“How do the whites people manage to have so many children?
— ... But they don’t have many children!
— Yes, they do—and without many wives.”
Old man from the Ithalie set.

1Neither contemporary demographic growth in Africa nor the models built to account for it give any real indication of the complex history of the people of these regions. As my research into the generation sets progressed, the informants—women and men alike, of their own full accord and in their own manner—described practices of late marriage, birth spacing, and barely married women who were no longer able to bear children. Eventually, reversing our roles, they in turn gave in to curiosity and began to ask their own questions on the subject of white sexuality. The idea of the “prolific white man”, evoked in the short dialogue cited above, remained a mystery until this figure was understood as the inverse of Tigania-Igembe fertility—a symbol of the major conflict between the administrators and the population over indigenous customs that came to seem increasingly bizarre following the rift created by colonial conquest.

2These former practices, linked to low fertility rates and described retrospectively by the informants, were a key element of the events of the colonial period as reflected in the archives. Their eventual replacement by a new set of practices was the result of multiple factors, and it is worth examining both the nature of these factors and their chronology. This process also dealt a fatal blow to ancient traditions, with their absence revealing various elements of the demographic regime by which the generation-based polity had operated. We will therefore conclude our study of this complex polity by considering how the Tigania-Igembe changed from a stationary to a growing population, via a largely invisible process involving a combination of immediate historical factors and the long-term effects of the demographic regime.

1. Retrospective accounts

3Our first case is that of a woman in a rather particular situation: born of a Meru mother and an Indian Sikh father who arrived to conduct trade at the beginning of the century, she has always lived in the region. In the 1930s, she became the second wife of a Sikh man whose first wife came from the Tigania section he had settled in. She shared the daily life of Meru women and her lifestyle is not much different from theirs but, because of her mixed parentage and her own family experience, her perspective is somewhat different. Her Meru co-wife has “only” three children whereas she has had at least seven. She begins by mentioning the small number of children born to Meru wives in earlier days, and moves on to discuss the practices of abortion and infanticide:

Five wives, five children, just one child [per wife] and then no more. Three wives, three children—I don’t know if they died. Look at M’M—his father had three wives: the mother of M’M with two children, the mother of M’I with two children, the mother of O with one child; three wives, five children. Nowadays, one woman—ten children. Before, women didn’t like to have many children. Now they’ll have ten children in ten years. Before they didn’t grow much, didn’t have to pay school fees, didn’t buy clothing, salt, paraffin oil, onions. Their activity was to cook and eat, and they didn’t like having many children—they called them nthuîrwa. You’d still see children this big [hand gesture indicating height] being breastfed by their mother. Children were very healthy—big and plump. Now, ten children in ten years [...] What’s happening these days is neither the white model [she uses the term mondeli derived from the English], nor that of the Meru—it’s a different model and I don’t know where it comes from. We know the white model with two children, and the other models are from the Indians and Meru. The Meru didn’t have many children, and having many was considered bad behavior. A mother wouldn’t have another child until her daughter was able to fetch wood from the forest, or her son was able to take the herd grazing as far Rwanda [the name of a pasture located fifteen kilometers downhill]. The mother would get old, and be unable to have any more children […]. What happened was that they married late, were initiated late, and their bodies were unable to have any more children. After [marrying], they’d have two or three of them, and then it was over, they were too old. That’s why there was a lack of children […]. They’d also have three, four, five abortions before marrying. The abortions were done badly and, afterwards, you couldn’t have any more children […] the reason for these abortions was that having children without being initiated and married was a source of shame. If you were initiated with a child, you’d be excluded from your own home. That’s why they’d throw them in the bush and kill them. You see, for gûtûra matû [ “ear piercing”], if a girl had a child, his father could not prepare the hydromel, nobody was allowed to drink any; if a girl is normal, she may present the hydromel and her father’s nthukî [generation] may drink some1 (December 1988).

4Another perspective is provided by Ithalie wives, now grandmothers who have converted to Catholicism and play an active role in parish life. They speak first about the ban on procreation by consecutive generations and then, spontaneously, provide an estimate of when things changed, before finally discussing the issue of non-initiated girls:

People said that if a child married and had children, that meant ‘I’m getting older—now I want to become a grandmother’. In the days of our nthukî, anyone who had their children initiated wasn’t allowed any more of them. You’d gotten old and were ashamed—ashamed in front of children. Nowadays, that’s changed […] but in the days of the Thirindi [Kiramunya wives], or the Cholubî [Gichungî wives], you’d never meet a mother who had had other children after her son’s initiation. If a woman accidentally became pregnant, she’d hide in her house, cover herself in clothes from head to toe, her children weren’t allowed to see her… [repetition]... she wouldn’t want to be seen by her daughter, because her daughter is a ngutu [an initiated girl in the reclusion stage]. When the child was born, she’d be very ashamed and even think about killing herself. When she went out, even then, she didn’t want her children to see her. But nowadays there’s none of that any more. Before, it would have been impossible to find someone who had other children after their children had children. It’s our children’s nthukî who started it, the Ratanya. We can’t insult the Mûkubu by saying that they did it—it’s the Ratanya who started all this […] If a girl wasn’t initiated, people would say her child was the child of a mukenye, and nobody would go near her. Nowadays, they become mothers. Before, the child of a non-initiate was called a kitheega, he would be suffocated and left in the bush, that’s what would happen […] Nowadays we say that all that was madness. Somebody was killing something created by Murungu [God], a real person, and even back then anyone who thought about it would wonder why something created by Murungu was being killed. The person doing that [the abortions] was put in prison and his hands were cut off. They asked why he did it; it was like taking a drug […] he was scorned by the others […] Once he was arrested by the [colonial] government, people couldn’t get abortions any more […] The children were hated—they were called kitheega and they were said to put others in danger. A segregated house was built for the ones who survived. Before, fathers and mothers made sure that all these children were killed. Nowadays, the same children are champions, the ones you heard me say would have been killed before. Even if he walked through lances, he wouldn’t be wounded, whereas the son of an initiated mother might die. And Murungu has helped them—now they are at the top, they govern us, these children who would have been aborted before. What happened before was bad. (December 1988).

5A kobia elder from the Michubu set explains the late marriages and birth spacing, a disclosure that is particularly interesting since men very rarely agree to be interviewed about such issues.

Why did our fathers have so few children, and even us [the Michubu] too? A warrior (mûthaka) remained a warrior for a long time because of the fighting; he remained a warrior until the time of his firstborn [men could sleep in the warriors’ gaaru until the birth of their first child and even after]. Girls also waited to marry. By the time they married, they’d already be old. They’d give birth to few children. And then, once a child was born, the father wouldn’t give the mother another child until the first one was able to take a gourd and fetch water. Nowadays, another child is born before the previous one has even left the mother. That’s why things are the way they are now; it changed with the Lubeeta, Miriti and Ratanya. But with the Michubu, there were few children, except for men with two wives. If another child comes [too quickly], that causes problems for the previous child. He isn’t strong enough; his hair turns brown. He’s been weaned too early. He can’t grow, because his father doesn’t have the strength (inya) to raise him. That’s why people said that a child shouldn’t be weaned, so that it could grow quickly. If the father had no herd for milk and was thus unable to do anything, people would say that the child shouldn’t be weaned, or it wouldn’t grow (April 1993).

6A Ratanya wife explains the changes that occurred in the 1950s; either spontaneously or in response to questions, she describes new habits relating to where spouses lived, changes in farming methods and the introduction of commercial farming that complement the above account, land-related problems, and the changes that have occurred for men:

Our mothers, the Ncencenga [Ithalie wives], were initiated when they were grown up, old (akûrû). By the time they were married, they weren’t able to have children any more. The Mûkubu [Michubu wives] came next and it was the same for them. Then came the Ratanya— they were initiated when they were still children (twaana). That’s why more children were born. Before, women would have two or three children. You didn’t have another child until the firstborn knew how to fetch water; he’d grow to the age of almost ten. The Lubeeta women didn’t even wait for their child to walk before having another. The Ratanya women waited for the child to be able to walk a little and get some strength (inya). And the Ithalie men had their gaaru [their own house] too. When a child was born, the father stayed in his house [i.e. his wife did not come to see him] for more than five years; for five years there was no more ngûgî [term meaning “activity”, referring in this context to sexual activity]; the husband couldn’t go near her until the child was old enough because of the smell of the child (ntuulî ya kaana). Now when the woman leaves the hospital, they go home together, there are no more gaaru […] Nowadays, there isn’t much land available (kithaka), but before somebody could have ten children without it being a problem. We started to run out of land around the time of the Ratanya. Now there isn’t any more land. For example, a man like my father had almost ten children; all his children had land. And now, for us, there’s no longer enough, because the land is divided up. My father was an only child; my father-in-law M’Licoro was an only child. He had a lot of land; he divided it up and there was enough for his children. Now it is difficult to have a place to build a house and we wonder where our children will go […] They used to use the cultivating stick to cultivate the land (kûrima na mîroo). There were many things back then [i.e. a diversity]: there were nchûgû, nchaabî, mweere, mwee, nthûroko, nthûroko cia mûniythû planted as vegetables. All that meant a big harvest. When they started to cultivate land using the hoe (mathitangi), that’s when tea arrived. We stopped cultivating a variety of things and now there’s only corn. Before, the kirûrû [a large basket in the wife’s hut used to store grain] would be filled to the brim with millet—and gourds like the ones you can see, we’d put millet gruel in them […] With the hoe you can go more quickly than with the cultivating stick, and with the hoe we also used the kalo [machete]. With the cultivating stick, it was a matter of chance whether we had food—only if the ground were fertile [nthî yaarî na into, literally: “the ground was with things”]. We would cultivate a few gardens and food was available. But we knew the advantages of the hoe; when the hoe arrived [she accompanies this by imitating the noise of the hoe], you turned the soil with the hoe but planted with the kalo. We all adopted the hoe and stopped using the stick. We knew that the whites had brought cash to cultivate land [i.e. they introduced both currency and cash cropping]. That was when people started to cultivate more and more land, and there was no more land available.

7A woman interrupts the interview and they begin talking in a group. To summarize: one describes a situation where the husband went to work elsewhere and left his wife with their children; when they grew older, they helped their mother to dig the fields and fetch water; another explains that some people earned money quickly and had many children because they married several women; another adds, grumbling, that no mûgiro occurred when births were less spaced—implying there was no actual rule obliging people to space births—and explains, in a critical tone, that “the women who didn’t have many children were just taking care of their body. Everyone knows that if you give birth often, you end up with problems. That’s what’s happened to me” (April 1993).

8A final testimony comes from Ncororo, an Ithalie wife, who describes how the kaaria was repressed:

Kaaria was a dance where the kaaria would be applied to a woman. It was thirikari [the government] that stopped the kaaria. They said that women were not working and wasting money because of the kaaria. That’s why the kaaria was shut down by the government. At that time, we were children […] the kaaria was stopped before the Mau Mau. Women liked it before it was stopped. The kaaria was stopped and this happiness (nkerû) was stopped. The women would use a lot of things for the kaaria, and the men said it was a waste, so the government stopped it. There’s no kaaria anymore; there was no way of keeping it going. If the government had let the women have their way, the kaaria would still exist nowadays. Yes, there were conflicts. Men could become divided because of the kaaria. If a wife did not know how to open it [implies kûrugûra kaaria, “open the kaaria”, i.e. to have obtained the kaaria on the public ground after the ground has been opened for this purpose], she might be humiliated. She’d go back to her house and come back out with a basket filled to the brim, which she took to the Owners of the kaaria. Then the kaaria was opened and she, in turn, could know this council [i.e. be a member of the kiama]. The njuri is what gave the government its strength. They said that the women were wasting money and the contents of the kirûrû [a store of grain kept in the wife’s house]. The njuri gave the government strength, and the njuri made the kaaria stop. The kaaria is not witchcraft, the kaaria is not witchcraft. That’s just a rumor. If a woman wasn’t a member of the kaaria, and her husband was opposed to it and said it was a waste, she would go mad. Then the woman would say: “I’ll never come back to this man”, and she might become the wife of another man [in this case, separation was legitimate]. Before, everyone valued becoming a member of the kaaria. Women valued it. A woman who didn’t join it couldn’t speak to anyone; people would say that she was not of the kaaria; she’d be scorned [agacimbwa, from the verb gûcimba: literally meaning “to spit through one’s upper teeth; to scorn”]. Then she’d go crying to her husband; he’d tell her to take grain from her kirûrû because it belonged to her. Then she’d carry a full basket to the “kaaria posts” [to the owners of the kaaria]; then she would be respected (kûgwatîrua nthoni) (January 1989).

2. The demology of the old Tigania-Igembe

9The rules and customs described in this book’s chapters are the cause of the low fertility rates indicated in these interviews, and have their basis in the generation-set structure which they perpetuate. We have seen them in a great variety of contexts, from the obvious to the more surprising, in the form of obligations, recommendations or prohibitions, in rituals, myths and minor everyday occurrences. Their combined effect helps contain the numbers of overaged and underaged people, and prevents too great a numerical imbalance between the successive sets and generational divisions, without which the sets and grades would no longer be clearly defined and the polity would break down. This extraordinary pervasiveness means we should examine the models and practices it implies, in relation to demographic reproduction, which we consider here from an ethnographical and ethnological perspective.

10Before taking this rather uncommon approach, which combines the intellectual methods of two disciplines, we must first recall certain basic notions which will help us to understand the demographic parameters by which the qualitative material collected by ethnographers can be analyzed. Demography looks at variations in the size of populations. From one census to the next, the population may have increased, decreased or remained constant. This dynamic depends on two figures, calculated as a rate per thousand people: the number of births between the two censuses, minus the number of deaths that occurred in the same period. The difference between the two rates reveals the so-called ‘rate of natural increase’, which may be negative (meaning more deaths than births) or positive (meaning more births than deaths). The rate of natural increase must then be corrected to account for net migration i.e. the number of emigrants in relation to the number of immigrants, with the difference between the two giving net migration that is either positive (more people entering than leaving) or negative (more people leaving than entering). Depending on the case, the net migration rate may either contribute to or reduce the rate of natural increase. The result gives the total increase.

11Among the various factors affecting the quantitative dynamics of a population, birth rate and fertility rate are particularly of interest here. The birth rate, or number of children born in a given population over a fixed period of time (usually one year), depends on two factors: firstly, the number of women of childbearing age, which in turn reflects the age pyramid and sex ratio—that is to say, the population divided up according to age (usually using intervals of five years) and gender; the greater the number of women of childbearing age, the higher the probability of children being born, and vice versa. The other factor is fertility: the average number of children per woman, either measured as a fertility rate or in terms of lifetime offspring, which depends on a range of variables (the “intermediate fertility variables” well known to demographers): to summarize briefly, these variables may be both material (the population’s general health, income, standard of living, and so on) and cultural (practices determining the beginning and end of a woman’s reproductive life, the way children are valued differently, and so on). The cultural—or, more precisely, qualitative—factors contributing to fertility are very difficult, impossible even, to analyses using the traditional methods of demometry and statistical analysis. On the other hand, they may be subjects for ethnography, provided the researcher bears in mind the issue of the quantitative effects of the practices and rules revealed by their investigation, without which the material gathered will come to nothing.

12The case of the age- and generation-set system is exemplary in this regard, and provides an almost experimental case for this kind of interdisciplinary exercise. The ethnologists studying these institutions whose work was mentioned in the first chapter, with the help of demographers and statisticians after returning from the field, had only scraped the surface of the issue by investigating the phenomenon of overageing and underageing—a factor that, as we have seen, can rapidly undermine any generation-set system. The fact that the “successive” generations are generations of fathers and sons, and not of mothers and girls or parents and children, did not influence their analysis, which remained androcentric despite the clearly gendered aspect of human reproduction. Nobody has examined the problem from the angle of women and the practices or rules which may affect their fertility, which is precisely what the elderly women and men in Tigania-Igembe encouraged me to do. I must additionally describe some elements of my methodology. I did not resort to demometrical analysis, in the usual demographic sense of population-measuring; an examination of historical censuses at the end of this chapter nonetheless enabled me to obtain a certain amount of numerical data. On the other hand, the major parameters of demography, recalled briefly above, are the basis for my analysis and guided my research. “Demology” here means the study of the models (i.e. representations) of norms and values, and the indigenous practices that result from these models and have a direct influence on demographic parameters, as well as the reverse effects of the latter on the former. In other words, it is an attempt to respond to the question implied by the woman cited at the beginning of the chapter: “What’s happening these days is neither the white model [she uses the term mondeli derived from the English], nor that of the Meru—it’s a different model and I don’t know where it comes from.” An understanding of demology allows me to employ the concept of a “demographic regime”, here envisaged in the widest possible sense as the intricate relationship between social structure and the factors affecting the reproduction of the population. Age- and generation-set systems represent an unparalleled case where this concept, which is of great interest to comparative study but not always easy to implement, may be adopted and refined. In the long term, the notion of a demographic regime also relates to a wider demological project providing the framework for a combined anthropological, demographical and historical study (Peatrik 1997, Cordell and Piché 1997, Kertzer and Fricke 1997).

13A primary element of the Tigania-Igembe demological model is the set of customs resulting from the stigmatization of simultaneous procreation by consecutive generations. A reproductive overlap between adjacent generations is regarded as an anomaly that brings mûgiro, illness and death. Consequently each generation, in the Tigania-Igembe sense of nthukî, has a limited window in which to reproduce. A text that is well known to ethnographers of this region’s populations (though its exact subject has perhaps never been fully understood) will help illustrate, by contrast, this stigmatization and the effect that a transgression of the rule can have on reproduction. An elder from the Macha, an Oromo group in western Ethiopia, whose gaada-type generation-set system is no longer in use, laments:

When gada was destroyed, they left gada. The bull refused to mount the cow, men no longer respected justice. There was no one who could be given the office of “abba biya”, father of the land. There was no one who could take the office of capturing criminals. There were no longer any real elders, and few children were born. The cows gave birth to deformed calves. Pregnant women gave birth to their children at the wrong time. They bore children without hands. Lambs were born without forelegs and without tails. And calves were born which had no tails. When the gada customs were destroyed, everything else was destroyed. When gada no longer existed, there was no justice. The crops that were cultivated no longer grew. And the oxen refused to fatten. The man who had formerly respected truth and justice abandoned them.
Knutsson 1967: 180.

14In this case, it is no longer merely a question of the negative effects of “suspended time”, denounced in the myth of Mbwaa, which can lead to a period of regression. The crisis lamented by this old man is in some ways similar to that of the Meru following colonial conquest, and during that period perhaps one might have found elders issuing similar complaints. In any event, the confusion resulting from the disappearance of the generation-set system and the associated avoidance rules had consequences for both public morals and actual procreation. These two elements, which are indissociable in the old man’s account, clearly demonstrate the effect of the system’s rules, in the eyes of the population, on ageing and accomplishment. The collapse of the system resulted in a general slowing in the growth of plants and animals, a disordered state of sexuality and reproduction, and the appearance of deformed humans and cattle: all things that, for the Meru, correspond to an increase in the number of the teratogenic kitheega they so fear. These negative effects allow us to better understand the positive elements of the system when it functions correctly.

15At the heart of any generational system—at least those with structured sets and grades, like the gaada and Tigania-Igembe systems—there is a model for the individual that includes a specific period for procreation; once a set reaches the appropriate grade, its members apply themselves to this task. This simultaneous engagement in reproduction is an inherent element of the bankiro bonds, which confer a sort of power by which reproductive accidents afflicting members of the set, no matter how serious, may be remedied. The notion of a window for procreation leads to practices marking both the beginning and end of the reproductive period that are supported by the banning or strong stigmatization of any conception outside these limits. While these temporal limits on procreation are not directly related to bodily changes such as puberty or the menopause, there is clearly some connection. This loose synchronization of the social and physiological “calendars” is probably the most intriguing aspect of the way this population deals with the criterion of age, since it implies the organization or regulation of the biological—the individual’s “body clock”—through social rules, as already suggested by the “life measures”, and the longevity or life expectancy that these suppose. Initiation signals the beginning of the reproductive period for girls; boys are considered procreators once they have progressed from the grade of warrior. The menopause occurs at the same time as acquisition of the kaaria, and one may also speak of a socially determined andropause for the Fathers of the country. This is because the effect of the set-based marriage rules is compounded by the late marriage of women, which, all other things being equal, reduces the number of women available to become second wives. The maintenance of an equal sex ratio, which might be threatened by the effect of selection whereby families with more boys are favored, appears to be ensured by the higher mortality rate for boys due to their participation in raids2. The model—or ideal, even—of polygyny certainly exists, but is difficult to implement because of factors which weigh in favor of monogamy, which is therefore encouraged by the system. There is a lack of statistical data on this subject, but the low occurrence of polygyny was conspicuous enough during the colonial era to fuel a rumor that monogamy was the result of the Hut and Poll Taxes imposed by the authorities, with the taxes seen to have discouraged Meru men from taking another wife or having more children (Kamunchulu 1975: 199). Taxes certainly had an effect in this area, as we shall see, but it was not the one assumed here.

16The idea of a specific period for reproduction is inseparable from the construction underlying the entire grade system whereby activities are divided up among society, which, more than the separation of individuals, involves the segmenting of life into a series of phases. Activities center on family life when individuals are at the “young father” and “young mother” stage and thus part of the gîtiba that is powerless at the time—a synchrony emphasized by the ear-piercing ritual being conducted at the same time as the grandparents become mwariki. The phase of political power, which men and women both experience, begins at the same time as the initiation of boys and, separately, of girls. The rebirth of their offspring within the system means that the parents progress to a new kind of paternity and maternity incompatible with the one that preceded it: one may not simultaneously conceive children within the family kraal and enable the rebirth of others on the public ground. In this phase of political power, natural fertility gives way to metaphorical fertility through the power of speech, which replaces it and assumes its role. This is part of a process that begins far in advance with the marking of the mouth and teeth and the transformation of the voice that is emphasized by the singing of songs in sequence. When they reach the following grade and become grandparents, men and women rediscover an activity and shared existence that mirrors their remarriage through the accomplishment ritual: they oversee the development of the grandchildren who have enabled their own rebirth, as well as that of the parents, their descendants, to whom the father, upon death, will hand down his supreme power to bless or curse—the highest form of the reproductive ability, where speech holds sway over life and death. In summary, there is a specific window for each kind of procreation; the first one is more specifically linked to the fertility rate studied by demographers, but together these different phases also raise the issue of life expectancy or its correlative, mortality. Unless we conclude that the Tigania-Igembe generation-set system is a complete fantasy—a view long held by ethnologists of these organizations, and one this monograph has sought precisely to dispel—we must assume that the means for the demographic reproduction of this structure lay in the Nyambene Hills.

17In conjunction with the rules designed to delimit and synchronize the reproductive period for women and men of the same set, certain notions and practices also help create a large gap—five years or more—between births, judged in terms of the child’s ability to accomplish particular tasks. Prolonged breastfeeding and late weaning are justified by the idea that children grow slowly, especially when the father has no herd to provide to milk, and that an additional child can endanger the previous one if (s) he is not yet old enough. The issue is that this reasoning, which seems logical enough, has been pushed to extremes by the Tigania-Igembe. There are several factors here. Unweaned children are associated with their life in the mother’s uterus, and the suspension of sexual activity between the parents is thus a corollary of the inter-generational avoidance rules. Just as sexual relations would be terminated after four months of pregnancy, they would remain so until the child was no longer attached to its mother through the activity of breastfeeding. Of course, women in this state often have amenorrhea and reduced fertility, but here it was the “smell of the child” which repulsed the father or husband wishing to approach his wife. Further, as we have seen, spouses had separate residences, an arrangement that meant the rule could be easily applied. In terms of the fertility rate, and accounting for the late marriage of women, this birth spacing meant each woman had around five children on average. In the event of a child’s death, sexual intercourse would resume with a bankiro after three months of mourning—the time considered necessary for the effects of the mûgirô responsible for the death to fade. Statistically speaking, two or three of these five children needed to reach adult age in order for the population to be replaced. There remains the issue of years when there was an excessive mortality rate. As there is no ethnographical indication that the avoidance rules were disregarded in any major way, we must assume that it was positive net migration, whether natural or forced, that enabled the Tigania-Igembe to survive periods where there was a risk of depopulation.

18Another dimension of the issue of low fertility rates is related to social values and the stigma of having numerous siblings. Rather than constituting transgression of an actual rule, frequent pregnancy was treated with criticism and scorn; it was seen to indicate a form of intemperance, or a concealed attack on the usual way of doing things. Having too many children was the sign of loose morals and animal-like behavior, of excessive sexual activity similar to the “hot” and ineffective sexuality of the Bushmen, or the confused vitality of female nkenye. The recourse to ritual coitus to remove mystical pollution is a telling sign of an attitude towards sexuality that even colonial administrators raised under Victorian morals found austere (Lambert 1952). The stigma attached to having numerous siblings also relates to the issue of links with the maternal side. At birth, the child belongs to the father and maternal uncle, and the establishing of patrilineality and paternal rights requires the offering of ndûgû payments, which, as we have seen, accumulate over the generations. This obligation has a dual effect on numbers of children. The ndûgû payments are costly in the long term, and while they cannot exactly be said to act as a deterrent, they certainly do not encourage additional births, since the man is producing children who belong equally to the wife’s brother and is thus in a sense acting in his favor. The implied content of the judgments— “too many children”, “loose morals”, and so on—can therefore be considered a condemnation of proliferation on the muntuetû’s side and at the father’s expense. True paternity is acquired through the initiation of boys, and, via marriage, that of girls; the preceding “virility” belongs to the maternal side. One cannot do without it, but one must keep one’s distance. In this context, we may re-examine the parallel so often drawn between the growth of the herd and the proliferation of humans, and the fact that they are seen to be part of the same cycle—that of the sun. For example, the milk of the cows and goats, given early to newborn infants, represents the paternal contribution to the child’s nourishment, without which infants are believed to grow too slowly. Once again, it is the father who causes growth. A large herd enables access to other wives; polygyny means that it is only fathers who may have a large number of children, since the rule of each woman having only a small number of children can then be maintained. The idea that a substantial herd means that more children can be nourished thus acquires even greater importance, but it also creates a paradoxical situation where the pastoral economy and gwato contracts lead to shortages, which, as I concluded in the chapter on the old economical system, turns the herd into an instrument of social division. At this point in our analysis, the Malthusian approach to herd management can also be seen as a way of maintaining low fertility rates and preventing the accumulation of children within families, since if such accumulation were to become widespread it would jeopardize the equilibrium and separation of the generation sets.

19Behind this condemnation of excess is also the idea that quantity can be detrimental to quality. Quantity and quality are in an inverse relationship and separated by a cognitive border that is of manifest importance, although ethnographical data is currently too scarce for us to provide anything more than an outline here. Life consists of a series of measured steps, and it is likely this idea is based on the notion of a form of energy that broadly corresponds the Kimeru term inya: the energy of Ngaï-Murungu, the supreme Principle, who rules the world and society, and allows man and woman to become Accomplished. A father explaining that he does not have enough energy to raise more children uses the same term, as if referring to a kind of energy beyond his control that exists in limited, exhaustible quantity. Ngaï, the supreme Principle, is a term closely related to nkaï, which indicates the testicles. In Kikuyu, hinya means sperm. When warriors went on a raid to kill enemies and acquire their nthukî (generation) name, they would obtain a form of energy at the expense—precisely—of the energy of enemies (maitha), categorized and killed as such for this purpose. The energy would allow them to become genuine fathers for as long as their generation was in this phase, as if the energy obtained in this way would eventually run out. C. Hallpike (1972: 19, 148-153), in his study of the ideas of the Konso in southern Ethiopia, whose generation-set system has many features in common with that of the Meru, and who similarly have relatively small families, describes an indigenous theory where men are born with a certain amount of “non-renewable” energy, which can be squandered through excessive sexual activity and must instead be channeled and put to good use. Stigmatization of proliferation and promotion of reserve are perhaps the Tigania-Igembe version of a comparable theory relating to both sexual economy and the rate at which individuals grow. In the myth of Mbugi, Ngaï sends death to remedy the excess of humans. People must avoid multiplying like ants, just as they must undertake things at the right time and grow at the right speed. This movement also depends on the contradictory actions of the sun and rain cycles. Excessive heat or rapid cooking can cause the ageing process to fail: observe the bodies of novices, which may be reduced to ash if they make certain gestures too early, or the blacksmith who dies because his ngerue is overcooked. Conversely, excessive cold or humidity symbolizes immobility, or time grinding to a halt; and without movement there can be no growth. Similarly, but in a less codified manner, and in the same way that plants cultivated at high and low altitudes grow at different rates, the Tigania, who live in the low-lying areas, are said to “ripen”—that is, age—slightly faster than their counterparts the Igembe, who are based in the cooler, more humid elevated areas, even though both are part of the same set system. Within the same set, the Michubu, for example, do appear to be younger in Igembe than in Tigania. In reality, the Tigania probably waited slightly longer before incorporating their descendants into sets, and this delay has accumulated from generation to generation, but even this rational emic explanation is interesting in that it highlights the contrasting effects of warmth and humidity on human growth.

20One last element of Tigania-Igembe demology lies in their appreciation of population size, and their ability to adapt when situations became too critical. The buildings which accommodated the age divisions, as well as their actual construction, provided the means for a rudimentary sort of census, revealing an ability and concern for estimating population numbers. The extinction of a people was also not an unfamiliar concept, as indicated by oral traditions recounting the situation of the Mûciî Mûkûrû people, or the rumors about the Chuka of Mount Kenya. Variations in the recruitment rules allowed certain changes to be made in order to repopulate the sets, but this was more for the purposes of preserving the system than for modifying it extensively or abandoning it. In this context, the almost entirely insular situation of the Tigania-Igembe becomes an important explanatory factor. Migrants arriving in search of food or women and children brought at lance-point were in both cases rapidly “Meru-ized”, and enabled the survival of the institutions. In this respect, the dual division between “insiders” and “outsiders”, as well as their external counterpart, the maitha-enemies, provided a way of rationalizing the necessity of turning to the outside world in order to compensate for shortages on the inside, and thus preserve the society.

3. New challenges, new practices

21In the light of this demological analysis, the various conflicts which gripped the population during the colonial era are more easily understood. The increase in abortions and infanticide and widespread unrest that are so well attested in the administration’s archives indicate the extent to which colonial conquest had undermined the most basic foundations and workings of Tigania-Igembe polity. The change of colonial policy instigated by Lambert was based on an analysis of the problems that was more incomplete than inaccurate: by explaining that it was necessary to lower the initiation age for boys and girls and implement a natalist policy, the administrator had certainly identified the problem but did not realize, or pretended not to realize, the complexities of the issue or the possible consequences of such a change. The change took place, directly triggering a shift in the demographic regime but the way in which the situation developed sheds new light on a process that is not normally so easily observed.

22In tandem with the conflict between healing associations and the njuri ncheke, women were protesting against the early initiation of the children, in particular that of their daughters, whom they were more able to protect. The generation-set system as political authority had become defunct, but the set-based matrimonial rules, the alliance rules that included the initiation of girls, and the domestic rules which maintained the distance between consecutive generations, continued to be applied nonetheless; they helped control the overlapping of generations and prolonged the statistical existence of the sets. Such a situation encourages comparative analysis, since it indicates that as long as the customs are upheld on the domestic level, the sets continue statistically to exist, and therefore to be sociologically relevant. The position of women and the organization of their councils had been less directly affected by colonial conquest. Certainly, the increase in abortions and infanticide is indicative of the disruption of normal behavior. But the initiation and marriage of girls—activities organized by women—continued as before, just like the forming of the sets of wives and the handing down of kaaria. Women remained attached to the hierarchy of successive generations, to the late marriages and small families that formed the basis for their power and around which the domestic economy had been constructed. The early initiation of girls promoted by the colonial administration called this entire structure into question. It entailed the overlapping of generations and the impossibility of handing the kaaria down at the correct time; with the sets of wives no longer clearly distinguished, the hierarchy of sets that had been the basis for the distribution of power disappeared, leading to a breakdown of the councils and powers of women. In terms of the family, the early initiation of girls, and of boys, led to the problem of the parents’ reproductive activity coming to an end, and related effects in terms of the father’s polygny and the resources of the domestic unit; it meant the early departure of girls from the home, since initiation is followed by marriage into another family, and thus the loss of a valuable helping hand in domestic work. The acceleration and shortening of rituals also disrupted the construction of marriage, whose long procedures could no longer be followed, beginning with the “tour” made by nkenye or non-initiated boys; this too meant that women lost some of their social rank and status, since the bonds passing through them were no longer consolidated, or even completely disregarded and thus likely to vanish entirely.

23In this context, the nature of the conflict with the njuri becomes clearer. This new form of council, which as we have seen eventually replaced the defunct kiama of Fathers, supported the initiation reform project and was an ardent proponent of the natalist policy implemented by Lambert. On the pretext that part of the ritual took place at night, the njuri banned kaaria and the process of handing it down, saying that the women were practicing witchcraft—a particularly serious accusation, and one from which it was difficult to recover. They also had the ritual condemned on the grounds of the unnecessary expense and waste that this “female witchcraft” imposed on the family economy. It is within this context that the ban on clitoridectomy should be considered. Earlier initiation had been imposed by the government. The girls were identified and then assembled on the public ground; a female circumciser appointed by the administration conducted the excision on the same day. Female elderly who describe these events explain that the girls gathered in this way did not know how to act or what to sing, mixed up and hastily initiated as they were. To compensate for this shift, which was unacceptable in their eyes, the parents who had been unable to prevent the operation on their daughter instead kept their ngutu in reclusion for longer than custom dictated. The ban on clitoridectomy to some extent defused this resistance, which had been entirely undermining the attempt to lower the age of marriage. Some claimed this provided the njuri with an easy way of imposing fines because it was impossible to apply such a ban in practice (Kinyua 1970). In reality, the ban on clitoridectomy followed that on kaaria and this time dealt a fatal blow to the construction of family relationships, and to the influence of affines and that of maternal kin. After the disappearance of their councils, women watched their position within the web of family relationships, which was itself coming undone, evaporate. The ntanîthio payment to enable initiation was to cease, as was the phase where the “wife-takers” came to bring the girl out of her reclusion hut and lead her to her new family kraal, and the related ndûgû payments. Payment of the rûracio continued, and the njuri made efforts to restrict it to a low sum; officially it remained at this low level, but in reality it increased because of the rising price, in equivalent heads of small cattle, of the standard bovine unit of one cow. Those in possession of herds, primarily men of the njuri, found it easier to fulfil their desire for polygyny now that this was no longer hampered by alliance obligations or the prospect of lengthy debt towards the maternal side. Within this upheaval, which had been triggered by the lowering of the age for the initiation of girls, as well as that of boys, women suffered a brutal loss of position, power and identity. Their initiatory grades were to disappear while those of men were modified and became relatively more important. At the same time as their sets system was collapsing, the alliance rules were vanishing and the filiation system was changing so as to promote patrilineality and benefit male descendants alone.

24Women of a certain age are fully aware of this turbulent history, which they summarize by saying they had been very active, “hot”, in bygone days, and that now that is all over, “the country has turned cold”—an expression which revives the old connection between the sun cycle and the exercise of power. By means of the sets—soon to become mere categories—they outline, in their own way, the sequence of events; this chronology overlaps with and is complemented by each woman’s own life story and reproductive “career”. For them, the ban on the transmission of kaaria was the most decisive event. The ban was introduced early in some territorial divisions, and late in others. Substitute versions were developed in secrecy; these related to the right to display one’s kaaria on the ground where boys are initiated, and the organization of clitoridectomy, which when it does still occur is conducted on girls of a much younger age. The Ithalie wives were the last to receive the kaaria (from the Kiramunya wives) and to have handed it down roughly according to the rules. The majority of the Mûkubu did receive it, but often in secrecy, and meanwhile sessions for the initiation of boys were becoming more and more frequent; this course of action constituted a fatal blow to the rules and idea of the deferred exchange of goods, and to related offerings. A greater and greater number of Ratanya wives refused to pay to receive the kaaria; they may have been yielding to the pressure of their husbands and the njuri, but they had also realized these customs were out of step with the expanding market economy. In refusing to pay, they broke the cycle of exchange and failed to reimburse an outstanding debt, which the female elderly consider the most tangible sign that the old world has vanished. With the Lubeeta wives, there is a new choice to be made between the kaaria and the karamu, i.e. the kaaria or the “pencil” (schooling). Over the last few years, the conversion of old women en masse is a sign that the Churches, in the widest sense of the term, have now replaced an institution that, in any case, had been reduced to merely surviving in secrecy and conferring an identity that held fewer and fewer benefits.

25Changes in fertility practices followed roughly the same pattern. Polygyny became a reality for the Ithalie and Michubu, although it remained at a moderate level—around three wives—and was only practiced by the richest and most important individuals. Five wives is the rarely attained maximum limit, beyond which it is thought that the husband will only neglect his wives3. The Ithalie and Michubu wives, roughly speaking, respected the traditional pattern of birth spacing. The major change came with the Ratanya wives, in the decade following the Second World War. Girls and boys were initiated far earlier, the reclusion period was shortened, and the young married as a consequence; couples formed in this way stopped adhering to the pattern of long gaps between births, and from this generation on the number of siblings began to increase (fig. 34). The number of family kraals increased, as did their size; overt polygyny flourished; husbands and wives began sleeping under the same roof4. The change in customs was rather sudden: it affected the generation set consisting of the children of the Ithalie, who had already begun these changes in their own day and ensured their continuation by initiating their children at a younger age. The movement gathered momentum with the Lubeeta generation, and was maintained with the Miriti. In the 1980s, the Miriti faced two related issues that had returned in a new form: teenage pregnancy, and marriages again being deferred due to the difficulty of obtaining a plot of land or finding work. But this is another story, yet to be investigated and written.

Fig. 34 — The offspring of a couple from the Ratanya set (documented 26th March, 1993).

26Other factors played a simultaneous role, leading to changes in traditional birth control and an increase in the number of children per woman. The fertility rate changed as a function of the increase in agricultural production, itself the result of a complex and long-term dynamic process occurring between 1920 and 1960 or beyond. Firstly, the obligation to pay the taxes imposed by the colonial authorities clearly made it necessary to produce more of a given commodity in order to sell it for cash, but this process did not become well established in the Nyambene Hills for as long as the Tigania-Igembe continued to use cultivating sticks and their ancestral farming techniques. The colonial administration was faced with problematic levels of livestock overpopulation, with overgrazing common, particularly in northern Tigania. In the 1930s, the Tigania, closely followed by the Igembe, finally yielded to the administration’s demands and adopted the hoe and machete. Realizing that these tools provided an advantage, men decided to take on this work—a significant development, since their job was to prepare and turn over the soil. More productive crop varieties proposed by the administration also came into use. There was an increase in farming as a result, although this necessitated double weeding—traditionally a feminine task—without which any extra sowing of land was pointless. Soon there was a shortage of manpower, and the advantages of having a greater number of children, which it would still be possible to feed, became clear. Changes to farming techniques therefore had a “populating effect”, increasing the population density—as with the pattern observed with irrigation rice farming in Asian societies, first identified by geographers and now considered a case study5.

27Migratory habits changed over the same period, and are also worth examining since they affected the entire population dynamic. The days of raids, and captives from raids, were over, but some migration flows persisted; in particular, the Kamba continued to seek refuge in the Nyambene Hills in times of drought, and eventually settled there. Separation from the people of the Northern Frontier and controls="true" on movement meant that the number of refugees arriving from those regions dwindled. The 1930s saw the arrival of the Kikuyu, who had found the patch of land granted them by the colonial administration too small, and with whom the Meru had always been on friendly terms. However, things quickly turned sour when the Kikuyu refused to “Meru-ize” themselves, i.e. to be adopted through “rebirth in a family kraal” like the Kamba, who had done so and in return been granted plots of land. There are at least two explanations for the attitude of the Kikuyu: firstly, the postpubescent initiation of Meru girls is strongly against the customs and values of the Kikuyu, who try to initiate their girls just before the beginning of menstruation; secondly, the low amount paid as rûracio by the Meru, especially relative to the bridewealth paid by the Kikuyu (which moreover had skyrocketed during these years), led the Kikuyu to believe the Meru were acting as if no marriage had taken place. The Kikuyu initiation process for boys was similar to that of the Meru and did not represent an obstacle to friendly relations between the two populations; on the other hand, differences in the status of girls, which pointed to marked differences in social structure, explain why the Kikuyu found it difficult to become Meru. Of course, the Meru did not accept such an attitude. Incidents soon occurred, particularly in the south of the District, leading to rampant conflict and new competition for territory. The short- and long-term consequences of the “landconsciousness” mentioned repeatedly in the administrative archives have already been described: the creation in 1956 of a Meru Land Unit distinct from the Kikuyu Land Unit and, in the same year, the ban on clitoridectomy; the clashes between the njuri and the Mau Mau regarding the notion of independent territory; conflict over the choice between community control and individual ownership of land—all these elements reveal the vectors of interest and issues still in play today.

28The combined effect of these different factors was ultimately to promote an increase in the fertility rate. The adoption of the hoe, which enabled an increase in the amount of land cultivated as long as weeding still took place, occurred just when competition for land was becoming an issue. Land became a means of subsistence and source of income like never before. Before an audience of njuri elders willing to listen to him, Lambert gave natalist speeches, linking birth rate and territory. The stigmatization of female solidarity and the banning of women’s institutions enabled the emergence of the family unit as economic unit and the future focus of activity and interest. As the traditions of the generation-set system began to fade, along with the links of kinship and affinity and the role of the maternal side, inter-family rivalry began to increase, highlighting the issue of land reforms. The decision to allocate individual title deeds (to men alone), made at the time of independence in order to spur the creation of a small- and medium-scale farming class that would protect its own interests, gave added momentum to a movement that had already begun, and is cited by some elderly who claim that land reform is responsible for the land shortage. They assert that communal management meant everyone had access to a plot of land; reform (which is still incomplete in these regions) accelerated the race for land, and, we might add, the subjection of women to men.

29The change in the hierarchy of power, as well as concurrent changes in the definition of the individual, gender- and generation-based relationships, and the nature of interpersonal bonds, are inseparable from the transformation of values and concepts that underlies the varied events of this period. Beginning in the 1950s, both the accelerated spread of the market economy that triggered, in this District, a somewhat delayed recognition of the advantages of schooling, and the accompanying spread of Christianity, were associated with more widely recognized changes that are often described or referenced ad nauseam. For this population, there is (was) a real desire to conform to a form of modernity that is no less appealing for being embodied by the whites people. However, the idea of the prolific white man, coming from afar in great number—remember, Kenya was a settlement colony—is a theme of modernity that has rarely been recognized; and yet it seems to have been one of the most intriguing for the Meru, to the point of becoming almost a reference model6.

4. Navigating through numbers

30The transformation or disruption of fertility-related rules and behavior has been investigated through ethnographic and historical research. The archives and censuses contain quantitative indications that allow us to analyze the way in which the population has changed and assess its size (fig. 35). Contrary to what one might expect, the oldest estimate, dated 1911-1912, in fact appears to be more accurate than subsequent ones. It is based on a survey made in the preceding years with the aim of collecting taxes. (The Tigania-Igembe, in particular, seem to have been unaware at this point of the burden such taxes would represent.) According to Fadiman (1993: 175-177), Horne had enlisted a significant number of “huts counters”, who worked under the aegis of the King African Rifles. The Meru District was described as being densely populated and with strong agricultural potential, which made the administrators doubly optimistic. It was easy to distinguish between the Tigania, Igembe and other groups in the District, since indigenous people were aware of and recognized these identities. The Tigania appear to have been more numerous than the Igembe, which corresponds with the fact that this part of the Nyambene Hills, traditionally more appealing than the cold, wet uplands of the Igembe, attracted a greater number of settlers. However, both major territorial sections were well populated, with a total of around 90,000 inhabitants—all governed by a single and unique generation-set system. This large figure is certainly conceivable: in a less favorable ecological setting, the generation sets of the Karimojong ethnic group numbered 60,000 individuals in the 1950s (Dyson-Hudson 1963). The figure of 90,000 reflects the particular unifying power of this form of polity, here uniting a population greater in size than many chiefdoms or kingdoms. This population size also causes one to wonder about the conditions for reproduction within generation-set systems, which are wider in scope than micro-populations; the low fertility rates caused by the system are also explained, since their potentially negative effect on reproduction is mitigated by the large number of inhabitants, and can be more easily compensated for. The corollary of this situation is that any polity based on generation sets will cease to be viable below a certain population threshold, which can only be assumed to exist and should be somehow estimated: after this point, it will be difficult for the group to survive, as it enters a pattern of depopulation that may be explained by the traditions we have studied here. These large population sizes also point to the ancient nature of this demographic regime and its related institutions, as well as what must have been the long-term inhabitation of the highlands.

Fig. 35 — The Meru District and the Tigania-Igembe in numbers, and with a certain margin of error.

31The 1920-1921 population count, which the administrators claimed was as accurate as was possible, revealed a total of 112,000 inhabitants in the Meru District whereas previous estimates, even allowing for the effects of epidemics and famines thought to have killed between a quarter and a third of the population, had indicated between 140,000 and 160,000 inhabitants (dc/mru. 1/1. 1920- 1921). The figure of 112,000 inhabitants is likely an underestimate: tax pressure had become a reality, and the way the Meru inhabited the area—spread throughout the forest cover, with a portion of the population living in gaaru—made it tempting and easy not to declare all the members of a family. From 1920-1921, the figures increase steadily; we must be aware of the respective roles played by improvements in counting techniques on one hand, and actual population growth on the other. The demometric and fiscal disappointments of the previous decade meant the commissioners were cautious. In his 1928 report, the commissioner notes: “The increase is partly due to the fact that a number of the old subheadmen have been replaced by younger men who have shown the huts counters a number of people who have not been previously enumerated. These figures however can only be approximately correct as it is probable that many people have still been able to evade inclusion on the Census list.” Similar reservations were expressed for the figures given for 1939 and, in 1948, the commissioner declared the time had come to conduct a genuine census. This took place a little later that year. The difference between estimate and census in 1948 matches the margin of error the commissioner had previously identified, demonstrating that over time the administration’s understanding of the issue had improved7. During this period, there was a switch in the relative sizes of the Tigania and Igembe populations. The 1948 census clearly shows the Igembe to be more numerous than the Tigania, a trend that previous counts had indicated but underestimated, due to the usual reasons of the population being spread out and hidden in the forest cover. The Tigania, less isolated than the Igembe in their uplands, perhaps suffered more as a result of the excessive mortality rate of 1918-1920 and malaria epidemics reported in the archives. Also, we cannot entirely exclude the effect of the administrative readjustments made to the boundaries of the two major territorial sections. These years also saw a reversal of age-old tradition, in that areas at higher altitude, which until then had been considered as of little value, became appealing because it was possible to grow coffee, tea and mirra in them; these unoccupied or underused areas of land received an influx of settlers arriving from the lowlands, where livestock farming was becoming less lucrative by comparison. Finally, from 1948, the population’s reaction to the questions in the census begins to vary from that of previous counts. The 1948 census took place with the support of the njuri elders, and it is not unrealistic to suppose that their participation strengthened their natalist convictions. With “landconsciousness” in full swing, population counting by the administration’s officers became a means of gauging the actual occupation of land—to be referred to as necessary. Another sign of such changes in attitude: whereas the burying of the dead had triggered protests and demonstrations when first enforced in the previous decade, it was now used in disputes as a way of proving one’s lengthy occupation of a piece of land.

32Having examined the issues of underestimation and change in the relative sizes of the Tigania and Igembe populations, we may now attempt to analyze the trends suggested by these figures. 1939 saw a return to 1911-1912 levels. During this period, it was as though the losses incurred during the 1918-1920 pandemic had been compensated for. What were the conditions for this? We have seen that, quantitatively speaking, net migration is a neutral factor. Health conditions improved, as demonstrated by the elimination of yaws. This non-venereal form of treponematosis, spread simply through contact, leads to skin problems and bone deformation, and mainly affects children between the ages of two and ten (Gentilini 1993: 342-348; Kiple 1993: 1025-1035). The elimination of this disease was a likely cause of the reduction in mortality for this age group. Further, the spates of death caused by drought and poor harvests were mitigated by measures taken by the administration (food aid, storage aid, anti-locust campaigns). The 1948 census uses gender and age categories, with boys over sixteen and girls over fourteen counted as adults, i.e. those born, generally speaking, after 1932. This corresponds primarily, although obviously not exclusively, to the offspring of the Ithalie, who comprised the future Ratanya. As the Michubu were initiated, beginning around 1930, their Kiramunya fathers gradually stopped having children. This rudimentary age pyramid still reflects the Tigania-Igembe generation-set pyramid. This analysis does not distinguish the age of those counted; the enumerators had certainly been instructed to make use of the age sets but, given the complexity of the system, were only able to make approximations.

1948 census

33Among both the Tigania and the Igembe, there is a higher number of females than males. This is due to the temporary absence of the men who left to work outside the Nyambene hills or the District: the obligation to pay tax was accompanied by a desire to earn money and take better advantage of the monetary economy, as well as the inclination simply to travel and see the outside world. The “youth”, i.e. those under fourteen or sixteen, represent 45 % of the Tigania and 46 % of the Igembe, without taking into account the adult men who had left to work elsewhere and whose inclusion in the calculation would lower these percentages, or any error in the assessment of ages by the enumerators. These percentages, while approximate, correspond to an age pyramid with a wide base, and suggest that the population, having compensated for the excessive mortality of 1919-1920, then grew even further. Here we can detect the effect of the changes in traditional birth control initiated by the Ithalie set. From the 1940s on, the population grew remarkably: roughly speaking, it doubled from 1948 to 1969, and again over the following twenty years. Within this dramatic increase, it is necessary to distinguish the following factors: the fall in mortality rates; the role of the birth rate, which depends on the number of women of childbearing age; and that of the fertility rate, i.e. the number of children per woman of childbearing age, which is directly affected by sociocultural change or upheaval. However crude these figures may be, they do confirm the idea of a “baby boom” indicated by qualitative investigation of the generation sets and the changes taking place within them. M. H. Dawson’s separate study (1987) of the effect of the fall in the mortality rate on the Kikuyu population dynamic in the first half of the century came to similar conclusions: after tracing the history of diseases in the population, Dawson concludes that the increased mortality rate caused by colonial conquest and the various new diseases that came with it was followed by a stabilization of the mortality rate and then an increase in the fertility rate linked to the lowering of the marriage age and changes to the Kikuyu initiation process. Following a critical examination of certain assumptions of the “demographic transition model”—in particular, that in the initial phase of the transition a high fertility rate compensates for a high mortality rate—Dawson outlines a process not dissimilar to that of the Tigania-Igembe. This kind of increase in population occurred across the whole of Kenya after the Second World War (Chesnais 1979), and it would be worth determining the extent to which the model identified for the Nyambene Hills applies for other groups. This could only be achieved through comparative studies, which would enable a better differentiation of the various factors involved.

34A proper study of micro-demography and historical demography would be beyond our scope here. Because I conducted my field research alone (except my research assistant) —a prerequisite for any decent immersion, but one that also limits certain kinds of investigation—I had neither the time nor the means to carry out sufficiently accurate or large-scale measurements of the standard required for demometry. In addition to the scattering of the population mentioned above, informants also show little interest in their family connections or genealogy. The exception is when they invent a genealogy to meet the requirements of the land reforms, according to which land is owned by men only and depends on agnatic kinship. My demo-historical enquiries ended up centering on people who had been more prolific, the others having been forgotten or having not had any children, which creates a frustrating bias, at least in ethnographical terms. The oldest women always claim to have had very few children. They were more likely to marry later, and therefore to undergo abortions that would prejudice their “reproductive career”, whereas younger women were able to marry earlier and have more children. The number of children declared, i.e. those who survive until they marry, increases from one generation to the next, and there is a clear contrast between the wives from older sets and those from more recent ones. The hundred or so interviews carried out on this subject provided information and numerical data that informed the research, but cannot be said to constitute a meaningful statistical survey. Obtaining information from the censuses is not a straightforward task, since counting methods and territorial boundaries vary from one census to the next. For example, it is difficult to establish an exact figure for the increase in inhabitants per square kilometer (i.e. population density) for a single “location”, although this information would be of great benefit to ethnographical, geographical and demographical study. Faced with such obstacles, I preferred to concentrate my efforts on spoken testimony from my informants, and the task of tracing the population processes that occurred during the colonial era. Fewer and fewer people survive from these bygone days. They represent vital banks of information for our understanding of the generation-set system—the functioning of which raises a number of questions, involving issues and concepts that are of interest to ethnologists and demographers alike.

35The case of the Meru is neither unique nor completely unprecedented. A collective study directed by A. Molnos (1973) with the aim of amassing cultural data in order to aid family planning in eastern Africa, provides a succinct but explicit summary of such practices among the Mount Kenya Meru (Mwambia 1973, Njeru 1973). Various publications dealing with eastern African populations also indicate low fertility rates—for example, C. R. Hallpike’s 1972 monograph on the Konso of Ethiopia, or F. H. Stewart’s 1977 study of the inherent logic behind age-based social organizations, in particular their demographical system and related rules, based on information gathered among the Rendille, Borana, and others. For lack of sufficient data, but also because of the barriers between disciplines, no author has yet been able to go as far as establishing a direct link between low fertility and the fabric of a generation-set system. The case of the Tigania-Igembe was of sufficient interest that I was encouraged to cross this threshold and construct a comparative model. However important the ecological setting and geographical position of the Nyambene Hills in balancing the system, this should not lead one to believe that, “in the final analysis”, ecology and geography alone can explain this remarkable demographic regime and its associated social structure. A connection between the case of the Tigania-Igembe and that of other populations in this part of Africa with systems of age- and generation-sets, for whom slavery was only a marginal factor, formed the basis for a comparison of these institutions, through analysis of their responses to the problem of overageing and underageing (Peatrik 1994, 1995a). This is a problem that will always reoccur, but the measures taken to reduce the number of overaged and underaged people can be classified under two main forms of generation-set organization, already mentioned with reference to kinship and social structure. The Karimojong system, with two political positions and long intervals between the handover of power, makes allowances for a certain degree of age spread within each generation, a relatively high number of children, and incorporates the practice of polygyny. On the other hand, the Tigania-Igembe or Borana system, with sets and fractions divided into grades, regulates the timing of reproduction to the extent that low fertility rates and monogamy are almost the norm. The first system is used by expanding populations, and the second by populations settled in a particular territory. Just as demographers distinguish populations of constant size as either “stable”, “stationary”, or “Malthusian”, so populations with generation-based fractions must combine their principles of social and economic organization in different ways in order to maintain the numerical balance of their sets. The Meru system represents an example which will help us to understand other such systems, and demonstrates the importance of taking all elements of social organization into account. Separate studies of populations with age-based sets located far from the Meru reveal a similar situation, where the fertility rate is regulated by the political system (Désveaux 1996). The concept of a demographic regime, implying an intricate and lasting connection between social rules and demographic parameters, here finds a new area of application, and is expanded to include issues and values connected to the ideal size of a population.