My ethnographic fieldwork in Africa was among the Kipsigis of the western highlands of Kenya. My research focused on household and community structure, male initiations, and ethnic identity. More recently I've become interested in Gregory Bateson's work connecting systems theory to the social sciences, Claude Shannon's mathematical theory of communication, Walter Ong's work on the psychodynamics of orality, etc,

Today I want to discuss some of African material and make some comments about the limitations of ethnography. Specifically I want to discuss the evolution, of ethnic identity among the Kipsigis and related groups, from the precolonial period to the early years of independence.

In writing a description of this development I have had to deal with three main problems which I will present in historical order (typically I mulled each over several times before it dawned on me that there was a common thread in all of them).

The first problem was to come to an understanding of the bewildering array of labels and misnomers for the variously named territorial groups, as they existed prior to colonization, that spoke dialects of the same language as the Kipsigis. For the time being I will refer to this language as Language X, and the set of peoples speaking X as Cluster X. The first problem then was basically a "who was who" question. It might seem that this task was simply deciding upon a list of the societies, cultures, or tribes which comprised Cluster X, but those terms do more harm than good in the early stages of the problem. Obviously in order to understand terms such as Kipsigis it is necessary to come to an appreciation of their referents, that is the nature of precolonial entities, and thus some of the basic principles of precolonial social dynamics on group levels.

The second problem arose in dealing with data from the colonial era -- that is to say most of the available data. Here the infamous word tribe is unavoidable since it was a key concept in colonial policy and the organization of colonial administration, and thus had a profound effect on the experiences and self-conception of the peoples of Cluster X and all other Kenyans. Basically, then, it became necessary to consider both the extent to which the dynamics of the colonial system interacted with indigenous modes of organization to shape and change ethnic awareness, and the extent to which the colonial mindset, or at least its organizational concepts, structured anthropological descriptions of Kenya, and at times obscured our understanding of the dialectical changes taking place.

Finally, the third problem, perhaps the easiest, had to do with how a self-consciously coined term for Cluster X and its associated concept of wider ethnic unity, which was promulgated by the first few educated men of these groups for what were basically political motives in the late colonial and proto-independent national contexts, was rapidly and universally embraced by the rural population, most of whom had very little if any understanding of the nature of modern politics.

Well, those are the questions. My contention is that in each case (understanding indigenous groupings, assessing colonial impact on ethnic identities, and explaining the rapidity of recent developments) the key is to focus on the media of communication, both native and introduced, and the uses to which they have been put. To put it simply,the nature of social dynamics is related to the channels of information flow in use. On another level I want to suggest that within academia (or at least anthropology) we are burdened with our own modes of communication, particularly the printed word, which can be misguiding in the analysis of preliterate and postliterate organization.

First, a few background facts:

Both oral traditions and linguistic evidence make it clear that populations ancestral to Language X speakers, as well as more distantly related groups now found in Tanzania, entered Kenya from the north. By 1000 AD Cluster X can be placed with some certainty around Mount Elgon and the Cherangani Hills. Mount Elgon is mentioned in the oral histories of each of the present groups as a place of dispersal. Glottochronology suggests that the Pokot dialect diverged from the others soon after 1000 AD. The Pokot remain in the northernmost of this cluster while the others expanded southward in the highlands and down into the floor of the Rift Valley.

During the 17th and 18th centuries various groups of Maasai pastoralists swept down the Rift Valley absorbing some Cluster X people and displacing many more from the plains of the valley floor back into the western highlands. At the height of their power Maasai groups together occupied a stretch of land 500 miles from the north to south and 150 miles from east to west at its widest. The Maasai expansion caused major reorganizations among surrounding populations. It is difficult, if not impossible, therefore, to trace the emergence of the various named subgroups of Cluster X back before the 17th or 18th century.

(refer to map)

The main groups in the cluster are the Sebei who ring the slopes of Mount Elgon, to their southeast the Nandi and an offshoot of them, the Terik, to the south the Kipsigis, to the east the Tugen, and along the slopes of the Elgeyo Escarpment, the Keyo and Marakwet. Finally, spreading down on the lower ground to the north are the Pokot. Additionally there are several bands of hunters and gatherers living in the high forest regions. They are known collectively as the Okiek (or in the older literature, the Dorobo). There are also a few tantalizing references to similar hunting bands, the Athi, who were isolated in the forests to the east of the Rift Valley by the Maasai expansion and survived to the beginnings of the historic period before being absorbed by the Kikuyu.

The basic ecological adaptation of Cluster X peoples, aside from Okiek hunters, involved a mixed economy of hoe cultivation of indigenous African millets and animal husbandry with much more emphasis placed on cattle than sheep and goats. All of the groups practiced the same general variety of pastoralism, the institutions and values of "the East African cattle-complex", "house-property complex" form. Unlike most pastoralists, however, Cluster X groups controlled land that ranged from grasslands at 5000 feet elevation to high forests over 8000 feet. Communities consisting of dozens of dispersed but closely adjacent homesteads occupied the higher ground where volcanic soils favored agriculture and defense was easier. Lower, open land was used for grazing the majority of cattle under the protection of young men, or "warriors," armed with iron spears and swords. Beyond that there was often a no-man's-land which had to be crossed before encountering culturally different people.

If ever there was a situation in which we might be tempted to apply the term tribe it was surely East Africa in 1900. But the reason it seemed easy to list the groups in Cluster X and locate them on the map is that I cheated -- I gave you, not indigenous reality, but the colonially recognized "tribes" and districts. Aidan Southall has discussed the "illusion of tribe" created in the colonial period. The problem, however, is that some of the groupings in Cluster X are less illusions than others.

Let us start our closer look with the people who ring Mount Elgon. They recognize six main segments: the Sabiny, Mbai, and Sor, now in Uganda and collectively called the Sebei, and on the Kenya side, the Bok, Bong'om, and Kony (called by the Maasai Elgoni, and erroneously by the British, the Elgon Maasai). In Kenya these groups are now collectively labeled the Sabaot.

I have yet to find a reference indicating that the various Sebei/Sabaot fought each other, though Goldschmidt mentions "mutual distrust and prejudice" between at least two segments (1976:63). On the other hand there is no evidence that they ever acted as a single polity. Goldschmidt comments (1976:63):

Note what Goldschmidt assumes ought to be (and for that matter what he assumes about the relationship between native social entities and what gets in his notes).

Just one other example of enigmatic ethnicity among the Sebei: Goldschmidt mentions in passing descendants of a small group of Bantu-speaking Bagwere, now called Bumachek who

The Kipsigis and the Nandi are the two largest groups in Cluster X, controlling the largest most coherent land masses. Their dialects are also the most closely related. The Kipsigis recognize three main territorially defined segments, though these distinctions seem not to be as salient as those among the Sebei/Sabaot. The people of the southern section of the Kipsigis, the Sotik, appear in many of the early colonial records as a separate tribe, no doubt because many people identified themselves as Sotik rather than Kipsigis when first contacted. There are minor dialect differences between areas within Kipsigisland, too subtle for me to catch (the Sotik for example are more influenced by Okiek bands in the mountains to the east and by large numbers of Bantu-speaking Gusii absorbed during their southward expansion in the 18th and early 19th centuries). Yet there is no question that the Kipsigis have always recognized their unity. Anywhere in Kipsigisland is "home" for any Kipsigis. While most raiding against enemy groups was small scale and locally organized, there were major campaigns in the late 19th century which combined many hundred men from Sotik and the other areas. In both internal affairs and military operations the Kipsigis were clearly one polity -- if that word can be used for a society which did not have centralized leadership and in which events which mobilized more than a local area were rare.

The Kipsigis and the Nandi occupied nearly adjacent highlands prior to colonization, people freely moved back and forth and, where they were in proximity to each other, intermarried, and jointly initiated their children. Indeed the Kipsigis and Nandi recognized one jural community between them. Being in close proximity only to the Nandi, among the groups in Cluster X, and having this "special relationship" with them, the Kipsigis never fought others of their language. But in this they were unique. The Nandi conducted raids against the Sabaot/Sebei as well as Cluster X groups to the northeast, the Keyo and Marakwet.

When one turns to the peoples living along the Elgeyo Escarpment, the "Keyo" and "Marakwet," the existence of a precolonial tribal entity becomes highly problematic. Here a number of local segments, or sub-tribes, or locations are strung along a narrow strip of highlands. The District Officer of Elgeyo, speaking in 1932, described the situation as follows:

Benjamin Kipkorir wrote a brief monograph on the Marakwet (started when he was still an undergraduate in Uganda, and revised while he was a doctoral student at Cambridge; he was later director of the Institute of African Studies at the University of Nairobi). This, then, is how a Marakwet opens his description of his own society:

Concerning precolonial unity Kipkorir states that:

There is little published ethnography on the Tugen, though a few dissertations were done in the 70s. Early colonial references describe a similar focus, not on "tribal-wide" organization, but on the complexities of history of small local groups moving from place to place and drifting from "tribe" to "tribe."

Glottochronologists have attempted to work out the various divergences of the main dialects in Cluster X from a single proto-language. No doubt this is a useful exercise, though, of course, no unique single dialect ever existed.

When one considers the social origins of the groups as reported in oral history, to which early colonial officers gave a fair amount of attention, the picture which emerges is that each group coalesced out of emigrants, refugees, or captives from all the other groups. For that matter there are also people in each group who trace their origins to one or another of all the other linguistic groups nearby (with the exception that no Kipsigis I ever met knew anyone who claims any connection with the Luo. Not surprisingly, the same clans, or descent categories, are often found in several groups. Characteristically, representatives of different clans observe minor differences in ritual behavior, some of it traceable to practices in non-X groups, such that a full description of a ceremony, for example a wedding, as practiced in a local area would have to encompass as many variants as one might find among the various Christian denominations in a town like Chapel Hill. It is not discrete boundaries, or grandfather clauses, or diagnostic traits which define these groups, but participation in a shared information process.

Beyond the dense network of kinship, these societies were held together by the cyclical age-set system and the statuses associated with membership in it. Age-set identity was a primary means of placing men from one group into the social fabric of another. Of course, age-sets also served as a means of recruiting and motivating young men into hostilities between groups. Having read just about everything available that mentions age-sets among Cluster X groups, I have become obsessed with three problems. First, it appears from fragmentary evidence that the Keyo, Marakwet and Tugen have eight age-sets while it is well established that the Nandi, Kipsigis, and Sabaot long ago dropped one out, and now have seven. Secondly, Peristiany reports that among the lowland Pokot, who unlike their fellows in the hills are much influenced by their neighbors, friends, and enemies, the Turkana, initiation into the age-set system is optional and may or may not be combined by some men with membership in the independently organized Turkana age-sets. Here the key institution which everywhere else in Cluster X is the sina qua non of ethnic identity, is one of the essential dimensions of all adult male interactions, and is, in large measure, what defines Cluster X as a wider social entity, becomes a matter of personal choice. Anyone attempting to get a typologically oriented grasp on indigenous social entities will find these data slipping right through his fingers.

The third problem concerning age-sets is the most puzzling. To make a long story short, it appears from a variety of rather equivocal sources, that the same age-set, Nyongi, was being initiated around the start of the century throughout Cluster X. I thought I was making progress when I understood how initiations were coordinated among 100,000 (or today 600,000) Kipsigis, but the possibility of some wider synchronization of age-sets has radical implications for our understanding of Cluster X, and perhaps social processes in other "multi-tribal" clusters in East Africa, and yet has not been addressed.

In recent years the traditional aspects of male initiation and the age-set system (or systems) have diminished greatly. I'm afraid I don't have a good grasp on the current situation throughout these groups. Unfortunately, recent fieldworkers, naturally interested in questions of economic development, seem less interested than I am in the continuation of information in the oral channel. Am I the only person who finds it remarkable that at least into the 1970s, the management of initiations and age-sets had been kept, separate from the modern media of communication? True, I've heard an age-set name mentioned on a 45 rpm record, but as far as I know elders did not correspond by mail to coordinate age-sets and no leader made pronouncements on the subject by newspaper or radio. I am fascinated that nearly two million people in a rapidly modernizing nation maintained their major institution with a high, possibly astounding, degree of consensus using our basic, preliterate medium of communication, word of mouth, a medium which, as I understand it, allows no other form of group action than what might be called a "natural democracy."

And so let me wind up this first part of my presentation (the next two parts are much shorter) with a note of frustration that anthropologists have not developed the implications of the nature of preliterate communication for social organization and the terms by which they describe it.

To belabor the obvious once more:

When the British entered the highlands at the turn of the century what they met were speech communities. All transportation was on foot; if you couldn't walk you only met those who came to you. All linguistic communication was by direct vocalization. One only heard what was said, or chanted or sung, within one's own earshot. There were no visual images of other people except those perceived through one's own face to face encounters. For that matter, without mirrors or photographs, one never really knew what one's own face looked like. Information about events removed in space and time reached one only through spoken repetition. There were no centralized political structures as found elsewhere in Africa, no mnemonic boards or other preliterate recording devices, and no official entrusted with memorizing an authorized version of history. All news was acquired either through first-hand experience or through what is indistinguishable from rumor and gossip. Nor, of course, was anyone keeping statistics.

Even though so many classical ethnographic analyses deal with such situations, I'm not sure that we have adequately confronted the nature of group dynamics in societies that did not have any technological means to amplify or multiply individual behavior. So I will turn the question around, back to its more familiar, comfortable form, the impact of mass media on society, and discuss colonial and post-colonial events. I hope, at least, that I have convinced you that it does not make a lot of sense to try to jam preliterate social entities into pigeon-holes of the written word.

But that, of course, is just what happened in the colonial era. It was only logical to draw up districts such that each encompassed, as nearly as possible, one of the many culturally and linguistically distinct groups that existed in Kenya (in fact, who can think of any other way to start?). Certainly, over the years, many a boundary had to be adjusted, and many a misnomer recognized as such and changed. But most of the colonial officers whose reports I have read, while they varied in sophistication and cross-cultural sensitivity, strike me as sincerely interested in trying to "get it right," particularly in the early years. The real problems arose, not from individual intentions but from systematic features inherent in colonialism. One does not see the exploitation and injustice of colonialism in reading 50 years of reports from District Officers in Kericho, but one certainly sees the insidious effects of literacy in its bureaucratic manifestation. As the standard outline for reports grew longer and longer the informational level dropped rapidly. Each new man inherited a greater body of "received knowledge" that inhibited original observation, much less the recognition of any social change in his district outside standard categories such as police actions, tax returns, road improvements, etc. It is also clear that there was increasing pressure to look good "on paper."

It has often been noted that the structure of colonial administration "crystallized" the population into tribes, creating them where they did not previously have meaning, as with the Marakwet, and dividing groups artificially, such as the Kipsigis and Nandi who found themselves in different provinces, or the Mount Elgon people who found themselves in separate colonies. As a general rule the organization of what is inside a social boundary is strongly influenced by relations to what is outside the boundary; but, of course, colonial policies worked counter to such evolved patterns. Existing patterns of interaction, both hostile and friendly, between ethnic units were disrupted. In their place inter-ethnic relations developed within new, colonial contexts, perhaps most importantly in migrant wage labor. In Kenya the work force, again for quite reasonable purposes initially, was organized into ethnic blocks. This in turn created and entrenched new forms of "tribal" competition.

The effects of a literate orientation can also be seen in social changes induced within districts. The judicial process, for example, was "rationalized" by having colonial officers write handbooks of tribal law to serve as precedent for Native Magistrates. I don't think I have to spell out the contradictions that entailed.

Anthropologists too are ruled by the written word and the need, if not to administer real people, to at least manage ethnographies about them. All of the same problems of classification come up: problems of initial unsophisticated observations becoming enshrined as truth, of terms that were inappropriately applied across a variety of realities becoming the basis for organizing data, of attention deflected away from processes and toward things which fit our categories, and of course that nightmare word, tribe.

Perhaps the ultimate contradiction in colonialism is that there is no way to keep the natives from getting in on the game. It is a logical step from labeling them and administering them in typed reports to putting labels on them individually, in the form of the kibande or identity card, and requiring them to make their mark on employment records, tax receipts, and the like. Soon one is promoting literacy through government schools in the hope of getting better clerks, or doing the same thing on the cheap by harboring missionaries who start schools for very different reasons. Missionaries, of course, are intensely interested in learning the venacular language and in training the brightest of their charges to assist in translating the Bible.

With large numbers of men from Cluster X thrown together in new settings, particularly the King's African Rifles which they joined in much greater proportion than men from other ethnic groups in World War II, and with the beginnings of an educated group sensitized to their linguistic unity, it is not surprising that ethnic awareness on the level of Cluster X took on a new urgency.

It just so happened, back in the very beginnings of the colonial era that the Nandi received much more attention than any other group in the cluster -- for the unfortunate reason that they alone mounted a stiff resistance to trade caravans and the building of the railroad line to Kisumu. In 1909 a senior colonial administrator (Hollis) published a book on Nandi Language and Folk-lore based on interviews with political prisoners and two trips to Nandiland shortly after the last major punitive expedition had secured the area with brutal efficiency. In all subsequent years of colonial administration no government sponsored writing on any other group in the cluster equaled it.

Thus through most of the colonial period Language X was called Nandi, and the people of cluster X were identified as Nandi-speakers. Of course they adopted the term as well: in the Annual Report for 1946 the District Commissioner, Kericho, expressed concern about a meeting of "tribal leaders" calling themselves "The Union of Nandi-Speaking Peoples." But this term was clearly unacceptable in the long run to all but the Nandi themselves.

The search, then, was on for a new term to use among those interested in the political mobilization of Cluster X in the post-war era.

During World War II the Kenya Broadcasting Commission (now the Voice of Kenya) produced radio shows in a number of vernacular languages. One set of broadcasts was done by John arap Chemallan, a Nandi. Chemallan made frequent use of the common interjections kale-i ("I say," used much as it is in British English), kalenjok ("I tell you," in the plural) and kalenjin ("I tell you," in the singular).

It is characteristic of Cluster X that a person who uses a particular word habitually is dubbed with it as a nickname. Thus the old man who was my adoptive father was widely known as Kiruchu, which translates "you bulls" because that is how, as a young war leader, he frequently exhorted his age-mates. Similarly a minor politician in the area in which I worked had the affectation of addressing people as "Mister" while speaking Kipsigis, and thus was known far and wide as "Mister Mister." Chemallan, however, (perhaps because radio is one-way conversation) came to apply the term kalenjin to his audience -- all those who listened and understood when he said "I tell you," in other words, the people of kalenjin. Undoubtedly some members of each group in the cluster heard him, quite possibly in places like the army barracks near Nairobi, but it is doubtful that many people in the rural areas had access to radios or even knew of these broadcasts. All that, however, was to change.

Meanwhile, by 1944, there were 14 young men from Cluster X groups attending Alliance High School near Nairobi. Alliance was the premier secondary school for Africans in the colonial period, drawing top students from all over Kenya. These men were obviously drawn to each other by their common language and background; their informal leader was a Kipsigis, Taita arap Towett. Much of the student life revolved around ethnic organizations such as the large Kikuyu Student Association and the Luo Student Association. In colonial terms the Kikuyu and Luo were tribes, though they were originally dialect clusters containing several smaller fairly autonomous segments that happened, unlike Cluster X, to occupy large continuous blocks of land. The obvious parallel for students from Kipsigis or Nandi was to organize on the level of Cluster X. Each of them proposed a series of names for their group, and the term Kalenjin was agreed upon because it is common to all dialects, no doubt also because they were familiar with Chemallan's broadcasts, and also, I am told, because it parallels in meaning the term Maasai, which is also a "cluster-level" label in a related language group.

Years later these men would rise to important elite positions. But for a while after their graduation the term Kalenjin had little currency. In 1948 Chemallan and a friend started the Kalenjin Union at Eldoret but Kipkorir reports that it was not widely known and faded for lack of any political issues.

In 1952 two events occurred that mark the real beginnings of Kalenjin identity as a mass phenomenon. One, the declaration of a State of Emergency and the ensuing campaign to defeat Mau Mau, led directly to the spread of Kalenjin identity among the more educated. The other event, the death of King George VI, contributed in a roundabout way, I contend, to the awareness of Kalenjin identity among the uneducated rural masses.

In suppressing Mau Mau the colonial government drew heavily on their loyal military sources, Cluster X. They also waged a very thorough public relations campaign both internationally and in the colony. Part of their program included establishing an Information Department office in Eldoret. The European officer was assisted by Kendagor arap Bett, one of the Alliance graduates. Bett replaced four district news-sheets (Loigoiywekap Nandi, Emostap Keiyo, Ng'alek ap Kipsigis, and Ng'alek ap Baringo) with one monthly paper, that to quote Kipkorir "rapidly gained considerable popularity among its readers and public" (1973:75). Bennett and Rosberg (1961:219) list its circulation (for 1960?) as 5,000. It was entitled, of course, Kalenjin.

If Mau Mau got Kalenjin awareness moving, it was the struggle to fill the political vacuum of impending independence that got it off the ground.

In 1959 Towett and Daniel arap Moi, from Tugen, both members of the Legislative Council, formed the Kalenjin Political Alliance. The KPA was one of a bewildering number of political parties started by a small number of elite men jockeying for position during conferences with the British Colonial Office to decide the structure of an independent government. KPA's primary political aim was "to make clear its prior claim to control over land in western Kenya, though it had no wish to interfere with European agriculture in the Highlands" (Bennett and Rosberg 1961:40). This is clearly a call to unity in the face of impending Kikuyu expansion, which was one of the main motives for Kalenjin involvement in the anti-Mau Mau campaign and an issue that lingered long after independence. In 1960 the KPA became a major part of the newly organized KADU, a coalition of small regional parties united primarily by their opposition to Kenyatta's Kikuyu dominated KANU (Kenya African National Union). When KANU became the party of power at independence, KADU assumed the role of "loyal opposition." KADU eventually disbanded but of the two leading spokesmen for "Kalenjinism," Towett served for a while in several high offices, including Minister for Education, while Moi, as you may know, is now the President of Kenya.

In addition to these few top elite, a middle range of literate people developed, including local politicians, school headmasters, junior administrative officers, and business entrepreneaurs. Charles Ng'elechei spent a year in London as a linguistic informant, and then developed a standardized orthography for Kalenjin which was adopted by the Kalenjin Language Committee, a self-created group with no budget for publishing. But hopes for the development of a Kalenjin literature remain unfulfilled (government schools no longer stress venacular languages as missions once did, and my data suggest that more Kipsigis have some literacy skills in Swahili than in their own language). With almost nothing to read in Kalenjin and a minority of the population able to read it, I do not think one can explain the rapid spread of the term Kalenjin being caused by the written word.

Yet, when I started fieldwork in 1965, people for whom Nairobi was just a word and political office a mystery had been using the term Kalenjin for years. My explanation, obviously, lies in another medium of mass communication. In 1952 when George VI died, his daughter Elizabeth was on holiday in Kenya. In 1953 the colonial government distributed free radios throughout the towns and smaller trading centers so that the populace could listen to the broadcast of her coronation. Clearly the event was memorable; the sub-set of Kipsigis male initiates who were circumcised at that time was named Lisabet, as were many girls born that year.

No doubt the authorities were interested in creating, beyond this one broadcast, a larger permanent audience for information promulgated by the Kenya Broadcasting Commission. Three stations, or services, were established: one in English, the others in Swahili, with a lot of time devoted to Hindi, and through regional transmitters, broadcasts in Arabic, other Asian languages, and a full range of African vernaculars. Since uhuru, of course, the Voice of Kenya pushes "nation building" full force.

Obviously one doesn't have to understand the spoken language to appreciate much of what is on the radio, especially music. The Kipsigis, for example, took an instant liking to Jimmy Rogers. But the effect of the radio in Kenya is, as I suspect it is everywhere, not simply to broaden one's perspective but also to sharply delineate and amplify local identities. For the Kipsigis as Kipsigis, and as Kalenjin, the key feature in this latter process of ethnic intensification has been the development of locally recorded music composed and sung by their fellow "tribesmen," referring to the familiar, decorated with idiomatic expressions, and that old standby for writing a hit, mentioning local place names.

During the 1950s musicologists recorded traditional tribal music throughout Kenya. Some of this is available on LPs for the small non-African market that is interested in such. Occasionally these traditional pieces, in Kipsigis, were played on the radio in the 1960s. But for a number of reasons they were soon displaced by "modern" records locally produced for the local market.

In the early 1960s a celebration in any remote corner of Kipsigisland might include a wind-up gramophone, with acoustic pick up, not electronic, playing brittle 78s that were recorded in the Asian owned studio over the bicycle repair shop in the district town. One master could press out 500 records. If the record sold out it was necessary to get the band back to cut another master for re-release. The musicians were Kipsigis, the songs often adaptations of traditional folktales, the instruments cheap guitars, poorly played, and the style a mix of traditional and rudimentary western forms. Within a few years many people had moved up to 45s played on battery operated machines. The music got better, the lyrics often focused on the trials of drink, or town life, or heartbreak in the big city, and the guitars were electrified. Meanwhile traditional music has virtually disappeared, except at initiations. When I asked one man what happened to the old music he replied "My father's generation played the lyre, my generation picked up guitars, and these young men walk around with transistor radios."

Each evening a 20 minute Kipsigis broadcast played local records interspersed with requests that amounted to messages to the folks back home beyond the range of postal service. People listened with intense interest to Kalenjin language broadcasts, tolerated some Swahili, particularly if it was music, and generally reacted with disgust if they ran across the cacophony of some other vernacular language such as Gusii or Turkana. There was very little interest in the national or international news.

Just as there is no society without music, I suspect there will soon be no society without its own developing tradition of recorded music -- provided of course that the group is large enough to be a viable market. But I warn you that that is possible among relatively small groups with surprisingly little cash reserves. As technology develops, the process will surely penetrate further and further into the third world. Records and radios are exceedingly portable, cheap, and infinitely easier to learn to appreciate than literature. They spread rapidly beyond the limits of literacy, ignoring time, space, and all the boundaries created through ages of social interaction. With the means of production now in almost everyone's hands, (and I've not even mentioned tape cassettes; these days for the equivalent of a few dollars anyone anywhere can buy a basic recording studio) radio is doing something far more complex than simply homogenizing the social world. It is difficult to describe in words the impact of suddenly hearing the recognizable and familiar in the midst of great linguistic and cultural diversity -- hence I've brought a tape that might give you a taste of what I mean.

I will conclude by simply saying that I'm not sure that academic disciplines whose basic concepts are still ill-suited to capturing the preliterate world will be able to keep up with the changes now going on in the postliterate, electronic age.