From: Goody, ed., Literacy in Traditional Societies (Cambridge University Press, 1968) - 1st part
The accepted tripartite divisions of the formal study of mankind's past and of his present are to a considerable extent based on man's development first of language and later of writing. Looked at in the perspective of time, man's biological evolution shades into prehistory when he becomes a language-using animal; add writing, and history proper begins. Looked at in a temporal perspective, man as animal is studied primarily by the zoologist, man as talking animal primarily by the anthropologist, and man as talking and writing animal primarily by the sociologist.
That the differentiation between these categories should be founded on different modes of communication is clearly appropriate; it was language that enabled man to achieve a form of social organization whose range and complexity were different in kind from that of animals; whereas the social organization of animals was mainly instinctive and genetically transmitted, that of man was largely learned and transmitted verbally through the cultural heritage. The basis for the last two distinctions, those based on the development of writing, is equally clear: to the the extent that a significant quantity of written records are available, the prehistorian yields to the historian; and to the extent that alphabetical writing and popular literacy imply new modes of social organization and transmission, the anthropologist tends to yield to the sociologist.
But why? And how? There is no agreement about this question, nor even about what the actual boundary lines between non-literate and literate cultures are. At what point in the formalization of pictographs or other graphic signs can we talk of 'letters', of literacy? And what proportion of the society has to write and read before the culture as a whole can be described as literate?
These are some of the many reasons why the extent to which there is any distinction between the areas and methods peculiar to anthropology and sociology must be regarded as problematic; and the difficulty affects not only the boundaries of the two disciplines but also the nature of the intrinsic differences in their subject matter.  The recent trend has been for anthropologists to spread their net more widely and engage in the study of industrial societies side by side with their sociological colleagues. We can no longer accept the view that anthropologists have as their objective the study of primitive man, who is characterized by a 'primitive mind', while sociologists, on the other hand, concern themselves with civilized man, whose activities are guided by 'rational thought' and tested by 'logico-empirical procedures'. The reaction against such ethnocentric views, however, has now gone to the point of denying that the distinction between non-literate and literate societies has any significant validity. This position seems contrary to our personal observation; and so it has seemed worthwhile to enquire whether there may not be, even from the most empirical and relativist standpoint, genuine illumination to be derived from a further consideration of some of the historical and analytic problems connected with the traditional dichotomy between non-literate and literate societies.
THE CULTURAL TRADITION IN NON-LITERATE SOCIETIES
For reasons which will become clear it seems best to begin with a generalized description of the ways in which the cultural heritage is transmitted in non-literate societies, and then to see how these ways are changed by the widespread adoption of any easy and effective means of written communication.
When one generation hands on its cultural heritage to the next, three fairly separate items are involved. First, the society passes on its material plant, including the natural resources available to its members. Secondly, it transmits standardized ways of acting. These customary ways of behaving are only partly communicated by verbal means; ways of cooking food, of growing crops, of handling children may be transmitted by direct imitation. But the most significant elements of any human culture are undoubtedly channelled through words, and reside in the particular range of meanings and attitudes which members of any society attach to their verbal symbols. These elements include not only what we habitually think of as customary behaviour but also such items as ideas of space and time, generalized goals and aspirations, in short the Weltanschauung of every social group. In Durkheim's words, these categories of the understanding are 'priceless instruments of thought which the human groups have laboriously forged through the centuries and where they have accumulated the best of their intellectual capital' (Durkheim, [Les formes élémentaires de la vie religieuse, 1912]). The relative continuity of these categories of understanding from one generation to another is primarily ensured by language, which is the most direct and comprehensive expression of the social experience of the group.
The transmission of the verbal elements of culture by oral means can be visualized as a long chain of interlocking conversations between members of the group. Thus all beliefs and values, all forms of knowledge, are communicated between individuals in face-to-face contact; and, as distinct from the material content of the cultural tradition, whether it be cave-paintings or hand-axes, they are stored only in human memory.
The intrinsic nature of oral communication has a considerable effect upon both the content and the transmission of the cultural repertoire. In the first place, it makes for a directness of relationship between symbol and referent. There can be no reference to 'dictionary definitions', nor can words accumulate the successive layers of historically validated meanings which they acquire in a literate culture. Instead, the meaning of each word is ratified in a succession of concrete situations, accompanied by vocal inflections and physical gestures, all of which combine to particularize both its specific denotation and its accepted connotative usages. This process of direct semantic ratification, of course, operates cumulatively; and as a result the totality of symbol-referent relationships is more immediately experienced by the individual in an exclusively oral culture, and is thus more deeply socialized.
One way of illustrating this is to consider how the range of vocabulary in a non-literate society reflects this mode of semantic ratification. It has often been observed how the elaboration of the vocabulary of such a society reflects the particular interests of the people concerned. The inhabitants of the Pacific island of Lesu have not one, but a dozen or so, words for pigs ..., according to sex, colour, and where they come from - a prolixity which mirrors the importance of pigs in a domestic economy that otherwise includes few sources of protein. The corollary of this prolixity is that where common emphases and interests, whether material or otherwise, are not specifically involved, there is little verbal development. Malinowski reported that in the Trobriands the outer world was only nemed in so far as it yielded useful things, useful, that is, in the very broadest sense;  and there is much other testimony to support the view that there is an intimate functional adaptation of language in non-literate societies, which obtains not only for the relatively simple and concrete symbol-referents involved above, but also for the more generalized 'categories of understanding' and for the cultural tradition as a whole.
In an essay he wrote in collaboration with Mauss, 'De quelques formes primitives de classification',  Durkheim traces the interconnections between the ideas of space and the territorial distribution of the Australian aborigines, the Zuni of the Pueblo area and the Sioux of the Great Plains. This intermeshing of what he called the collective representations with the social morphology of a particular society is clearly another aspect of the same directness of relationship between symbol and referent. Just as the more concrete part of a vocabulary reflects the dominant interests of the society, so the more abstract categories are often closely linked to the accepted terminology for pragmatic pursuits. Among the LoDagaa of northern Ghana, days are reckoned according to the incidence of neighbouring markets; the very word for day and market is the same, and the 'weekly' cycle is a six-day revolution of the most important markets in the vicinity, a cycle which also defines the spatial range of everyday activities. 
The way in which these various institutions in an oral culture are kept in relatively close accommodation one to another surely bears directly on the question of the central difference between literate and non-literate societies. As we have remarked, the whole content of the social tradition, apart from the material inheritances, is held in memory. The social aspects of remembering have been emphasized by sociologists and psychologists, particularly by Maurice Halbwachs.  What the individual remembers tends to be what is of critical importance in his experience of the main social relationships. In each generation, therefore, the individual memory will mediate the cultural heritage in such a way that its new constituents will adjust to the old by the process of interpretation that Bartlett calls 'rationalizing' or the 'effort after meaning'; and whatever parts of it have ceased to be of contemporary relevance are likely to be eliminated by the process of forgetting.
The social function of memory - and of forgetting - can thus be seen as the final stage of what may be called the homeostatic organization of the cultural tradition in non-literate society. The language is developed in intimate association with the experience of the community, and it is learned by the individual in face-to-face contact with the other members. What continues to be of social relevance is stored in the memory while the rest is usually forgotten: and language - primarily vocabulary - is the effective medium of this crucial process of social digestion and elimination which may be regarded as analogous to the homeostatic organization of the human body by means of which it attempts to maintain its present condition of life.
In drawing attention to the importance of these assimilating mechanisms in non-literate societies, we are denying neither the occurrence of social change nor yet the 'survivals' which it leaves in its wake. Nor do we overlook the existence of mnemonic devices in oral cultures which offer some resistance to the interpretative process. Formalized patterns of speech, recital under ritual conditions, the use of drums and other musical instruments, the employment of professional remembrancers - all such factors may shield at least part of the content of memory from the transmuting influence of the immediate pressures of the present. The Homeric epics, for instance, seem to have been written down during the first century of Greek literature between 750 and 650 B.C., but 'they look to a departed era, and their substance is unmistakably old' (Finley [M.I. Finley, The World of Odysseus, New York, 1954], p.26).
With these qualifications, however, it seems correct to characterize the transmission of the cultural tradition in oral societies as homeostatic in view of the way in which its emphasis differs from that in literate societies. The description offered has, of course, been extremely abstract; but a few illustrative examples in one important area - that of how the tribal past is digested into the communal orientation of the presnt - may serve to make it clearer.
Like the Bedouin Arabs and the Hebrews of the Old Testament, the Tiv people of Nigeria give long genealogies of their forebears, which in this case stretch some twelve generations in depth back to an eponymous founding ancestor.  Neither these genealogies, nor the Biblical lists of the descendants of Adam, were remembered purely as feats of memory. They served as mnemonics for systems of social relations. When on his deathbed Jacob delivered prophecies about the future of his twelve sons, he spoke of them as the twelve tribes or nations of Israel. It would seem from the account in Genesis that the genealogical tables here refer to contemporary groups rather than to dead individuals;  the tables presumably serve to regulate social relations among the twelve tribes of Israel in a manner similar to that which has been well analysed in Evans-Pritchard's work (1940) on the Nuer of the southern Sudan and Fortes' account (1945) of the Tallensi of northern Ghana.
Early British administrators among the Tiv of Nigeria were aware of the great importance attached to these genealogies, which were continually discussed in court cases where the rights and duties of one man towards another were in dispute. Consequently they took the trouble to write down the long lists of names and preserve them for posterity, so that future administrators might refer to them in giving judgement. Forty years later, when the Bohannans carried out anthropological field work in the area, their successors were still using the same genealogies.  However, these written pedigrees now gave rise to many disagreements; the Tiv maintained that they were incorrect, while the officials regarded them as statements of fact, as records of what had actually happened, and could not agree that the unlettered indigenes could be better informed about the past than their own literate predecessors. What neither party realized was that in any society of this kind changes take place which require a constant readjustment in the genealogies if they are to continue to carry out their function as mnemonics of social relationships.
These changes are of several kinds: those arising from the turnover in personnel, from the process of 'birth and copulation and death'; those connected with the rearrangement of the constituent units of the society, with the migration of one group and fission of another; and lastly those resulting from the effects of changes in the social system itself, whether generated from within or initiated from without. Each of these three processes (which we may refer to for convenience as the processes of generational, organizational and structural change) could lead to alterations of the kind to which the administration objected.
It is obvious that the process of generation leads in itself to a constant lengthening of the genealogy; on the other hand, the population to which it is linked may in fact be growing at quite a different rate, perhaps simply replacing itself. So despite its increasing length the genealogy may have to refer to just as many people at the present time as it did fifty, a hundred, or perhaps two hundred years ago. Consequently the added depth of lineages caused by new births needs to be accompanied by a process of genealogical shrinkage; the occurrence of this telescoping process, a common example of the general social phenomenon which J.A. Barnes has felicitously termed 'structural amnesia', has been attested in many societies, including all those mentioned above...
Organizational changes lead to similar adjustments. The state of Gonja in northern Ghana is divided into a number of divisional chiefdoms, certain of which are recognized as providing in turn the ruler of the whole nation. When asked to explain their system the Gonja recount how the founder of the state, Ndewura Japka, came down from the Niger Bend in search of gold, conquered the indigenous inhabitants of the area and enthroned himself as chief of the state and his sons as rulers of its territorial divisions. At his death the divisional chiefs succeeded to the paramountcy in turn. When the details of this story were first recorded at the turn of the present century, at the time the British were extending their control over the area, Jakpa was said to have begotten seven sons, this corresponding to the number of divisions whose heads were eligible for the supreme office by virtue of their descent from the founder of the particular chiefdom. But at the same time as the British had arrived, two of the seven divisions disappeared, one being deliberately incorporated in a neighbouring division because its rulers had supported a Mandingo invader, Samori, and another because of some boundary changes introduced by the British administration. Sixty years later, when the myths of state were again recorded, Jakpa was credited with only five sons and no mention was made of the founders of the two divisions which had since disappeared from the political map. 
These two instances from the Tiv and the Gonja emphasize that genealogies often serve the same function that Malinowski claimed for myth: they act as 'charters' of present social institutions rather than as faithful historical records of times past [B. Malinowski, Myth in Primitive Psychology, London: 1926, pp.23, 43]. They can do this more consistently because they operate within an oral rather than a written tradition and thus tend to be automatically adjusted to existing social relations as they are passed by word of mouth from one member of the society to another. The social element in remembering results in the genealogies being transmuted in the course of being transmitted; and a similar process takes place with regard to other cultural elements as well, to myths, for example, and to sacred lore in general. Deities and other supernatural agencies which have served their purpose can be quietly dropped from the contemporary pantheon; and as the society changes, myths too are forgotten, attributed to other personages, or transformed in their meaning.
One of the most important results of this homeostatic tendency is that the individual has little perception of the past except in terms of the present; whereas the annals of a literate society cannot but enforce a more objective recognition of the distinction between what was and what is. Franz Boas wrote that for the Eskimo the world has always been as it is now.  It seems probable, at least, that the form in which non-literate societies conceive the world of the past is itself influenced by the process of transmission described. The Tiv have their genealogies, others their sacred tales about the origin of the world and the way in which man acquired his culture. But all their conceptualizations of the past cannot help being governed by the concerns of the present, merely because there is no body of chronologically ordered statements to which reference can be made. The Tiv do not recognize any contradiction between what they say now and what they said fifty years ago, since no enduring records exist for them to set beside their present views. Myth and history merge into one: the elements in the cultural heritage which cease to have a contemporary relevance tend to be soon forgotten or transformed; and as the individuals of each generation acquire their vocabulary, their genealogies, and their myths, they are unaware that various words, proper names and stories have dropped out, or that others have changed their meanings or been replaced.
 Some writers distinguish the field of social anthropology from that of sociology on the basis of its subject matter (i.e. the study of non-literate or non-European peoples), others on the basis of its techniques (e.g. that of participant observation). For a discussion of these points, see Siegfried F. Nadel, The Foundations of Social Anthropology (London, 1951), p.2.
 Bronislaw Malinowski, 'The Problem of Meaning in Primitive Languages', in C.K. Ogden and I.A. Richards, The Meaning of Meaning (London, 1923), pp.296-336, esp. p.331. But see also the critical comments by Claude Lévi-Strauss, La Pensée sauvage (Paris, 1962), pp.6, 15-16.
 L'Année sociologique, VII (1902-3), pp.1-72. See also S. Czarnowski, 'Le morcellement de l'étendue et sa limitation dans la religion et la magie', Actes du congrès international d'histoire des religions (Paris, 1925), I, pp.339-59.
 Jack Goody, unpublished field notes, 1950-2. See also E.E. Evans-Pritchard, The Nuer (Oxford, 1940), chap. 3, 'Time and Space', and David Tait, The Konkomba of Northern Ghana (London, 1961), pp.17ff. For a general treatment of the subject, see A. Irving Hallowell, 'Temporal Orientations in Western Civilisation and in a Preliterate Society', American Anthropologist, XXXIX (1937), pp.647-70.
 Les Cadres sociaux de la mémoire (Paris, 1925); 'Mémoire et société', L'Année sociologique, 3e série, I (1940-8), pp.11-177; La Mémoire collective (Paris, 1950). See also Frederic C. Bartlett on the tendency of oral discourse to become an expression of ideas and attitudes of the group rather than the individual speaker, in Remembering (Cambridge, 1932), pp.265-7, and Psychology and Primitive Culture (Cambridge, 1923), pp.42-3, 62-3, 256.
 Laura Bohannan, 'A Genealogical Charter', Africa XXII (1952), pp.301-15; Emrys Peters, 'The Proliferation of Segments in the Lineage of the Bedouin of Cyrenaica', Journal of the Royal Anthropological Institute, XC (1960), pp.29-53. See also Godfrey and Monica Wilson, The Analysis of Social Change (Cambridge, 1945), p.27.
 Ch. 49; further evidence supporting this assumption is found in the etymology of the Hebrew term Toledot, which originally denoted 'genealogies', and assumed also the meaning of 'stories and accounts' about the origin of a nation. 'In this sense the term was also applied to the account of the creation of heaven and earth' (Solomon Gandz, 'Oral Tradition in the Bible' in Jewish Studies in Memory of George A. Kohut, ed. Salo W. Baron and Alexander Marx, New York, 1935, p.269).
 'A Genealogical Charter', p.314.
 Jack Goody, unpublished field notes, 1956-7; the heads of the divisions who could not succeed to the paramountcy also claimed descent from sons of the founding ancestor, Jakpa, but this was not an intrinsic part of the myth as usually told, and in any case their number remained constant during the period in question.
 Franz Boas, 'The Folklore of the Eskimo', Journal of American Folklore LXIV (1904), p.2. Lévi-Strauss treats the absence of historical knowledge as one of the distinctive features of la pensée sauvage in contrast to la pensée domestiquée (La Pensée sauvage, p.349).