Religious memory,between orality and writing

Since the groundbreaking studies of the French sociologist Maurice Halbwachs (1997 [1950]), we have learned to call “religious memory” those aspects of collective memory dealing with religious practice as well as with religious beliefs (p. 96). Like memory itself, the concept of religious memory is strikingly polyvalent, or even metaphoric, and is expressed very differently in non-literate and literate societies. The question of the nature of religious memory in non-literate societies is not at all a trivial one.

What do we mean when we speak of religious memory? The purpose of the following pages is to offer a tentative answer to this question. We shall do that partly through the angle of the birth and early development of Christianity in late antiquity, which obviously represents a major chapter in the history of religions, one to which Halbwachs devoted his last monograph. This approach permits us to reflect on the birth of a new religious tradition, established as the reinterpretation of an older one, and thus in times of dramatic change in the relationship between oral and written literatures in the Near East and the Mediterranean. I shall argue that there are two kinds of religious memory, and that this dual nature is strikingly reminiscent of semantic and episodic (long-term, declarative—or explicit) memory, as revealed by contemporary neuroscientists, who speak of two (and sometimes of more than two) systems of memory, used simultaneously (Squire and Kandel, 2009). To be sure, the claim that there are multiple systems of memory is far from new. It is already present in the dual conception of memory in Greek: mnēmē, continuous memory, versus anamnēsis, recollection, a duality reflected in modern languages such as German (Gedächtniss/Erinnerung) or French (mémoire/souvenir) (Cancik and Mohr, 1990; Carruthers, 1990). I shall seek to show that these two kinds of religious memory function simultaneously, complementing one another.

For most neuroscientists, “memory” fully exists only within an individual mind, although some new studies seriously challenge this perspective (Dudai and Edelson, 2016). As a complex set of contacts between neuronal systems, memory, no static conservatory of things past, is essentially polyvalent and involved in both reconstructing the past and offering guidelines for action. Memory’s plasticity permits it to constantly remodel itself. The reconstruction effected by memory is in constant evolution.

On the other hand, sociologists and historians have learned, for almost a century, to speak of the “collective memory” of a society (Halbwachs, 1997 [1950]), sometimes referring to its “social” or “cultural” memory (Assmann, 2000a, 2000b; Welker, 2016). The question of whether we can speak of “memory” beyond the individual mind in a non-metaphorical way, which is not a trivial one, is not always clearly addressed. While there is a plethora of studies of collective memory, the relationship between the personal and the collective dimensions of memory remains strikingly understudied. Oddly enough, there have been relatively few attempts to deal with the connection between individual and collective memory.

Religious memory is a special case of cultural memory. As a concept, it was first coined by Halbwachs, almost a century ago (Halbwachs, 1925: 178–221). A social phenomenon by nature, collective memory is intrinsically related to the invention and use of writing and to the ability to retain archival knowledge outside a single mind. As Yadin Dudai, a leading neuroscientist working on memory, points out, little is known about the acquisition of collective memory (Dudai, 2002). For Halbwachs, there is a symmetry between the passage from personal to collective memory and that from collective memory to history.

Among all aspects of culture, religion, in both its aspects of storytelling (mythos) and ritual action (praxis), is indeed most strikingly marked by memory. In all societies, religion is an essentially conservative social phenomenon, which seeks to preserve, through time, the memory of an ancient period (Halbwachs, 1925: 193). Hence, it insists on tradition, on remembering the past, and on cultivating tradition. For Halbwachs, who early in life was under the double influence of both Bergson and Durkheim, memory is at once private and public, personal and political. For him, religions aim at retaining intact, through time, the pure memory of an ancient past. Like any collective memory, religious memory does not conserve the past but reconstructs it, through texts, traditions, and the present.

It is in Les cadres sociaux de la mémoire, originally published in 1925, that Halbwachs referred for the first time to “[collective] religious memory” (mémoire [collective] religieuse) in order to identify the essential ability of religion to retain intact the memory of the past. For him, the transformation, through time, of both dogma and ritual enriches religious memory. Halbwachs adds that religious memory follows the same laws as collective memory, reconstructing rather than conserving the past. The latter is reconstructed through rituals, texts, and traditions (Halbwachs, 1925: 193, 221).

According to Halbwachs’ (1997 [1950]) Bergsonian taxonomy, personal (or interior) and social (or exterior) memory should not be conceived as radically distinct from one another, since individual memory is never totally isolated and closed, hence always retaining a social dimension (p. 99). He insists on the fact that memories are always collective, as one is reminded by others even of events which one alone experienced, of objects which one alone saw (Halbwachs, 1997 [1950]: 52). Individual memory, indeed, is not totally isolated and closed (Halbwachs, 1997 [1950]: 98). By nature, then, the concept of memory refers to a number of different phenomena. For Halbwachs, religion is essentially tradition, and the process of “traditionalization” is directly related to memory. Hence, it is possible to speak of a “Halbwachsian reduction of religion to memory” (Hervieu-Léger in Halbwachs, 2008 [1941]: 37). Now if there is no such thing as a totally individual memory, then the use of “memory” for a collective phenomenon is not simply metaphorical; it is in fact as legitimate as its traditional (and contemporary) use for certain activities within the brain. In this respect, the French sociologist of religion Danièle Hervieu-Léger speaks of religion as a chain of memory (Hervieu-Léger, 2000). This is true in non-literate as well as in literate societies, although it is expressed, of course, in vastly different ways in various forms of societies, where the past can be preserved in writings and monuments.

Halbwachs’ last work, La topographie légendaire des évangiles, argues that sociology reflects a deeper reality than history. This work represents in a sense a global answer to Marc Bloch’s criticism and remains to this day a very important study of memory in early Christianity. More than two generations after its publication and despite a new edition, this study oddly remains almost ignored. Halbwachs was able to show how early Christian memory transformed the Jewish memory in which it was grounded. It is only in Jewish memory that Christian memory could find its roots. In a sense, he almost wrote, the Christians were attempting to steal Jewish memory (Halbwachs, 2008 [1941]: 47). One has yet to analyze the consequences of such an attempt on the redeployment of religious memory in Rabbinic Judaism. My guess is that it could only have strengthened the oral trend mentioned above.

Jan Assmann has done much in the last 20 years or so to deepen our reflection on the religious dimensions of cultural memory—a concept developed, after Halbwachs, by the German Jewish art historian Aby Warburg (Assmann, 2000b). Assmann argues that it takes about three generations for actual, personal memories to be transformed into cultural memory. During that period, he proposes to speak of “communicative memory” (kommunikative Gedächtnis), when memories are transmitted by the witnesses to their children and grandchildren (Kirk and Thatcher, 2005). For Assmann, stabilization of religious memory occurs through institutions, and the rituals remained for a long time the decisive medium of reproduction of religious memory (Assmann, 2000a; cf. Atran, 2002; Severi, 1993). I shall offer here a slightly different perspective, insisting that in religions of the book, scriptural hermeneutics play a role at least as important as that of rituals in the stabilization process.

One may at least mention here the question of transgenerational religious memory. Can the deep impact of a striking event on the mind be passed to the next generation(s)? This is a major question, which until very recently remained in the domain of science fiction. But neuroscientists have recently been able to show how long-term memory can bring to a modification of neuronal structures, and epigenetics is today referring to a “cellular memory” and to the transmission between generations of non-biological characters. Despite intense efforts of scientists, their results are still preliminary, and little is accepted by a consensus. In any case, it is not far-fetched to imagine that powerful memories of parents have an effect on the next generation.

The Oxford social anthropologist Harvey Whitehouse has proposed a theory of religion differentiating between two fundamental modes of religiosity, which he calls, respectively, doctrinal and imagistic, to be found, in different ways, in all societies (Whitehouse, 2002). The doctrinal mode of religiosity, which is highly routinized (in a Weberian sense of the word), is essentially characterized by high frequency and low arousing, while the opposite is true of the imagistic mode. Of direct interest for our present purposes are the different kinds of memory Whitehouse attributes to these two modes of religiosity. He proposes to follow a broadly accepted differentiation between explicit (or declarative and conscious) and implicit (or non-declarative and unconscious) memory and between two kinds of explicit memory—short-term and long-term memory—the latter being in its turn divided between semantic and episodic memory. For him, the doctrinal mode of religiosity makes use of semantic memory, while it is episodic memory that is mainly activated in the imagistic mode. Whitehouse does not claim that religions follow either the doctrinal or the imagistic pattern and insists on the fact that all patterns of religious life offer different mixtures between the two modes.

A number of attempts have been made to apply the categories developed by Whitehouse to the anthropology and history of religion. Two relevant volumes of essays were recently edited. One dealt with major religious trends in the Roman Empire. Another one, edited by Whitehouse and James Laidlaw, was entitled Ritual and Memory (Whitehouse and Laidlaw, 2004). In one of his numerous studies of oral and written cultures, the anthropologist Jack Goody (1986) compares image and doctrine (in Whitehouse’s parlance) to speech and writing and notes that in the interface between religions of oral cultures and “religions of the book,” the latter have an advantage and tend to absorb the former (ch. 1).

It seems difficult, or even dangerous, as claimed by Jack Goody, to speak of collective memory before the development of literacy, as in such societies, where religious memory remained by necessity oral, there were no written archives outside the human mind (Goody, 2000). One obvious track, then, is to call attention to the essentially different organization of memory in non-literate and in literate societies. In literate societies, the existence of archives (outside the mind) and of literary texts permits a less intensive use of memory in order to recall the past as well as some of the other essential elements of collective identity. Non-literate societies, on the other hand, make much use of generative memorizing (based on repetition and variation). Enigmas, proverbs, myths, fables, and stories are all dependent on social memory—a memory more or less shared by the whole society (Goody, 2000; Ong, 1988 [1982]). Historians of ancient religions have called attention to the different ways in which ancient societies present different articulations of oral and written memory. The existence of two types of religious memory, related to orality and writing, can be detected in various archaic societies, in Greece as well as in Mesopotamia (Borgeaud, 1988; Stolz, 1988).

In ancient and pre-modern societies that we may call semi-literate, oral and written memory functioned together, as literacy remained the privilege of small elites. A characteristic of such societies was the interface between oral and written culture (Ong, 1988 [1982]). In the history of religion, the most important distinction in this respect is the appearance of scriptural religions, starting with the formation of the Hebrew Bible and of the Iranian Avesta. In contradistinction to what happens in other ancient religions, the existence of revealed texts (usually written, but sometimes oral, as in the case of the Avesta) puts these texts at the very core of both religious thought and ritual. Hence, we may speak of a real paradigm change in religious memory with the rise of the scriptural religions. This paradigm change in religion parallels that in culture inaugurated by the emergence of writing.

By focusing on the formation of religious memory in early Christianity, I hope to be able to grasp some of the processes through which collective memory is formed and evolves. The first centuries of Christianity represent a particularly interesting and important chapter in the history of religions. A number of facts make early Christianity a particularly clear case of the birth of a new religion and the invention of a new religious memory, or rather of the drastic reformulation of an earlier religious memory, that of Israel (the Christians call themselves verus Israel), thanks to a series of new hermeneutical thought patterns. Moreover, we shall see how this new system of religious memory was closely linked to some momentous changes in the relationship between orality and literacy, in the attitudes to reading, and in the blossoming of scriptural religions in late antiquity. We shall refer in succession to each of the various elements involved in the formation of early Christian memory if we hope to eventually see them together, synthetically.

In The End of Sacrifice (Stroumsa, 2009 [2005]), I have argued at length that some major transformations of religious life can be distinguished in late antiquity. These transformations brought to a strong internalization of religion, most clearly exemplified through the eventual prohibition of blood sacrifices in the Roman Empire before the end of the fourth century. This most striking transformation of ritual was combined with the growth and development of what has been called “the scriptural moment” of late antiquity, a period going roughly from Jesus through Mani to Muhammad, or rather from the making of the New Testament to that of the Qur’an (Smith, 1993). To some extent, one can say that the scriptures replaced blood sacrifices in ritual. The “scriptural movement” in the Mediterranean world and the Near East saw the emergence and development of the “religions of the book” (i.e. Judaism, Christianity and their offshoots, such as Manichaeism, and then Islam), the transformation of Greco-Roman traditional education, or paideia, by the development of Christian culture. From now on, the literary heritage of Greco-Roman culture would be subsumed to scripture, to the book of divine revelation. The parallel hermeneutics of these two literary corpora would from now on constitute the core of Christian cultural memory.

The complexity of Christian memory is not only related to the intricate Christian conception of scriptures: two main levels (Old and New Testaments), each to be read and understood in light of the other, the first announcing the second, the second interpreting the first. This complexity also reflects the cohabitation of what was an essentially Jewish scripture with Greco-Roman cultural tradition (paideia) and the fact that this cohabitation entailed the establishment of a set of correlations between these two literary corpus (Lössl, 2014; Stroumsa, 2005, 2009 [2005]). The deep transformation of memory in early Christianity is conditioned by the new duality between religious memory (essentially the reinterpretation and appropriation of that of Israel) and cultural memory (essentially the reinterpretation and appropriation of that of Greco-Roman civilization). Hence, early Christianity permitted, for the first time, the emergence of something which did not really exist in ancient societies and which has become a central aspect of modernity: a clear distinction between the dimensions of the sacred and of the secular.

The situation, however, was even more complex, as the Christian scriptures themselves were a two-tiered corpus, encompassing both the Old and the New Testaments. The New Testament offered the key to the correct understanding of the Old Testament, which represented a typos, or figura, of the momentous events described in the Gospels: under the Christian text, the Jewish sub-text. Like the Greco-Roman culture, the ancient history of Israel should be deciphered as the hidden text in the divine palimpsest, under the fresher ink of the deeds of Jesus and of the early Church, verus Israel. We should at least note here briefly that the Christian attitude to the Hebrew Bible (which was always read in translation) entailed a new attitude to historia sacra itself, and hence a new pattern of historical memory. Memory, in such a system, became less the remembrance of things past than the uncovering of things hidden and retrieving secret truth under lies apparent, the work of the devil. Gnosis, in that sense, and the dualist, esoteric movement that it represented, sought to retrieve the hidden memory of the divine message of salvation (Tardieu, 1988). In the complex hermeneutical system of early Christianity, religious memory meant actualization, and transformation, rather than conservation. Some key texts, such as Psalms, were memorized and recited both in public ritual and in individual prayer and meditation—in particular, by religious virtuosi, the monks, to use Max Weber’s parlance. This type of “reading” can be characterized as “intensive” or meditative, rather than as “extensive” or discursive, as reading had usually been in the pagan world (Stroumsa, 2008).

The paradigm shift we can discern in late antique religion is paralleled by no less dramatic cultural changes, through the technologies of both writing and reading. These too would have a major impact on the transformation of religious memory. Literary texts had usually been written on scrolls, of parchment or papyrus. From the first century CE on, the codex would more and more become the more common support for such texts. It should be emphasized that the Christians seem to have adopted the codex much faster than any other group in the Roman Empire. At the same time, people were learning to read silently, rather than aloud, as had been the norm in the past—although moving from loud to silent reading, as a process, would take centuries. Silent reading also meant its privatization and internalization and facilitated memorization of selected texts, while codices permitted a much more efficient intertextuality, through easy moves between passages distant from one another in the text. The varied body of religious texts, composed of both the revealed scriptures and their interpretive texts (such as commentaries, theological tractates, spiritual works, etc.), is the object of what we may call explicit religious memory.

But there is also another system of religious memory, which is not established upon written texts, but upon oral traditions and rituals. Indeed, we should not forget that in antiquity, there was only a small minority of literate people in the population. In such a world, oral and written religious traditions functioned simultaneously, at different levels, or in different registers, as it were. While oral religious traditions (as reflected in myths, enigmas, proverbs, fables, and stories) had been the norm in pre-literate societies, a mixture of oral traditions and written texts composed the treasure to be stored in explicit religious memory (Borgeaud, 1988b). Moreover, since the times of pre-literate societies, rituals were performed through a memory of the tradition as it was accomplished in previous generations. This second system of religious memory may be called implicit religious memory.

The two systems of religious memory function together, as two complementary systems, the one helping to the consolidation of the other. Ritual patterns are reinforced by thought patterns, while theological conceptions are reinforced by ritual behavior. In a sense, the two systems of religious memory function similarly to the two systems of declarative memory, explicit religious memory being parallel to episodic memory, while implicit religious memory is parallel to semantic memory.

Throughout late antiquity, the two systems of religious memory functioned side by side, complementing one another. Moreover, in the first Christian generations, we can observe the passage from personal to collective memory—in particular, of the life and death of Jesus as well as his teachings, which had until then remained oral. Jesus’ first disciples had a personal memory of their hero, and this memory soon became the kernel of the religious memory of the religion in the making (Thomassen, 2010). This memory as told orally was put to writing in the various Gospels, those which became canonical and the many more which were rejected by the orthodox tradition as apocryphal. It first became, for two or three generations, what Assmann has proposed to call communicative memory (as long as there is someone alive who still possesses a personal memory of the event or the person), before it became cultural (or religious) memory.

Late antiquity brought a deep transformation of religious memory, in particular, with the birth and growth of both Christianity and Rabbinic Judaism (Mendels, 2004). The importance of memory in ancient Israel, since the Biblical injunction: zachor! (Remember!), and then in Judaism through the centuries has long been noted and has often been studied (Yerushalmi, 1996 [1982]). As is well known, the Eucharist, the first and central Christian ritual, a representation of Christ’s sacrifice, was called anamnēsis (memory, in the sense of “recollection”) in the ancient church. The use of this word is significant, as it highlights the power invested in memory in antiquity (Basset, 1988; Itter, 2009).

While the first generations of Christians were learning to transform their oral memories into texts (Gerhardsson, 1961), the Jews were accomplishing a no less alchemical transformation of their religious memory. Now that their temple in Jerusalem, with its daily blood sacrifices, had been destroyed, they reverted to oral teachings in order to preserve the uniqueness of their scripture, the Torah. While Christian late antique intellectuals, whom we call the Church Fathers, were busy writing the huge corpus of Patristic literature, in a number of languages (Greek first, then, mainly, Latin and Syriac), the Rabbis were redacting orally and memorizing their oral commentary on the Torah, the Mishna, and then their legal discussions and other hermeneutical interpretations on the Mishna over a few centuries, the Gemara (both texts constituting the Talmud, the Oral Torah). Thus, both sister religions were developing two highly different conceptions of religious memory. For the Christians, memory is essentially the “memory of God” (mnēmē theou), while for the Jews it remains, fundamentally, that of God’s “great deeds” in history since the creation of the world. For the Christians, Jewish memory was all storage, no recollection, and was therefore frozen, irremediably belonging to the past. For the Jews, the Christians could not have any “storage” memory of Israelite history; hence, their “recollection” was pure imagination, no memory. In fact, both Jews and Christians were able to develop a complex system of religious memory, but the emphasis was very different for the ones and the others.

To summarize, I have argued that there exist, concomitantly, two different kinds of religious memory, implicit and explicit religious memory, and that these two kinds of memory are similar to the two types of declarative memory (semantic and episodic). Moreover, they seem to be also parallel to Whitehouse’s two modes of religiosity (imagistic and doctrinal). In these different couples, the two systems of memory function together, one consolidating the other. In late antiquity, the development of the religions of the book (those we call today the Abrahamic religions) both permitted and demanded the emphasis of religious memory to move from ritual and orality to hermeneutics and textuality. This passage, it seems to me, represents the major transformation of religious memory in late antiquity. Religious memory, from being essentially ritual, or imagistic, and intuitive, became consolidated as essentially theological, or doctrinal, and discursive. This transformation took different shapes in Judaism and in Christianity, as in Rabbinic Judaism hermeneutics remained to a great extent oral (the Talmud, which was not redacted for generations, is also called “Oral Torah”), while the Church Fathers produced a huge literary production: one has even been able to speak about the early Christian “logorrhea” (Harris, 1991). If my claim above about the dialectical relationship between personal and social memory is convincing, then the history of religions might be read, like the psyche, as a “divine palimpsest” retaining traces of the whole past. The early Christians, who thought of themselves as verus Israel and claimed to replace vetus Israel on the divine parchment, never erased the ancient text. In late antiquity, a number of religious groups sought to define their identity through the making of the Abrahamic religions in late antiquity (Stroumsa, 2015). The Qur’an can only be understood as a palimpsest written over the erased texts of both the Old and the New Testaments.

In a recently published bestseller, the psychologist Daniel Kahneman has shown that there are two different systems of thought patterns, one more intuitive, or fast, while the other more analytical, or slow. Kahneman argues that we use these two different systems of thought together, simultaneously, and that they are constantly interacting (Kahneman, 2011). It would be a promising task to check whether a similar dual functioning exists between two complementary modes of religious memory, one helping the other in the consolidation of religious memory.

Although they are not always conscious of it, the neuroscientists who speak of two systems of memory are deeply indebted to Freud’s study of the unconscious and to his suggestions on how to bring back to consciousness facts and feelings repressed into the unconscious. There are clear historical and personal links between Freudian psychoanalysis and the contemporary biological study of memory, as revealed in Eric Kandel’s fascinating autobiography (Kandel, 2006). For Freud, repression (Verdrängen) and neurosis at the personal level were reflected in religion at the collective level, with its conscience of guilt entailing monotheism. Freud thought that the memory of the original murder of the father was never totally eradicated, and that it remained, in the unconscious, a most powerful motor of both personal psyche and collective identity. For him, neurosis was the core of what makes us civilized, and religion is to society what neurosis is to the individual. Freud, indeed, knew that nothing is ever fully erased in the psyche. What had become illegible always left traces, and only their decipherment, the work of memory (Erinnerungsarbeit), could reconstitute the original text under the palimpsest (Freud, 1974). Neurosis might be synonymous with the human condition. But if it cannot be avoided, it may perhaps be managed. There exist two major illnesses of religious memory: atrophy, or amnesia, when there is too little of it, and cancer, when it overgrows and leaves no place for forgetting and building of the new. An essential character of religious memory is its elasticity. The elasticity of religious memory in scriptural religions is provided by their system of hermeneutics and by the interaction between the two systems of memory. The constantly changing balance between remembering and forgetting, between the old and the new, is the secret motor of transformation in religious history.