A history of the secularization issue

Appendix: An Important Technical Remark
Before turning to the historical part proper, a brief - but very important - technical remark is in order. How will I proceed concretely in order to avoid writing a "whiggish" account every time I mention an exemplar? Conversely, how will I proceed to avoid totally isolating a given author in his own context every time I apply a Skinnerian methodology? Some of the ways in which I will do this cannot be set forth a priori, but will depend on a number of circumstances, which will have to be dealt with by improvisation. At this point, using the formulation of one of Skinner's critics, I would contend that sometimes, indeed, there is "remarkably little that can profitably be said in general" (Minogue 1981, p. 186). But one of the techniques I will use is fairly simple, and can be presented immediately.

Confusion in the forthcoming analysis between my own analytical categories (the "exemplars" defined in WPs # 2 and 3) and the authors' categories can be avoided by the use of a simple editorial convention. Every time I will use one of my own categories while the author is not using it himself, this category will be followed by an asterisk (i.e., differentiation*). This device will allow us to become aware, as the analysis proceeds, of the emergence of the elements contained in the exemplars which interest us while avoiding any risk of terminological - and substantive - confusion. Exemplars that are not followed by an asterisk were used by the author - either in the same sense as I do, or in a different sense⁵⁰.

Finally, one last technical comment: Throughout the following discussion, all the italics within citations are the authors'. I have added no italics. From here on, I will no longer note: "emphasis in original".
CHAPTER 6

A BRIEF OUTLINE OF THE GENESIS OF THE CISR PARADIGM
The whole methodology of the research having thus been set forth, we can now embark upon the historical part by giving a first very brief overview of the history we are going to cover in the rest of this part of the book.

The whole history of the development of the CISR paradigm must be seen against a general background: the apparent crisis of religion which started at the time of the Enlightenment. This crisis has gone through different phases, some of which can be summarized under familiar headings: French Revolution, British secularism, German Kulturkampf, liberal theology, oecumenism and Vatican II Council. Although they are in most respects very different, all these events share some common elements: they are all related to the transformations of institutional religion resulting from the advent of modernity.

The first important event against this background is the rise of positivist social science, which held a view of secularization in which religion was doomed to disappear. In this context, the word "secularization" itself was sometimes employed, but more often in the political field than in the "scientific field". Thus, Comte thought that traditional religion would disappear, but did not use the word secularization. But some of the ideas which were later going to be accepted in the CISR paradigm (i.e.: differentiation) appeared in this period (see chap. 7).

The second important event against this background is the emergence of classical sociological theory at the turn of the century. In contrast to the positivistic period, this period witnessed the emergence of some of the most important ideas which were later to be taken over into the secularization paradigm (i.e.: rationalization, worldliness), reformulated some of the others which had already been expressed in the positivistic tradition (i.e.: differentiation), and used the word secularization on some occasions, but very unsystematically (see chap. 8 and 9).

The third important event is the rise of an empirical sociology of religion in the period preceding and following World War II, mainly under the impulse of Catholic sociologists. The approach typical of this period stands in sharp contrast to the later CISR paradigm, and can best be understood through its genesis. First, Catholic sociology depended on a world view which we could loosely define as "classical (or neo-classical) theology", that is, a theology which had not yet drawn all the conclusions of the crisis of religion. However, Catholic sociology was, in a different way, a response to this same religious crisis, for one of the factors explaining its rise was a deep concern over the sharp decline in practice characterizing this period. Finally, Catholic sociology was characterized by a set of scientific habits directly derived from American empirical sociology, that is, a largely a-theoretical gathering of "facts". The combination of these factors enables us to define Catholic sociology as an attempt to use empirical research methods in order to find a solution to the "crisis of the Church" (see chap. 10).

The fourth important event, again situated against the general background of the perceived crisis of religion, is the emergence of secular theology, in the 50s and 60s. This theology provided a new world-view in which, by contrast to classical theology, modernization was no longer considered as intrinsically evil. Secular theologians held that the crisis of the church should not be resolved through a return to the previous situation, but through an adaptation of religion to the new prevailing conditions (see chap. 11).

The fifth important event is the religious revival which occurred in the United States in the 50s. This revival suddenly brought religion back to the center of the stage as an important element for sociological investigation. The study of religion, which had hitherto been left to Catholic researchers, became a topic of interest for professional sociologists, and sociology of religion very rapidly acquired a recognized, although not very prestigious status, in the sociological field (see chap. 10).

We are now in a position to understand the emergence of the secularization paradigm in the 60s. First, a new development in the "religious crisis", which had found its provisional epilogue in Vatican II, resulted in a loosening of the grip of the Church on Catholic sociology. Catholic sociologists, in turn, feeling partially demoralized by the evolution of the situation, lost some of their belief in the virtues of empirical religious sociology. Second, the arrival of professional sociologists in the field of religious studies as a result of the American religious revival led to a reconsideration of the importance of the theoretical aspect of sociology of religion, which took the form of a return to the classics. This arrival of new actors in the field also resulted in a shift of emphasis in the scientific habits, from a-theoretical empirical investigation to theoretically informed empirical investigation, or event to pure theory. Third, the development of secular theology made it possible to consider the fate of religion in modern society under a new light, no longer as a decline, but as a transformation, an adaptation to the new conditions, which could even yield positive results. Thus, in a few years, a completely new way of studying religion in modern society appeared, and a new generation of scholars sent the aging Catholic sociologists back to their churches. What happened in these years deserves to be called an intellectual revolution (the equivalent of the Kuhnian scientific revolution, but in the social sciences). This revolution was made possible not through a conversion of sociologists to new evidence, but through a power struggle resulting in the elimination of an older generation and the rise of a new one (see chap. 10).

The new names which became dominant in the field in the 60s are still very dominant today. By the same token, the new paradigm they imposed is still the dominant paradigm today. Although it shows some signs of aging, it still holds an excellent explanatory value, and no new generation of sociologists of religion has made its appearance on the scene yet (see chap. 12 and 13).

CHAPTER 7

THE PREHISTORY OF THE SECULARIZATION ISSUE
In the present chapter, we will cover the history of the secularization issue from the first mention of the term "secularization", in 1646, to the turn of the 20^th century. The use of the term "prehistory" to designate this phase is justified to the extent that present-day sociologists never refer to the writings of this period in their theories. These writings thus constitute, as it were, an "unconscious", or "unwritten" substratum within the CISR secularization paradigm.

I will make no attempt at providing an all-encompassing history of these early developments. For one thing, they are sufficiently numerous to provide material for - at least - a whole dissertation⁵¹. For another, we are interested in this early debate only to the extent that it allows us to gain a better understanding of the genesis of the CISR paradigm and of the approaches surrounding it. For these reasons, I will only marginally touch upon the developments that took place either in the social-political sphere or in intellectual fields such as anthropology or political theory, and devote most of this chapter to a discussion of three figures central to the development of sociology and of the secularization issue: Auguste Comte, Herbert Spencer, and Ferdinand Tönnies.

Auguste Comte and Herbert Spencer
The early debate about the place of religion in a modern society was cast in terms very different from those prevailing in the CISR paradigm. The debate largely revolved around the question of the conflict between religion and science. In the political turmoil accompanying and following the French revolution, science was seen by many as the only solution to humanity's problems. Past societies were doomed because they were based on religion, but the new society emerging to replace the old was to be based on a rational understanding of the social order.

This was a central problem for both Comte and Spencer. The problem these two thinkers saw themselves as addressing can be summarized as follows. Given the fact that the traditional legitimation of society is rapidly vanishing⁵², how is the new social order envisioned by sociology to be sustained? As we will see presently in more detail, Comte's and Spencer's answers were very different. But on one point at least, they fully agreed: religion must be replaced by science.

Indeed, not only did Comte and Spencer want to replace religion by science, but their "rational" systems themselves took on characteristics which, in retrospect, appear more religious than scientific. Both thinkers reacted to the diminishing role of religion in their societies - which entailed a good measure of anxiety as to the place of man in the universe - by proposing a new answer of their own to the old religious question of the relation between "this world" and the cosmos. In doing so, both of them went much farther than pure reason and science (at least as we understand them today) would have allowed, proposing new schemes in which every item "here below" found its correspondent "over above". Both Comte and Spencer proposed schemes in which the most insignificant features of this world were somehow in harmony with the global structure of the cosmos. According to Comte, "our social evolution is [...] in reality nothing but the most extreme development of a general progression, continued without interruption in the whole organic realm, from the simple plants and the lesser animals" (Comte 1839, p. 202). In Spencer, even more perhaps than in Comte, the subsuming of all social processes under a universal law provided a reassuring ordering of every individual's daily experience⁵³. Everything fell nicely into place under a single law: complexification. As Lewis Coser correctly stresses, "Spencer formulated a universal explanatory scheme for those in quest of certainty, which in a previous age only religious doctrines could provide" (Coser 1971, p. 123). As a matter of fact, not only did the theories put forward by Comte and Spencer provide a new way of mapping the world; such a mapping was implicit in the sheer form in which these theories were planned and executed. Both attempts were encyclopaedic, extremely systematic and dogmatic, as is exemplified by the fact that both of them were set out in multi-volume works which had been planned well in advance, and which covered all the areas of knowledge.

Now, as we will see, the CISR paradigm is concerned only marginally with such questions. From a broad historical perspective, the genesis of the CISR paradigm can be interpreted as the gradual emergence of a more distinctively "social-structural" view of the secularization process, as contrasted with the previous, more "anthropologically" (in the sense of a philosophical anthropology, or general conception of "man") inspired view. In effect, to the extent that it is based mainly on the differentiation exemplar, the CISR paradigm is inherently social-structural: differentiation (as defined in the paradigm) is strictly a social-structural process. In contrast, by modern standards, the views of the founders of sociology were more anthropological (in the sense indicated above) than sociological.

Does this imply that there is no connection between XIX^th century sociology and the CISR paradigm? Strictly speaking, there is no direct link between Comte and Berger, between Spencer and Parsons⁵⁴, or between Tönnies and Wilson⁵⁵. However, as we will see, a number of indirect links are evident. In chapter 1, we have discussed how exemplars acquired a "timeless" quality. To give an obvious example, Spencer directly influenced Durkheim, who in turn influenced the carriers of the CISR paradigm. It is not my intention to provide even a tentative map of these links between early secularization theory and the paradigm. What I will do instead will be to focus on one particularly important locale where a first evolution in the direction of the new approach is evident. The first evidence of a differentiation between a "anthropological" and a more "social-structural" approach can be noted in the gap between the views of Comte, who is literally obsessed by self-reference, and the views of Spencer, who relies more on a structural-functional analysis.

To my knowledge, Comte never used the word secularization. Spencer used it at least once in a context which will be mentioned below, but he did not give it a central theoretical significance. As is the case with most other thinkers in the XVIII^th and XIX^th centuries, discussions concerning the fate of religion are placed in the framework of the religion/science problematic. This is the set of problems that must be examined to understand Comte's and Spencer's positions on the subject of what we today call secularization.

The first difficulty in any attempt to compare these two authors is that, even though they were contemporaries and focused their attentions on very similar questions, Comte and Spencer were not much interested in each other's work. From a modern vantage point - and even for some of their contemporaries - this mutual neglect is rather surprising. As Spencer published his first paper twenty years after Comte's first book, the Plan des travaux scientifiques nécessaires pour réorganiser la société (1822), there could be some reason to suspect that he had been influenced by Comte. Allegations of this type, which were repeatedly made, particularly infuriated Spencer, who was very sensitive as to all matters of priority. He went out of his way on several occasions - most notably in his Genesis of Science (1854), in his Reasons for Dissenting from the Philosophy of M. Comte (1864a), and in his Autobiography - to stress that their views were essentially at variance, and that where they overlapped, this was only because, all things considered, Comte's theories were nothing but a systematization of quite common conceptions. I am not going to engage into a discussion of the mutual influences these men had on each other. What I will do instead, will be merely to compare their views, centering on the problem of the relationship between religion and science.

For the superficial observer, there are many reasons indeed to assume a close similarity between these two thinkers. They were, each in their own country, the first to propose a complete social philosophy under the new heading of "sociology": Comte coined the word sociology, and Spencer was the first to follow suit in the English-speaking world (Spencer 1864a, p. 133; Peel 1971, p. vii). Both were dogmatic thinkers, who read very little, and wrote dozens of volumes containing complete systems of philosophy. Both were academically marginal scholars - neither of them ever held on official position in a university. Finally, both were, albeit to different degrees, positivists, who conceived of at least part of their work as an alternative to the rapidly decaying (as they saw them) religious traditions. This overall impression of similitude is compounded by the fact that these two founders of sociology have fallen into a common disgrace, and are today considered to be representatives of an almost prehistorical, and abortive, phase of sociological thinking⁵⁶. They are the most outstanding examples of what sociology for the most part no longer wants to be: positivist and evolutionary.

However, important as these similarities appear to be, they are mostly superficial. Comte's evolutionary theory was very different from Spencer's, and his positivism - especially in its later version - was of an altogether different breed. But before turning to these divergences, a rapid assessment of the areas of convergence in their theories is called for.
Similarities between Comte and Spencer
The first similarity between Comte and Spencer concerns the epistemological status of sociology. None of them conceived of any discontinuity between the realm of the natural and of the social sciences. This first convergence is important for the religion/science problematic: Long after having abandoned their claim to authority on natural phenomena, theologians continued to assert that human affairs were their reserved domain. Comte and Spencer both rejected this stance. Human affairs were to be studied by science exclusively, and could no longer be left to the authority of religion.

This is best illustrated in Comte's classification of the sciences. The only reason why the knowledge of society had hitherto been considered to be of a different nature from that of natural phenomena was that the first sciences to become positive were the simplest. Only at the last stage of evolution could the study of man and of society, the most complex of the sciences, reach the same level of positivity. The traditional separation between the knowledge of nature and the knowledge of human society, between science and politics, was nothing but the manifestation of an immature state in social evolution. Similarly, and although he strongly disagreed with Comte's classification of the sciences (Spencer 1854), Spencer's constant practice of simply drawing parallels, without any sort of qualification, between biological and social "organisms" bears testimony to his undifferentiated approach.

If we turn to more substantive aspects of their sociologies, we also find many similarities between Comte and Spencer. The most striking, perhaps, is that both of them contended that society must be considered as an organism. In his Reasons for Dissenting, Spencer recognized that his own conception of society as an organism was essentially the same as his opponent's: "The analogy between an individual organism and a social organism, which was held by Plato and by Hobbes, is asserted in [my] Social Statics, as it is in the Sociology of M. Comte. Very rightly, M. Comte has made this analogy the cardinal idea of this division of his philosophy" (Spencer 1864a, p. 136). Furthermore, both thinkers contended that society must be considered from two complementary points of view: as an organism that could be studied on a synchronic level on the one hand, and as an organism in evolution on the other hand. The first aspect, Comte called "static", whereas Spencer analyzed it in terms of function and structure. The second aspect, called "dynamic" by Comte, was studied by Spencer under the heading of the law of differentiation. Although present-day sociologists tend to cast doubts on the compatibility of these two divisions as they were presented by these thinkers, there is no mistaking the fact that they both heavily insisted on this complementarity which, in their view, derived necessarily from the biological analogy. This is obvious in the case of Spencer (Burrow 1966, pp. 190-96), but it also applies to Comte: "[The notions of order and of progress] must be as rigorously inseparable in social physics as the notions of order and of life in biology, whence, from the point of view of science, they obviously derive" (Comte 1839, p. 16).

One of the main consequences of this continuity between the natural and the social sciences, as it was put forward by Comte as well as by Spencer, was that history became subjected to natural laws. It was no longer to be considered a human creation, but the necessary unfolding of a cosmic necessity, which operated with the same ruthless and blind precision as in nature itself. In both thinkers, the development of society necessarily followed an universal pattern, which was molded on the development of the individual organism. "Does not each of us, in looking back upon his own history - asked Comte rhetorically - remember that he was successively, as to his most important notions, theologian in his childhood, metaphysician in his youth, and physicist in his mature life?" (Comte 1830, p. 6). Similarly, in Spencer, the progressive differentiation of the social system was only a reflection of a universal law, observable in the evolution of the universe as well as in all biological organisms.

As we will see, Comte's and Spencer's grand schemes of social evolution were largely at variance, but on one point at least, they converged: Both theorists used a dichotomy between the "military" (Comte), or "militant" (Spencer) societies of the past, which directed their aggressivity toward other societies on the one hand, and their own societies, which they both referred to as "industrial", and which directed their energies toward the exploitation of nature on the other hand⁵⁷. Even more importantly, in both systems, as a result of natural evolution, religion had been markedly weakened. And, in both cases, this weakness was to be compensated by the intervention of the authors of the theories themselves - Comte wanted to base society on scientific knowledge, Spencer on a rational morality. Thus, in both these thinkers, religion was being replaced by science or, speaking more broadly, by some form of rationality. To this extent, as noted above, Comte and Spencer both belong to the autoreference tradition. The writings of both these men can best be understood against the XIX^th century background of apparent decline in the social influence of religion, as compared with the rising prestige of science. Both men considered that the decline of religion was the necessary consequence of a natural law of evolution, and both hailed the rise of science as the sign of the advent of a new age for humanity. Furthermore, both considered that the knowledge of society must be based entirely on a positive approach, and that science must replace religion in shaping the destinies of humanity.
First divergence: The law of evolution
So much for the assumptions explicitly or implicitly shared by Comte and by Spencer. Their divergences will be presented in three parts. The first concerns their different evolutionary schemes. Although they shared some conclusions, their theories were based on very different assumptions. The second area concerns their epistemologies. Although both Comte and Spencer proposed systems which, to different degrees, possessed religious overtones, their explicit assessments of the ultimate cognitive validity of religion differed markedly. Finally, we will turn to their differing appraisals of the way in which reason was to be used to insure an harmonious relation between the individual and the social order. If both thinkers agreed in their dismissal of traditional religion as a guide to social life, the "rational" solutions they proposed could not have differed more widely.

To turn to the first of these issues, whereas Comte's evolutionary scheme was concerned mainly with the evolution of ideas, Spencer's was based almost entirely on the evolution of the social structure. Spencer himself took pains to underscore this divergence: "What is Comte's professed aim? To give a coherent account of the progress of human conceptions. What is my aim? To give a coherent account of the progress of the external world. Comte proposes to describe the necessary, and the actual, filiation of ideas. I propose to describe the necessary, and the actual, filiation of things. Comte professes to interpret the genesis of our knowledge of nature. My aim is to interpret [...] the genesis of the phenomena which constitute nature" (Spencer, quoted in Coser 1971, p. 89).

Comte contended that one cannot find a single example of a change in society which had not been preceded and made possible by "theoretical works". "Every social regime is an application of a philosophical system and [...] consequently, it is impossible to institute a new regime without having previously established the new philosophical system to which it must correspond" (Comte, quoted in Arbousse-Bastide 1966, p. 5). Thus for example, the feudal system had been made possible by theoretical works in the Christian tradition since the foundation of the school of Alexandria, and the French revolution had been made possible by the works of the philosophes (Comte 1822, pp. 81-82). As a result of this conception, Comte's law of evolution was based on a series of changes in human thinking, to which correlative social changes were only subordinated: In the stage of theological thinking, humans ascribed the observed phenomena to supernatural agencies; in the stage of metaphysical thinking, these supernatural agencies were replaced by abstract forces; and in the stage of positive thinking, humans renounced explaining these phenomena, and restricted themselves to observing them and to establishing the laws to which they were submitted (Comte 1830, pp. 3-5).

Spencer explicitly refuted this conception of progress: "The progress of our conceptions, and of each branch of knowledge, is from beginning to end intrinsically alike. There are not three methods of philosophising radically opposed; but one method of philosophising which remains, in essence, the same" (1864a, p. 125). Evolution, Spencer argued, must be considered at the level of the social structure. At this level, it immediately becomes clear that the evolution of the social organism is based on the same law as the evolution of any organism, and of the universe in general: "The series of changes gone through during the development of a seed into a tree, or an ovum into an animal, constitutes an advance from homogeneity of structure to heterogeneity of structure. In its primary stage, every germ consists of a substance that is uniform throughout, both in texture and chemical composition. The first step is the appearance of a difference between two parts of this substance; or, as the phenomenon is called in physiological language, a differentiation. [...] This law of organic progress is the law of all progress. Whether it be in the development of the Earth, in the development of Life upon its surface, in the development of Society, of Government, of Manufactures, of Commerce, of Language, Literature, Science, Art, this same evolution of the simple into the complex, through successive differentiations, holds throughout" (1857, pp. 9-10).

Thus, Spencer's law of progress was based on the idea of differentiation. One of the early manifestations of this process appears in the area of religion. Even though the first differentiation that arises in society is the one between the "governing" and the "governed", and even though, in this stage, the rulers are regarded as divine, this initial fusion between the political and the religious realms does not last: "No sooner does the originally-homogeneous social mass differentiate into the governed and the governing parts, than this last exhibits an incipient differentiation into religious and secular - Church and State" (1857, pp. 20-21). Subsequently, other spheres of life differentiate themselves from religion. Thus for instance, the arts - "Dancing, Poetry, and Music" - become separate spheres (1857, p. 31). As we will see later on, science itself slowly became differentiated from religion.

The consequences of these divergent evolutionary theories with respect to the role of religion in social life are extremely important. We will return to them at the end of this assessment of the thoughts of Comte and Spencer; for the time being, let it suffice to be noted that whereas, in Comte's view, traditional religion, which was once all-powerful, was eventually totally eliminated from social life, in Spencer's view, religion merely changed its location and its function in the social system.

Second divergence: The epistemological status of religion
The second area of divergence with respect to the matter for discussion is the different assessments on the ultimate validity of religious explanations of the world. Comte asserted that philosophy and science were radically incompatible (1844, p. 33). He was convinced that all views as to the ultimate causes of phenomena, in so far as they lay beyond empirical observation, could not be part of science. "As to the problem of determining what [...] attraction and [...] weightiness are in themselves, what their causes are, these are questions which we all regard as insoluble, which are no longer part of the domain of positive philosophy, and which we rightly abandon to the imagination of the theologians, or to the subtleties of the metaphysicians" (Comte 1830, p. 13). Even when, in his later life, Comte founded the Religion of Humanity, he remained truthful to this insight; his religion was to be totally rational: In his Catéchisme positiviste, Comte asserted that the aim of his theology was not to explain the causes of phenomena, but only to describe phenomena (1852a, p. 57).

Spencer explicitly refuted this position: "The consciousness of cause remains as dominant to the last as it was at first; and can never be got rid of. The consciousness of cause can be abolished only by abolishing consciousness itself" (1864a, p. 127). Religion, Spencer argued, cannot simply be discarded as irrelevant. Even if it contains a great deal of error, it cannot be entirely false. In the first pages of his Synthetic Philosophy, before setting out to explore the realm of the "knowable", Spencer carefully carved out a special space for the Unknowable. Contrary to Comte, he argued that we must make the effort to understand "what there was in this belief which commended it to men's minds" (1864b, p. 3). Thus for instance, in primitive societies, leaders and kings were considered of divine origin. We now know that this view is untrue. But the grain of truth it contained, and which is still valid today, is that every individual action must, in some way, be subordinated to social requirements. The reason for subordination given in the religious views of primitive societies was incorrect, but the sheer necessity of subordination implicit in them remains as a timeless truth (1864b, p. 9).

Primitive religion, then, can be regarded as a form of speculation on the order of things, undertaken by an as yet immature intelligence. As human intelligence progresses, so does the quality of its conceptions. Thus, positive thought comes to occupy an ever more important place in these conceptions. But does this imply that, one day, positive conceptions will be able to totally eliminate religious views? Spencer thinks that this is impossible: "Positive knowledge does not, and never can, fill the whole region of possible thought. At the uttermost reach of discovery there arises, and must ever arise, the question - What lies beyond" (1864b, pp. 16-17). The regress in causal analysis cannot be indefinite: somewhere it has to stop. Somewhere, we are bound to say: This question I cannot answer. Religion alone is able to answer the question of the First Cause of things.

Up to this point, Spencer's conception seems not very different from Comte's. The latter also acknowledged that something, which he referred to as metaphysics, passed scientific understanding, and that it must be "left to the theologians". But Spencer diverged from Comte in that he contended that it was wrong to assert that, because it passed scientific understanding and empirical observation, a knowledge of the ultimate cause of things was impossible. We cannot know the Absolute rationally; but this does not mean that we cannot know it by other means: "Besides that definite consciousness of which Logic formulates the laws, there is also an indefinite consciousness which cannot be formulated. Besides complete thoughts [...] there are thoughts which it is impossible to complete; and yet which are still real, in the sense that they are normal affections of the intellect. [...] To say that we cannot know the Absolute, is, by implication, to affirm that there is an Absolute" (1864b, p. 90).

As there will always be unanswerable questions, there will always be a legitimate place for religion. Thus, instead of speaking of a warfare of science of religion, we must seek the best way to reconcile them, for they are not antithetical, but complementary. Far from having suffered losses in its confrontation with science, religion has progressed toward a fuller recognition of its true essence: "The truly religious element of religion has always been good: that which has proven untenable in doctrine and vicious in practice, has been its irreligious element; and from this it has been ever undergoing purification" (Spencer 1864b, p. 104). The same applies to science: as it progressed, it gradually abandoned metaphysical explanations to limit itself to establishing laws. In other words, "religion and science have been undergoing a slow differentiation" (1864b, p. 108). Thus, "the disagreements between [religion and science] have throughout been nothing more than the consequences of their incompleteness; and as they reach their final forms, they come in entire harmony" (1864b, p. 107).

Finally, Spencer also explicitly refused the idea of a Religion of Humanity: "No such thing as a 'Religion of Humanity' can ever do more than temporarily shut out the thought of a Power of which Humanity is but a small and fugitive product - a Power which was in course of ever-changing manifestations before Humanity was, and will continue through other manifestations when Humanity has ceased to be" (1873, p. 284).

Third divergence: The individual, the social order, intellectual knowledge, and religion
The previously mentioned problem of social engineering, which lay at the root of Comte's and Spencer's attempts, can perhaps best be formulated as follows: What, given the demise of religion, is to be the role of intellectual knowledge in the establishment of a new harmony in the relation between the individual and the social order? To understand the discrepancies between Comte and Spencer on this point, we must consider their views as to the relation between the individual and the social order. For Comte, the social order was to be predominant, whereas for Spencer, the individual was to retain predominance⁵⁸. Comte wanted every individual to be completely subordinated to the needs of society. In his view, the natural evolution of humanity leads to "sociocracy", that is, a society in which all individuals come to recognize the primacy of the collective order. This tendency finds its logical development in the Religion of Humanity, where individuals worship this society which stands so much above them. It should be stressed that this view was not a late development in Comte's thought. As early as 1825, in his Considérations sur le pouvoir spirituel, Comte argued that man's natural anti-social tendencies could be checked only by an organized moral force, which only could impose to humans the constant sacrifices demanded by social life ([1825] 1968, p. 154).

For Spencer, this view of society as requiring sacrifices from individuals was not wrong, but it applied only to the militant type: "In a society organized for militant action, the individuality of each member has to be so subordinated in life, liberty, and property, that he is largely, or completely, owned by the State; but in a society industrially organized, no such subordination of the individual is called for. [...] With the absence of need for that corporate action by which the efforts of the whole society may be utilized for war, there goes the absence of need for a despotic controlling agency" (1876-96, pp. 538-39). In an industrial society, order is not based on coercion, but on voluntary cooperation and individual self-restraint⁵⁹.

Thus, "while Comte stressed that men should aim at discovering the laws of society in order to act collectively on the social world, Spencer argued with equal conviction that we should study them in order not to act collectively. In contrast to Comte, who wanted to direct society through the spiritual power of his sociologist-priests, Spencer argued passionately that sociologists should convince the public that the society must be free from the meddling of governments and reformers" (Coser 1971, pp. 99-100). Thus, neither Comte nor Spencer advocated a society primarily based on political power. For both men, civil society must be the base of the social order. Consequently, they both lay great stress on the role of morality. They diverged only in the sense that for Comte, this civil society was very holistic (to use a modern terminology), whereas for Spencer, it could only be individualistic.

The second point which must be considered to understand the divergence is the role attributed to scientific knowledge by the two thinkers. The shortest way to put the difference would be to argue that, whereas Comte's reorganization efforts were to bear solely on intellectual ideas, Spencer's also aimed at reforming sentiments. This is the view taken by Spencer in his rebuttal of Comte: "So far from alleging, as M. Comte does, that society is to be re-organized by philosophy; [my Social Statics] alleges that society is to be re-organized only by the accumulated effects of habit and character" (1864a, p. 136). In other words, according to this view, the divergence between Comte and Spencer would boil down to different views as to the forces that lay behind history. As noted above, Comte contended that ideas were the driving force in any society. To this, Spencer answered: "Ideas do not govern and overthrow the world: the world is governed or overthrown by feelings, to which ideas serve only as guides. [...] Not intellectual anarchy, but moral antagonism, is the cause of political crises" (1864a, p. 128). Spencer argued that a sound morality was a very important basis for social order. So much so that, with religion declining, it was urgent that a scientific system of morality be established. It is in this context that the word "secularization" appears in Spencer's writings: "The establishment of rules of right conduct on a scientific basis is a pressing need. Now that moral injunctions are losing the authority given by their supposedly sacred origin, the secularization of morals is becoming imperative. Few things can happen more disastrous than the decay and death of a regulative system no longer fit, before another and fitter regulative system has grown up to replace it" (1882, p. vi). The Principles of Morality, which constituted the last part of the huge Synthetic Philosophy, were very important in Spencer's work. Indeed, according to J. Peel, "The fundamental purpose of Spencer's whole lifework was to provide a scientific morality" (1971, p. 84).

There thus seems to be a fundamental divergence between Comte and Spencer as to the role of morality. However, this divergence is mostly superficial. Even tough Comte contended that society could be changed through a reform of intellectual knowledge, at the same time, like Spencer, he was aware that this could be achieved only by an appeal to moral authority. As early as 1822, in his Plan des travaux scientifiques, Comte argued that society must be reorganized and presided over by men holding spiritual power⁶⁰ (1822, p. 83). This spiritual power was to be exercised by scientists, because they were the only ones who were professionally trained to think theoretically (1822, p. 86). It is not altogether clear, however, how this power of a spiritual nature is derived from the exercise of reason. In the Plan, these two aspects are simply juxtaposed. This fundamental ambiguity in Comte's proposal found expression in his own life by the succession of two phases which are considered by most scholars as being of very different nature. In the first part of his life, Comte founded positivism, and proposed to transform politics, or the direction of society, into a positive science, based on observation (1822, p. 94). In the second part of his life, after his encounter with Clotilde de Vaux, he founded the Religion of Humanity, and proposed to transform politics into a religious affair. Now, Comte always denied having shifted his views. Considering the ambiguity already contained in several of his early writings, this claim cannot be dismissed a priori (Manuel 1962, p. 266). But even if Comte did not change his mind, there remains the question of how these two aspects of his proposal can be reconciled.

I think that this question can be resolved through consideration of a rather machiavelian passage found in the Plan. Comte makes a distinction between the intellectual processes which lead to decisions concerning the future of humanity, and the practical means by which these decisions can be applied. As the decisions are taken by scholars, in accordance with their best knowledge of the constraining nature of the laws of evolution, we must conform to these laws regardless of whether they appear to be "good" or not. The question whether evolution is ethically good is simply irrelevant: it is neither good nor bad, but necessary. It is clear, however, that these decisions cannot be presented to the general public in this fashion: "In order that a new social system establishes itself, it is not sufficient that it be conceived correctly, it is also necessary that the mass of society impassions itself to constitute it" (1822, p. 129). This can obviously not be done by an appeal to the inevitability of that new system: "One will never be able to rouse the passions of the masses of men for any system, by proving to them that it is the system for which the walk of civilization, from its inception, has prepared the advent, and which is today called to direct society. Such a truth is within the reach of too small a number of minds, and even for these minds, it requires too long a series of intellectual operations, to be apt to ever rouse the passions" (Comte 1822, p. 129). Comte goes on to argue that, to achieve this implementation, we must revert to giving imagination free play, for instance in favoring art.

This is probably the solution of the riddle. Although Comte did not advocate an appeal to religious sentiments in this particular piece of writing, he clearly set forth the dilemma which later led him to create the Religion of Humanity. This is the dilemma inherent in the autoreference paradox: one must hide the intellectual operations that are necessary to establish the new social order. Although the new society was to be planned rationally, it was to be implemented through an appeal to the emotions. In the first part of his life, Comte designed the rational plan; in the second part, with the help of a biographical accident which allowed him to tap his emotional resources to foster his own enthusiasm⁶¹, he sought to implement it. As Anthony Giddens explains, "at first sight the call to establish a Religion of Humanity seems quite inconsistent with the positive philosophy advocated in the Cours, and many commentators have supposed that there is a major hiatus between Comte's earlier and later works. But it is perhaps more plausible to argue that the Système de Politique Positive brings fully into the open the latent substratum of the positivist spirit: we see that science cannot, after all, provide its own commitment" (1972, p. 242).

By now, the reader must have the impression that we have strayed too far from our basic question: the religion/science problematic. But this long parenthesis has been necessary in order to establish what the effective divergence between Comte and Spencer was. This divergence can now easily be summarized. Comte contended that science could eliminate traditional religion, but must at the same time find a way to replace it. In order to do this, science must take on two functions traditionally attributed to religion: the exercise of central moral authority, and the capability to arouse the passions of the population. Hence, the necessity of the creation of the Religion of Humanity. As for Spencer, he contended that science could not eliminate religion, and needed not replace it. Furthermore, he argued that the social order in an industrial society needed no central moral authority, and that it was sufficient for science to provide a new foundation for ethics in order to guarantee social order. In other words, where Comte envisioned the possibility of a totally autoreferent reconstruction of the social order, Spencer was much more modest, proposing for autoreference a much more limited role: only morality could be reconstructed autoreferentially.

But how can we explain these divergent positions? To do this, we must again draw on some of the conclusions reached in the preceding analysis, and take into account two factors. The first of these is extrinsic to the religion/science problematic; the second is intrinsic. First, whereas Comte favored a view of social order in which the whole determined the parts, Spencer contended that the parts are determining the characteristics of the whole. Organized religion being by definition a collective phenomenon, it fitted easily into Comte's scheme, but certainly not into Spencer's. Where Comte saw, as the only solution to the problem of social integration, a strong central spiritual authority, Spencer preferred a diffuse, secular, and consensual form of morality.

Second, as noted above, in the first part of his life, Comte held the view that religion must be totally eradicated, and replaced by science. When he addressed the problem of social integration, however, he recognized that this posed a problem: Society needed spiritual power, and men could be motivated only by passion. Thus, for practical purposes, religion had to be brought back into the picture. The new religion, however, was to be totally subordinated to science. As Comte saw it, the Religion of Humanity was a totally rational creation, purged of all mystical elements. As we can read in the opening sentence of the Système de politique positive, "First spontaneous, then inspired, then revealed, religion finally becomes demonstrated" (1852b, p. 7). Thus, society was to be held together by science, which was translated into a religious appearance only in order to appeal to the sentiments and the passions of men. Having taken a radical position on the religion/science problematic, Comte was compelled to resort to a dramatic strategy when faced with the "practical" problems of social engineering.

Spencer's position with regard to the religion/science problematic was not as radical as Comte's; consequently, his strategy to cope with the problem of social integration needed not be as dramatic. In Spencer's view, we unmistakably see at work some of the exemplars central to the CISR paradigm. In the course of social evolution, religion and ethics have undergone a progressive differentiation⁶² (differentiation). Furthermore, in the realm of knowledge, religious views have progressively given way to more empirical forms of knowledge (scientization*). Thus, only non-empirical questions continued to fall under the purview of religion. Where Comte, having eliminated religion in the first part of his life, was left with nothing but science, Spencer retained three distinct entities: science, ethics, and religion. As the world was not governed by ideas, but by morality, religion needed not play any role for social integration. All that was needed to insure social integration was the creation of a scientific ethic. This could be done without the help of religion, which could thus safely be left alone, concerned with matters of no practical importance (autonomization*).