The Methodology
Introduction
ringa, who may be called the father of the socio-historical approach to early Christianity, argued that the earliest church patterned itself after house synagogues.46 In 1753, Mosheim presented a similar line of argument.47 As we have seen, most subsequent scholars focused on the parallels Vitringa made between the Jewish and Christian presbyterate. Kist, however, cited Vitringa in support of his contention that the earliest Christians met in the houses of its wealthier members who acted as its first leaders.48 Baur49 and Weingarten50 presented similar arguments. This approach was short-lived. Although scholars such as Kirsch,51 Hauck,52 and Leclerq53 continued to argue that the early church met in houses, deductions regarding leadership were not drawn from this fact.54
After the discovery of the house church at Dura-Europas in 1930-31,55 however, Filson postulated that the development of the ministry could not be understood without reference to the house church.56 In 1945 Dix noted that the early eucharist took place in private houses.57 And Farrer in 1946 argued that the apostolic church's hospitable patrons were its bishops.58 In 1962, Telfer argued that bishops lived in "church houses" commonly owned by the community, and were responsible for offering hospitality to travellers.59 In 1949,
Wagner, in a thesis which unfortunately has remained unpublished, assessed the importance of the household in the development of the eucharist, again without attempting to determine how it influenced the ministry.60
Social-historical studies since 1960 have attempted to assess the importance of the household for leadership in Pauline churches,61 but there is often little reference to the extra-canonical evidence. Hans-Josef Klauck, in what is becoming the authoritative work on house churches in the N e w Testament, provides a fruitful account of the impact of the household on the earlier churches, but is disappointingly reticent about identifying a similar setting in later canonical and extra-canonical evidence.62 Most recently, Ernst Dassmann has painted in very broad strokes a house-church setting of Ignatius' letters.63 Literary and archaeological evidence suggests that Christians met in private homes at least until the beginning of the third century.64 This field is relatively untouched; although the N e w Testament data has been investigated with enthusiasm in recent years, no scholars of whom I am aware, Dassmann excepted, have ventured outside its limits.
The Methodology Subsequent chapters employ insights gleaned from the social sciences, especially sociology. Sociological explanations of early Christianity have become increasingly popular amongst N e w Testament scholars.65 The application of sociological tools to early Christian evidence is not an entirely new venture; in 1925
Cullmann argued that the success of Form Criticism was dependent on "une branche spéciale de la sociologie [soit] consacrée à l'examen des lois de Vévolution des traditions populaires."66 In the first quarter of this century, Shailer Mathews, Shirley
Jackson Case, and Donald Riddle, members of the so-called Chicago School, argued for careful attention to the social background and dynamics of early church history.67
Cullmann's request, however, remained unheeded, and the work of the
The Social Setting of the Ministry group at Chicago is often burdened by a crude form of sociological functionalism. 8 Sociologists and historians over the last two centuries have often been uneasy bedfellows, the historian accusing the sociologist of butchering the English language with dressed-up pseudo-scientific language which only states the painfully obvious, the sociologist accusing the historian of being, as Auguste
Comte (1798-1857) charged, "the blind compilers of sterile anecdotes,"69 unable to transcend their parochialism and stubbornly resisting potential insights offered by investigating comparable social phenomena.
This is surprising: sociology initially developed out of the attempts by Comte, Dürkheim, Weber, and others to understand historically the development and structures of European civilization.71 Max Weber, for example, was a noted Roman historian and regarded the collection of historical data as integral to his sociological task. In fact, he did not distinguish between sociology and history. In the 1950s and 60s, it became generally accepted to regard the tasks of both disciplines as distinct but complementary. Lipset, a central figure in the creation of this détente describes the new relationship as follows:
History must be concerned with the analysis of the particular set of events or processes. Where the sociologist looks for concepts which subsume a variety of particular descriptive categories, the historian must remain close to the actual happenings and avoid statements which, though linking behaviour at one time or place to that elsewhere, lead to a distortion in the description of what occurred in the set of circumstances being analysed.72
When the two tasks are divorced from each other there is a danger, on the one hand, of ending with an armchair "great-men" approach to history, wholly atomistic and isolated from the flesh and blood of real life, and, on the other, a crude reductionism which ignores the creativity of individuals to transcend roles and social expectations.
The sociologist concerns himself less with what is historically idiosyncratic than with what is "typical, recurrent, general,"73 less with "singular conditions of a specific situation than with structural relationships which apply to several situations."74 The problem for the investigator who hopes to identify such phenomena in early Christian texts is that early Christian writers and the composers of the N e w Testament wrote for the most part not about typical, recurrent or general issues, but about special concerns. Moreover, an educated person such as Paul or Clement of Rome may not be a typical member of his community. The researcher using sociological methods to investigate Christian origins must therefore learn "to read between the lines" and recognize what are the typical social variables underlying a text and what are not. Gerd Theissen identifies three procedures to facilitate this task: constructive, analytic, and comparative.75
Constructive methods pay particular attention to expressions in the text which describe the social situation in which the document has been composed.
Such expressions include descriptions of groups, institutions, and organizations and those of individuals — their roles, background, and status.
Analytic methods infer underlying social realities from historical events, social norms, or religious symbols. By describing conflicts, for example, a writer often inadvertently identifies differing attitudes, customs, and social assump-
Introduction
tions which shape opposing opinions. Norms such as those contained in the Haustafeln or truisms such as "the love of money is the root of all evils" often help one to identify the assumed role expectations and more general social structures exerting an influence on the group. The analysis of religious symbols facilitates one's understanding of the group's identity, challenges, and ideals.
Finally, comparative procedures analyse texts which are not derived from the Christian group being investigated. Here the investigator looks outside the Christian milieu to its contemporary society, to such institutions as synagogues, philosophical schools, and Graeco-Roman clubs in the hope of gaining insights into what Christianity has in common with its greater environment and how it differs from it. Or he or she seeks what early Christianity has in common with sociological phenomena of other periods. Maurice Duverger identifies two categories of comparison: close and distant.76
Close comparison is used to compare phenomena relatively proximate temporally and culturally; it aims at precision, thoroughness, and detail; it seeks to identify differences between data and to explain those variations by reference to possible differences of context and structure. For historical research, close comparisons are only helpful where there is a wealth of detailed information concerning the social phenomenon under investigation. Because this is not available for investigators of the early church, they must rely upon Duverger's sccond kind of comparison: distant comparison.77 Here, rather than differences between phenomena, resemblances are sought:
Different types of structures or institutions from different cultural contexts or of different dimensions or of different significance, are compared. They are either historical comparisons covering widely separated periods or ethnological comparisons.78
The researcher is interested in the degree to which they are similar and the significance of their resemblances. The comparison is not rigorous (it cannot be), but there must be sufficient proximity to be able to draw analogies. Such analogies then become the basis of hypotheses for further research. The point of a distant comparison, therefore, is not so much to identify social phenomena of various epochs and cultures, but to provide, as Duverger says, "one means amongst others of provoking shock which produces discovery."79
Researchers attempting to make distant comparisons between phenomena have two tools available to them: ideal types and models. The ideal type is especially associated with Max Weber (1864-1920). Weber, as one historian among many who attempt to understand contemporary society by comparison with past and culturally different societies, developed this device to locate differences between groups which could then be subjected to further hypotheses and research. "An ideal type/' Weber explained, is formed by the one-sided accentuation of one or more points of view and by the synthesis of a great many diffuse, discrete, more or less present and occasionally absent concrete individual phenomena, which are arranged according to those one-sidedly emphasized viewpoints into a unified analytical construct.
In its conceptual purity, this mental construct cannot be found anywhere in reality 80
The Social Setting of the Ministry More simply, ideal types are logically integrated mental constructs of specific phenomena arising from the observation of a wide range of similar data drawn from a wide variety of periods and cultures. They are purely heuristic devices designed to enable a researcher to focus his attention on similar phenomena, identify differences between them, and generate hypotheses for further research. They are very useful in the construction of distant comparisons. In subsequent chapters we shall make use of Bryan Wilson's sect type and Weber's type of charismatic leadership. Used as heuristic tools, these ideal types will help to discover social aspects associated with the leadership of the early church.
The word "heuristic" is used with emphasis; it is unsound to allow types to generate data or make predictions where the text is silent. This is a point Robin
Scroggs forgets when he sets out to treat early Christian communities as sectarian communities.81 He first identifies seven aspects of sects and then seeks to locate them in the available data. The methodologically fateful step comes when he argues that where evidence is lacking it may be assumed that the typical elements were present82 Thus he uses the type to generate evidence and argues in a vicious circle. The proper course is to note that there are enough similarities to draw an analogy between contemporary sects and early Christian groups, and then to use contemporary evidence about sects to formulate questions to bring to the ancient sources. Another instance of a misuse of the methodology is committed by James A. Wilde in his discussion of Mark's gospel. He uses Bryan
Wilson's sevenfold typology of sectarianism to help to understand the group behind the gospel by arranging characteristics inferred from the text according to the different types of sects Wilson identifies. He then argues that Mark's community constituted a "vector," i.e. a combination of characteristics drawn from different sect types.83 The problem is that Wilde's attempt fits together too closely contemporary types and ancient data, and confuses description with explanation. In order for types to be fruitful in analysis it is imperative that they be used as more than descriptive boxes: their chief value lies in their ability to stimulate the investigator's imagination, to see the data in a new light, and to ask questions which have not been raised before.
The second tool available to researchers studying a text using a comparative procedure is the model. Models are schematized abstractions formulated on the basis of theories.84 They are created to help formulate hypotheses to test theories. Bruce Malina describes models as abstract, simplified representations of more complex real world objects and interactions.... [T]hey are approximate, simple representations of more complex forms, processes, and f unctions of physical and non-physical phenomena.85
Like ideal types, they cannot be used to generate data — otherwise theories would never be able to be falsified or verified — but unlike ideal types, they predict relationships between entities and outcomes when elements of the model are changed. They are "blueprints"86 for further study. As in the case of ideal types, models are also heuristic devices when applied cross-culturally; they help to generate questions concerning relationships between various phenomena adduced from the text. The social models potentially available to the historian are manifold, but it is necessary to recall Duverger's warning before distant
Introduction
comparisons are drawn between phenomena: there must be good reason to suppose that the data which the model describes and the topic of investigation are analogous. Most, if not all, historians make use of models of one form or another (although not all use comparative models); as Carney has noted, the question is not so much whether historians employ models, as which models to use and whether they will choose to be conscious of them in the future,87 One of the aims of this study is to show that the traditional ways in which the Shepherd and 1 Clement are treated lead to inaccurate conclusions because of the use of inappropriate models.
The chief model which we shall employ is Berger and Luckmann's concept of legitimation.88 Again, there is a timeless aspect to their model; examples they use to illustrate various points are drawn widely from different cultures and periods. Legitimation is defined by them as a process which arises after institutions have been created. It arises because of the need to justify and explain those institutions to individuals whom they affect but who did not initially construct them, or to explain their validity to those who are questioning their validity.
Berger and Luckmann describe four levels of legitimation; the most important is that of the symbolic universe.89 It is described as a "canopy" under which the individual, his past and future, his relationships, and all institutions are placed, and which gives them all a place and a meaning,90 By referring to it an individual knows who he or she is, where he or she belongs, and why. When the legitimacy of institutions is questioned, this results in the creation of new symbols or the explication of old ones. This model has great value when one attempts to understand the development of the ministry. Clement and Ignatius, for example, may be described as legitimators of leadership institutions which have come under attack. We shall be interested to see how they use symbols shared with the group to justify the authority of local leaders, and how they contributed to the building of further institutions, i.e., the extension of the canopy.
As in the case of ideal types, models have also been misused in the attempt to apply insights from the social sciences to the early church. John G. Gager is a good example of a scholar who has forgotten that models are a heuristic means to help understand early data 91 In his book entitled Kingdom and Community:
The Social World of Early Christianity, Gager argues that the social sciences can provide new data concerning the early Christians.92 This is false: they can only help raise new questions which then may facilitate the discovery of data which were always there but which had been passed over. More importantly, however, he forces the N e w Testament data to fit the model he chooses for comparative analysis. The model he uses is borrowed from L. Festinger and others. Festinger argues that contemporary, socially deprived millenarian groups often engage in greater mission activity rather than disintegrate when their prediction of a certain date for the Second Coming is proved incorrect; this results in what he calls
"cognitive dissonance" and he argues that this is overcome by increased proselytism.93 Gager argues that a similarly disappointed expectation in the early church accounts for its emphasis on mission. He contends that the death of Jesus presented a crisis for his disciples, which disconfirmed their belief in him as Messiah and resulted in a vigorous mission campaign to overcome their
"cognitive dissonance."94 For the model to fit he must first establish that the