ARE DISTRIBUTIONS REALLY STRUCTURES? A CRITIQUE OF THE METHODOLOGY OF MAX WEBER
Harriet Friedmann*
Research Paper No. 63
*Harvard University and Centre for Urban and Community Studies University of Toronto
The first draft of this paper was much improved through the criticisms and suggestions of Stephen Berkowitz, Ron Brieger, George Homans, and Harrison White. The final version emerged after several drafts, each of which benefited greatly from the detailed criticism and enthusiastic encouragement of Karen Anderson, Barry Edginton, Leslie Howard, Jack Wayne, and Barry Wellman. I wish also to express my appreciation to William Michelson and L.S. Bourne who have generously made available the facilities of the Centre for Urban and Community Studies, University of Toronto.
July, 1974 Abstract
Two conceptions of social structure can be distinguished:
structural approaches analyze the patterns of relations among units,
while distributional approaches seek to depict social structure through
determining the distributions of characteristics of the units. Through
explicitly developing the analytic method of constructing organized
social forms from probabilistically conceived social actions, Weber
systematically defined the aggregative and distributional logic which
underlies much contemporary sociology. An examination of some of his
most important work -- the study of bureaucracy, class and status,
and charisma -- reveals that Weber abandoned his methodological strictures
in favour of a more structural approach when the substantive and theore
tical analysis required it. An understanding of the limitations of dis
tributional conceptions of social structure points to theoretical and methodological approaches through which sociologists can build on both
the distributive and the structural aspects of Weber's work. Are Distributions Really Structures?
A Critique of the Methodology of Max Weber
The most fundamental concept in sociology is social structure.
Although it is rarely defined explicitly, one or another meaning of
the concept provides the basic assumption underlying all sociological
analysis. At least two fundamentally different conceptions of social 1 structure can be identified, which I shall call structural and
distributional. The first defines structure as changing patterns of
relations among units, whether individuals, organizations, or groups.
The second, usually implicitly, seeks to depict social structure
through determining distributions of characteristics of the units,
whether they be shared norms or values, or socio-economic attributes.
Both approaches have been taken, for example, in mobility studies,
and a simplified contrast may clarify the distinction. A structural
approach determines both the existing paths of mobility within the organized set of positions (not occupational categories) in society
and the career lines determined by the patterns of recruitment into
those positions. A distributional strategy approaches social structure less directly, through its reflection in the distribution of rewards (such as income or status) , and seeks to determine the distribution of characteristics (such as education, motivation, or father's occupation) of mobile and non-mobile individuals. I shall return to this example in my conclusions.
This paper is a critical examination of one of the classical sources of distibutional, or aggregative, strategies of social - 2 -
analysis, the methodology of Max Weber. Weber's ideal-typical
social actions are rooted in probabilities of the occurrence of
given meaningful behaviors; he bases his construction of organi
zational forms sometimes on the simple aggregation of these actions, and other times, on the distributions of types of action among the population of social actors. Although the importance of Weber's work is universally recognized among sociologists, and a wide range of researchers, from ethnographers2 to survey analysts, draw upon his social action methodology, the probabilistic underpinnings of Weberian analysis have rarely been examined. Nor have either the implications of restricting the unit of analysis to probabilistically conceived ideal-typical social actions, or the derivation of structure to the aggregations of these units. In what follows, I will attempt to explicate Weber's explicit methodology and to examine the degree to which he applies it consistently in his substantive work.
What will emerge from the analysis is that Weber both system atically defined the aggregative and distributional logic which underlies much contemporary sociology, and at crucial points in his substantive work, abandoned it in favour of a more formal and structural approach. As Zeitlin (1968) convincingly argues, Weber should not be read in opposition to thinkers seeking to analyze objectively observable social dynamics. But this substantive and theoretical compatibility exists precisely because Weber was willing to sacrifice his methodological strictures when the analysis required it. His commitment was to understanding basic social dynamics, such as the development of rationality as the basis of social action in - 3 - modern Western society and not in other times or places. There fore, when his method failed him, he implicitly adopted another.
Some of his most significant work for later generations of scholars
--- the study of bureacracy, class and status, and charisma -- will provide the material for an examination of his own adaptations to the limitations of the distributional methodology implied by the social action approach. This critical examination, in turn, will lead to conclusions about the appropriate uses of each strategy of analysis. - 4 -
Probability and Aggregation: From the ideal-typical actor
to organized social life
Weber's emphasis on the subjective meaning of social action uniquely defines the character of his analysis and differentiates 3 him from other classical thinkers. This methodological requirement restricts the object of analysis to the individual, as the only entity which can think and act. Organized social life must, according to Weberian methodology, derive from the cumulated social actions of individuals. While the larger consequences may not be predicted or even understood by the set of social actors, their individual actions must be subjectively meaningful and thus subject to the interpretive understanding of the analyst. In this way, as Wrong (1970: 22) puts it, the concept of verstehen, is the link between Weber's methodological reflections on the social sciences and his general sociological concepts.
But it is precisely these larg.er social consequences of cumulated social actions which are of interest to sociology. The central methodological problem, then, is how to derive the organizational outcomes of a set of individual social actions in a consistent way.
Weber argues for two related strategies. First, in order to isolate subjective meanings which are important analytically, the analyst must adopt a nominalist stance; he chooses a point of view from which to select and order relevant observations and constructs ideal- typical social actions from the probabilities that actors will attach a given meaning to a given behavior. Second, and more problematically, - 5 - he attempts to aggregate these social actions in such a way that he can explain organization. In analyzing Weber's methodological argu- ments, I will focus on: (1) the element of probabilistic reasoning in his concepts of "verstehen" and "ideal-type" and
(2) the derivation of different organizational units out of typical social actions.
Verstehen, Weber's fundamental methodological concept, is probably best translated as "interpretative understanding". (Freund
1968: 93) Such understanding may refer either to common-place rational activity (such as arithmetic operations) or to the motive or "meaning" of action as subjectively experienced by an actor.
However, it is only the latter --- understandable, subjectively meaningful behavior which Weber considers "action". If this action is oriented toward others, it is "social action". In contrast to historical explanation, a truly sociological approach attempts to interpret the meaning of mass4or aggregated, social actions. (Weber 1968: 4 - 9)
For Weber, the interpretation of mass phenomena of this kind requires an abstraction from complex ideas and motivations of those elements which are interpreted as central from the point of view the analyst chooses to adopt. This abstraction is an "ideal type"; i.e., the typical meaning, which as a shared individual attribute, defines a set of actors.
An ideal type is formed by the one-sided accentuation
of one or more points of view and by the synthesis of - 6 -
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.
(Emphasis in the original.) (1949: 90)
The epistemology which underlies Weber's concept of the ideal
type posits an infinite number of "facts", which are ordered by
analytic constructs. This nominalism implies only a relative
correspondence between any given construct and actual historical reality. For this reason, understanding that seeks to go beyond the typical individual, even to the simplest relationships, must be verified in terms of probabilities. (Freund 1969: 50)
Weber consistently defines his simplest relational concepts in this way. For example, the "probability of social action" is the basis of his definition of "social relationship":
The social relationship thus consists entirely and
exclusively in the existence of a probability that
there will be a meaningful course of social action
irrespective, for the time being, of the basis for this
probability. (My emphasis.) (1968: 27)
Also, "legitimacy" is defined in terms of the probability of a type of social action:
Action, especially social action which involves a social
relationship, may be guided by the belief in the
existence of a leg~timate order. The probability that - 7 -
action will be so governed will be called the
"validity" of the order in question. (My emphasis.) (1968: 31)
"Types of meaning" governing social relationships are in turn defined by the probability of the use of different mechanisms of enforcement:
An order will be called (a) convention so far as its
validity is externally guaranteed by the probability
that deviation from it within a given social group will
result in a relatively general and practically significant
reaction of disapproval; (b) law if it is externally
guaranteed by the probability that physical or psycho
logical coercion will be applied by a staff of people
in order to bring about compliance or avenge violation.
(My emphasis.) (1968: 34)
Finally, relations of power, command, and obedience depend on the probabilities of types of reciprocal action:
"Power" is the probability that one actor within a social
relationship will be in a position to carry out his own
will despite resistance, regardless of the basis on which
this probability rests. "Domination" is the probability
that by virtue of habituation a command will receive
prompt and automatic obedience in stereotyped forms,
on the part of a given group of persons. (My emphasis.)
(1968: 53)
Given the construction of ideal typical social action through - 8 -
a determination of the probability of its occurrence, Weber's
task is to derive its organizational consequences and to characterize
these as ideal types. His aggregative strategy, however, is general,
not specific to social actors. Thus, when he speaks of "arranging
• concrete individual phenomena" (1949: 90), he is referring
not only to individual social actors, but to "individual" organizational
units, and to whole societies as well. He is generalizing his units
of aggregation, thus making the requirements of his verstehen approach
problematic.
The use of the term "individual" which fits most consistently
with his concept of verstehen is that of the ideal-typical historical
actor. However, Weber also uses other examples in discussing the ideal
type: the "city economy", and "capitalist culture", (1949: 91).
While the term "individual" is used, by both Weber and his English
translators, in each case, the term in fact implies different units of analysis. "Individual historical occurrences" (e.g., modern capitalism) , "individual concepts" (e.g., Calvinist theology) , and individual "constituent elements of social occurrences" (e.g., types of domination) (Parsons 1949: 604-605; Abel 1969: 155), are not equivalent to "individuals" as real historical people, however typified. Thus, although these different analytic units involve no methodological inconsistency per se, it is only the last which can be analyzed through "interpretive understanding". Only people, actual or typical, engage in action based on subjective meaning. - 9 -
Problems of Aggregation:
Integration of Levels of Analysis
Here it becomes possible to distinguish three levels of analysis, which may be called individual, institutional, and societal. Weber is not interested in the individual per se, either as a concrete historical person, or as a typical actor. Rather he is interested in determining how probabilistic social actions become aggregated:
(a) The aggregation of individuals into institutional structure is
based on the logical consequences of "specific action-patterns
of separate individuals". (Abel 1969: 121) An example of this
is his general treatment of domination, specifically bureaucracy
based on legal-rational action.
(b) The relationship between institutional structures and organization
at the societal level is analyzed through assigning probabilities
to each of the possible logical combinations of institutional
structures, and then establishing the existence of empirical
correlations. Thus, collegial bodies and the market, as well
as bureaucracy tend to occur with legal-rational social action.
(c) The aggregation of individuals into inclusive sets of categories
of the whole society is based upon the distribution of individuals
with respect to one or a series of cross-cutting variables. This
is exemplified in Weber's treatment of "class" and "status".
(d) The "routinization of charisma" is a case in which Weber attempts
to introduce social change into his analysis. - 10 -
Weber is most successful methodologically when constructing
intermediate structures from one ideal typical social action. Thus,
in his studies of religion, he attempts to demonstrate that the
aggregate consequences of individual social action would be the same
whether this action were typified by the analyst on the basis of
material or ideal interests. As Bendix puts it:
One corollary of this starting point was Weber's tendency
to treat all concepts of collectivities or larger social
aggregates as convenient labels for tendencies of action .
• • • This conceptualization was also a method of analysis,
in that Weber would inquire into the ideas and affinities
associated with the apparently most single-minded pursuit
of gain, and into the economic interests associated with
the apparently most otherworldly pursuit of religious
salvation. Even then, the approach was limited to social
relationships based on a "coalescence of interests",
arising from actions that were construed as a reasoning,
emotional, or conventional pursuit of "ideal and material
interests" (1962: 476-477).
Therefore, "tendencies of action" are the basis for the establishment
of correlations among meaning systems; as, in this case, between "the
spirit of capitalism" and "the Protestant ethic".
In contrast, Weber consciously evades the difficulties of
constructing organizational forms characteristic of whole societies
through the aggregation of individual social actions. The common collective form within which the individual pursuit of gain in capitalist society takes place is the market. In its modern highly - 11 -
developed stage, its central characteristic is the anonymity of
the relations within it. Weber describes the market in terms reminiscent
of the Durkheirnian idea of "constraint":
Formally, the market conununity does not recognize direct
coercion on the basis of personal authority. It produces
in its stead a special kind of coercive situation which,
as a general principle, applies without any discrimination
to workers, enterprisers, producers, and consumers, viz.,
in the impersonal form of the inevitability of adaptation
to the purely economic "laws" of the market .•. The private
enterprise system transforms into objects of "labor
market transactions" even those personal and authoritarian
hierarchical relations which actually exist in the
capitalistic enterprise. While the authoritarian
relationships are thus drained of all normal sentimental
content, authoritarian constraint not only continues but,
at least under certain circumstances, even increases. (1968: 731)
Weber's task here is to derive this anonymous structure from
"instrumentally rational social action". It is by no means clear that such an involved derivation is possible; therefore, Weber focuses instead on the derivation of types of intermediate "organizational structures", based on their "specific way (s) of distributing the powers of conunand" . (Emphasis mine . ) (1968: 953)
This point must be emphasized because of Weber's own explicit recognition of the limitations of his methodology in analyzing large scale social structure. If society is pictured through aggregation, - 12 - then structures within society are pictured through distributions of variables. In this case, "power" ("the probability that one actor within a social relationship will be in a position to carry out his own will despite resistance") is an individual attribute whose distribution will reveal the distribution of positions from which differential amounts of coercion can be exercised. Since "power" is a quantitatively distributed attribute of individuals resisted by other individuals, Weber recognizes the difficulty in deriving impersonal systems of constraint of the market type. He sidesteps the problem by focusing instead on intermediate structures of direct command, which are by definition personal and direct:
In the following discussion we shall use the term
domination exclusively in that narrower sense which
excludes from its scope those situations in which power
has its source in a formally free interplay of interested
parties such as occurs especially in the market. In
other words, in our terminology domination shall be identical
with authoritarian power of command. (1968: 496) - 13 -
Legal-Rational Social Action, Bureaucracy, and Capitalism:
From Individual to Institution to Society
Given this self-imposed limitation of Weber's methodology in the case of power and domination, let us examine the ideal type of one of his institutions of domination: bureaucracy. Consistent with his methodological rules, Weber asserts that "the legitimacy of a system of domination may be treated sociologically only as the probability that to a relevant degree the appropriate attitudes will exist, and the corresponding practical conduct ensue". (1968: 214) It is upon this probability that "a certain minimum of the assured power to issue commands" rests. (1968: 215)
In the case of legal-rational social action, the structure of domination rests on the probable collective acceptance by a set of actors of the following specific "ideas":
1. That any given legal norm may be established by
aggreement or by imposition, on grounds of expediency
or value-rationality or both, with a claim to
obedience at least on the part of the members of the
organization.
2. That every body of law consists essentially in a
consistent system of abstract rules which have
normally been intentionally established. Further
more, administration of law is held to consist in
the application of these rules to particular cases - 14 -
3. That thus the typical person in authority, the
"superior", is himself subject to an impersonal
order by orienting his actions to it in his own
dispositions and commands • . •
4. That the person who obeys authority does so, as it
is usually stated, only in his capacity as a "member"
of the organization and what he obeys is only "the law" . . .
5. In conformity with point 3, it is held that the members
of the organization, insofar as they obey a person in
authority do not owe this obedience to him as an
individual, but to the impersonal order. Hence, it
follows that there is an obligation to obedience only
within the sphere of the rationally delimited juris
diction which, in terms of the order, has been given
to him. (1968: 217-218)
In sum, the "meaning" of legal-rational social action is belief in the validity of a particular set of procedures for both the making and the administration or rules and policy. Given such a shared belief, it is logically consistent but not necessary that the ideal typical structure of bureaucracy should emerge. In fact, Weber sees several different structures as depending on legal-rationaJ action, bureaucracy being the one which is "most unambiguously a structure of domination". (1968: 219) Thus, the characteristic of the "purest - 15 -
type of exercise of legal authority" is a hierarchy of offices, each with a clearly defined sphere of competence, filled by personally free officials on the basis of technical qualification and through a free contractual relationship. 5 The official does not appropriate his position, but is salaried, He is subject to strict discipline in the conduct of his office, which is his primary occupation, and which is part of a regular patterned career. (1968: 220-221)
Weber reaches the high point of his methodological consistency in the construction of such intermediate levels of organization from types of social action. This consistency primarily rests upon the logically derived and formally described structures which correspond to types of social action. There is no clear direction of cause here, but an assertion of expected correlation between legal rational action and bureaucracy, comparable to the correlation between religious and economic action mentioned above. Thus, "collegial bodies", just as bureaucracies, rest on legal-rational action. So, for that matter, does the market.
What, then, is the relationship among these structures? It seems fair to say that Weber intends to characterize whole societies in terms of the types of social action in which their members engage.
This is evident in the example of legal-rational social action, at least to the extent that the formal analysis of bureaucratic structure implies the existence of control from the top, and a clientele at the bottom. Given a perfectly hierarchical flow of commands, the - 16 - method of recruitment to the top of the structure becomes crucial theoretically. (1968: 987) Further, even given legal-rational action throughout a society, not simply in a single bureaucracy, there are numerous alternative rational procedures for determining the person at the top of the hierarchy. In addition, bureaucracy, as opposed to other forms of organized legal-rational action, assumes a clientele; i.e., something is administered for, or to, other actors.
These actors, as well as the officials in the bureaucracy, must at least passively accept the legitimacy of legal-rational domination.
Weber sees the ideal-typical bureaucracy (as opposed to patri monial bureaucracy or other types) as a central institution of modern society. It is "fully developed in political and ecclesiastical communities only in the modern state, and in the private economy only in the most advanced institutions of capitalism." (1968: 956)
More generally, "bureaucracy is the means of transforming social action into rationally organized action". (Emphasis in original) •
(1968: 987). Bureaucracy is related to the formation of collegial bodies, especially in the legal profession and parliament, which are also based on legal-rational action. Bureaucratization, as the most rational example of modern social action, is related to the leveling of social differences in the interests of technical efficiency and universal administration. Finally, all of this is related to the emergence of modern capitalism. (1968: 296)
Thus, legal-rational social action results in several different, but related, ideal-typical forms: the market, bureaucracy, and collegial bodies. These forms, in turn, are composed of various - 17 -
factors, some or all of which have appeared independently or in
various combinations in other historical circumstances. Further
more, legal-rational social action is only one of at least three
ideal-typical social actions. Any of these can occur in principle
in any conceivable combination; for example, feudalism is based
simultaneously on both traditional domination (personal fealty)
and legal-rational domination (free contract) • (1968: 255-256)
Weber makes this combination of factors an explicit methodological
statement:
The type of social science in which we are interested
is an empiricial science of concrete reality. Our aim
is the understanding of the characteristic uniqueness
of the reality in which we move. We wish to understand
on the one hand the relationships and the cultural sig
nificance of individual events in their contemporary
manifestations and on the other the causes of their being
historically ~ and not otherwise . . . (It) concerns itself
with the question of the individual consequence which the
working of these laws (universally valid causal relation
ships) in a unique configuration produces, since it is
these individual configurations which are significant for us.
(Emphasis in original.) (1949: 72-73)
These "configurations" are derived from an examination of all combinations of all "objective possibilities", which in turn are - 18 - intermediate levels of organization based on types of social action.
Weber (1949: 181-182) asserts that "the judgment of 'objective' possibility admits gradations of degree and one can form an idea of the logical relationship which is involved by looking for help in principles which are applied in the analysis of the 'calculus of probability'". (Emphasis in original.) He develops this logic of analysis still further as a model. Just as the bias of a loaded die is established through frequent repetition, comparable, if less reliable, methods can be used in social science:
The only difference is that it is precisely here in the
sphere of concrete causality that ability to assign a
numerical measure of chance is wholly lacking since this
presupposes the existence of "absolute chance" or certain
measurable or countable aspects of phenomena or results
as the sole object of scientific interest. But despite
this lack, we can not only very well render generally
valid judgments which assert that as a result of certain
situations, the occurrence of a type of reaction,
identical in certain respects, on the part of those
persons who confront these situations, is "favored"
to a more or less high degree . . . And we can finally
estimate the degree to which a certain effect is
"favored" by certain "conditions" --- although we cannot do
it in a way which will be perfectly unambiguous or even
in accordance with the procedures of the calculus of
probability. We can, however, well enough estimate the - 19 -
relative "degree" to which the outcome is "favored" by
the general rule by a comparison involving the consider
ation of how other conditions operating differently "would"
have "favored" it. When we carry through this comparison
in our imagination by sufficiently numerous conceivable
modifications of the constellations of conditions, then
a considerable degree of certainty for a judgment of the
degree of objective possibility is conceivable, at least
in principle --- and it is only its conceivability in principle
which concerns us here primarily. (Emphasis in original.)
(1949: 183)
Weber's logic of inquiry has significant implications. First,
as Parsons notes, a given phenomenon is, for Weber, "capable of description only in terms of a specific combination of the values
of analytical elements (ideal types) • . . (which) introduces an
element of rigidity that may issue in a methodological atomism .
(and) a 'mosaic' theory of history". (Parsons 1949: 621) In other words, history becomes a "process of shuffling of ideal types as units". However, Parsons' proposed solution is further atomization, i.e., "generalized theory which breaks down the particular element combinations in the ideal types". (Parsons 1949: 626)
While Parsons correctly asserts that such a breakdown adds "flexibility",
the "mosaic" problem is compounded by the use of finer units and cross-cutting variables. I shall examine this later in Weber's analysis of class and status, where Weber does what Parsons argues for. - 20 -
The difficulty with the "mosaic" is that it is at once timeless and static. While Weber is consistently (though certainly not exclu sively) seeking to trace "developmental stages" through differential combinations of "types", each of which is a difference of "degree"
(1968: XXXI), in principle all possible combinations may occur at any given time for any given society. While such "stages", especially in the emergence of rational action as the dominant type, are empirically hoped for, they may or may not be "discovered" as an historical sequence. Even if they are, these "stages" will be based on a series of combinations of ideal types, which as nominalist constructs are highly dependent on the "facts" selected by the analyst in the initial construction of each type. Furthermore, even if such stages could be both empirically and theoretically justified, the actual processes of change (rather than the identification of which elements changed) could not be incorporated into the analysis in a methodologically consistent way. As we shall see, one of Weber's most impressive analyses of the processes of change, the "routinization of charisma" departs from this methodology in favor of a more formal and dynamic, if reductionist, approach. - 21 -
The Confusion of Distribution and Structure:
Class and Status
The implications of this strategy of aggregation become clear when Weber studies stratification. His analysis of class and status is an important example of the conception of structure as distributions of individually-attributed and cross-cutting variables. Here also, in his contrast between status as a property of groups and class as a shared individual characteristic, the relationship between intermediate organization and the depiction of society as a whole can be examined.
First, Weber asserts that "classes" and "status groups" (as well as "parties") are "phenomena of the distribution of power within a community". "Power" is defined here as "the chance (elsewhere,
"probability") of a man or number of men to realize their own will in a social action even against the resistance of others who are participating in the action". (My emphasis ) (1968: 926-927). Like
"power", "class" for Weber is an individual attribute. A particular
"class" is a set of individuals defined by the operator of "class situation". In Weber's terms, "'class' means all persons in the same class situation"; and "class situation" is defined as a potentially shared attribute of individuals, i.e., "the typical probability of
1. procuring goods, 2. gaining a position is life, and
3. finding inner satisfactions". (1968: 927)
"'Class situation' and 'class' refer only to the same (or similar) interests which an individual shares with others" (1968: 302) - 22 -
Classes "represent possible, and frequent, bases for social action."
(1968: 927) At the same time, classes 'emerge only on the basis of
social action among members of different classes" (i.e., "the
labor market", "the commodities market", and "the capitalistic enter-
prise"). (1968: 930) (My emphasis). What Weber seems to be saying
here is that the distributions of economic chances of individuals
simultaneously results from the social action upon which markets are
based, and serves as the basis for that same social action. This is
neither inconsistent, nor tautological, but evidence that Weber is
not seeking to determine causes or consequences of stratification.
Instead, he is seeking to establish correlations between types of
social action based on interest, and types of distributions of interest
defined in all logically possible ways.
The picture of society that emerges has two important character
istics: First, all actors in society are distributed according to a
series of cross-cutting variables; and second, each of these distri
butions serves as the basis for the division of all social actors into
mutually exclusive categories. The population is, or may be, distri
buted according to the value of property ("property classes") ,
the marketability (price) of commodities and/or skills ("commercial
classes") , or degree of potential mobility ("social classes") , or any combination of these classes simultaneously. Thus, any given individual
social actor is analytically divided into a series of variables, any of which may give meaning to his social action. Each of these distri
butions then serves as a basis for the creation of sets or categories;
for example, "commercial classes" are categorized into various types of entrepreneurs, professionals, skilled and unskilled workers. - 23 -
Furthermore, these variables cross-cut each other; for example,
Weber describes certain "commercial classes" as being "middle classes"
which are "in between" "positively and negatively privileged property
classes". (1968: 303-304)
Thus far, these categories of "class situation" appear to be
purely nominalist. This interpretation is supported by Weber's assertion
that in order to avoid ambiguity, "class interest" ("interest" being
the operator defining "classi1) is understood to be "the factual
direction of interests following with a certain probability from
the class situation for a certain average of those people sujected to
the class situation". These interests are shared individual
properties which may result in either "mass behavior" or "amorphous
social action". (1968: 928-929). It seems clear that the creation,
choice, and combination of categories would be the sole basis for the
construction of ideal-typical interests among various sectors of the
population.
However, stratification, for Weber, takes on an implicitly
independent reality in his discussion of the formation of intermediate
organizations, or associations, for the purpose of organized social
action on the basis of class interest. His discussion of this is
remarkably similar to Marx's theory of class structure and conscious
ness, in which an analysis of the objective structure provides a vantage point from which to study the adequacy of meaning (approximation
to a correct understanding) for the actor: - 24 -
The degree in which 'social action' and possibly
associations emerge from the mass behavior of the
members of a class is linked . • . to the extent of the
contrasts that have already evolved, and is especially
linked to the transparency of the connections between
the causes and consequences of the class situation.
For however different life chances may be, this fact
in itself, according to all experience, by no means
gives birth to "class action" (social action by the
members of a class). For that, the real conditions
and the results of the class situation must be
distinctly recognizable. For only then the contrast
of life chances can be felt not as an absolutely given
fact to be accepted, but as a resultant from either
(1) the given distribution of property, or
(2) the structure of the concrete economic order. (My
emphasis) . (1968: 929)
The reason for Weber's shift here seems to lie in his own recognition of the limitations of simple aggregation in determining structure at the societal level. If individuals in a particular society are indeed differentiable through a potentially infinite series of combinations of cross-cutting variables, it is difficult to posit stable sets of categories without being extremely arbitrary. He never actually confronts this problem in his substantive studies, however, because he - 25 - never attempts to use this aggregative methodology to analyze an entire society. He does use it, however, to contrast two basically different types of stratification which may characterize ideal- typical societies: "class society" or "status society". (1968: 306,
937) In order to do this, he employs the concept of status and
"status group" which may be interpreted both as an additional variable and as an intermediate level of structure.
Weber defines "status" as "an effective claim to social esteem in terms of positive or negative privileges; it is typically founded on (a) style of life, hence (b) formal education .•. or .
(c) hereditary or occupational prestige". (1968: 305-306).
Thus, like class, status is an individual attribute. Weber goes on to say, however, that "in practice, status expresses itself through
(a) connubium, (b) commensality, possible (c) monopolistic appropriation of privileged modes of acquisition or the abhorrence of certain kinds of acquisition, (d) status conventions
(traditions) of other kinds" (1968: 306). The first two of these are relational, or group, rather than individual characteristics.
Similarly, his "social classes" have a relational component as bounded categories within which individual and generational mobility typically occurs.
Weber is contrasting types of societies here in terms of the appropriateness, however determined, of classifying individuals as individuals, or as actors organized into intermediate structures.
The definition of class situation as "ultimately market situation"
(1968: 928) contrasts with the definition of status groups as at - 26 -
least to some extent self-defining (1968: 306). This contrast is
sometimes evolutionary, the market ("commercial classes") being
based on instrumentally rational social action, and status groups
being based on traditional action (1968: 306). At other times, there
is a more explicit notion of process: Since status groups involve consumption at a level to maintain a specifically defined style of life, given a "relatively stable" (1968: 938) distribution of goods, property classes may "create" status groups. (1968: 307)
In his analysis of class and status, Weber confronts the difficulty of timelessness and stasis. I have attempted to show the problems in volved in analyzing any given society through selection from all logic ally possible combinations of ideal-types based upon an infinite number of possible variables. Weber attempts to resolve this difficulty through constructing intermediate levels of structure --- class associations and status groups. In order to do this, he abandons his methodology in describing the processes of group formation. Specifically, since he recognizes that similar social action does not imply organi zation, he posits, albeit implicitly, the existence of large-scale social organization which does not derive from social action, but which provides an "objective" framework within which subjective meanings develop. - 27 -
The Problem of Social Change:
New Strategies of Analysis
This methodological shift is even more clear in Weber's discussion
of the routinization of charisma. Bureaucracy and patriarchalism, based
on legal-rational and traditional types of social action, seem for Weber
to cover the basic possibilities for continuous domination. But it is
this very continuity which does not allow for the prediction or explan-
ation of change. Since change from one form of domination to another
must be accounted for within the set of logical possibilities of dam-
ination, Weber constructs an ideal-typical form of domination, charisma,
which is "the specifically creative revolutionary force of history".
(1968: 1117). In order to describe an unstable form of social action, Weber resorts to a kind of reductionism and to formalism , both of which are logically inimical to his explicit methodology.
First, as opposed to both traditional and legal-rational bases
of legitimate domination, charismatic social action rests on the
recognition of the extraordinary qualities of a particular "individual
personality" (1968: 241). In fact, this variable of "personality"
is the basis for the fragility and impermanence of charismatic domination. This psychologism applies not only to the gifted leader, but also to the followers:
In traditionalist periods, charisma is the great
revolutionary force. The likewise revolutionary
force of "reason" works from without: By altering - 28 -
the situations of life and hence its problems,
finally in this way changing men's attitudes toward
them; or it intellectualizes the individual. Charisma, on
the other hand, may effect a subjective or internal
reorientation born out of suffering, conflicts, or enthus
iasm. It may then result in a radical alteration of
the central attitudes and directions of action with a
completely new orientation of all attitudes toward the
different problems of the "world". (Emphasis in original.)
(1968: 245)
Thus, while Weber's methodology requires that all social action be interpreted with respect to the subjective "meaning" which orients it, charisma, as "the great revolutionary force" is specifically contrasted with other types of social action as being "internal".
At the same time, such a psychological interpretation of change requires, for a thinker as systematic as Weber, that this action occur within some more or less implicit "objective" framework. Thus, the situations in which "'natural' leaders" arise and have their charisma recognized are "moments of distress --- whether psychic, physical, economic, ethical,religious, or political" (1968: 1111-1112).
Furthermore, charisma is antithetical to rational, sustained economic activity, although the situation from which it emerges must have such activity. Therefore, the "principal motives" underlying the"routinization of charisma", either traditionalization or rational ization, are: "(a) the ideal and also the material interests of the - 29 -
followers in the continuation and the continual reactivation of the
community, (b) the still stronger ideal and also stronger material
interests of the members of the administrative staff, the disciples,
the party workers, or others in continuing their relationship"
(1968: 246). These interests derive from the "objective necessity"
for routine economic activity and administration (1968: 252) •
Thus far Weber has rooted his major ideal type of social change
in "internal" reactions to "objective" conditions. The process of
routinization is directed, to the extent possible, by those members
of the "administrative staff" or "disciples", who act in an instru
mentally rational way, that is, in terms of self-interest. The new
form of domination must be created through the dominating group~s
separating itself organizationally from the followers:
When a charismatic movement develops in the direction of
prebendal provision, the 'laity' becomes differentiated
from the 'clergy' ... , that is, the participating members
of the charismatic administrative staff which has now
become routinized. These are the priests of the developing
'church'. Correspondingly, in a developing political body
--- the 'state' in the rational case --- vassals, benefice
holders, officials or appointed party officials (instead
of voluntary party workers and functionaries) are
differentiated from the "tax payers". (1968: 251)
The central characteristic of this analysis is its formalism.
Weber constructs a single series of formal stages through which any type - 30 -
of society must go in order to change into any other type of society.
Specifically, the use of psychological characteristics and vaguely
defined objective conditions leads to the constructing of a formal
process which occurs independent of historical or social conditions.
Neither "natural leaders" nor "times of stress" need be rooted in time,
place, or organizational structure.
While an abstract formalism of this kind resolves the methodo
logical problems of aggregation of individual units, it is not adequate
for a sociology which seeks to understand both organization and process.
Weber's concerns in developing his methodology were both to draw bound
aries around the neighboring disciplines, creating a special province
for sociology, and to develop a unique logic of analysis, avoiding the pitfalls of analogy into which such schools as the functionalists had
fallen. However, in looking to probability theory as an abstract system with which to order the infinite variety of "facts" and "events", his analysis could only lead to more or less arbitrary sets of individual actors, individual ideal types, or unique configurations of variables.
His substantive interests in social organization and social change -- in systems of domination, in the relationships between idea systems and systems of economic activity, in the historical development of modern capitalism with its particular institutions --- were ultimately not sacrificed. He abandoned or modified his methodology when it failed him. But he did not develop a coherent theory of society based on the interpretative understanding of subjectively meaningful social action. - 31 -
Conclusions: Notes on a Structural Approach
Weber's substantive work rests on the assumption that collective
entities, or structures, exist in their own right. His social action
methodology, however, has led to the logically consistent conclusion
that "to see them (collective entities) as possessing a supra
individual reality of their own is to reify them", and consequently
that "a society is essentially a set of broadly warranted predictions
made by its members about one another's behavior. In this sense, it
exists only in people's minds," even if it results from observable
interaction (Wrong, 1970: 22, 25). A restricted interpretation of
Weber's emphasis on social action leads to the formulation of problems
in terms of "the causal determinants of behavior" (Wrong, 1970: 22),
for which a distributional approach is appropriate. It is, for example,
appropriate for the study of certain critical aspects of social
structure, such as the socio-biological constraints analyzed by
demographers.
However, the major thrust of Weber's methodology asserts that
social structures which exist in their own right can be derived from
the cumulation of social actions. Therefore, the study of social
structure cannot be limited to the causes of behavior, but must analyze, according to Weberian methods, the ways in which this cumulation of social actions occurs, and its consequences in social structure. Indirectly, Weber's intellectual descendants include all - 32 - those who build their conceptions of social structure on the distributions of individual characteristics among populations.
Consequently, it is crucial to examine the limitations of his logic of analysis.
This is not an argument against the use of probabilistic reasoning or aggregative or distributional strategies. Instead, it is an attempt to clarify what such approaches are actually measuring and analyzing. Any regularly occurring event can be measured, and probabilities assigned to its occurrence under specified conditions. Similarly, any attribute shared by a population in different degrees, such as income, can be described in terms of its distribution in that population. Having done this in a particular study, the next question is why do these distributions occur? The answer is usually sought in the measured distributions of other variables, and "causation" is formulated in the language of independent and dependent variables. Thus, income and education vary together; depending on the framework of the analyst, education leads to higher paying jobs, or high income increases access to education, or both.
This model assumes one of two things: either the distribution of one variable will change with a change in the distribution of the other
(e.g., if more people go to university, more people will have high incomes), thus changing the social structure; or, more compellingly, there exists an unexamined social structure which somehow rewards people differentially (as measured in our income distribution) , and we can predict where any given individual will fall on this distribution - 33 -
if we know certain facts about him or her, such as years of education.
Distributions, then, measure selected consequences of structural
dynamics as these consequences differentially affect individuals. As
such, they are useful indicators of questions to be asked in analyzing
social structure directly, but are neither descriptions nor analyses
of the structure itself. Thus, the answer to the question posed in
the title of this paper is that distributions are not really structures,
but are reflections of social structure.
The distinction between the measurement of the differential
individual consequences of social structure and the analysis of
structure itself, can be clarified by returning to the earlier
example of mobility studies. Assuming that at any given moment there
is a fixed number of jobs, each of which has attached to it a fixed
set of rewards, and that only one person can fill a job at a time,
there are two fundamentally different questions which can inform
research into occupational mobility. The first, or distributional
approach asks what characteristics of individuals best predict
individual mobility between occupational categories. (Blau and
Duncan, 1967). That is, it assumes a social structure which is reflected in a given set of occupational roles, which can themselves be categorized in terms of their individual characteristics, such as
type of activity (e.g., white collar) and reward. It can thus determine
the probability of any given career line in the aggregate, by discovering what proportion of the population in category 1 moves to category 2
(similarly with inter-generational mobility). It can further determine the distributions of characteristics of individuals, such as education - 34 -
and father's occupation, as well as the combinations of these, which
predict the occurrence of a given career line.
The second, or structural, approach examines the actual career
lines existing in a concrete structure of jobs. While the aggregation
of jobs represented by occupational categories allows for the measure
ment of the amount of mobility in a population, and for the identifi
cation, in general, of the kinds of people who are mobile, it points
to the need for an analysis of the actual organization of jobs, the
patterns in which job openings occur, and the mechanisms of recruitment
into them. Assuming that at least some of the people recruited into
a given job thereby leave another, thus creating another opening into
which yet another individual is recruited, and so on, in a continuous
but finite process, it becomes possible to analyze the actual
structural dynamics which underlie the aggregative measurements
of the first approach, as White (1970), so elegantly demonstrates. ·~ t Similarly, Howard (1974) compares the two strategies in his analysis
of the incorporation of rural migrants into the occupational structure
of an Indian city. He finds that networks of kinship, friendship,
and neighbourhood are more powerful predictors of successful job
search than are individual characteristics usually associated with
"modernity".
As I have attempted to demonstrate in this paper, Weber constructed
a methodology for social analysis based on probabilities and distrib
utions. While he did not confront the limitations of this approach
directly, as I am seeking to do here, neither did he allow his analysis - 35 -
of social organization to suffer these limitations. He proceeded as
far as he could with his methods, and then went beyond them to
analyze social structure directly. He in fact, in his best work, used
his aggregative and distributional findings to direct him to important
structural concerns.
Weber's direct descendants, especially in the field of formal
organization, have inherited some of the best products of his
substantive analysis. The formal structure of bureaucracy is one
of the best examples of objectively defined and observable patterns
of relationships. The value of the legacy, however, increases with an understanding of its source and limitations. Its source, as I have argued, lies in a modest departure from Weber's self-imposed methodological requirements, in the sense that it is consistent with, but not drived from legal-rational social action. Its limitations,therefore, consist not in the problems of aggregative or distributional logic, as do the limitations of his analysis of stratification, but in the static nature of the formal structure, a problem Weber inadequately addressed through the concept of charisma.
As a consequence, the predominant explanation of change in the field has had to rely on dynamics separate from the formal organization itself --- either "outside", in a vaguely defined "external situation",
(Parsons, 1962), or "inside", in a relatively less structured "informal organization". (Friedmann, 1973) . Both take on meaning only as contrast conceptions to "formal organization". They are nonetheless the focus for studies of bureaucratic dynamics (Blau, 1963), or are seen as the source of changes in the formal structure itself (Blau and Meyer, 1971:
138-146) . - 36 -
The task confronting students of social organization, then is to develop models which simultaneously analyze structure and change or in more precise language, structural dynamics. While such models can be only hinted at here, some exploratory suggestions emerging from this perspective on the field of formal organization may be of value. For instance, it should be possible to reinterpret the empirical findings and theoretical insights in the field so that they form an integral base for further analysis. Instead of three related problems --- formal bureaucratic structure, informal organization, and the relations of organizations to the larger society --- a structural approach would define a consistent over-all theoretical perspective for posing research questions. All types of relations among people and roles should be treated simultaneously over time, and the patterns of relationships which emerge should define the theoretical unit.
This contrasts with the ~priori definition of the theoretical unit as the formal organization, within which and outside of which diff- erent, if related, sets of events occur. Thus, ties across the boundaries of formal organizations as currently conceived could be treated simultaneously with intra-organizational ties, "formal" and
"informal". Such an approach would broaden the range of important questions within the field, and provide a theoretical framework for integrating and building on the findings of each of these areas. Further- more, this approach should return to Weber's emphasis on historical . 6 ana 1 ysis. It is through tracing changing types and patterns of relation- ships over long periods of time that sociologists gain insight into 7 important structural dynamics. - 37 -
Two conclusions emerge from this analysis of Weberian methodol ogy. First, those areas of sociological inquiry which rely on Weber's verstehen approach and the aggregative, distributional logic which he so clearly laid out, should be re-evaluated in the light of its methodological and theoretical implications. Thus, the utility of empirical studies which assume that distributions of variables measured on individual units are synonymous with social structure increases with a clear theoretical understanding of what they are and are not measuring, and of what conclusions about social organi zation can be drawn from them. Similarly, the common conceptual origins of supposedly "objective" empiricism and areas such as phenomenology, which emphasize the interpretive methodology fundamental to Weber's approach, point to new interpretations of these fields. Weber's subjective methodology does not imply subjective results (Zeitlin, 1968: 117). Neither, I have argued, does empirical measurement of individual attributes imply objective results. Funda mentally, whatever their differences, both can be contrasted to a structural approach.
Second, sociologists seeking to develop a more dynamic, structural method of analyzing social organization should build on those aspects of Weber's work in which he departs from his explicit methodology.
Certain of his ideal types, such as bureaucracy, are formal models which provide one basis for this development. But to analyze structural dynamics, we will need more closely and explicitly to - 38 -
integrate theoretical and empirical work. It will no longer
suffice to develop new concepts, such as "informal organization",
to account for empirically observed discrepancies from the ideal
type. Instead, we can build on what is best in the work of the
classical theorists by explicitly adopting and developing structural
strategies of analysis, which analyze changing patterns of relations
among individuals, organizations, groups, classes, and nations. To
take the concept of social structure seriously, we must take as our analytic units the actual ties among individuals or organizations.
The patterns of these ties over time is social structure. - 39 -
Footnotes
1 This use of the term structuralism is different from and more limited than its uses by, for instance, the schools following Talcott Parsons and Claude Levi-Strauss. Both of these schools are seeking universal social laws, the former through the functional integration of differentiated subsystems defined a priori, and the latter through the human mental processes underlying social organization. It is closer to Etzioni's description of structuralism (1964: 41-49).
2 Primarily through the work of Alfred Schutz (Wagner, ed., 1970: 5-11). Compare also Weber's discussion of motives in Economy and Society (1968: 8-13) with C. Wright Mills' "Situated Action and the Vocabulary of Motives" (1963).
3 Weber developed his methodological approach in dialogue with other classical sociologists. Weber saw the utility of functional analysis for "provisional orientation", but attempted to avert the dangers of "reification" inherent in organic analogies. (Weber 1968: 15). While accepting the formalist distinction between "form"and "content", Weber was primarily concerned with the "content" or "meaning" of social action. (Weber 1968: 86) . This focus is the reverse of Simrnel's, who advocated the exclusive investigation of "forms of sociation". (1971: 124-125). Weber recognized the utility of Marxian "laws and developmental constructs" if used as ideal types (Weber 1949: 103), but his focus on "subjectively meaningful social action" did not allow him to confront directly the problem of "objective reality". For instance, Weber's definition of "an economy" rests upon the subjective views of actors (1968: 63). In contrast, Marx emphasized the"circumstances directly encountered, given, and transmitted from the past" in terms of which people act. (1968: 15). Marxian theory, roughly, interprets idea systems which guide action as more or less close approximations to a "true" under standing of social organization --- based either on a correct perception of class interest ("consciousness"), or on a "rational" but obfuscating description of society ("ideology"). There is no comparable vantage point in Weberian theory from which to judge the adequacy of meaning for the actor. In principle, "reality" is understood only through the subjective meanings of actions. - 40 -
4 Weber's use of the term "mass" actions here is reminiscent of current concepts such as "collective behaviour", in con trast to his more organizational concepts. This is clearly antithetical to the major thrust of Weber's work, and his use of such language in his methodological discussions is indicative of the problems of attempting to construct social organization through cumulative social actions. I am indebt ed to Les Howard for this point.
5 Thus Weber sees bureaucracy as a structure in its own right, and does not, as Wrong (1970: 23), argues, "identify bureaucracy as a form of social organization with the bureaucrat as a human type".
6 Aron (1970: 86-89), discusses the problems of Weber's approach to causality in history. It is not his specific methodology of historical analysis referred to here, but his general theoretical and substantive interest in historical explanation.
7 Abner Cohen (1969), in his study of the political and economic leverage in Nigerian society gained through changing Hausa ethnic organization, provides an exciting example of the potentials of such an approach. See, especially, his theoretical arguments in his introductory and concluding chapters. I am indebted to Jack Wayne for suggesting this example. REFERENCES
Abel, Theodore 1969 Systematic Sociology in Germany. New York: Octagon Books.
Aron, Raymond 1970 "The Logic of the Social Sciences", in Dennis Wrong (ed.), Max Weber. Englewood Cliffs: Prentice-Hall.
Blau, Peter M. 1963 The Dynamics of Bureaucracy. Chicago: The University Press •
• and O.D. Duncan ~~~- 1967 The American Occupational Structure. New York: Wiley.
~~-· and Marshall W. Meyer 1971 Bureaucracy and Modern Society. New York: Random House.
Cohen, Abner 1969 Custom and Politics in Urban Africa. Berkeley: University of California Press.
Etzioni, Amitai 1964 Modern Organizations. Englewood Cliffs, N.J.: Prentice-Hall.
Freund, Julian 1968 The Sociology of Max Weber. New York: Random House.
Friedmann, Harriet 1973 "Networks of Dynamic Structure: The Creation of a World Wheat Market in the Nineteenth Century". MS.
Howard, Leslie L. 1974 "Industrialization and Community in Chotanagpur", unpublished doctoral dissertation, Cambridge, Harvard University.
Levi-Strauss, Claude 1967 Structural Anthropology. Garden City, N.Y.: Doubleday & Company, Inc.
Marx, Karl 1963 The 18th Brumaire of Louis Bonaparte. New York: Inter- national Publishers.
Mills, c. Wright 1963 Power, Politics, and People. I.L. Horowitz, (ed.). New York: Ballantine B-oks.
Parsons, Talcott 1962 "Sociological Aspects to the Theory of Organization", in Amitai Etzioni, Complex Organizations. New York: Holt, Rinehart, & Winston, pp. 32-47.
1949 The Structure of Social Action. Glencoe: The Free Press. Simmel, Georg 1971 "The Problem of Sociology", in Donald N. Levine (ed.) Georg Simmel on Individuality and Social Forms. Chicago: University of Chicago Press.
Wagner, Helmut R., ed. 1970 Alfred Schutz on Phenomenology and Social Relations. Chicago: University of Chicago Press.
Weber, Max 1968 Economy and Society (Wirtschaft und Gesellschaft), An Outline of Interpretive Sociology. Guenther Roth and Claus Wittich (eds.). New York Bedminster Press.
1949 The Methodology of the Social Sciences. Edward A. Shils and Henry A. Finch (eds.). Glencoe: Free Press.
White, Harrison C. 1970 Chains of Opportunity. Cambridge, Mass.: Harvard University Press.
Wronq, Dennis, ed. 1970 Max Weber. Englewood Cliffs, N.J.: Prentice-Hall, Inc.
Zeitlin, Irving M. 1968 Ideology and the Development of Sociological Theory. Englewood Cliffs, N.J.: Prentice-Hall, Inc.