corm is uprooted. Second, the suckers (funta) are divided from the corm. Third, suckers (funta) are transplanted at about 15-20 cm diagonal spacing (about 50 suckers/sq. m). Men plant suckers (funta) using a chest-height, single-poled hoe (wenencho). Once the suckers (funta) are planted thay are referred to in Sidamic as “awulo”. Small, weak, or inferior suckers (funta) can be planted in groups of 2 or 3 plants in the same hole at the same spacing along with the vigorous suckers (funta). These stunted suckers planted in groups of 2 or 3 in the same hole once planted are referred to in Sidamic as “mugicho”.

May-June, year 3 – Sub-script: Do first transplanted-sucker (awulo) weeding/manuring.

Female labor: 0.5 hour/2 sq. m plot

Male labor: 1 hour/2 sq. m plot

The first weeding (karkara) is typically done about two or three months after sucker (awulo) transplanting. There are three steps to this weeding. First, manure is placed between the plants by women. Approximately 20, 4 to 6 kg baskets of manure (gimbola) are typically applied to an area of 50 sq m. Second, manure is tilled-in by men using a short handled hoe with a single pointed head (helako or tike) during the weeding process. Large weeds are pulled-up by hand, the dirt is knocked-off, and the weeds are left as mulch. Small weeds are simply hoed under along with the manure. Third, for the plants that are growing vigorously, men turn down the leaves to prevent their suppression of the surrounding plants. Turning down the leaves at this time helps to equalize the growth across the cohort of plants. The turning down of leaves is done by grasping the leaf midrib near the base and rapidly snapping the midrib in a downward motion. The leaf is not necessarily totally removed. It can be left loosely attached to the plant, left as a mulch at the base of the plant, or fed to animals.

July-August, year 3 – Sub-script: Do second transplanted-sucker (awulo) weeding/manuring.

Female labor: 0.5 hour/2 sq. m plot

Male labor: 1 hour/2 sq. m plot

This weeding/manuring is done exactly the same as the first weeding/manuring operation except that the rate of manure applied is slightly reduced and leaves are managed slightly differently. Approximately 15, 4 to 6 kg baskets of manure (gimbola) are typically applied to 50 sq. m. During the second weeding, the leaves of all plants are turned down. Turning down the leaves at this time allows water to infiltrate better during the rainy season and helps to equalize the growth across the cohort of plants. Any plants that have died are removed and suppressed enset plants from the same cohort are transplanted into their place. The stunted suckers planted in groups of 2 or 3 (mugicho) are typically used to replace dead suckers.

September-October, year 3 – Sub-script: Do third transplanted-sucker (awulo) weeding/manuring.

Female labor: 0.5 hour/2 sq. m plot

Male labor: 1 hour/2 sq. m plot

Box 4. Sub-script: Do small primary processing (small hassa).

Starting October-February – Sub-script: Process pseudostem and corm.

Female labor (hired or household): 30 person days

Male labor (hired or household): 4 person days/

The purpose of this stage is to uproot the enset plants from the enset field and transport the plants to the processing areas and extract the pulp from the pseudostem and pulverize the corm. Women conduct the processing portion of this stage. Men assist women in the uprooting and transport of plants to the processing areas due to the large number of plants required for primary processing (hassa). The plants must be dug out of the ground and the leaves are removed using a knife (konchora). The following numbers of each stage are uprooted and transported to the processing areas. Ten, three-years-after-transplanting stage plants (malancho) are harvested per day for the first five days. Twenty-five, four-years-after-transplanting stage plants (etancho) and 25, five-years-after-transplanting stage plants (bujancho) if available. Enset plants are continually harvested and partially processed over approximately a 15 day period. Two women typically work together to process the enset plants. Each day as the plants are harvested and brought to the processing areas, the corms of the plants are processed in one area (corm processing area, CPA) and the pseudostems are processed in another area (pseudostem processing area, PSPA). In the PSPA, the starchy pulp (abicho) is extracted using a bamboo (sisicho) or metal scrapper to scrape the pseudostem while it lays on a slanting board (meeta) that is supported by a ‘y’ shaped vertical post (dawe). A pit is dug and lined with enset leaves under the slanting board (meeta) where the extracted pulp is placed. High quality fiber for making ropes, mats, etc. can be extracted from the inner portions of the pseudostem, while the tough outer portions of the pseudostem are saved and dried (shigido) for use as wrapping, tying, and for building material. At the CPA the outer soil-covered portion of the corm is pared-off and discarded. The corm is pulverized using a serrated tool made of cattle bone (keho). The pulverized corm (dassa) is placed into a two pit system lined with enset leaves that allows the liquid from pulp stored in an upper pit to drain into a lower pit when it is squeezed and mixed through treading with the feet. Each evening the pulp from the PSPA extracted that day (abicho) is brought to the CPA and mixed with the pulverized corm (dassa) in the higher pit. The combined mixture of processed pseudostem (abicho) and processed corm (dassa) pulp is then referred to as “abicho” (abicho + dassa = abicho). The starch-rich liquid that drains into the lower pit is allowed settle overnight, and in the morning the water is drawn off of the top and a fine, high quality starch is left behind (bulla). This routine continues until the appropriate number of enset plants have been harvested, requiring two women about 15 full days to complete.

5 days after pseudostem and corm processing started – Sub-script: Prepare fermentation starter (gamancho).

Female labor (hired or household): 0.5 person days

Male labor (hired or household): none

The purpose of this stage is to prepare a fermentation starter (gamancho) made from fermented sections of mature enset corms that will be used to help innoculate the processed pseudostem and corm to allow it to properly ferment to maturity. Women solely conduct this stage of processing. This stage runs concurrently with pseudostem and corm processing and is started about 20 days after harvesting and processing has begun. This stage takes about 8 to 10 days to complete and is therefore started about 10 days before the beginning of the next step must begin. Only corms from four-years-after-transplanting stage plants (etancho) and five-years-after-transplanting stage plants (bujancho) work properly for making the fermentation starter (gamancho). A section of the corm is cut off, rubbed with a piece of dried enset pseudostem (shigido) or other local plant (hecho, maraca, etc.). The corm is then tightly wrapped with pieces of dried enset pseudostem (shigido) to make an airtight wrapping (first sala) and left for 5 days. On the fifth day the wrapping is removed in the morning and allowed to air until evening. At this stage the corm is checked for proper ripening. The corm is fed to animals if not ripening properly and is not used for enset processing. On the evening of the fifth day the ripening corm is re-wrapped (second sala) and allowed to continue ripening for another 3 to 5 days.

After pseudostem and corm processing completed – Sub-script: Add fermentation starter (gamancho) to processed pseudostem and corm (abicho).

Female labor (hired or household): 1 person day