MEMORIZATION , as the act of storing information in the memory, is distinguished by the fact that it can be either mechanical or deliberate. It is through practice and imitation, through the mechanical repetition of the traditional gestures and speech of his social group, that the individual, without actually realizing it, memorizes most of the information necessary for proper social and religious behavior. Taken in this sense, memorization culminates in the acquisition of the innumerable actions, of behavior, thought, and sensibility, that define a social and cultural identity. From the classic texts of Maurice Halbwachs on social memory and Marcel Mauss on bodily techniques to the more recent studies of André Leroi-Gourhan on mechanical operatory chains and Erwin Goffman on interaction rites, this type of memory acquisition has been the object of numerous investigations that need not be considered here. It is sufficient to emphasize that, in contrast to this kind of memorization, there exists another, deliberate form, the techniques of which become especially prominent when certain individuals are momentarily separated from their usual social group in order to take part in an initiatory ritual or to become part of an educational institution. These extreme cases do not apply to all members of a community, however, and those to whom they do apply are never required to memorize everything, but only those gestures, techniques, and special narratives that are of particular importance, as for example certain ritual formulas, declarations of faith, religious chants, prayers, and rules of religious behavior. Deliberate memorization thus appears to be a specialization of the more natural process of acquiring knowledge and techniques, religious or otherwise, that unconsciously determine a person's membership in a particular tradition.
To this initial distinction, between mechanical and deliberate memorization, can be added another, which does not coincide with it, but applies to each term independently: the techniques and practices of memorization, be they mechanical or deliberate, vary according to whether they are associated with orality or writing. Studies by Laura Bohannan, E. A. Havelock, and Jack Goody have established that memory is organized differently when written records and models are available; without writing, memory does not function as exact reproduction, but rather as generative recollection that ties repetition to variation. It would be wrong to think that this second distinction is historical. Oral memory and memory determined by writing can easily coexist in the same culture, as the Greek, Jewish, Celtic, and Hindu examples to be mentioned below will show. This is also still the case in contemporary cultures. In the exposition that follows, which must be limited to only a few examples, will be traced a line that leads from the oral to the written. At each stage it is necessary to respect the double contribution of mechanical memorization and deliberate memorization.
In societies without writing, riddles, proverbs, myths, fables, and stories depend upon a memory that is more or less shared by the entire community. In this sense, one can speak of "social memory" or "shared knowledge." However, memorization is often an activity left to the free choice of individuals, to their tastes, affinities, and personal gifts. Henri Junod (1936) recalls a woman among the Tsonga who could tell riddle after riddle until late into the night. He met storytellers of every age and of both sexes: "Such a narrator might know only one story, and repeat it on every occasion, as did Jim Tandane, who told the story of an ogre, Nwatlakoulalambibi, with such enthusiasm that he was nicknamed after his hero! But others can tell six, ten, or twenty stories" (p. 159).
In certain societies, in particular among the Native Populations of North America, the knowledge and the possession of a myth or chant may be the privilege of an individual, who alone may pronounce it. It is for this reason that a Navajo of New Mexico may give as a sign of his poverty the fact that he does not own a single chant. A chant thus becomes a piece of "property" that concerns his own social and spiritual identity.
Most often, however, it is because certain stories are of an important collective interest that they are entrusted to the vigilant memory of one or more persons. The task of memorization is then taken up by a specific institution, often religious. These institutions are generally controlled by an elite close to power. In Rwanda, the oral tradition of the Ubwiiru, in which the rites to be performed by the king were described, was divided into eighteen rituals that were kept strictly secret. In an essay on this oral tradition Pierre Smith (1970) notes that "the individuals in charge of remembering and repeating it word for word—errors could be punished with death—were the most important dignitaries in the kingdom, and the most important three among them, the only ones who knew the text as a whole, partook of the sacred character of royalty" (p. 1385). Such "memory specialists" can be found wherever a community expresses in narrative its needs to preserve its identity. In Oceania, the experts in oral tradition, the "holders of memory," were assembled in colleges analogous to religious confraternities. The most famous among them, portrayed by Victor Ségalen in Les immémoriaux, were the harepo of Tahiti, who were the keepers of the genealogies, myths, and epics.
These orators were given true responsibility only after a serious examination, composed of difficult tests. The least mistake in memory was enough to eliminate a candidate, whose preparation was the responsibility of the priests. It is said that the harepo practiced in complete isolation, during long nocturnal walks. The transmission of ancestral knowledge rested with them. These story tellers were surrounded by a whole set of religious rituals. (O'Reilly and Poirer, 1956, pp. 1469–1470)
On Easter Island, the rongorongo, from noble families often attached to the king, used to teach chants and oral traditions in special huts. Alfred Métraux (1941) describes how this oral tradition is learned: "The student's memory was perfectly trained. During their first years of schooling, they had to learn certain psalms by heart, which they recited while playing cat's cradle: each figure … would correspond to a chant to be recited" (p. 168).
Among the Inca, the education of the nobility was the responsibility of the amauta s, who were of aristocratic descent. Their instruction lasted four years. The first year was devoted to the learning of the Quechua language; the second year to learning the religious traditions; and the third and fourth years to the handling of the famous knotted strings, the quipu.
Memorization, as it is practiced by such specialists, becomes a technique that can be taught, and that has its appropriate equipment. The Peruvian quipu, the kou-hau made by the rongorongo on Easter Island, the skeins of coconut fiber adorned with knots made in the Marquesas Islands, the wooden tablets of the Cuna Indians in Panama, and the pieces of bark of the Ojibwa Indians of North America do not, strictly speaking, constitute writing systems, but they do represent mnemotechnical means pertaining to oral memory. The same is true of certain systems of pictographic notation, such as the Aztec ideograms. Fernandon de Alva Ixtlilxóchitl recalls that the Aztec used to have writers for each type of history:
Some would work with the Annals [Xiuhamatl], putting in order the things which took place each year, giving the day, the month, and the hour. Others were charged with the genealogies and ancestries of the kings and lords and persons of lineage.… Others took care of the paintings of the boundaries, the limits, and the landmarks of the cities, provinces, and towns, and [recorded] to whom they belonged. (quoted in Léon-Portilla, 1963, p. 157)
These "writers" used pictographs to construct a mnemonic system that later historians could refer to, provided that they also referred to the purely oral tradition of the chants (ibid., p. 156), since as a system of notation it was not sufficient in itself for the total preservation of information. It was necessary in addition to have recourse to the memory that was transmitted by word of mouth through the traditional chants. One finds a similar situation, mutatis mutandis, in the early days of Islam, when to read the Qurʾān it was necessary that one already know it, since writing was still too rudimentary to be the sole means of transmission.
In oral cultures, memorization remains closely tied to the conditions of performance, despite the use of mnemonic techniques. Between listening and repeating, the absence of a fixed model does not allow for exact word-for-word repetition. Variability is essential, even though the transformations from one speaker to another often go unnoticed. There is no original version that others could reproduce, or from which they could depart. Claude Lévi-Strauss suggests that there is nevertheless a logical model, which follows certain laws of transformation. Although reproduction is not determined by the ideal of fidelity to an original (a "text"), this does not mean that it thereby becomes prey to arbitrariness. Its flexibility, its adaptability, respects certain formal conditions. "Understood in this way," notes Dan Sperber,
the facts presented by Lévi-Strauss, these peculiar correspondances and regularities, represent the intellectual capital available for primitive thought, and more particularly … for storing and retrieving information in the absence of the external memory which writing provides. Thus the study of myths can clarify the nature of human thought itself, in some of its least known aspects. (Le savoir des anthropologues, Paris, 1982, p. 115)
As logical as these rules of transformation can be, and as apt be enlightening on the workings of the human mind, they are not incompatible with trivial motives. Take, for instance, what Edmund Leach (1954) reports of the Kachin of Burma:
Kachins recount their traditions on set occasions, to justify a quarrel, to validate a social custom, to accompany a religious performance. The story-telling therefore has a purpose; it serves to validate the status of the individual who tells the story, or rather of the individual who hires a bard to tell the story, for among Kachins the telling of traditional tales is a professional occupation carried out by priests and bards of various grades (jaiwa, dumsa, laika). But if the status of one individual is validated, that almost always means that the status of someone else is denigrated. One might then infer almost from first principles that every traditional tale will occur in several different versions, each tending to uphold the claims of a different vested interest. (Leach, 1954, pp. 265–266)
This amounts to saying that the priestly bard adjusts his stories to the requirements of the audience who hired him. The horizon of expectation, the "reception," appears to be a constitutive component of oral memory, a component that conditions the very notions of fidelity and truth.
Oral memory does not like writing; there are numerous examples of this. This is not simply because it knows that writing can place it in contradiction with itself. It is primarily because the standard of truth is different for each. To understand this phenomenon better, one may turn to cultures where the two types of memory coexist. First the Celts, where the specialists of the sacred, the druids, ran their own schools, in which the main subject was memorization. According to an Irish judicial treatise, the ollam (the highest ranking scholar) was considered the equal of a king; he could recite 350 stories, 250 long ones, and 100 short ones. "As for the tenth-ranked oblaire, who makes do with leftovers at a feast, and whose escort is small, only seven stories suffices." The druids, who were the only Celts who knew how to write, refused to use their skill for religious purposes. "They say," wrote Caesar,
that they learn a great number of verses by heart: some spend twenty years at their school. They believe that religion forbids the use of writing for this purpose, unlike any other purpose such as recording public or private stories, for which they use the Greek alphabet. It seems to me that they established this usage for two reasons. On the one hand, they did not want their doctrine to spread among the people; on the other hand, they did not want those who study to rely on writing and neglect their memory, since it often happens that the use of texts has the effect of reducing efforts to memorize by heart and weakens the memory. (Gallic Wars 6.13)
Georges Dumézil (1940) comments on this testimony as follows: "knowledge is reincarnated in each generation, in each student; it is not received as a deposit; it assumes a form which, even while retaining its meaning and its essential traits, rejuvenates it and in a certain measure actualizes it." It is this dynamic, flexible, and adaptable character of oral memory that is threatened by writing. This is apparent from recent testimonies as well, such as that of a native of New Guinea (Humboldt Bay), who told an ethnologist, "in putting down our myths and legislative rules in writing you just kill them." According to Freerk C. Kamma (1975) "he meant to say: to fix or stabilize a progressing living reality means to cut it off from accompanying the living community."
In India, the brahman s who teach the Vedas are specialists in the techniques of memory, even though the Vedas have for a long time been fixed in writing. Louis Renou has noted that
there is something fascinating in the process of memorizing the verses. The master stares at the student while feeding him the verses, so to speak, with an implacable regularity, while the student rocks back and forth in a squatting position. After looking on for a few moments in such a recitation class, one better understands the hymn of the Rgveda (7.103) in which this monotonous delivery has been likened to the croaking of frogs. (Renou, 1950, p. 36)
A precise description of the techniques of memorization in the Vedic schools can be found in the fifteenth chapter of the Rk Pratisakhya, an old phonetic and grammatical treatise.
E. A. Havelock and Marcel Detienne have insisted on the coexistence of two types of memory in ancient Greece up until the time of Plato: (1) written memory and (2) social memory that is still dependent on oral tradition. Thus it is noteworthy that, although archives were available from the end of the fifth century bce, it never occured to Greek historians to refer to them as historical sources more reliable than the tradition transmitted by the works of their predecessors (appraised according to their degree of verisimilitude) or transmitted by the experience of sight (autopsía) or hearing (testimony). And yet, already from about 470 bce, Pindar and Aeschylus employ the metaphor that represents memory as an inscription, on the tablets of the soul, of what is fit to be remembered. Shortly before, the poet Simonides is said to have invented the art of memory, a technique built upon the metaphor of writing, which will undergo an important development, passing by way of Roman rhetoric (Quintillian) to the Renaissance. At the beginning of the fourth century bce, Plato is obviously preoccupied with the negative effects of the invention of writing on memory. And Antisthenes of Athens recommends according more trust to personal memory than to the external memory of written annotations.
Although Homer appears to have been a necessary reference point in ancient Greece, since his written text was learned by heart in the schools and was recited by specialists at religious festivals, there was no religious text that had authority over others. Nor was there a class of religious specialists, comparable to the pontifices, flamines, and other Roman colleges, or to the Celtic druids or Vedic brahmans. Essentially pluralist and political, Greek religion was a religion without dogmas. It obeyed customs, which varied from region to region, and from one sanctuary to the next. As a result, correct practice depended on diverse forms of information derived from a variety of sources: the family, the tribe, the town, and so on. Certain religious practices, such as those connected with the mysteries or with divination, were sometimes reserved for certain families or circles of initiates (for example, the Eumolpides and the Ceryces, the Iamides, the Trophoniades), but every Greek, regardless of social status, was capable of addressing a prayer to the gods or performing the actions indispensable to a sacrifice. Deliberate memorization, and for that matter writing as well, appeared as religious practices only in the context of such marginal devotions as Orphism and Pythagoreanism.
In the Judaic tradition, memorization plays a different role in the study of the written Torah than it does in the study of the oral Torah. The written Torah is taught through reading. The transmission of the text, teaching of the scriptures, and public readings, must all be done from a book. Even if these activities eventually result in the memorization of the text, and in fact many rabbis do know the text by heart, it is specified that the written Torah must never be copied from memory. On the other hand, the oral Torah is taught through repetition from memory, even though written notes may be used as a mnemotechnic device, and even though, at an early date, the Mishnah, and then the Talmud, was committed to writing. The masters of the oral Torah, the tannaim ("teachers"), were like living memories, capable of reproducing an impressive number of traditions. Their knowledge, often mechanical and lacking in reflection, was used as a reference source by the rabbis and colleges. A famous example is Natronai ben Ḥavivai (eighth century), who wrote down the entire Talmud from memory after immigrating to Spain.
In the Christian tradition, the role of memorization seems to be much less important, although from the fourth century there are references to religious schools where the Psalms, the words of the apostles, prayers, and passages from the Old Testament, were learned by heart. In the Divine Office, for instance, the use of a breviary, even though required to be recited aloud, served as a substitute for memorization. Thus blindness could relieve a monk of the obligation of reciting the hours, save for what he knew from memory.
In Islam, which is a religion of the word as much as a religion of the book, memorization was essential from the very beginning. The words of the Prophet, which repeated the Archangel Gabriel's reading of the archetypal book, were transmitted orally by a group of the companions of the Prophet and by specialists in memorization before the Qurʾān was finally written down. From the time of the third caliph, writing made possible the fixation of the tradition, but it never did away with recourse to memory. In effect, to read the Qurʾān in its primitive form, it was necessary to know its contents. Later, writing and memorization continued to be closely related practices. The Qurʾanic schools (madrasah s) were tied to a mosque. Children came to learn the Qurʾān by heart, even before they could read. These schools also taught the ḥadith s, the tradition that was guaranteed by a chain of authorities, or isnad. Before being written down in such texts, such as that of al-Bukhari, this tradition was transmitted orally. The information it gives about the acts and words of the Prophet are used to regulate daily life down to the smallest details, in profane as well as in religious matters. The tradition represents the Prophet himself, sitting in the mosque and teaching the ḥadith s. His words are repeated three times by all present, until they are known by heart.
For a general discussion of memorization and of method, see Maurice Halbwachs's Les cadres sociaux de la mémoire (Paris, 1925) and La mémoire collective (Paris, 1950); Marcel Mauss's "Les techniques du corps," Journal de psychologie 32 (March–April 1935): 271–293, reprinted in Mauss's Sociologie et anthropologie (Paris, pp. 365–383); André Leroi-Gourhan's Le geste et la parole, vol. 2, La mémoire et les rythmes (Paris, 1965); Laura Bohannan's "A Genealogical Charter," Africa 22 (October 1952): 301–315; Jack Goody's The Domestication of the Savage Mind (Cambridge, 1977); Jan Assmann, Das Kulturelle Gedächtnis. Schrift, Erinnerung und politische Identität in frühen Hochkulturen (München, 1992); Jan Vansina's Oral Tradition: A Study in Historical Methodology, translated by H. M. Wright (Chicago, 1965); Ruth Finnegan's Oral Poetry: Its Nature, Significance and Social Context (Cambridge, 1977); W. Kelber, "Modalities of Communication, Cognition, and Physiology of Perception: Orality, Rhetoric, Scribality," Semeia 65 (1995): 193–216.
The following works discuss the role and nature of memorization in specific regions and religious traditions.
Henri A. Junod, Mœurs et coutumes des Bantous, vol. 2, Vie mentale (Paris, 1936). Pierre Smith, "La lance d'une jeune fille: Mythe et poésie au Rwanda," in Échanges et communications: Mélanges offerts à Claude Lévi-Strauss a l'occasion de son soixantième anniversaire, edited by Pierre Maranda and Jean Pouillon, vol. 2 (The Hague, 1970), pp. 1381–1408.
Marcelle Bouteiller, "Littérature indienne d'Amérique du Nord," in Histoire des littératures, edited by Raymond Queneau, vol. 1 (Paris, 1956), pp. 1513–1523. Robert H. Lowie, Primitive Society (New York, 1961), pp. 224–232.
Patrick O'Reilly and Jean Poirer, "Littératures océaniennes," in Histoire des littératures, edited by Raymond Queneau, vol. 1 (Paris, 1956), pp. 1461–1492. Alfred Métraux, L'Ile de Pâques (Paris, 1941), pp. 165–179.
M. L. Locke, The Ancient Quipu or Peruvian Knot-Record (New York, 1923). Rafael Karsten, A Totalitarian State of the Past: The Civilization of Inca Empire in Ancient Peru (1949; reprint, Port Washington, N. Y., 1969).
Miguel León-Portilla, Aztec Thought and Culture: A Study of the Ancient Nahuatl Mind, translated by Jack Emory (Norman, Okla., 1963).
Edmund Leach, Political Systems of Highland Burma: A Study of Kachin Social Structure. (Cambridge, Mass., 1954).
Freerk C. Kamma, trans. and comp. Religious Texts of the Oral Tradition from Western New-Guinea (Irian Jaya), pt. A (Leiden, 1975).
Georges Dumézil, "La tradition druidique et l'écriture: Le Vivant et le Mort," Revue de l'histoire des religions 121 (March–June 1940): 125–133. Françoise Le Roux and Christian-J. Guyonvarc'h, Les druides, 3d ed. (Rennes, 1982).
Louis Renou, Sanskrit et culture (Paris, 1950). Louis Renou, Les écoles védiques et la formation du Véda (Paris, 1957).
E. A. Havelock, Preface to Plato (Cambridge, Mass., 1963). Marcel Detienne, L'invention de la mythologie (Paris, 1981). M. Simondon, La mémoire et l'oubli dans la pensée grecque jusqu'à la fin du cinquième siècle avant J.-C.: Psychologie archaïque, mythes et doctrines (Paris, 1982). Marcel Detienne (ed.), Les savoirs de l'écriture en Grèce ancienne (Lilles, 1988).
Birger Gerhardsson, Memory and Manuscript: Oral Tradition and Written Transmission in Rabbinic Judaism and Early Christianity, translated by Eric J. Sharpe (Uppsala, 1961).
Theodor Klauser, "Auswendiglernen," in Reallexikon für Antike und Christentum, vol. 1 (Stuttgart, 1950).
Dale F. Eickelman, "The Art of Memory: Islamic Education and Its Social Reproduction," Comparative Studies in Society and History 20 (October 1978): pp. 485–516. Pierre Crapon de Caprona, Le Coran: Aux sources de la parole oraculaire (Paris, 1981), pp. 147–162; Alfred-Louis de Prémare, Les fondations de l'Islam. Entre écriture et histoire, Paris, 2002.
Phillipe Borgeaud (1987 and 2005)
Translated from French by Marie-Claude Hays-Merlaud