/ Rathei than cause the triumph of Chiisti.mm the emperor Con-stantine s "{■dut of Milan" was an astute response to rapid Christian hroutil thai had alread\ made them a mai<>i tual torce. * CHAPTER 1 * Conversion and Christian Growth Jl inally, all questions concerning the rise of Christianity are one: How was it done? How did a tiny and obscure messianic movement from the edge of the Roman Empire dislodge classical paganism and become the dominant faith of Western civilization? Although this is the only question, it requires many answers—no one thing led to the triumph of Christianity. The chapters that follow will attempt to reconstruct the rise of Christianity in order to explain why it happened. But in this chapter I will pose the question in a more precise way than has been done. First, I shall explore the arithmetic of growth to see more clearly the task that had to be accomplished. What is the minimum rate of growth that would permit the Christian movement to become as large as it must have been in the time that history allows? Did Christianity grow so rapidly that mass conversions must have taken place—as Acts attests and every historian from Eusebius to Ramsay MacMullen has believed? Having established a plausible growth curve for the rise of Christianity, I will review sociological knowledge of the process by which people convert to new religions in order to infer certain requirements concerning social relations between Christians and the surrounding Greco-Roman world. The chapter concludes with a discussion of the legitimate uses of social scientific theories to reconstruct history in the absence of adequate information on what actually occurred. Since this book is a work of both history and social science, I have written it for a nonprofessional audience. In this way I can make sure that the social science is fully accessible to historians of the early church, meanwhile preventing social scientists from becoming lost amidst obscure historical and textual references. Before I proceed, however, it seems appropriate to discuss 3 CHAPTER 1 \ whether an attempt to explain the rise of Christianity is not somewhat sacrilegious. If, for example, I argue that the rise of Christianity benefited from superior fertility or from an excess of females who made possible high rates of exogamous marriage, am I not, thereby, attributing sacred achievements to profane causes? I think not. Whatever one does or does not believe about the divine, obviously God did not cause the world to become Christian, since that remains to be achieved. Rather, the New Testament recounts human efforts to spread the faith. No sacrilege is entailed in the search to understand human actions in human terms. Moreover, I do not reduce the rise of Christianity to purely "material" or social factors. Doctrine receives its due—an essential factor in the religion's success was what Christians believed. The Arithmetic of Growth Studies of the rise of Christianity all stress the movement's rapid growth, but rarely are any figures offered. Perhaps this reflects the prevalence among historians of the notion, recenüy expressed by Pierre Chuvin, that "ancient history remains wholly refractory to quantitative evaluations" (1990:12). Granted, we shall never discover "lost" Roman census data giving authoritative statistics on the religious composition of the empire in various periods. Nevertheless, we must quantify—at least in terms of exploring the arithmetic of the possible—if we are to grasp the magnitude of the phenomenon that is to be explained. For example, in order for Christianity to have achieved success in the time allowed, must it have grown at rates that seem incredible in the light of modern experience? If so, then we may need to formulate new social scientific propositions about conversion. If not, then we have some well-tested propositions to draw upon. What we need is at least two plausible numbers to provide the basis for extrapolating the proba- 4 CONVERSION AND CHRISTIAN GROWTH ble rate of early Christian growth. Having achieved such a rate and used it to project the number of Christians in various years, we can then test these projections against a variety of historical conclusions and estimates. For a starting number, Acts 1:14-15 suggests that several months after the Crucifixion there were 120 Christians. Later, in Acts 4:4, a total of 5,000 believers is claimed. And, according to Acts 21:20, by the sixth decade of the first century there were "many thousands of Jews" in Jerusalem who now believed. These are not statistics. Had there been that many converts in Jerusalem, it would have been the first Christian city, since there probably were no more than twenty thousand inhabitants at this time—J. C. Russell (1958) estimated only ten thousand. As Hans Conzelmann noted, these numbers are only "meant to render impressive the marvel that here the Lord himself is at work" (1973:63). Indeed, as Robert M. Grant pointed out, "one must always remember that figures in antiquity . . . were part of rhetorical exercises" (1977:7-8) and were not really meant to be taken literally. Nor is this limited to antiquity. In 1984 a Toronto magazine claimed that there were 10,000 Hare Krishna members in that city. But when Irving Hexham, Raymond F. Currie, and Joan B. Townsend (1985) checked on the matter, they found that the correct total was 80. Origen remarked, "Let it be granted that Christians were few in the beginning" (Against Celsus 3.10, 1989 ed.), but how many would that have been? It seems wise to be conservative here, and thus I shall assume that there were 1,000 Christians in the year 40. I shall qualify this assumption at several later points in the chapter. Now for an ending number. As late as the middle of the third century, Origen admitted that Christians made up "just a few" of the population. Yet only six decades later, Christians were so numerous that Constantine found it expedient to embrace the church. This has caused many scholars to think that something really extraordinary, in terms of growth, happened in the latter 5 CHAPTER 1 . half of the third century (cf. Gager 1975). This may explain why, of the few numbers that have been offered in the literature, most are for membership in about the year 300. Edward Gibbon may have been the first to attempt to estimate the Christian population, placing it at no more than "a twentieth part of the subjects of the empire" at the time of Con-stantine's conversion ([1776-1788] 1960:187). Later writers have rejected Gibbon's figure as far too low. Goodenough (1931) estimated that 10 percent of the empire's population were Christians by the time of Constantine. If we accept 60 million as the total population at that tíme—which is the most widely accepted estimate (Boak 1955a; Russell 1958; MacMul-len 1984; Wilken 1984)—this would mean that there were 6 million Christians at the start of the fourth century. Von Hert-ling (1934) estimated the maximum number of Christians in the year 300 as 15 million. Grant (1978) rejected this as far too high and even rejected von Hertling's minimum estimate of 7.5 million as high. MacMullen (1984) placed the number of Christians in 300 at 5 million. Fortunately, we do not need greater precision; if we assume that the actual number of Christians in the year 300 lay within the range of 5-7.5 million, we have an adequate basis for exploring what rate of growth is needed for that range to be reached in 260 years. Given our starting number, if Christianity grew at the rate of 40 percent per decade, there would have been 7,530 Christians in the year 100, followed by 217,795 Christians in the year 200 and by 6,299,832 Christians in the year 300. If we cut the rate of growth to 30 percent a decade, by the year 300 there would have been only 917,334 Christians—a figure far below what anyone would accept. On the other hand, if we increase the growth rate to 50 percent a decade, then there would have been 37,876,752 Christians in the year 300—or more than twice von Herding's maximum estimate. Hence 40 percent per decade (or 3.42 percent per year) seems the most plausible estimate of the rate at which Christianity actually grew during the first several centuries. 6 conversion and christian growth Table 1.1 Christian Growth Projected at 40 Percent per Decade Year Number of Christians Percent of Population* 40 1,000 0.0017 50 1,400 0.0023 100 7,530 0.0126 150 40,496 0.07 200 217,795 0.36 250 1,171,356 1.9 300 6,299,832 10.5 350 33,882,008 56.5 a Based on an estimated population of 60 million. This is a very encouraging finding since it is exceedingly close to the average growth rate of 43 percent per decade that the Mormon church has maintained over the past century (Stark 1984, 1994). Thus we know that the numerical goals Christianity needed to achieve are entirely in keeping with modern experience, and we are not forced to seek exceptional explanations. Rather, history allows time for the normal processes of conversion, as understood by contemporary social science, to take place. However, before we take up the topic of conversion, it seems worthwhile to pause and consider the widespread impression that Christian growth speeded rapidly during the last half of the third century. In terms of rate of growth, it probably did not. But because of the rather extraordinary features of exponential curves, this probably was a period of "miraculous-seeming" growth in terms of absolute numbers. All of this is clear in table 1.1. Progress must have seemed terribly slow during the first century—the projected total is only 7,530 by 100. There was a greater increase in numbers by the middle of the second century, but still the projection amounts to only slightly more than 40,000 Christians. This projection is in extremely close agree- 7 CHAPTER 1 ^ ment with Robert L. Wilken's estimate of "less than fifty thousand Christians" at this time—"an infinitesimal number in a society comprising sixty million" (1984:31). Indeed, according to L. Michael White (1990:110), Christians in Rome still met in private homes at this time. Then, early in the third century, the projected size of the Christian population picks up a bit and by 250 reaches 1.9 percent. This estimate is also sustained by a prominent historian's "feel" for the times. Discussing the process of conversion to Christianity, Robin Lane Fox advised that we keep "the total number of Christians in perspective: their faith was much the most rapidly growing religion in the Mediterranean, but its total membership was still small in absolute terms, perhaps (at a guess) only 2 percent of the Empire's total population by 250" (1987:317). But even more compelling is how the absolute number (as well as the percent Christian) suddenly shoots upward between 250 and 300, just as historians have reported,1 and recent archaeological findings from Dura-Europos support this view. Excavations of a Christian building show that during the middle of the third century a house church was extensively remodeled into a building "entirely devoted to religious functions," after which "all domestic activities ceased" (White 1990:120). The renovations mainly involved the removal of partition walls to create an enlarged meeting hall— indicative of the need to accommodate more worshipers. That my reconstruction of Christian growth exhibits the "sudden spurt" long associated with the second half of the third century adds to the plausibility of the figures. The projections are also extremely consistent with Graydon F. Snyder's (1985) assessment of all known archaeological evidence of Christianity during the first three centuries. Snyder determined that there really isn't any such evidence prior to 180. He interpreted this to indicate that before then it is impossible to distinguish Christian from non-Christian culture in "funerary art, inscriptions, letters, symbols, and perhaps buildings . . . [because] it took over a century for the new community of 8 CONVERSION AND CHRISTIAN GROWTH faith to develop a distinctive mode of self-expression" (Snyder 1985:2). That may be, but it must also be noted that the survival, of Christian archaeological evidence would have been roughly proportionate to how much there could have been to start with. The lack of anything surviving from prior to 180 must be assessed on the basis of the tiny number of Christians who could have left such traces. Surely it is not surprising that the 7,535 Christians at the end of the first century left no trace. By 180, when I project that the total Christian population first passed the 100,000 mark, there would finally have been enough Christians so that it is probable that traces of their existence would survive. Thus Snyder's findings are very compatible with my estimates of a very small Christian population in the first two centuries. As an additional test of these projections, Robert M. Grant has calculated that there were 7,000 Christians in Rome at the end of the second century (1977:6). If we also accept Grant's estimate of 700,000 as the population of Rome for that year, then 1 percent of the population of Rome had been converted by the year 200. If we set the total population of the empire at 60 million in 200, then, based on the projection for that year, Christians constituted 0.36 percent of the empire's population. This seems to be an entirely plausible matchup, since the proportion Christian should have been higher in Rome than in the empire at large. First of all, historians assume that the church in Rome was exceptionally strong—it was well known for sending funds to Christians elsewhere. In about 170, Dionysius of Corinth wrote to the Roman church: "From the start it has been your custom to treat all Christians with unfailing kindness, and to send contributions to many churches in every city, sometimes alleviating the distress of those in need, sometimes pro-riding for your brothers in the mines" (Eusebius, Ecclesiastical History 4.23.6, 1965 ed.). Second, by 200 the Christian proportion of the population of the city of Rome must have been substantially larger than that in the whole of the empire because 9 CHAPTER 1 Christianity had not yet made much headway in the more westerly provinces. As will be seen in chapter 6, of the twenty-two largest cities in the empire, four probably still lacked a Christian church by the year 200. Although I have estimated the overall number of Christians in the empire, I am fully aware that Christian growth was concentrated in the East—in Asia Minor, Egypt, and North Africa. Moreover, there is general agreement among historians (Harnack 1908; Boak 1955a; Meeks 1983) that the Christian proportion of the population was substantially higher in cities than in the rural areas at this time—hence the term paganus or "countryman" came to refer to non-Christians (pagans). In any event, here too the projections closely agree with estimates based on independent sources. Now, let us peek just a bit further into the future of Christian growth. If growth held at 40 percent per decade for the first half of the fourth century, there would have been 33,882,008 Christians by 350. In an empire having a population of at least 60 million, there might well have been 33 million Christians by 350—for by then some contemporary Christian writers were claiming a majority (Harnack 1908: 2:29). Looking at the rise of a Christian majority as purely a function of a constant rate of growth calls into serious question the emphasis given by Eusebius and others to the conversion of Constantine as the factor that produced the Christian majority (Grant 1977). So long as nothing changed in the conditions that sustained the 40-percent-a-decade growth rate, Constantine's conversion would better be seen as a response to the massive exponential wave in progress, not as its cause. This interpretation is entirely in keeping with the thesis developed by Shirley Jackson Case in his 1925 presidential address to the American Society of Church History. Case began by noting that attempts by the emperor Diocletian in 303, and continued by his successor Galerius in 305, to use persecution to force Christians to support the state had failed because "by the year 300 Christianity had become too widely accepted in Roman so- 10 CONVERSION AND CHRISTIAN GROWTH ciety to make possible a successful persecution on the part of the government" (1928:59). As a result, Case continued, by 311 the emperor Galerius switched tactics and excused the Chris- • tians from praying to Roman gods, and asked only that they pray to "their own god for our security and that of the state" (Case 1928:61). Thus Constantine's edict of toleration, issued two years later, was simply a continuation of state policy. Case's assessment of Constantine's edict stressed the impact of Christian growth on this policy: In this document one perceives very easily the real basis of Constantine's favor for Christianity. First, there is the characteristic attitude of an emperor who is seeking supernatural support for his government, and secondly, there is a recognition of the fact that the Christian element in the population is now so large, and its support for Constantine and Licinius in their conflict with rivals who still opposed Christianity, is so highly esteemed, that the emperors are ready to credit the Christian God with the exercise of a measure of supernatural power on a par with the other gods of the State. (1928:62) It is reassuring to have the projections of Christian membership in table 1.1 fit so well with several independent estimates, with major historical perceptions such as the rapid increases during the latter part of the third century, and with the record of Mormon growth achieved over the past century. Keep in mind, however, that the numbers are estimates, not recorded fact. They seem very plausible, but I would be entirely comfortable with suggestions that reality may have been a bit lumpier. Perhaps growth was somewhat more rapid in the earliest days and my beginning number of 1,000 Christians in 40 is a bit low. But it also seems likely that there were periodic losses in the early days, some of which may have been very substantial for a group still so small. For example, following the execution of James and the subsequent destruction of Jerusalem, the Christian community in Palestine seems to have died out (Frend 11 CHAPTER 1 1965, 1984). And while Tacitus's claim that "an immense multitude" (Annals 15.44, 1989 ed.) was butchered by Nero in about 65 is much exaggerated (see chapter 8), even the deaths of several hundred Christians would have been a very serious setback. I have tried to offset such bumps and lumps in the growth curve by starting with a very conservative number. Moreover, my purpose in generating these numbers was not to discover "facts," but to impose needed discipline on the subject. That is, by resorting to simple arithmetic I believe I have demonstrated adequately that the rise of Christianity required no miraculous rates of conversion. Several years after I had completed this exploration of the arithmetic of early Christian growth, when this book was nearly finished, my colleague Michael Williams made me aware of Roger S. Bagnall's remarkable reconstruction of the growth of Christianity in Egypt (1982, 1987). Bagnall examined Egyptian papyri to identify the proportion of persons with identifiably Christian names in various years, and from these he reconstructed a curve of the Christianization of Egypt. Here are real data, albeit from only one area, against which to test my projections. Two of Bagnall's data points are much later than the end of my projections. However, a comparison of the six years within my time frame shows a level of agreement that can only be described as extraordinary—as can be seen in table 1.2. Bagnall's finding no Christians in 239 can be disregarded. Obviously there were Christians in Egypt then, but because their numbers would still have been very small it is not surprising that none turned up in Bagnall's data. But for later years the matchups are striking, and the correlation of 0.86 between the two curves borders on the miraculous. The remarkable fit between these two estimates, arrived at via such different means and sources, seems to me a powerful confirmation of both. Although the projections seem very plausible through 350, the rate of Christian growth eventually must have declined rap- 12 conversion and christian growth Table 1.2 Two Estimates of Christianization Compared Projected Percent Christian in Year the Greco-Roman World Percent Christian in Egypt* 239 1.4 0 274 4.2 2.4 278 5.0 10.5 280 5.4 13.5 313 16.2 18.0 315 17.4 18.0 r = 0.86 Bagnall 1982, 1987. idly at some point during the fourth century. If nothing else, the empire would have begun to run out of potential converts. This is evident when we realize that had the 40 percent growth rate held throughout the fourth century, there would have been 182,225,584 Christians in the year 400. Not only is that total impossible, growth rates must always decline when a movement has converted a substantial proportion of the available population—as the pool of potential converts is progressively "fished out." Or, as Bagnall put it, "the curve of conversion becomes asymptotic, and incremental conversion becomes slight after a time" (1982:123). Clearly, then, the projections from my model are invalid after the year 350. However, since my concerns only involve the rise of Christianity, it is not necessary to venture beyond this point. On Conversion Eusebius tells us that early Christian missionaries were so empowered by the "divine Spirit" that "at the first hearing whole multitudes in a body eagerly embraced in their souls pietv towards the Creator of the universe" {Ecclesiastical History 3.37.3, 13 CHAPTER 1 ^ 1927 ed.). Not only do many modern historians of the early church accept Eusebius's claims about mass conversions in response to public preaching and miracle working, but they often regard it as a necessary assumption because of the rapidity of Christianity's rise. Thus in his distinguished study, Christianizing the Roman Empire, Ramsay MacMullen urged acceptance of the reports of large-scale conversions as necessary to explain better the rate of change we are observing. In the whole process, very large numbers are obviously involved ... [I] t would be hard to picture the necessary scale of conversion if we limited ourselves to . . . evangelizing in private settings ... [If this mode of conversion], however, is combined with evidence for successes en masse, the two in combination do seem to me adequate to explain what we know happened. (1984:29) MacMullen's views reflect those of Adolf Harnack (1908: 2:335-336), who characterized the growth of Christianity in terms such as "inconceivable rapidity" and "astonishing expansion," and who expressed his agreement with Augustine's claim that "Christianity must have reproduced itself by means of miracles, for the greatest miracle of all would have been the extraordinary extension of the religion apart from any miracles" (335n.2). This is precisely why there is no substitute for arithmetic. The projections reveal that Christianity could easily have reached half the population by the middle of the fourth century without miracles or conversions en masse. The Mormons have, thus far, traced the same growth curve, and we have no knowledge of their achieving mass conversions. Moreover, the claim that mass conversions to Christianity took place as crowds spontaneously responded to evangelists assumes that doctrinal appeal lies at the heart of the conversion process—that people hear the message, find it attractive, and embrace the faith. But modern social science relegates doctrinal appeal to a very secondary role, claiming that most people do not really become very 14 CONVERSION AND CHRISTIAN GROWTH attached to the doctrines of their new faith until after their conversion. In the early 1960s John Lofland and I were the first social scientists to actually go out and watch people convert to a new religious movement (Lofland and Stark 1965). Up to that time, the most popular social scientific explanation of conversion involved the pairing of deprivation with ideological (or theological) appeal. That is, one examined the ideology of a group to see what kinds of deprivation it addressed and then concluded (mirabile dictu!) that converts suffered from those deprivations (Glock 1964). As an example of this approach, since Christian Science promised to restore health, its converts must disproportionately be drawn from among those with chronic health problems, or at least those who suffer from hypochondria (Glock 1964). Of course, one could as plausibly argue the reverse, that only people with excellent health could long hold to the Christian Science doctrine that illness was all in the mind. In any event, Lofland and I were determined to watch people go through the process of conversion and try to discover what really was involved. Moreover, we wanted to watch conversion, not simply activation. That is, we wanted to look at people who were making a major religious shift, as from Christianity to Hinduism, rather than examine how lifelong Christians got themselves born again. The latter is a matter of considerable interest, but it was not our interest at the time. We also wanted a group that was small enough so that the two of us could provide adequate surveillance, and new enough so that it was in an early and optimistic phase of growth. After sifting through many deviant religious groups in the San Francisco Bay area we came upon precisely what we were looking for—a group of about a dozen young adults who had just moved to San Francisco from Eugene, Oregon. The group was led by Young Oon Kim, a Korean woman who had once been a professor of religion at Ewha University in Seoul. The movement she served was based in Korea, and in January 1959, she arrived in Oregon 15 CHAPTER 1 «j to launch a mission to America. Miss2 Kim and her young followers were the very first American members of the Unification Church, widely known today as the Moonies. As Lofland and I setüed back to watch people convert to this group, the first thing we discovered was that all of the current members were united by close ties of friendship predating their contact with Miss Kim. Indeed, the first three converts had been young housewives, next-door neighbors who became friends of Miss Kim after she became a lodger with one of them. Subse-quendy, several of the husbands joined, followed by several of their friends from work. At the time Lofland and I arrived to study them, the group had never succeeded in attracting a stranger. Lofland and I also found it interesting that although all the converts were quick to describe how their spiritual lives had been empty and desolate prior to their conversion, many claimed they had not been particularly interested in religion before. One man told me, "If anybody had said I was going to join up and become a missionary I would have laughed my head off. I had no use for church at all." We also found it instructive that during most of her first year in America, Miss Kim had tried to spread her message directly by talks to various groups and by sending out many press releases. Later, in San Francisco the group also tried to attract followers through radio spots and by renting a hall in which to hold public meetings. But these methods yielded nothing. As time passed, Lofland and I were able to observe people actually becoming Moonies. The first several converts were old friends or relatives of members who came from Oregon for a visit. Subsequent converts were people who formed close friendships with one or more members of the group. We soon realized that of all the people the Moonies encountered in their efforts to spread their faith, the only ones who joined were those whose interpersonal attachments to members overbalanced their attachments to nonmembers. In effect, conversion is 16 r CONVERSION AND CHRISTIAN GROWTH not about seeking or embracing an ideology; it is about bringing one's religious behavior into alignment with that of one's friends and family members. This is simply an application of the highly respected control theory of deviant behavior (Toby 1957; Hirschi 1969; Stark and Bainbridge 1987; Gottfredson and Hirschi 1990). Rather than asking why people deviate, why they break laws and norms, control theorists ask why anyone ever does conform. Their answer is posed in terms of stakes in conformity. People conform when they believe they have more to lose by being detected in deviance than they stand to gain from the deviant act. Some people deviate while others conform because people differ in their stakes in conformity. That is, some people simply have far less to lose than do others. A major stake in conformity lies in our attachments to other people. Most of us conform in order to retain the good opinion of our friends and family. But some people lack attachments. Their rates of deviance are much higher than are those of people with an abundance of attachments. Becoming a Moonie today is an act of deviance, as was becoming a Christian in the first century. Such conversions violate norms defining legitimate religious affiliations and identities. Lofland and I saw many people who spent some time with the Moonies and expressed considerable interest in their doctrines, but who never joined. In every instance these people had many strong attachments to nonmembers who did not approve of the group. Of persons who did join, many were newcomers to San Francisco whose attachments were all to people far away. As they formed strong friendships with group members, these were not counterbalanced because distant friends and families had no knowledge of the conversion-in-process. In several instances a parent or sibling came to San Francisco intending to intervene after having learned of the conversion. Those who lingered eventuallyjoined up too. Keep in mind that becoming a Moonie may have been regarded as deviant by outsiders, but 17 CHAPTER 1 it was an act of conformity for those whose most significant attachments were to Moonies. During the quarter century since Lofland and I first published our conclusion—that attachments lie at the heart of conversion and therefore that conversion tends to proceed along social networks formed by interpersonal attachments—many others have found the same to be true in an immense variety of religious groups all around the world. A recent study based on Dutch data (Kox, Meeus, and 't Hart 1991) cited twenty-five additional empirical studies, all of which supported our initial finding. And that list was far from complete. Although several other factors are also involved in the conversion process, the central sociological proposition about conversion is this: Conversion to new, deviant religious groups occurs when, other things being equal, people have or develop stronger attachments to members of the group than they have to nonmembers (Stark 1992). Data based on records kept by a Mormon mission president give powerful support to this proposition. When missionaries make cold calls, knock on the doors of strangers, this eventually leads to a conversion once out of a thousand calls. However, when missionaries make their first contact with a person in the home of a Mormon friend or relative ofthat person, this results in conversion 50 percent of the time (Stark and Bainbridee 1985). ö A variation on the network proposition about conversion is that successful founders of new faiths typically turn first to those with whom they already have strong attachments. That is, they recruit their first followers from among their family and close friends. Thus Muhammad's first convert was his wife Khadijah; the second was his cousin Ali, followed by his servant Zeyd and then his old friend Abu Bakr. On April 6, 1830, the Mormons were founded by Joseph Smith, his brothers Hyrum and Samuel, and Joseph Smith's friends Oliver Cowdery and David and Peter Whitmer. The rule extends to Jesus too, since it appears that he began with his brothers and mother. 18 CONVERSION AND CHRISTIAN GROWTH A second aspect of conversion is that people who are deeply committed to any particular faith do not go out and join some other faith. Thus Mormon missionaries who called upon the Moonies were immune, despite forming warm relationships with several members. Indeed, the Moonie who previously had "no use for church at all" was more typical. Converts were not former atheists, but they were essentially unchurched and many had not paid any particular attention to religious questions. Thus the Moonies quickly learned that they were wasting their time at church socials or frequenting denominational student centers. They did far better in places where they came in contact with the uncommitted. This finding has received substantial support from subsequent research. Converts to new religious movements are overwhelmingly from relatively irreligious backgrounds. The majority of converts to modern American cult movements report that their parents had no religious affiliation (Stark and Bainbridge 1985). Let me state this as a theoretical proposition: New religious movements mainly draw their converts from the ranks of the religiously inactive and discontented, and those affiliated with the most accommodated (worldly) religious communities. Had we not gone out and watched people as they converted, we might have missed this point entirely, because when people retrospectively describe their conversions, they tend to put the stress on theology. When asked why they converted, Moonies invariably noted the irresistible appeal of the Divine Principles (the group's scripture), suggesting that only the blind could reject such obvious and powerful truths. In making these claims converts implied (and often stated) that their path to conversion was the end product of a search for faith. But Lofland and I knew better because we had met them well before they had learned to appreciate the doctrines, before they had learned how to testify to their faith, back when they were not seeking faith at all. Indeed, we could remember when most of them regarded the religious beliefs of their new set of friends as quite odd. I recall one who told me that he was puzzled that such nice 19 CHAPTER 1 'r people could get so worked up about "some guy in Korea" who claimed to be the Lord of the Second Advent. Then, one day, he got worked up about this guy too. I suggest that this is also how people in the first century got themselves worked up about someone who claimed to be the Lord of the First Advent. Robin Lane Fox suggests the same thing: "Above all we should give weight to the presence and influence of friends. It is a force which so often escapes the record, but it gives shape to everyone's personal life. One friend might bring another to the faith. . . . When a person turned to God, he found others, new 'brethren,' who were sharing the same path" (1987:316). Peter Brown has expressed similar views: "Ties of family, marriages, and loyalties to heads of households had been the most effective means of recruiting members of the church, and had maintained the continued adherence of the average Christian to the new cult" (1988:90). The basis for successful conversionist movements is growth through social networks, through a structure of direct and intimate interpersonal attachments. Most new religious movements fail because they quickly become closed, or semiclosed networks. That is, they fail to keep forming and sustaining attachments to outsiders and thereby lose the capacity to grow. Successful movements discover techniques for remaining open networks, able to reach out and into new adjacent social networks. And herein lies the capacity of movements to sustain exponential rates of growth over a long period of time. Some readers may suspect that the rapid rise in the absolute number of new Christians between 250 and 350 would require mass conversions even though the rate of conversion remained constant at 40 percent per decade. Admittedly, exponential growth curves are counterintuitive and easily seem incredible. Nevertheless, the dynamics of the conversion process are not changed even as the absolute numbers reach a rapid growth stage along an exponential curve. The reason is that as movements grow, their social surface expands proportionately. That 20 CONVERSION AND CHRISTIAN GROWTH is, each new member expands the size of the network of attachments between the group and potential converts. As noted above, however, this occurs only if the group constitutes an open network. Thus if we are to better understand and explain the rise of Christianity, we must discover how the early Christians maintained open networks—for it would seem certain that they did. This last remark sets the stage for a brief discussion of the appropriate scope of social scientific theories and whether it is possible even to apply propositions developed in one time and place to other eras and cultures. On Scientific Generalization Many historians believe that cultures and eras verge on the unique. Thus in his very thoughtful response to my use of the network theory of conversion to discuss the success of the mission to the Jews (see chapter 3), Ronald F. Hock noted that I seem to think that networks, for example, are not "all that different from period to period, society to society" (1986:2-3). He then pointed out that the networks utilized by Mormons are those consisting of a member's family, relatives, and friends, but are ancient networks the same? Ancient cities are not modern ones, and ancient networks that were centered in aristocratic households included more than family and friends: domestic slaves, freedmen, and perhaps parasites, teachers, athletic trainers, and travelers. In addition, urban life was lived more in public, so that recruitment could proceed along more extensive and complex networks than we find among Mormons in our more nuclear and anonymous cities and suburbs. I am certain that Hock is correct, but I am unrepentant. What he is noting are details that might tell us how to discover networks should we be transported to ancient Antioch, but 21 CHAPTER 1 that have no implications for the network proposition per se. However people constitute structures of direct interpersonal attachments, those structures will define the lines through which conversion will most readily proceed. The definition of network is not locked to time and space, nor is the conversion proposition. Many historians seem to have considerable trouble with the idea of general theories because they have not been trained in the distinction between concepts and instances. Proper scientific concepts are abstract and identify a class of "things" to be regarded as alike. As such, concepts must apply to all possible members of the class, all that have been, are, shall be, or could be. The concept of chair, defined as all objects created to seat a lone individual and support his or her back, is an abstraction. We cannot see the concept of chair. It is an intellectual creation existing only in our minds. But we can see many actual chairs, and as we look at some, we discover immense variation in size, shape, materials, color, and the like. Moreover, when we look at chairs used in the ancient world, we perceive some very noticeable differences from the chairs of today. Nevertheless, each is a chair so long as it meets the definition set out above—other somewhat similar objects belong to other object classes such as stools and couches. These points apply as fully to the concept of social network as to the concept of chair. The concept of social network also exists only in our minds. All that we can see are specific instances of the class—networks involving some set of individuals. As with chairs, the shapes and sizes of social networks may differ greatly across time and space, and the processes by which networks form may vary as greatly as do techniques for making chairs. But these variations in details never result in chairs' becoming pianos, nor do variations in their makeup ever turn social networks into collections of strangers. It is only through the use of abstract concepts, linked by abstract propositions, that science exists. Consider a physics that CONVERSION AND CHRISTIAN GROWTH must generate a new rule of gravity for each object in the universe. And it is precisely the abstract generality of science that makes impossible for social science to contribute anything to our understanding of history, let alone to justify efforts to reconstruct history from social scientific theories. Let me now turn to that important issue. Social Theory and Historical Reconstructions During the past several decades historians of the New Testament era have become increasingly familiar with social science and have become increasingly inclined to use social scientific models to infer "what must have happened" in order to fill blanks in the historical and archaeological record. As Robin Scroggs pointed out in an influential essay, "there may be times when a sociological model may actually assist our ignorance. If our data evidence some parts of the gestalt of a known model, while being silent about others, we may cautiously be able to conclude that the absence of the missing parts is accidental and that the entire model was actually a reality in the early church" (1980:166). Since those lines were published, the practice Scroggs suggested has become common (Barton 1982, 1984; Holmberg 1980; Elliott 1986; Fox 1987; Gager 1975, 1983; Green 1985; Malina 1981, 1986; Meeks 1983, 1993; Kee 1983; Kraemer 1992; Sanders 1993; Theissen 1978, 1982; Wilken 1984; Wire 1991). I have quite mixed reactions to this literature. Some studies I have read with pleasure and admiration. Other examples have made me very uncomfortable because the social science "models" utilized are so inadequate. Some of them are merely metaphors—as Durkheim's "discovery" that religion is society worshiping itself is merely metaphor. How would one falsify that statement, or assertions to the effect that religion is a neurotic illusion or the poetry of the soul? The 23 CHAPTER 1 problem with metaphors is not that they are false, but that they are empty. Many of them do seem to ooze profundity, but at best metaphors are merely definitions. Consider the term charisma. Max Weber borrowed this Greek word meaning "divine gift" to identify the ability of some people to convince others that their authority is based on divine sources: "The holder of charisma seizes the task that is adequate for him and demands obedience and a following by virtue of his mission. His success determines whether he finds them. His charismatic claim breaks down if his mission is not recognized by those to whom he feels he has been sent. If they recognize him, he is their master" (1946:246). Charisma is commonly observed in religious leaders, and surely no one would dispute that Jesus and many of the apostles and early evangelists had it. Thus the literature on the early church is saturated with the term. Unfortunately, charisma is too often understood as a nearly magical power possessed by individuals rather than a description of how they are regarded. That is, their power over others is attributed to their charisma, and it is often suggested that particular religious leaders are so potent because they had charisma. Roy Wallis, for example, claimed that Moses David (David Berg), founder of the Children of God, maintained control over his followers because of his "charismatic status" (1982:107). But this is entirely circular. It is the same as saying that people believed that Moses David had divine authority because people believed he had divine authority. Because Weber's discussions of charisma did not move beyond definitional and descriptive statements, and said nothing about the causes of charisma, the concept is merely a name attached to a definition. When we see someone whose authority is believed by some people to be of divine origin, we have the option of calling this charisma, but doing so will contribute nothing to our understanding of why this phenomenon occurs. Hence when studies of the early church utilize the term charisma, what we usually confront is only a name that too often is thought to explain something, but does not. 24 CONVERSION AND CHRISTIAN GROWTH Besides metaphors and simple concepts, other "models" used in this literature are nothing but typologies or sets of concepts. One of the most popular of these consists of various definitions to distinguish religious groups as churches or sects. The most useful of these definitions identifies churches and sects as the end points of a continuum based on the degree of tension between the group and its sociocultural environment (Johnson 1963; Stark and Bainbridge 1979, 1987). Sects are religious groups in a relatively high state of tension with their environment; churches are groups in a relatively low state of tension. These are very useful concepts. Unfortunately, they are often used, even by many social scientists, as if they explained something. All such efforts are circular. Thus it is circular to say that a particular religious body rejects the world because it is a sect, as Bryan Wilson (1970) often does, since bodies are classified as sects because they reject the world. The concepts of church and sect do nothing more (or less) than allow us to classify various religious bodies. But theories using these concepts do not reside in the concepts themselves. For example, it is well known that religious bodies, especially if they are successful, tend to move from a higher to a lower state of tension—sects often are transformed into churches. But no explanation of this transformation can be found in the definitions of church and sect. Instead, we must use propositions to link the concepts of church and sect to other concepts, such as upward social mobility and regression to the mean, in order to formulate an explanation (Stark and Bainbridge 1985, 1987). Let me emphasize: concepts are names, not explanations. The act of naming some objects or phenomena tells us nothing about why they occur or what they influence. Explanation requires theories: abstract statements saying iuhya.ná hoxv some set of phenomena are linked, and from which falsifiable statements can be derived (Popper 1959, 1962). Metaphors, typologies, and concepts are passive; they cast no light of their own and cannot illuminate the dark corners of unrecorded history 25 CHAPTER 1 (Stark and Bainbridge 1979, 1985, 1987). Granted, concepts may permit some useful comparisons among some sets of phenomena—comparisons of the social class composition of two religious movements, for example, can be very revealing. But if a model is to provide more than classification, if it proposes to explain, then the model must include not simply concepts, but propositions. The difference here is that between a parts catalog and a working diagram of an engine. That is, a model must include a fully specified set of interrelations among the parts. Such a model explains why and how things fit together and function. For this task, only a theory, not a conceptual scheme, suffices. It is not surprising that scholars trained in history and in textual interpretation might find themselves more comfortable with an older generation of social "scientists" who dealt in metaphors rather than scientific theories, if for no other reason than that their work abounds in literary allusions and is redolent of ancient library dust. But let it be noted that in science, unlike papyrology, older seldom is better. And I regard it as an essential part of my task in this book to familiarize historians of the early church with more powerful and modern social scientific tools, and particularly with real theories rather than with concepts, metaphors, and typologies pretending to have explanatory power. However, even if we use the best social science theories as our guide for reconstructing history, we are betting that the theories are solid and that the application is appropriate. When those conditions are met, then there is no reason to suppose that we cannot reason from the general rule to deduce the specific in precisely the same way that we can reason from the principles of physics that coins dropped in a well will go to the bottom. Even so, it is better when we can actually see the coins go down. Need is the only justification for the application of social science to fill in historical blanks. But we must be very cautious not to fill the blanks with fantasy and science fiction. 26 CONVERSION AND CHRISTIAN GROWTH In this book I shall attempt to reconstruct the rise of Christianity on the basis of many inferences from modern social scientific theories, making particular use of my own formal theorizing about religion and religious movements (Stark and Bainbridge 1979, 1980, 1985, 1987; Stark and Iannaccone 1991, 1992). I will frequendy employ the arithmetic of the possible and the plausible to test various assumptions. To guard against error I shall test my reconstructions against the historical record whenever possible, as I have done in this chapter.3 27