Marcel Mauss’s essay On Prayer : an important contribution on the nature of sociological understanding

Although there has been increasingly sophisticated discussion of Mauss’s work in recent years [1], his uncompleted thesis ‘On Prayer’ has not featured prominently in these accounts.

This neglect is unsurprising, for the text was printed privately in 1909, and remained extremely rare until it was put into the public domain by Karady’s edition of Mauss’s Oeuvres in 1968. A recent translation [2] has drawn attention to the text, and raises the question (for English-speaking sociologists) of what significance should be attributed to the essay. Does this work add anything beyond the already-published materials, or is its neglect deserved ?
The work consists of two parts, a lengthy attempt to define the topic and an investigation of the relevant Australian ethnography with this definition in mind.

The interest of the piece lies principally in the first part, an essay that rehearses Mauss’s views on the objects and methods of sociology while applying them to the subject of prayer. Mauss indeed portrays the Australian material as an exemplary case study rather than a topic in its own right ; furthermore, subsequent ethnographies demanded its revision [3].

In this paper, then, I shall confine myself to a consideration of method, and only consider prayer insofar as it helps us understand this topic (except in an Appendix). For the force of his observations on prayer is entirely bound up with considerations of method, as Mauss proposes a dialectical relation between his method and object. By ‘dialectical’, I mean his thought is critical and recursive. My aim is to expound Mauss’s text to bring out the structure of the argument, following this dialectic at work. I offer a careful textual reading, although the reader should not underestimate the degree of interpretation in the account proposed. There is indeed a dilemma in such a reading : it is my contention that in this text Mauss makes a decisive move towards identifying a logic specific to the nature both of sociological objects and their understanding. Yet it is important not to exceed the text and the intentions of the author ; my task has been therefore to follow the moves being made, to do justice to them, and to point to their wider significance. For this reason, I do not make much reference to wider debates, either of the period in question or contemporary ; my aim is narrower, to establish a case for the importance of the text as a contribution on the nature of sociological understanding.

I will make two related claims. First, this essay represents Mauss’s most sustained engagement with issues of theory and method, and as such commands our interest, for Mauss’s contemporary significance as a social theorist is increasing rather than diminishing. It is worth remarking at the outset that Mauss develops a technical language adequate to his purposes, but one always expressed in terms borrowed from ordinary language. It is only once we have mapped out this language and the structure of argument it embodies that we can begin to explore the role the essay plays in the development of Mauss’s thought.

Second, although Mauss’s thought is presented in a scientific, positivist language, it in fact radically breaks with this framework, proposing a sophisticated historical account of social organization and its understanding, with a revisable dialectic of observation and explanation. In order to explore this break I make a parallel with the work of the philosopher of science, Gaston Bachelard, who was attempting a similar project in the same period. This detailed understanding of the historical nature of social intelligibility constitutes Mauss’s potential to contribute to contemporary debates ; I shall sketch an outline summary as a conclusion. This reading may also allow us, in the realm of the history of ideas, to see a mutation in the development of the Durkheimian project, a shift from the framework of the Rules to that of the Elementary Forms, and to estimate Mauss’s share in it.
This essay then is a preliminary sketch, with these aims in mind.


Mauss’s essay attempts to claim ground and to construct on it a controlled, reasoned, and therefore reliable, way of understanding social phenomena. Mauss is concerned with the possibility of establishing sociology as a ‘science’, although his concern is with an interpretative discipline appropriate to its object, rather than with a positivist science imagined on the basis of a model of the natural sciences.

Even this initial act of clarification points to the polemic edge borne by Mauss’s project : although he rarely describes his opponents in detail, a considerable part of his work is an effort to distinguish what he is saying from the claims of contemporary theorists. I will begin by following him in this task of identifying targets, for this is usually a good way of understanding an author’s intentions and to gain a sight of the core of his project.

Prayer considered as an institution.

Mauss’s first move is to identify prayer as ‘an institution’, a social reality comparable in kind to such collective facts as language, economic practices, or the family. A single institution, he writes, ‘can fulfil the most diverse functions… [and] take on multiple forms while still being itself and without changing its nature’ [4].

This identification is in part a polemic against an individualistic or psychological definition of prayer (as a person’s communication with God or the fulfilment of some mental need). But the concept of an institution is conceived initially against other targets, on the one hand the reduction of human practices to a concept, accessible to reason, and on the other, the notion that human realities are incapable of comprehension, being too complex and without an underlying order. Or, to give them labels, the one an idealism that seeks an essence, the other a materialism that cannot go beyond a crude empiricism.

On the basis of the conception of prayer as an institution Mauss makes countervailing claims against these two reductions. He proposes, first, prayer may be the object of a sociological demonstration, showing how the sociologist’s powers of abstraction can legitimately be employed on the concrete and changing material of human practices. In other words, he claims that the human real is intelligible, beyond the projections of idealism or the confusions of materialism.

But second, prayer is also ‘one of the central phenomena of religious life’ [5]. Mauss makes two points concerning its centrality. First, prayer is a religious form in which representation (and, in particular, speech) is intimately connected with action and, because of the role of speech, their interrelation can be investigated with greater precision and, indeed, ease than is the case with most religious phenomena. The relation is characteristic, and opens up to examination the solidarity of myth and ritual.
And then, its centrality is revealed as an indicator of the stage of advancement of a religion. Most rites, Mauss suggests, regress as religions develop ; he mentions dietary prohibitions and sacrifices in this regard. But prayer ‘shows an uninterrupted line of development and ends by overrunning the whole ritual system’ [6]. We might therefore, he suggests, claim that the history of religion is more or less that of the advance of prayer. Mauss identifies two major trends in the development of religion, from the mechanical to the spiritual on the one hand, and from the collective to the individual on the other, and argues that prayer is a major cause of this double evolution.

Such a developmental perspective allows Mauss considerable benefits in terms of clarity of presentation, but at the same time, offers hostages to an evolutionist interpretation. We should note two things at this stage. Mauss proposes, within such a complex process, to confine himself to the elementary forms of prayer, to see how prayer comes into being, and what are its first transformations (i.e. its principle divisions). That is the significance of his recourse to the Australian materials. And he sets aside the other topics – the trends of spiritualization and individualization, which he attributes to Indian and Semitic religions respectively – as well as the additional matter of regressions or reversals of these trends.

But further, it is worth noting that his concept of an institution, with its negative definitions, sets aside the grosser charges of evolutionism. We are dealing neither with the transformations of an essence, nor with the formless permutations of a contingent history. The elementary forms have a logical status which will have to be clarified, and will emerge in the course of the definition of method, but is not easily to be subsumed under the evolutionist rubric. Indeed, to apply any simple notion of evolution to Mauss’s thought in an uncritical fashion would be to offer an example of the recuperation of a scientific concept by a philosophical notion, a topic to which we will return.

The limits of theological and other accounts of prayer.

This largely negative work of clarification continues when Mauss reviews previous writings on the topic. He concentrates on the failure of these various attempts to grasp the nature of the object – as institution – and therefore on consequent questions of method. Accounts in the human sciences, he claims, fall into one of three groups. The first, anthropologists and ethnographers, have largely neglected the topic of prayer. The second, philologists and historians, have approached the topic, but never with an eye to what we may call necessary connection. Philologists look to the meaning of words, and not to their efficacy. Similarly, historians look to antecedents rather than causes, and do not have recourse to theory, principles or laws (even when, on occasion, they resort to dating from internal evidence), working instead by relations of resemblance and contiguity. These concepts of necessity and efficacy will recur.

The third group comprises theologians and philosophers, who have concerned themselves directly with the topic as such ; nevertheless, Mauss has three criticisms to offer of their approaches. First, their accounts are not scientific, but rather document the states of mind of the believer. The theories they produce are put to work in specific contexts, for example, in liturgical acts, and so their theories are phenomena, voicing indigenous beliefs instead of offering analysis. We might say, the accounts neglect their status as social products, assuming their necessity and efficacy of which there is question. Second, the accounts refer to specific historical situations, and so treat individual belief as characteristic, which they take to be the single modality in which religion expresses itself. The contemporary context is thereby projected as a universal condition. Third, these approaches create ‘historical’ accounts by illustrating general opinions by examples.

Putting these criticisms in general terms, these theories do not respond to the facts ; the facts they choose to demonstrate these theories are not connected by necessary connections (and so do not constitute ‘proof’) ; and the terms in which the arguments are drawn up therefore reflect neither the necessary consequences of the method nor the natural relations of things, but instead reflect what Mauss calls subjective criteria, even current prejudices [7].

He criticizes in these terms contemporary theological accounts of prayer, whether considered as conversation with God (and hence a secondary activity) or as the central activity of religion. In either case, the theory derives from within a contemporary Christian understanding, emphasizing the individual and states of belief, which approach vitiates the methods adopted in reading the material and, above all, permitting the ignoring of counter-examples. This is the context in which individualistic and psychological explanations emerge.

Against those who look to individual belief and experience as the subject matter defining prayer, Mauss presses the following two objections. In the first place, if we rely on religious experience, the subject matter is elusive, to be discovered by introspection, together with speculation about the introspection of others. And in the second place, these experiences, however well described, may mislead us. ‘For we are always performing actions whose rationale, meaning, scope and nature elude us’ [8], and bringing these actions to consciousness may only end in self-deception. In brief, individualistic and psychological explanations cannot deliver a satisfactory sociological explanation.

Mauss’s relationship to philosophy.

It is worth clarifying our understanding of Mauss’s achievement and, in particular, his relationship to philosophy. He proceeds dialectically, with successive refinements of his object. He starts with the notion of prayer as an institution, moves to prayer as a social reality, and will conclude in due course with a definition. Along the way, a number of philosophical targets have served to give shape to his ends, most notably with polemic against idealism, essentialism, evolutionism, individualism, psychologism, empiricism, and materialism.

We might say from this list that Mauss does not precisely engage in philosophy as such ; rather, he identifies positions from which he differs. In this, he resembles a contemporary, Gaston Bachelard [9], who wrote on the relation of science to philosophy and similarly offered a ‘non-philosophy’, lying in another dimension or axis. Bachelard announced the time of the ‘Anabaptist philosophers’ [10] – Anabaptist in that they forswear all the beliefs and dogmas of traditional philosophy. Mauss is far more circumspect in expression, but his aim is likewise to establish thinking elsewhere, on the territory of ethnographic knowledge, which has its own demands and practices. Scientific knowledge of the social is produced in opposition to existing philosophies, not with their aid, and the truth of a (scientific) sociological statement is not founded on any philosophical guarantees.

The role of this non-philosophy is not, then, as a spokesman on behalf of various ideologies vis-à-vis the sciences ; rather its task is to distinguish within a given account what derives from scientific practices and what originates in what we might call ideological discourse. Mauss expends a good deal of effort outlining how this is to be done in practice, as we shall see. The point of the polemic is to identify the various forms of philosophical accounting, and see how they condense at various points and in different forms as knowledge of social institutions progresses. A marker of philosophy is that it tends to form a system, to effect a closure, generating a series of questions and answers, and by so doing, ignores or represses the structure and life of the real – of history and the social order.

Yet these accounts arise in continuity with what we may call everyday life, or common sense ; effectively, philosophical accounts of social life extrapolate from and systematize aspects of the social experience of the theoretician. Philosophers create broad theories out of local experience, and the theories generated thereby are neither widely applicable nor reliable. These theories do not see to the mainsprings and underlying patterns of human behaviour, and so cannot create sociological understanding – and yet, they are plausible, in that their accounts may be recognized in part by others who share the experience being projected.

It is for these reasons of continuity with experience and plausibility that Mauss does not wish to break abruptly with such positions, and continues to employ ordinary language rather than invent new terms. This is an aspect of the subtlety of his approach. Yet the ordinary language must be employed decisively in another fashion, if it is to yield understanding rather than simply testifying to experience. There has to be an effective break with common sense, and new bases of understanding created.

To develop this idea, we may pursue the parallel with Bachelard by borrowing an image he uses of a ‘philosophical spectrum’, which in his account runs from idealism through conventionalism, formalism, positivism and empiricism to realism, each form changing into the next. And he marks the as-yet-unformulated concepts he wishes to mobilize by putting a pair of brackets in another axis, to either side of this spectrum [11]. We might compose a similar image for Mauss, in like fashion :

( ) Individualism ( )

Once again, the contents of the two empty brackets are to be determined.

An ‘open’ approach.

The new function of philosophy is to escort the progress of the sciences, which in the human sciences fall under two principle heads : history and sociology. The approach will be ‘open’, in contrast to the closures effected in the other axis. For as a human science progresses (and, as we have seen, Mauss is attentive to the range of human sciences), it offers new opportunities for being misread by philosophers, as well as producing new values and understandings. Its progress cannot be anticipated by any theory or ideology. The open approach also takes the various human sciences seriously, in that the different branches will develop unevenly.

What is the basis of this openness ? It is by making understanding thoroughly historically situated, in both human context and time. Instead of the relation subject-object, the comprehension of an inert entity by an active mind wholly external to it, Mauss substitutes what we may call the relation ‘comprehension-extension’. In a science, in his sense, the object of thought is constructed, and comprehension is a function of the concept which allows the object to be thought.

We can understand, then, how this approach avoids any trace of ‘scientism’, of imagining that a positive knowledge can be achieved about the world, natural or social, independent of the processes of gaining that understanding. In this approach, there is no philosophically defined world of things in themselves which empirical science apprehends piecemeal or asymptotically, nor is there any philosophically defined consciousness to which all scientific statements can be reduced.

In considering the physical sciences, Bachelard fills out the two empty brackets with a dialectic between theoretically informed experimentation and experimentally informed theory, which he calls respectively an ‘applied rationalism’ and a ‘technical materialism’. Scientific phenomena are produced in a ‘phenomenotechnics’. Likewise in this respect, the human sciences themselves produce their objects and phenomena, in a dialectic in time between theoretically formed experience and experientially shaped theories, the objects being given material form in textual (ethnographic) ‘proofs’. Mauss gives the two empty brackets names taken from common language, in accordance with his distrust of neologisms : ‘observation’ and ‘explanation’.

The role of definition.

Mauss is able to situate his thinking in another dimension to that of the ideologies supported by traditional philosophy (the spectrum summed up here by the term ‘positivist’) because he follows Durkheim, whom he believes has constituted sociology as a science. This is the point at which Mauss’s historical thought sits most uncomfortably with the positivist language of contemporary sociology, for the science he pursues is precisely not a ‘natural’ science. Yet he refers to ‘the principles laid down…in Règles de la méthode sociologique [12], and adopts the tripartite division of definition, observation and explanation outlined there. What is the purpose of ‘definition’ with respect to the other two terms ?

The role definition plays is quite precise : it constitutes the initial claim to be a science, and thus forms a break with philosophical accounts. Its purpose is to stimulate research, and it achieves this through preliminary delimitation of the field of observation, in order to deploy and test hypotheses. Research will allow alteration of the definition, extension of the field of facts and refinement of hypotheses.

By definition, at this stage, Mauss simply refers to the identification of a few objective features, ‘according to outside signs’ [13], which will allow us to avoid impressionistic and vague debate over imprecise objects. He turns to the task of defining prayer – selecting its objective features or outside signs – in the third chapter (see Appendix). But first, he fills in the two concepts that lie on either side of the philosophical spectrum : the dialectic of observation and explanation. In doing so, he abandons natural science as a principle point of contrast, and turns instead to history as a conversation partner.

Observation : the tasks of criticism.

Facts are never crude, then, but always constructed. In the case of social studies, the facts are contained in documents, in historical and ethnographic records. A special procedure is needed to select and construe them, and this procedure, borrowing from history to apply to sociological thought, is criticism. Mauss would have considered sociology as a discipline impossible to conceive without the advances made by contemporary historical thinking. He claims to offer examples of such critical thinking in the second section, concerning Australian ethnography [14].

Sociology, however, while adopting critical methods from history also elaborates them, making them clearer, more explicit and more rigorous, and extending them while directing their use towards clearly defined goals. Historians, in contrast, tend to use these methods as ends in themselves ; they establish facts at the expense of systematizing what they have discovered.

Mauss elaborates the tasks of criticism. On the one hand, there is the business of external criticism, or evaluation of the document : establishing how much error it contains by taking into account its present condition, the means of transmission, its date, sources and so forth. With specific respect to ethnographic documents, this evaluation will focus upon the author : the role of interpretation in the observation, evidence of the author’s views, competence (for example, in translation), and conscientiousness, and whether there is countervailing evidence within the text. And one has to assess the authenticity of the document, the conditions under which it was acquired and the status of the informant (person, place, time and occasion). All these protocols represent good scholarly practice, but they were not universally practised at the time Mauss was writing, and there is still a point to enumerating them now.

Then, on the other hand, there are matters of internal criticism. One must subject the facts reported by the document to scrutiny, to see how the date of the document relates to the age of the facts reported ; to situate the event reported with respect to its milieu ; and to break it down into its component parts. Taken together, these pages serve as a methodological preface to The Essay on the Gift [15].

What permits criticism ?

More strikingly, Mauss points out that, in order to perform internal criticism, history has to use a number of techniques all of which are based in the sociological principle of the interdependency of social phenomena. In other words, the kind of understanding that sociology achieves is a function of the nature of its object ; this is a significant formulation. For example, textual criticism may date a prayer according to the antiquity of the verbal and syntactic forms it uses, appealing to the institution of language. Or it may perceive a text as presupposing certain events which themselves demonstrate a certain level of development. Or it deduces an evolution from impure to pure forms (or vice versa). In each case, the assumption is that other social institutions leave their mark, and that they influence the thing in question. Observation, therefore, contains two sources of insight : the active power of imaginative thought on the part of the investigator is matched by the capacity of social institutions to yield themselves to investigation.

This sociological principle that permits criticism is better employed when articulated, for it permits a different interpretation of the data. Mauss cites biblical criticism in this regard, which tends to separate out contradictory texts as being of different date and origin, and hence interpolations. This, he claims, misunderstands how social phenomena are joined together : one religious idea or act can carry multiple meanings, and the one institution can serve varied functions and produce opposed effects. He therefore situates himself in opposition to source criticism – one should rid oneself of useless questions about authorship – and favours what will come to be called form criticism [16]. Prayer is a social institution ; the question is ‘which group used a prayer ?’ (not who composed it), under what conditions, and at what stage of religious development ? We are concerned with the received text, not with the original, with the ideas of a group of people, not one person’s ideas, and with the liturgical situation of a text, rather than its date of origin [17].

Mauss sums up the question of observation in this way : the task of the sociologist is to strive to see facts in their detail and to relate them to a well-described context. That context is primarily the whole range of social institutions within which any one fact takes its place, and from which it derives its meaning. Here we have foreshadowed the ‘total social fact’. In this perspective, criticism leads not to commentaries and discussion for its own sake (which is where much history stops), but prepares the way for explanation.

What kind of knowledge is produced by critical thought ?

Before we turn to explanation, it is worth elaborating the characteristics of the kind of knowledge produced by critical thinking, for it allows us to distinguish very clearly Mauss’s account of sociological knowledge from positivist (common-sense or scientific) accounts. Texts, and the facts they contain, are the products of social contexts, to be deciphered by the process of criticism, which is an activity of a comparable kind. Perhaps the crucial point of critical thought is that it actively produces its object in a dialectical process of intelligent reading. There is no real to be apprehended : critical thought instructs itself by what it constructs. This is very different to a common-sense approach that claims to apprehend the world of phenomena.

Because of this, criticism has to arise in a break with its social context ; this is the importance, as we have seen, of the act of definition, which in turn allows the identification of obstacles to criticism, secreted by the uncritical thought habits of everyday life and experience. Two things follow. First, such a break is not once and for all ; rather, there will be successive re-castings of sociological understanding, once constituted, as it overcomes obstacles. The criticism produced will always be provisional, and the understanding gained will be of its moment, not of universal validity. This contrasts with a view of scientific knowledge as certain and universal, as well as clear. And the clarity produced by criticism will be an intellectual gain, rather than an experience of transformed vision. Understanding is constructed, and contextual, and not to be confused with common sense.

And second, given this understanding of the nature of criticism, critical thinking is primarily a collective possibility, rather than a heroic individual act. Critical thought will progress, through observation and definition, but by so doing it constitutes collective forms of investigation, a common practice that goes without saying. Critical thought is itself a social practice, one that develops and evolves, neither determined nor arbitrary in its forms. Mauss gives a sociological description of the sociologists’ vocation.

As it proceeds, this social practice of criticism will, in the first place, be able to characterize the nature of the obstacles that recur, and that will, because of the inter-penetration of social institutions, continue to emerge within the constituted science. These notions - such as the unity of thought (as opposed to the plurality of thought), the unbroken development of knowledge (as opposed to the break), or a simple empirical object to be grasped (as opposed to its social construction) – permit the versions of reality proposed by the various philosophical accounts, and also permit the positing of an individual psyche which is to grasp that reality.
In the second place, the social practice of criticism should be able to think through the transition from one mode of understanding to another. In other words, it can trace the histories (in the plural) of the sciences. The history of sociology, for example, is born by constituting a body of concepts, with their rules of production, and develops through forming concepts and theories. The history will be recurrent, for newly discovered values and solutions call back into question past judgements and facts. In this view, history progresses by reorganizing its principles. It is what we might call a current history, biased towards the future ; the past is the history of recognized errors and the present account is what has been sanctioned : obstacles identified and epistemological acts of explanation defined. What are we to say of the latter ? What is an explanation ?


Just as observation is the social production of critical concepts to be applied in description, so explanation concerns the construction of the interrelations of the facts described, the discernment of the successive orderings contained in the situation under consideration. ‘To explain is to establish a rational order between the facts, once they have been ascertained’ [18]. In this way, the total social fact belongs simultaneously to the indigenous society and to the investigating community ; it is an effect of the interaction of the two. The solutions found to the problems of the assemblage of concepts must integrate the theoretical conditions of these concepts’ formulation.

Mauss proposes, then, that the sociologist goes beyond descriptions of prayer in various societies, and investigates the relations that link the facts of prayer both to one another, and to the other facts that condition them. ’It is a matter of constructing a hierarchy of mutually illuminating facts which, taken as a whole, constitute a theory of prayer’ [19].

Mauss suggests that there are two sorts of systematization. The first is to formulate a generic idea or schema, and to see how different species arise within it. This is the method employed in his ‘Essai sur la nature et la fonction du sacrifice’ [20], where in practice a specific instance (Vedic sacrifice) is taken as a particularly complete exemplar, which will serve as a model for other cases.

The second, instead of starting from the genus and proceeding to the species, starts with the most rudimentary forms, and passes from these to more developed forms. The first, Mauss says, is a single logic, a hierarchical series of ideas set outside time and space. The second is a genetic explanation, where we are dealing with successions in time, and where the aim is to retrace their order. He terms the two approaches ‘schematic’ and ‘historical’ (or ‘genetic’).

We need to be alert to in what sense ‘historical’ is meant. For we are not reconstructing facts but explanation. The schematic method works best when phenomena assume definitive forms so that their essence can readily be discerned [21], and variations are limited. Prayer is not like this, but is in a constant state of development, and the historical approach is appropriate to forms that vary greatly over time. And, indeed, the two approaches are not so opposed, for a genetic approach can prepare the ground for a schematic explanation [22].

This is what Mauss means when he uses the term ‘evolution’, as when he writes : ‘the first step in a genetic explanation is to establish a genealogical classification of the types of prayer, that is to say, to organize the types by arranging them in the order of their evolution’ [23]. Evolution describes a logic of connection of forms. The criterion Mauss adopts is the interpenetration of forms and their progressive separation : superior forms are not more complex than inferior, for the same elements are all present in the latter, but conflated [24].

This conception of a genetic series of societies, constructed along an axis of variation concerning the dispersion / condensation of institutions, was outlined in Durkheim’s Rules of Sociological Method as the ‘method of concomitant variations’ ; it formed the basis of Primitive Classification, and indeed of the Elementary Forms, and it was to have been the organizing principle of the thesis On Prayer, alas never completed.

Comparison and necessary connection.

The second step, then, is to determine the causes of development of the series. These forces are not shown by a genealogical classification, even if it points to the original causes ; they are not internal. Nor can they be discovered in the properties of the human psyche ; the general laws of representation are the same everywhere, and cannot describe the necessities that give rise to such a diversity of types.

These causes lie instead in the social milieu : there is a necessary link between a given prayer, a given society, and a given religion. This is the other side of the interdependency of social institutions invoked as the possibility of criticism, and we should draw attention to it as the complement corresponding to the latter. Together they form the heart of Mauss’s exposition of the basis of social intelligibility. There are therefore patterns to be discerned in the material, and to make sense of genetic explanations one has to resort to comparison. This is the other face of explanation. We might say that the two faces of explanatory procedures, systematization and causation, correspond in a formal way to the internal and external aspects of criticism. And comparison is embodied, for Mauss, in the carefully formulated, theoretically dense sociological terms he develops that serve to organize the indigenous practices and categories from which we begin the process.

In the processes of comparison, differences are more important than similarities. The latter allow the sociologist to establish types, but differences are crucial to establishing causes. Attention to differences allows the sociologist to go beyond superficial resemblance (and difference) in the establishment of types. But more important, in the establishment of causes, the sociologist looks to differential features of the wider social milieu to discern the causes of a local phenomenon such as prayer. This method is all the more appropriate when considering developments, for one looks to parallel changes in the corresponding social milieux.
Even here, one must distinguish between the usually indirect influence of the general institutions of society on prayer and the direct action of other religious institutions (such as myth, or the priesthood). In brief, in the comparative method, there are processes of separation of elements and their reintegration, in a discernment of causes or necessary connection.

The necessary connection or cause discerned, then, lies neither in the empirical facts, nor in the mind of the observer. Nor does it lie precisely in the constructive construal by the collectively formed mind of the sociologist of the significant constructions of the subjects in whom he or she is interested, although we are concerned with the dialectic of the construction of understanding and the recognition of processes of organization discerned in the material. Rather, it lies in the experienced interpretation of the differences between comparable cases.

To summarize these principle concepts in a phrase, the two most characteristic forms of the activities of observation and explanation are criticism on the one hand and comparison on the other. And we can claim that, by proceeding as he does, Mauss creates a distinctive framework for these commonplace ideas, giving them a closely defined content that merits our continued attention.


In practice, as Mauss points out, the terms of the dialectic are ‘naturally combined’ [25]. Only the provisional definition, on which all else follows, needs to be considered separately – and this Mauss does in the last chapter of the section. I consider this account of prayer in the Appendix, for it has its intrinsic interest.

We may sum up Mauss’s achievement following the three headings he takes over from the Rules. First, he presents the act of definition as a break with the various modes of common-sense knowledge, which are conceived as obstacles to understanding, although social products. In brief, social understanding has a different form to that modelled on (natural) scientific thinking.

Then, concerning observation, Mauss claims that this different form is termed ‘criticism’. Criticism is permitted by the interdependency of institutions, which he makes the central feature of his argument. In my introductory remarks, I suggested that Mauss is identifying a logic appropriate to the human sciences ; the overlap of classes that describes this interdependency points to a non-traditional logic, one that does not depend upon the basic principles of identity and non-contradiction. This is not the place to elaborate the argument, but simply to note some of the consequences in the text. In terms of the object of sociological knowledge, one does not simply break down social phenomena into their simple elements, but is concerned also with totalities and relations. Further, the investigation of such phenomena shares in the characteristics of the object, so that both social life and knowledge of it are revisable and recast at intervals, and the transition of forms may be understood from the perspective of the present as a history of error discarded and gain of understanding.

Finally, explanation relates to the formation of what we may call a scale of such forms. As the causes of development are to be found in the interdependency of institutions, explanation is generated by comparison between adjacent forms on the scale. This account then builds upon the feature implied by the overlap of classes, that differences in degree are not wholly distinct from differences in kind. In The Gift, Mauss will develop this kind of analysis ; the central characteristic of ‘total’ social phenomena concerns the degree of overlap of their institutions, and kinds of social formation are distinguished by the degree of separation out of these. This overlap structures the later essay as a whole. The point here is that Mauss has developed an outline account of necessary connection, conceived as causality or obligation which, as Durkheim remarks, is the mark of the social.

This summary is necessarily condensed ; I do not wish to go beyond Mauss’s text but to draw attention to what is at stake in it, using terms – non-traditional logic, overlap of classes, scale of forms – that he does not employ, and that are drawn from another source [26]. In order to judge the value of the essay with respect to method, a certain preliminary labour of exploration and exposition has proved necessary, both in order to understand what is at stake and, negatively, to avoid recuperating Mauss’s originality by reading him through philosophical lenses. Once this reading is begun, it becomes possible to place this essay in the development of Mauss’s thought, and to evaluate its significance, with respect both to earlier and to subsequent work. In particular, the essay may be read as a methodological manifesto fulfilled in the essay on The Gift, while at the same time offering re-readings of the earlier essays on the nature of sociology, magic, and sacrifice.

But the significance of this essay in particular is the complex historical account of social intelligibility it contains which, while apparently expressed in the terms of the Rules of Sociological Method, in fact proposes a more subtle and far-reaching view. Mauss offers a description of social understanding based in another form of logic to that of the dominant positivist account, a logic in which understanding is revisable, created through the dialectical interplay of observation and explanation, the interplay between them deriving in the end from the specific feature of social formations – the interplay or overlap of their elements – which is repeated in the form of their apprehension.

A tentative judgement on the work as a whole might then be as follows : the exposition of method is outstanding, unrivalled elsewhere in the corpus of Mauss’s work, and worthy of detailed study. It is, moreover, of continuing importance, for the ideas it contains could contribute to contemporary debate on the nature of the social sciences and their objects.

Appendix : the ‘Initial Definition’ of Prayer – Part I, chapter 3.

Prayer as a social phenomenon.

What then is the key to grasping the positive nature of prayer as an institution ? In brief, it lies in the early formulation ‘prayer as a social phenomenon’ [27]. Mauss advances the following five ‘proofs’ that prayer is a collective phenomenon.

First of all, prayer is full of all sorts of elements whose origin and nature escape us. Myth and rite - or dogma and history - converge in even the simplest formula. As Mauss says of the phrase ‘In the name of the Father…’ ‘nearly the whole of Christian dogma and liturgy can be found closely entwined in this formula’ [28]. Dogmas, doctrines, historical claims, Church order (or the division of labour), and the task of the Church are all condensed in this phrase and its right use. Prayer is the product of a long history ; it is not just the effusion of a soul, but a ‘fragment of religion’.

Then, even when prayer is free, its content shows every sign of its social origin, in the language that is used, the topics it dwells on, the ideas it employs, and the sentiments it expresses.

Furthermore, prayer is social not only in content but in form ; its forms are social in origin and do not exist apart from some kind of ritual or formulary. The circumstances, times and places where prayers are to be said, and the posture to be adopted, are all laid down. The form of words, their meaning and order are all determined as appropriate by the community, and this remains true, in the sense of foundational, of the most individual of prayers.

Mauss draws attention, fourthly, to the prohibitions and other orderings that demonstrate the social nature of prayer : which fashions of speaking and chanting are authorized, and which are not ; who may say them, and who may not (women, children, lay persons…) ; which secular forms are taken up, and which are not.
Finally, Mauss indicates the relations of prayer to other collective phenomena, especially legal and moral formulae, in such forms as the oath, the solemn contract, and turns of phrase required by etiquette (whether in the context of leaders, kings, law courts, parliaments, or common courtesy). We are concerned here with the power of formal language both to express relationships that bind and to create them from nothing. There is, he states, ‘hardly any sphere of social life where prayer does not, or has not, played some part’ [29] : in family organization, through initiation and marriage ; in legal situations, through the oath ; in morals, through confession and expiation ; and in economic contexts, through effects of accumulation and productivity.

‘Prayer, Mauss concludes, is above all a social reality outside the individual, and in the sphere of ritual and religious convention’ [30]. And, he points out, the individual’s possibilities – for personal prayer, for example – are derived from this social reality. Some such possibilities may indeed fulfil a social function, such as in the place of free prayer in an established ritual, or meeting the need for innovation. And even the most individual figures concerned with prayer – priests, prophets, or seers – are not simply individuals, but are inspired, themselves expressing social forces.

Definition redefined.

Mauss’s first essay concludes with an attempt to offer a definition of the institution of prayer, based on his work of clarification, in which he has amply demonstrated that object and method are logically interdependent, and that the object can only be refined through the recasting of earlier steps.

The aim, he states, is to discover an external but objective sign by which the phenomenon of prayer may be recognized [31]. As we know, we are not supposing a social entity (or essence), distinct from the facts, but rather their system. And further, their system is neither precisely in the facts themselves, nor is it simply contributed by the ethnographer’s mind ; it is the effect of the collective intelligence of the people in question discerned by the ethnographer. It is a moral or social fact. Mauss therefore says, the word ‘prayer’ is simply a noun by which a set of phenomena is designated. And through comparison, since the term is general, one may abstract the certain particular characteristics that various sets of human activities share in common.

So Mauss can say, definition is a question of clarifying and distinguishing. It is not a matter of doing violence to current ideas of prayer, but of replacing personal impressions with an objective sign that will dispel ambiguity and confusion. This approach, he claims, avoids neologisms, the coining of new terms, on the one hand, and on the other, forestalls mere playing with words. It is an analysis conducted with ordinary language.

To define, however, means to classify, to place the term in question in relation to other, previously defined ideas. And Mauss concludes this reprise on the problem of definition by pointing out that the science of religion has not yet produced a methodological classification of the facts, so we also have tentatively to define the phenomena by means of which we may express prayer. Mauss therefore turns to consider the rite.

A definition of a religious rite.

Prayers are among the rites of religion. But what is a rite ? Mauss proceeds dialectically. Rites are actions ; they are traditional actions, in that they take a form adopted by the group or recognized by an authority. And they are efficacious, in that they achieve material ends (for example, making crops grow). ‘A rite is therefore an efficacious traditional action’ [32].

In the course of this progression, Mauss distinguishes rites from individual actions, and from other repeated, structured activities (such as play, or dancing). But by focussing on efficacy he has also encountered another anthropological crux, that between ritual and technical action. The latter is an effect of the material civilization of a society, the division of labour, technology, ownership and so forth, and is therefore a social characteristic just as much as is language, the legal system or religion. But because rites and techniques are so closely intertwined, they need to be distinguished.

Mauss suggests that the efficacy attributed to rites cannot be assayed by observing the nature of the actions and their real effects. Rather, the difference lies in the manner in which the efficacy is conceived : the effects are not mechanical, but of their own kind – sui generis. The effects proceed entirely from special powers that rites have the ability to set in motion [33].

Within the class of ‘efficacious traditional actions’, we therefore need to distinguish the nature of the forces mobilized. Mauss proposes essentially a distinction between magic and religion, between forces created by the rite, on the one hand, and on the other, forces that exist apart from the rite and that are worked upon by it. In the latter case, the power in the rite sets in motion other creative principles. In the case of magic, the rites involved are more coercive and deterministic, and the practitioners more individual. In the case of religion, the rites are more contingent, owing to the ability of the independent powers to resist them, and contain solicitations through offerings or respect paid, made by the collective group (or a representative) responsible for the cult of sacred things.

These two forms represent species of the same genus, and there are in practice many intermediate forms. Though they differ, one cannot completely separate magical from religious phenomena. But the rites of religion relate to the exclusively sacred character of the forces to which they are addressed. ‘We can define them [the rites of religion] – Mauss concludes - as efficacious, traditional actions which have bearing on things that are called sacred’ [34].

Prayer as a religious rite.

Mauss then claims, concerning prayer, that ‘a considerable group of acts which everyone calls prayers, show all the characteristics of the religious rite as we have just defined it’ [35]. He justifies this claim on four grounds. First, that every prayer is an act, ‘an effort of physical and moral energy in order to produce certain effects’. Second, it is a traditional act, in as much as it forms part of a ritual. This has been argued above. Third, prayer is efficacious ; this characteristic covers a range of possibilities, from constraint to adoration. But even the latter assumes a god with powers upon whom the prayer has effects. Lastly, its efficacy is addressed to religious powers, and so can be distinguished (in theory) from magical incantation, although the two cannot fully be separated (consecration, for example, invokes gods through magical acts).

Prayer, he concludes, can be said to exist every time we meet a text that specifically mentions a religious power, or which is performed in a sacred place, or in the course of a religious ceremony, or by a religious person [36].

Mauss moves to consider a further distinction within religious rites, between manual and oral rites. The former consist in movements of bodies and objects, the latter in ritual utterances. Prayer is clearly an oral rite, exploiting the power of expression that speech enjoys over gesture. But not all oral rites are prayers ; for example, an oath, the verbal contract in a religious marriage, the religiously founded wish, blessing or cursing, vowing, and dedication. In each case, although prayer may be involved, the principal effect is to modify the state of some profane body, which is to be endued with a religious character. A new quality is given to what is declared, wished or promised.

Prayer, by contrast, is a means of acting on sacred beings, which are influenced and changed by prayer. It is true that prayer has repercussions in the everyday domain, even benefits for the worshipper, but these are by-products. For prayer is addressed to religious powers, and it is by their mediation that the rite affects ordinary life. It may be concerned simply with the world of the powers : no result may follow. Although the distinction between prayer and other oral rites, as before, may not be so clear in practice, Mauss concludes his work with this definition, as follows : ‘prayer is a religious rite which is oral and bears directly on sacred things [the sacred]’ [37].


With respect to the work on prayer, his is by no means a radical account ; nevertheless, the clarifications produced by the path taken are instructive, and the claim Mauss makes as to the centrality of prayer to the development of the history of religions has yet to be explored. The text should serve as an encouragement to further ethnographic work. Furthermore, the challenge he poses to theological accounts of prayer in particular, as testimonies to experience from within the categories rather than analytic explanations, stands to be answered.

When we consider the relation of this essay to Durkheim’s The Elementary Forms of Religious Life, however, the balance sheet is less easy to draw up – although we know that Hubert, at least, thought Mauss should have been credited as co-author. Mauss’s thesis remained incomplete in two senses. In part, he never completed the work on prayer in societies other than the Australian Aboriginals’, allowing him to formulate a series by which to test his hypotheses. And in part, even with respect to the Aboriginal materials, because of new work produced by Strehlow and Spencer and Gillen. Durkheim drew heavily on these ethnographies. Perhaps it is significant that in so doing he rejects a central tenet of Mauss’s account, that religion may be defined in any primary way in relation to the idea of divinity. Durkheim states that ‘religion is more than the idea of gods and spirits’ ; this departs from Mauss’s account of religion ‘as an organic system of collective ideas and practices relating to acknowledged sacred beings’ [38]. Durkheim puts his emphasis on the power and structuring properties of classificatory systems, and moreover – whatever else he derives from Mauss’s work – scarcely mentions the topic of prayer. We might conjecture that the limitation that is expressed in the unfinished nature of Mauss’s project lies in this difference.

On this score, we conclude that while the topic discussed – prayer – may be flawed in its conception, it raises a number of extraordinarily useful points that need to be pursued.

Timothy Jenkins

Faculty of Divinity

University of Cambridge

Bibliography :

Bachelard, G. 1940. La philosophie du non. Paris : Presses Universitaires de France.
Bachelard, G. 1949. Le rationalisme appliqué. Paris : Presses Universitaires de France.
Chimisso, C. 2001. Gaston Bachelard. Critic of Science and the Imagination. London, Routledge.
Collingwood, R.G. 1933. An Essay on Philosophical Method .Oxford, Clarendon Press.
Derrida, J. 1991. Donner le temps. Paris : Editions Galilée.
Didier, G. 1993. Bachelard et la culture scientifique. Paris, Presses Universitaires de France.
Dzimira, S. 2007. Marcel Mauss, savant et politique. Paris, Éditions de la découverte.
Durkheim, E. 1907 [1895]. Règles de la méthode sociologique. 3è.éd. Paris : Alcan.
Durkheim, E, 2003 [1912]. Les formes élémentaires de la vie religieuse. Paris, Presses Universitaires de France.
Durkheim, E. & Mauss, M. 1903. ‘De quelques formes primitives de classification : contribution à l’étude des representations collectives’. Année sociologique vi : 1-72.
Fournier, M. 1994. Marcel Mauss. Paris : Fayard.
Hallward, P. 2003. The one or the other. French Philosophy today. Angelaki. Journal of the Theoretical Humanities 8 : 1-32.
Hénaff, M. 2002. Le Prix de la verité. Le don, l’argent, la philosophie. Paris, Seuil.
James, W. & Allen, N. (eds) 1998. Marcel Mauss : A Centenary Tribute. Oxford & New York : Berghahn.
Jenkins, T. 1998. Derrida’s reading of Mauss. James & Allen (eds) 1998 : 83-94.
Jenkins, T. 2003. Review of Mauss 2003. Durkheimian Studies/Etudes Durkheimiennes 9 : 90-96.
Karsenti, B. 1997. L’homme total. Sociologie, anthropologie et philosophie chez Marcel Mauss. Paris : Presses Universitaires de France.
Lecourt, D. 2002. L’epistémologie historique de Gaston Bachelard. Paris, Vrin.
Mauss, M. 1898. Essai sur la nature et la fonction du sacrifice. Mauss 1950.
Mauss, M. 1904. L’Origine des pouvoirs magiques dans les sociétés australiennes. Mauss 1968, II.
Mauss, M. 1909. La Prière. Paris : Alcan. Mauss 1968 I : 357-477.
Mauss, M. 1925. Essai sur le don : forme et raison de l’échange dans les sociétés archaïques. In L’Année Sociologique, ns 1 : 30-186. Mauss 1950.
Mauss, M. 1950. Sociologie et anthropologie. Paris : Presses Universitaires de France.
Mauss, M. 1968. Marcel Mauss : Oeuvres. 3 volumes, présentation de Victor Karady. Paris : Les Editions de Minuit.
Mauss, M. 1998. An intellectual self-portrait. James & Allen (eds) 1998 : 29-42.
Mauss, M. 2003. On Prayer. Oxford & New York : Berghahn. Translation of Mauss 1909.
Mauss, M. & Fauconnet, P. 1901. La Sociologie : objet et méthode. Grande Encyclopédie 30 : 165-76.
Morphy, H. 2003. Some concluding anthropological reflections. Mauss 2003 : 139-54.
Papilloud, C. 2003. Le don de relation. Georg Simmel, Marcel Mauss. Paris, L’Harmattan.
Parinaud, A. 1996 (1966). Gaston Bachelard. Paris, Flammarion.
Tarot, C. 1999. Du Durkheim à Mauss. L’invention du symbolique : sociologie et science des religions. Paris : La Découverte.
Tiles, M. 1984. Bachelard : Science and Objectivity. Cambridge, Cambridge University Press. Wunenburger, J.-J. 2003. Bachelard ou l’epistémologie française. Paris, Presses Universitaires de France.

Un message, un commentaire ?

modération a priori

Ce forum est modéré a priori : votre contribution n’apparaîtra qu’après avoir été validée par un administrateur du site.

Qui êtes-vous ?
Votre message
  • Pour créer des paragraphes, laissez simplement des lignes vides.

// Article publié le 6 novembre 2008 Pour citer cet article : Tim Jenkins, « Marcel Mauss’s essay On Prayer : an important contribution on the nature of sociological understanding », Revue du MAUSS permanente, 6 novembre 2008 [en ligne].

[1For example, James, W. & Allen, N. eds 1998 ; Tarot, C. 1999 ; Karsenti, B. 1997 ; Fournier, M. 1994. Derrida, J. 1991, although on the gift, is also an important addition. Of these, Tarot at least is well aware of the importance of the thesis On Prayer as an exposition of Mauss’s ideas on method. More recent publications include Hénaff, M. 2002, Papilloud, C. 2003 and Dzimira, S. 2007.

[2Mauss, M. 2003. On Prayer. Oxford & New York : Berghahn : a translation of Mauss, M. 1909. La Prière. Paris : Alcan, by S. Leslie, edited by W. Pickering with the aid of N. Allen and R. Parkin, on behalf of the British Centre for Durkheimian Studies and the Durkheim Press. Pickering has contributed an introduction, and H. Morphy an essay concerning the Australian ethnography contained in Mauss’s second section. A review of Mauss 2003 containing some of the material in this article has appeared in Jenkins (2003).

[3See Mauss, M. 1998. An intellectual self-portrait. James & Allen 1998 : 39 and the comments in Morphy, H. 2003. Some concluding anthropological reflections. Mauss 2003 : 140.

[4Mauss 2003 : 21, = Mauss, M. 1968. Marcel Mauss : Oeuvres, présentation de Victor Karady. Paris : Les Editions de Minuit. Volume I : 357.

[5Mauss 2003 : 21/ Mauss 1968 : 358

[6Mauss 2003 : 23/ Mauss 1968 : 360

[7Mauss 2003 : 30/ Mauss 1968 : 373

[8Mauss 2003 : 32/ Mauss 1968 : 376

[9Bachelard’s dates are 1884-1962, compared with Mauss 1872-1950. Recent discussions include : Tiles 1984 ; Didier 1993 ; Parinaud 1996 (a biography) ; Chimisso 2001 ; Lecourt, D. 2002 ; Wunenburger, J.-J. 2003.

[10Bachelard, G. 1940. La philosophie du non.

[11Bachelard, G. 1949. Le rationalisme appliqué.

[12Durkheim, E. 1907. Règles de la méthode sociologique ; see footnote 72, Mauss 2003 : 103/ Mauss 1968 : 385, which refers also to Mauss, M. & Fauconnet, P. 1901. La Sociologie : objet et méthode. Grande Encyclopédie 30 : 165-76.

[13Mauss 2003 : 38/ Mauss 1968 : 386

[14And in his early essay, Mauss, M. 1904. L’Origine des pouvoirs magiques dans les sociétés australiennes – see footnote 76, Mauss 2003 : 103/ Mauss 1968 : 388.

[15Mauss, M. 1925. Essai sur le don : forme et raison de l’échange dans les sociétés archaïques. In L’Année Sociologique, ns 1 : 30-186. On the protocols adopted in that article, cf. Jenkins, T. 1998. Derrida’s reading of Mauss. James & Allen 1998 : 83-94.

[16Mauss cites O. Dibelius, Das Vaterunserer 1903 – note 47, Mauss 2003 : 101/ Mauss 1968 : 377 – and Loisy, Les évangiles synoptiques, possibly = L’évangile selon Luc – see note 84, Mauss 2003 : 104/ Mauss 1968 : 392.

[17Mauss 2003 : 42/ Mauss 1968 : 392

[18Mauss 2003 : 42/ Mauss 1968 : 393

[19Mauss 2003 : 42/ Mauss 1968 : 393

[20Mauss, M. 1898. Essai sur la nature et la fonction du sacrifice – see footnote 87, Mauss 2003 : 104/ Mauss 1968 : 394.

[21One might be uncomfortable with the word ‘essence’, but it should be taken in an ordinary language sense – rather than as a technical philosophical term – as pointing to what is broadly at issue in the institution in question.

[22Mauss 2003 : 44/ Mauss 1968 : 395

[23Mauss 2003 : 44/ Mauss 1968 : 395-6

[24It is important to grasp the specificity of this concept of ‘simplicity’ in order to evaluate the contribution Mauss makes to twentieth century French thought. Peter Hallward has recently identified how important and widespread in contemporary French philosophy is a concern with ‘a sort of simplicity’, an (ahistorical) singular principle of individuation that generates differences without concern for ‘the mechanisms of mediation and interaction’ (Hallward 2003 : 2). He concludes that ‘the task of tomorrow’s generation of thinkers may be to develop an equally resilient relational alternative’ to this ‘singular or non-relational thought’ (op. cit. 23). Yet Mauss is clearly already seeking to differentiate himself from this contemporary thought, both in the earlier, historical conception of ‘simplicity’ outlined here and in the later concept of the ‘total social fact’. This crucial theme can only be noted, and not explored, in this paper.

[25Mauss 2003 : 47/ Mauss 1968 : 400

[26See Collingwood, R.G., An Essay on Philosophical Method, 1933.

[27Mauss 2003 : 32/ Mauss 1968 : 377

[28Mauss 2003 : 33/ Mauss 1968 : 377

[29Mauss 2003 : 36/ Mauss 1968 : 383

[30Mauss 2003 : 36/ Mauss 1968 : 383-4

[31Mauss 2003 : 49/ Mauss 1968 : 401

[32Mauss 2003 : 50/ Mauss 1968 : 404

[33Mauss 2003 : 52/ Mauss 1968 : 407

[34Mauss 2003 : 54/ Mauss 1968 : 409

[35Mauss 2003 : 54/ Mauss 1968 : 409

[36Mauss 2003 : 54/ Mauss 1968 : 410

[37Mauss 2003 : 57/ Mauss 1968 : 414

[38Mauss 2003 : 33/ Mauss 1968 : 337

Suivre la vie du site RSS 2.0 | Plan du site | Espace privé | SPIP | squelette