CHAPTER 6 
CHAPTER 6 IN HIS brilliant study of the early church, Wayne Meeks (1983) uses the title of his book to emphasize that Christianity was first and foremost an urban movement. Or, as he put it early in the first chapter, “within a decade of the crucifixion of Jesus, the village culture of Palestine had been left far behind, and the Greco-Roman city became the dominant environment of the Christian movement” (1983:11). In the remainder of the book Meeks offers many insights about the spread of Christianity; his primary emphasis, however, is not on cities, but on urbanites. His aim is to help us recognize who embraced the new movement and why.
My concern in this chapter is not so much with who or why as with where. What characteristics of cities were conducive to Christianization? To this end I will apply some standard tools used by urban sociologists and conduct a quantitative analysis using a data set consisting of the twenty-two largest cities of the Greco-Roman world circa 100. I will develop and test some hypotheses about why Christianity arose more rapidly in some places than in others. However, rather than present the hypotheses first and then move to the statistical analysis, I shall develop and test each seriatim. The reason for this format is that each variable must be discussed at some length as it enters the analysis, and each variable reflects a hypothesis.
Christianity was an urban movement and, like many new faiths before it, it arose in the east and spread westward, as can be seen in this map showing the twenty-two largest cities of the Roman Empire (ca. 100)
Despite libraries stuffed with books on many Greco-Roman cities, the sad fact is that, as Lewis Mumford has pointed out, “the city itself remains a shadow” (1974:vii). Indeed, it has required Herculean efforts even to estimate such an elementary and essential fact as the population of these cities. Fortunately, Tertius Chandler made it his life’s work to try to discover the populations of ancient cities. Assisted by Gerald Fox (with Kingsley Davis performing the vital role of midwife), he was finally able to publish his extraordinary work (Chandler and Fox 1974). In Three Thousand Years of Urban Growth, Chandler and Fox offer a plausible and well-documented basis for estimating the populations of the world’s largest cities in 100. Among these are twenty Greco-Roman cities. However, because Chandler and Fox chose to list only those cities having a population of 40,000 or more, they provided no population estimates for Athens or Salamis, although these are usually included in lists of important Greco-Roman cities. I added these two cities, bringing the total to twenty-two. I was unable to muster any Chandler-like scholarship to determine their populations at this time. But, after a good deal of poking around, I settled on 35,000 for Salamis and 30,000 for Athens.1 If these prove faulty, be assured that removing them from the analysis has no effect on the results I report below.
Here are the twenty-two cities and their estimated populations:
| Rome | 650,000 |
| Alexandria | 400,000 |
| Ephesus | 200,000 |
| Antioch | 150,000 |
| Apamea | 125,000 |
| Pergamum | 120,000 |
| Sardis | 100,000 |
| Corinth | 100,000 |
| Gadir (Cadiz) | 100,000 |
| Memphis | 90,000 |
| Carthage | 90,000 |
| Edessa | 80,000 |
| Syracuse | 80,000 |
| Smyrna | 75,000 |
| Caesarea Maritima | 45,000 |
| Damascus | 45,000 |
| Cordova | 45,000 |
| Mediolanum (Milan) | 40,000 |
| Augustodunum (Autun) | 40,000 |
| London | 40,000 |
| Salamis | 35,000 |
| Athens | 30,000 |
How can we measure the receptivity of cities to Christianity, that is, their relative degree of Christianization at various times? My method is neither original nor (I would hope) particularly controversial. I have simply followed Adolf Harnack (1908, vol. 2) in using the notion of the expansion of Christianity—for in order to rise, a movement must spread. In his masterwork, Harnack identified those communities in the empire that possessed local Christian churches by the year 180. Later scholars have added much to Harnack’s original reconstruction, drawing on the many important archaeological finds of recent decades. However, owing to the lack of quantitative inclination among scholars in this area, it is only in the numerous historical atlases that this literature has been pulled together. Through study of many an atlas, I found four that seemed to reflect solid scholarship on this particular topic, and these are shown in table 6.1, along with Harnack’s original findings (Blaiklock 1972; Aharoni and Avi-Yonah 1977; Frank 1988; Chadwick and Evans 1987; Harnack 1908).
TABLE 6.1
Coding Christianization
| Source 1a | Source 2b Source | 3c | Source 4d | Source 5e | Codef | |
| Caesarea Maritima | 2 | 2 | 2 | 2/1 | 2 | 2 |
| Damascus | 2 | 2 | 2 | 2/1 | 0 | 2 |
| Antioch | 2 | 2 | 2 | 2/1 | 2 | 2 |
| Alexandria | 2 | 2 | 2 | 2/1 | 2 | 2 |
| Pergamum | 2 | 2 | 2 | 2/1 | 2 | 2 |
| Salamis | 2 | 2 | 2 | 2/1 | 2 | 2 |
| Sardis | 2 | 2 | 2 | 2/1 | 2 | 2 |
| Smyrna | 2 | 2 | 2 | 2/1 | 0 | 2 |
| Athens | 2 | 2 | 2 | 2/1 | 2 | 2 |
| Corinth | 2 | 2 | 2 | 2/1 | 0 | 2 |
| Ephesus | 2 | 2 | 2 | 2/1 | 2 | 2 |
| Rome | 2 | 2 | 2 | 2/1 | 2 | 2 |
| Apamea | 1 | 1/0 | 1 | 0 | 2 | 1 |
| Cordova | 1 | 1/0 | 1 | 0 | 1 | 1 |
| Edessa | 1 | 1/0 | 1 | 2/1 | 1 | 1 |
| Syracuse | 1 | 1/0 | 1 | 2/1 | 1 | 1 |
| Carthage | 1 | 1/0 | 1 | 2/1 | 1 | 1 |
| Memphis | 2 | 1/0 | 1 | 2/1 | 1 | 1 |
| Mediolanum (Milan) | 0 | 1/0 | 0 | 0 | 0 | 0 |
| Augustodunum (Autun) | 0 | 1/0 | 0 | 0 | 0 | 0 |
| Cadir (Cadiz) | 0 | 1/0 | 0 | 0 | 0 | 0 |
| London | 0 | 1/0 | 0 | 0 | 0 |
a Aharoni and Avi-Yonah 1977 (map of cities with churches by end of first century and map of cities with churches by end of the second century).
b Chadwick and Evans 1987 (map of cities known to have had churches by the end of the first century).
c Frank 1988 (map of cities known to have had a church by the end of the first century and by 180 C.E.).
d Harnack 1908 (map of cities known to have had a church by 180 C.E.).
e Blaiklock 1972 (map shaded to show cities known to have had a church by end of first century and by end of second century).
f In the codes used when the chapter appeared as an essay in Sociological Analysis, Memphis was scored 2 rather than 1 and Cordova was scored 0 rather than 1. Upon further research I decided to make these corrections. However, they did not alter the statistical results in any important way.
I have quantified the expansion of Christianity in terms of three thresholds. The cities most receptive to Christianity are those known to have had a church by 100. They receive a score of two. The next most receptive are those cities known to have had a church by 200. They receive a score of one. The least receptive cities are those still lacking a church by 200. Their score is zero. The result is a three-value, ordinal measure of Christianization.
Cities scored two are: Caesarea Maritima, Damascus, Antioch, Alexandria, Pergamum, Salamis, Sardis, Smyrna, Athens, Corinth, Ephesus, and Rome.
Cities scored one are: Apamea, Carthage, Cordova, Edessa, Memphis, and Syracuse.
Cities scored zero are: Augustodunum (Autun), Gadir (Cadiz), London, and Mediolanum (Milan).
Let us pause here for our first hypothesis. Is there any reason to suppose that city size would have influenced Christianization? Harnack thought so: “The larger the town or city, the larger (even relatively, it is probable) was the number of Christians” (1908: 2:327). Moreover, there is a solid theoretical basis for such a hypothesis in the sociological literature. In his wellknown subcultural theory of urbanism, Claude S. Fischer offered this proposition: “The more urban the place, the higher the rates of unconventionality” (1975:1328). Fischer’s thesis is that the larger the population, in absolute numbers, the easier it is to assemble a “critical mass” needed to form a deviant subculture. Here he specifically includes deviant religious movements. During the period in question Christianity obviously qualifies as a deviant religious movement in that it clearly was at variance with prevailing religious norms. Therefore, Fischer’s theory of urbanism predicts that Christians would have assembled the critical mass needed to form a church sooner, the larger the city.
As can be seen in table 6.2 below, there is a positive correlation in support of Fischer’s thesis. Although this correlation falls slightly short of significance at the .05 level, it is not at all clear that significance is an appropriate standard here, since the data are not based on a random sample. Indeed, it was with great reservations that I included significance levels in the table.
One thing we know with certainty about these cities is where they were. And that means we can measure travel distances from one to another. Therefore, I have determined the distance of each of these twenty-two cities from Jerusalem.
We know where Christianity began. If we want to discover how it spread, we ought to take into account how far it had to go to get to various cities. The issue here is not simply that missionaries had to go farther to get from Jerusalem to Mediolanum than to Sardis. Indeed, anyone could cross the empire from one end to the other in a summer, and travel was common. Meeks (1983:17) reports a merchant’s grave inscription found in Phrygia that attests to his having traveled to Rome seventy-two times, a distance of well over a thousand miles, and Ronald Hock (1980) estimates that Paul covered nearly ten thousand miles on his missions. As Meeks put it, “the people of the Roman Empire traveled more extensively and more easily than anyone before them did or would again until the nineteenth century” (1983:17).
My interest is in the primary consequences of all this travel and trade: communication, cultural contact, and networks of interpersonal relationships based on kinship, friendship, or commerce. As I will discuss below, these were vital factors in preparing the way for the Christian missionaries—in determining what kind of reception awaited them. I propose to use distance from Jerusalem as a gradient of these factors.
Given this interest, simple distance as the crow flies is inadequate. Instead, in making the actual measurements I have attempted to trace known travel routes. Moreover, the bulk of trade and of long-distance travel was by boat—Paul traveled as much or more by sea as by land. Therefore, I have assumed sea travel whenever it was feasible and measured distances along the commonly used routes. First, I sketched the route from Jerusalem to a given city on a map. Then, I used a map meter to measure the distance—this makes it easy to follow curves and turn corners. Each measurement was made several times. And the final measurement was then converted to miles on the basis of the map’s legend. The routes themselves may be subject to modest errors, but I would not expect errors in excess of plus or minus 10 percent. In order to see the potential impact of errors of that magnitude, I created an additional mileage measure by flipping a coin and adding or subtracting 10 percent depending on whether the coin came up heads or tails. The distorted measure in fact correlated .99 with the original, and each yields identical results with other variables.
In addition to seeking to use travel distances to estimate the degree to which the way was prepared for Christians by the prior ties to Jerusalem and to Jewish culture, we can also use distance as a way to measure the degree of Romanization and the tightness of Roman control. Using the same tactics described above, I therefore measured the travel distance to each city from Rome. Finally, I created a ratio of the two sets of distances to summarize the relative weights of Roman and Jewish influence. The actual distances in miles follow. (The distance given between Athens and Rome is calculated on the assumption that the traveler’s boat did not take the portage at Corinth.)
| From Jerusalem, | From Rome | |
| Alexandria | 350 | 1,400 |
| Antioch | 250 | 1,650 |
| Apamea | 280 | 1,600 |
| Athens | 780 | 1,000 |
| Augustodunum | 1,920 | 525 |
| (Autun) | ||
| Caesarea Maritima | 60 | 1,575 |
| Carthage | 1,575 | 425 |
| Cordova | 2,440 | 1,225 |
| Corinth | 830 | 800 |
| Damascus | 130 | 1,600 |
| Edessa | 550 | 1,775 |
| Ephesus | 640 | 1,185 |
| Gadir (Cadiz) | 2,520 | 1,200 |
| London | 2,565 | 1,190 |
| Mediolanum | 1,900 | 260 |
| (Milan) | ||
| Memphis | 360 | 1,500 |
| Pergamum | 840 | 1,200 |
| Rome | 1,480 | 0 |
| Salamis | 270 | 1,450 |
| Sardis | 700 | 1,300 |
| Smyrna | 820 | 1,150 |
| Syracuse | 1,100 | 375 |
In chapter 3, I stressed the importance of cultural continuity in the success of new religious movements. Specifically, people are more willing to adopt a new religion to the extent that it retains cultural continuity with conventional religion(s) with which they are already familiar In the instance at hand, the way was paved for Christianity to the extent that people were already familiar with Jewish culture—the “God-Fearers” being an apt example. Here were people familiar with Jewish theology, who accepted the idea of monotheism, but who were unwilling to become ethnic Jews in order to fully participate in the Jewish religion. Presumably, the principle of cultural continuity as a facilitating factor in the spread of Christianity can, to some extent, be estimated by distance from Jerusalem. And table 6.2 shows an immense negative correlation between distance from Jerusalem and Christianization (-.74), which is highly statistically significant.
In chapter 3, I argued that, in fact, the mission to the Jews was quite successful and that a steady and significant flow of Hellenized Jewish converts to Christianity probably continued into the late fourth or early fifth century. To recapitulate, my case rests on several sociological propositions. The first is cultural continuity. Not only was Christianity highly continuous with the Jewish heritage of diasporan Jews, it was also highly congruent with their Hellenic cultural elements. The second proposition is that social movements recruit primarily on the basis of interpersonal attachments that exist, or form, between the convert and members of the group. And who were the friends and relatives of the early Christian missionaries setting out from Jerusalem to spread their faith? The Jews of the diaspora, of course. In fact, many of the missionaries were, like Paul himself, diasporan Jews.
Even if I am wrong about how late Jewish conversion continued, everyone is agreed that Jews were the primary sources of converts until well into the second century. As Harnack put it:
The synagogues of the Diaspora . . . formed the most important presupposition for the rise and growth of Christian communities throughout the empire. The network of the synagogues furnished the Christian propaganda with centres and courses for its development, and in this way the mission of the new religion, which was undertaken in the name of the God of Abraham and Moses, found a sphere already prepared for itself. (1908: 1:1)
So, in addition to using distance to measure Jewish cultural influences, we ought to seek a measure of Jewish presence in cities. There simply is no good way to calculate the probable size of the Jewish population in these cities. The best substitute I could obtain is information on which of these cities are known to have had a synagogue in about 100. The data come from many of the atlases noted above and others, and from MacLennan and Kraabel (1986). The result is a dichotomous variable scoring cities with a synagogue as one, and the others as zero. The following cities received a score of one: Caesarea Maritima, Damascus, Antioch, Alexandria, Sardis, Athens, Rome, Corinth, and Ephesus. Table 6.2 shows that there is a powerful, positive correlation between synagogues and Christianization (.69). Clearly, then, Christianity took root sooner where there were Jewish communities.
TABLE 6.2
Pearson Product Moment Correlations
| Pop. 100 C.E. | From Jerusalem | Synagogues | From Rome | Romanized | Gnostics | |
| Christianization | .32 | -.74a | .69a | .42b | -.71a | .59a |
| Population 100 C.E. | -.06 | .41b | .21 | -.29 | .48a | |
| Miles from Jerusalem | -.46b | -.54a | .68a | -.49b | ||
| Synagogues | .28 | -.44b | .41b | |||
| Miles from Rome | -.84a | .37 | ||||
| Romanized | -.43b |
a Significance level: <.01.
b Significance level: <.05.
Now what about Roman culture and power? In the beginning Christianity did best in Greek cities and soon incurred considerable official Roman antagonism. It seems realistic to treat Roman power as a function of distance—the farther from Rome, the less the local impact of Roman policy. Once again we may simply trace the trade routes to each city from Rome and measure the distances. However, since our interest really centers on the interaction of Roman and Eastern culture and influence, we can divide the distance to Jerusalem by the distance to Rome. The higher the ratio, the greater the relative weight of Roman influence; hence this variable is identified as Romanization (Rome, of course, is excluded from this analysis). Table 6.2 shows that distance from Rome is negatively correlated (-.42) with Christianization, but the really potent effect is from Romanization (-.71). The more Roman and the less Eastern (Greek and Jewish) influence on a city’s culture, the later its first church—Rome itself being the obvious exception.
TABLE 6.3
Regression: Dependent Variable Is Christianization
| Independent Variables | Unstandardized Betas | Standardized Betas | Standard Error | T |
| Synagogues | 0.731 | 0.466 | 0.236 | 3.099a |
| Romanized | -0.220 | -0.499 | 0.077 | -3.317a |
| Multiple R-Square = 0.672 | Y-Intercept= 1.374 | |||
a Significant beyond .001.
Table 6.3 shows the result of entering Christianization, synagogues, and Romanization into a regression equation. Each of the independent variables displays a robust effect, and together they explain an amazing 67 percent of the variance in Christianization.
Not only were there many new religious movements active in the urban empire in this era, there were many Christianities. Almost from the start, factions espousing rather different views of Christ and of Scripture arose, each seeking to be the Christianity. Since the discovery of the Nag Hammadi manuscripts, there has been immense interest in the groups known as the Gnostics or the Christian Gnostics (Layton 1987; Williams 1985). A map published by Layton (1987:6-7) makes it possible to create a measure of Gnostic presence similar to the measure of Christianization. Cities known to have had active Gnostic groups prior to the year 200 were scored two. Those having Gnostic groups prior to the year 400 were scored one. Cities not known to have had Gnostic groups by 400 were scored zero. Those scored two were Alexandria, Antioch, Caesarea Maritima, Carthage, Ephesus, Pergamum, Rome, Sardis, and Smyrna. Those scored one were Apamea, Damascus, Edessa, and Memphis. The others were scored zero.
Table 6.2 shows a substantial positive correlation between Gnostics and Christianization (.59). Moreover, Gnostic presence is significantly correlated with population size, in agreement with Fischer’s theory. The table also shows a significant correlation between Gnostic presence and synagogues. These findings are worth pursuing at greater length.
For the past century there has been a serious debate about the connections between Gnosticism, on the one hand, and Christianity and Judaism, on the other. Late in the nineteenth century, Harnack (1894) was content to classify the Gnostics as a Christian heresy, as an acutely Hellenized brand of Christianity. Soon thereafter, however, many scholars (e.g., Friedlander 1898) began to trace the origins of the Gnostics to Jewish roots, and to regard Christianity and Gnosticism as parallel offshoots of first-century Judaism. Despite the Christian content of most of the manuscripts discovered at Nag Hammadi, the debate continues, and the view that the two movements were parallel probably has more support these days than does the view that the Gnostics were primarily a competing stream within Christianity. Birger Pearson, echoing Friedlander, puts this position most forcefully: “Gnosticism is not, in its origins, a ‘Christian’ heresy, but . . . it is, in fact, a ‘Jewish’ heresy” (1973:35).
If we bear that in mind, table 6.4 is of more than passing interest. Here regression analysis is used to assess the net effects of Christianization and Jewish presence on the rise of Gnosticism. The results are very conclusive—at least from the statistical point of view. When the effects of Christianization are held constant, no direct Jewish effects remain. Whereas Jewish presence has a very substantial impact on the spread of Christianity (see table 6.3), only Christianity seems to have any impact on the rise of Gnosticism. This suggests a causal order fully in keeping with Harnack’s original position: that Christianity began as a Jewish heresy and its initial appeal was to Jews, but that Gnosticism began subsequently as a Christian heresy, appealing mainly to Christians (from whom it adopted its very stridently anti-Jewish content).
TABLE 6.4
Regression: Dependent Variable Is Gnosticism
| Independent Variables | Unstandardized Betas | Standardized Betas | Standard Error | T |
| Christianization | 0.678 | 0.578 | 0.299 | 2.262a |
| Synagogues | -0.022 | -0.012 | 0.470 | -0.047 |
| Multiple R-Square = 0.344 | Y-Intercept = 0.067 | |||
a Significant beyond .05.
Such statistical evidence is not, of course, conclusive proof that Gnosticism was a Christian heresy. But it seems worth noting here that participants in this debate may well be talking past one another. As I understand the proponents of the Jewish origins of Gnosticism, their concern has been to trace the origins of some of the central mystical notions of the Gnostics to prior Jewish writers. But as sociologists understand these matters, heresy per se has little to do with the pedigree of ideas and consists primarily of the embodiment of “deviant” ideas in a social movement. Put another way, writings can be heretical, but only human beings can be heretics. Moreover, the origins of ideas and of movements need not, and often are not, the same. Consider the many modern groups with spurious claims to unbroken descent from ancient pagan cults. Judged by their doctrines, their claims to be of ancient origin are true. But an examination of their “human history” reveals them to be of contemporary origins. Thus Gnostic writers could have been profoundly influenced by the writings of earlier Jewish mystics without representing a social movement coexisting with Christianity and deriving from pre-Christian origins among Jews. Although the data in table 6.4 have no implications for the origins of Gnostic ideas, they do encourage the conclusion that Gnosticism, as a social movement, was a Christian heresy.
Whatever the impact of table 6.4 on historical interpretations of Gnosticism, it is obvious that the other findings reported in this chapter are not going to greatly revise social histories of the rise of Christianity. Even without quantification, every competent historian has known that the Christian movement arose most rapidly in the Greco-Roman cities of Asia Minor, sustained by the very large communities of the Jewish diaspora. Indeed, the findings of greatest substantive interest are probably those lending support to Fischer’s propositions about city size and subcultural deviance.
In my judgment, the real surprises are statistical, not substantive. The magnitudes and stability of the statistical outcomes are amazing and strongly testify that a data set based on these twenty-two cities can be of great scholarly value—to the extent that we are able to identify and secure interesting variables.
It strikes me that many additional variables could be created by scholars with the proper training. It would be very interesting to build rates for these cities (and perhaps other aggregate units) based on the immense collections of inscriptions. For example, since historians are agreed that Christianity was one of many new religions to come out of the East, can we use inscriptions to estimate when and where these “Oriental cults” gained followings? In chapter 9 we will examine a variable based on when and whether the Isis cult had a temple in these cities, but many more could be coded.
Initially, I had hoped to create measures of social disorganization of these cities, especially factors that disrupt integration by reducing the strength of interpersonal attachments. It is axiomatic that conformity to the norms is the result of attachments—to the extent that we value our relationships with others, we will conform in order to retain their esteem. When people lack attachments, they have much greater freedom to deviate from the norms. In modern studies, unconventional behavior is strongly correlated with various measures of population turnover and instability. For example, where larger proportions of the U.S. and Canadian populations are newcomers or have recently moved from one residence to another, rates of participation in unconventional religious activities are high (Stark and Bainbridge 1985).
I began by examining data on when and how a city was founded, or refounded, and the ethnic heterogeneity of its population. I was fascinated that both Corinth and Carthage had stood empty when Caesar decided to refound them in order to transport large numbers of Rome’s “undesirable” population. To this he added a bunch of retired legionnaires who, in turn, drew numbers of women to the city from nearby villages. Talk about Dodge City, or some other wild and wooly place. As I proceeded, however, I began to realize that all the cities of the empire were incredibly disorganized, even compared with rapidly growing and industrializing cities of the nineteenth century, the ones that caused early sociologists to express endless gloom and doom. What Rome had achieved was political unity at the expense of cultural chaos. No one has captured this fact more lucidly that Ramsay MacMullen in the opening sentences of his remarkable work on paganism:
It was a proper melting pot. If we imagined the British Empire of a hundred years ago all in one piece, all of its parts touching each other, so one could travel. . . from Rangoon to Belfast without the interposition of any ocean, and if we could thus sense as one whole an almost limitless diversity of tongues, cults, traditions, and levels of education, then the true nature of the Mediterranean world in [the Roman era] . . . would strike our minds. (1981:xi)
For these reasons I ceased my efforts to compare cities in terms of disorganization and shall, instead, devote the next chapter to tracing how the acute disorganization of Greco-Roman cities in general eased the rise of Christianity.
Greco-Roman cities were covered with graffiti—this anti-Christian example (ca. 200) was found in Rome. It shows an ass-headed figure on the cross and reads, “Alexamenos worships his god.”
A preliminary version of this chapter appeared as “Christianizing the Urban Empire,” in Sociological Analysis 52 (1991): 77-88.