Arguing from different premises, we can never improve our understanding unless we
examine and reformulate our assumptions. (Douglas 1986. 8)

There is a world of differences between those who wish to reduce authority so as to
promote individual differences, and those who reject authority so as to reduce individual
differences. (Wildavsky 1982, 17)

In the years since World War II political science has frequently looked to other social sciences for inspiration and guidance. With some oversimplification could say that the 1950s and 1960s represented the heyday of sociology and social psychology, while the 1970s became the period of economics. After a period in which individual values and behavior and their measurement were highly fashionable, we have now witnessed a time in which models of rational actors have been at the core of much theorizing in our discipline. It appears that the time has now come for anthropology. Political science, and the social sciences more generally, seem to have seriously discovered "cultural analysis'.

Why cultural analysis and why anthropology? Political scientists are
looking for a theoretical framework which can make sense out of the

Cultures in a Comparative Perspective. Paris 10-15 April 1989. I thank Einar Berntzen. Richard Ellis, Gunnar Grendstad, Stein Kuhnle. Ulf Lindström. Johan P. Olsen. Kåre Strom. Aaron Wildavsky and one anonymous reviewer lor comments. A more extensive monograph. Werner Jann. The ideas behind this article are as much his as mine.

Resumé

This article is an introduction to a new and promising alternative to dominant reductionistic, utilitarian and instrumental political theories. 'Cultural Theory', or 'Grid-Group' analysis has been developed by the British anthropologist Mary Douglas and transferred to political science by the American political scientist Aaron Wildavsky. Starting from two dimensions - 'group' (group strength) and 'grid' (the number of rules that influence the individual space of action). there exist four main forms of human interaction or ways of life - individualism (markets). sectarianism (egalitarism), collectivism (hierarchy) and fatalism (slavery) which can be applied to all kinds of societies, organizations and individuals. These cultures are then used to predict political action (politics), institutions (polity) and results (policies). The article shows how the theory is built up and where it comes from, and then concentrates on important logical and empirical problems that so far remain unsolved.

confusing developments of recent years. Why do political regimes seem to lose the support of large parts of the population, especially among the younger members? Why are some nations apparently better able to cope with the economic difficulties of the 1980s than others? Why is it becoming more difficult to govern while governments allegedly are doing better than ever before? Why have environmental movements suddenly become so much stronger than they were only a few years ago while our environment seems to be getting better? Why does the overall consensus about our political institutions, procedures and outcomes seem to disappear while we have a more active political debate? And so on.

In this paper no attempt will be made to investigate whether these statements and worries are true or false. But it seems obvious that quite a few social scientists have turned to cultural theories in order to find more satisfying explanations of what is going on around us. Theories come in all methodologies and ideologies, from Berger's phenomenology to Foucault's neo-structuralism, from Habermas's critical theory to Inglehart's social psychology.

The reasons for the growing interest in anthropology seem straight forward enough. There is a long and distinguished tradition in anthropological about culture. Thus, in times of distress, why not try a hitherto neglected field? But there is more to it than this. Newer developments in cultural anthropology promise to solve two main problems of current mainstream political theory. On the one hand, political science in the behavioral or social psychological tradition has great difficulties in connecting values and behavior. If you share certain values, does that predict your behavior in all kinds of circumstances? Many have expressed doubts about this form of 'idealistic theory', in which values come first and action follows suit. On the other hand, political theory in the economic or rational tradition cannot explain where preferences come from either. They are treated as 'external' to social life. These are supposed to explain and guide all or at least most of our choices, but nobody knows where they come from or how they stick together.

Coming from the British tradition of cultural anthropology, Mary Douglas, of the leading 'Durkheimian' anthropologists of our time, has developed a theoretical and typological framework which tries to solve these and other problems. Her 'grid-group analysis' promises not only to put individual preferences (values, attitudes, etc.) and individual behavior into one common theoretical framework, but includes organizational forms and their problems as well as whole cultural systems within the same framework (Douglas 1970, 1975, 1982a, 1982b, 1986; Douglas & Isterwood 1979; Douglas & Wildavsky 1982).

The new cultural theory aims at cultural and social explanations, not
psychological ones. The break with social psychology is a break with

political culture as used (if used) in most of the political science and sociology literature (Almond & Verba 1963; Inglehart 1977; Barnes & Kaase 1979). In these studies the focus is on individual values and beliefs and it is the 'psychological orientation towards social objects' that is of interest. These theorists view the transformation of political culture in western society in light of changing conditions influencing basic socialization.In tradition political culture is not only 'the set of attitudes, beliefs, and feelings about politics current in a nation at a given time' (Almond & Powell 1978, 25), but political culture determines action. The subjective dimensions of politics are given explanatory power (Almond & Powell 1978, 25; Almond 1980, 27).

By denning culture as 'way of life' and arguing the coherence between organizational forms and subjective perceptions, the new cultural theory is different. Beliefs are no longer separated from structure or action, but are part of the action itself (Douglas 1982a, 199-200). The question of what comes first, belief:; or actions, becomes irrelevant. Furthermore, the theory does not accept the dominant perspective in the social science literature of the isolated actor, but emphasizes instead human interaction and integration. and norms cannot be understood without understanding the social context. Cultural theory is similar to modern network analysis in emphasizing the individual placed in a set of social relations (Burt 1979: Wellman 1983). However, in cultural theory you can explain neither choice nor institutions without understanding the normative and moral aspects of human life. Central to understanding the perspective is the emphasis upon culture as moral order (Wuthnow et al. 1984, 252).

In the following, the application of grid-group analysis as it has been put forward by Aaron Wildavsky (Douglas & Wildavsky 1982; Thompson & Wildavsky 1986; Thompson et al. 1990; Wildavsky 1982. 1983. 1984a, 1984b, 1984c, 1985, 1987) will be examined. Attention will focus on the writings of Wildavsky from the early 1980s not only because his development the theory is one of the most elaborate and provocative applications of and additions to political science, but also because it shows very clearly the promises and pitfalls of applying cultural anthropology to current problems. Of crucial importance in understanding "grid-group" theory is the book. Cultural Theory (1990). Here Wildavsky, together with Richard Ellis and Michael Thompson, gives a concentrated expression of what cultural theory is all about. The work of Douglas will be referred to only in order to clarify the theoretical background and implications of the theory.

Wildavsky has chosen "to break [his] bones on the rocky shoals of a general theory", by trying to develop a "global cultural theory in order to explain political phenomena' (Wildavsky 1984c. 1). The intent here is to give the reader an impression of why this theory is of importance to political

science and the social sciences in general, and, at the same time, to put fortn some doubts about its present formulation. First a brief introduction to how cultural theory or 'grid-group' analysis is built up will be presented. The second and main part of the article will be a critical but benevolent evaluation of the main parameters the theory builds upon, and an assessmentof consequences of what may be seen as problematic assumptions.The part emphasizes what is new about this perspective and how it is linked to other ongoing debates in the social sciences.

The Composition of 'Grid-Group' Theory

'Grid-group' theory starts from the basic assumption 'that what matters most to people is their relationships with other people and other people's relationships with them' (Wildavsky 1984c, 1). The major decision for people is therefore the form of social order they adopt. 'Social order' is defined as, 'shared values legitimating social practices'. A key proposition of the theory is that there are only a limited number of such 'social orders', subsequently called 'cultures', but which may also be called 'ways of life'.

The construction of the theory is essentially as follows: from two basic dimensions of 'social order' four distinctive 'cultures' are constructed, which can be combined to form eight political 'regimes', which in turn can be used to predict innumberable political processes, structures and outcomes (see Figure 1). The two basic dimensions of all possible social orders, cultures or 'ways of life' are: (1) the outside boundaries that people have erected between themselves and the outside world ('group' in Douglas's terminology); and (2) all the other social distinctions and delegations of authority that they use to limit how people behave toward one another ('grid' in Douglas's terminology). Why there are only these two dimensions is not explained. They are 'generalizations of generalizations' and chosen to yield the 'smallest number of dimensions with the largest payoff in predictive and retrodictive power' (Wildavsky 1984c, 5).

If one combines these two dimensions, one gets four basic cultures: (1) hierarchical collectivism (strong groups, many prescriptions), (2) competitiveindividualism groups, few prescriptions), (3) egalitarian sectarianism (strong groups, few prescriptions), and (4) powerless fatalism (weak groups, many prescriptions). It seems fair to assume that fedualism, laissez-faire capitalism, a small New England settlement and finally a colony which is totally controlled from the outside would correspond to those four cultures. But perhaps cultures are smaller, perhaps they do not encompass whole societies? In this case the examples could be different forms of organizations: a tightly controlled bureaucracy, a network of brokers, a small egalitarian political organization and finally, slaves on a big

plantation. But perhaps the theory is really about individuals? Then the cultures could be represented by the hierarchical bureaucrat, the individual entrepreneur, the egalitarian sectarian and the powerless fatalists. Obviously we are dealing here with ideal types" in the Weberian sense, and will have to return to the problems which this entails.

But that is not all there is to it. These four basic cultures can be combined to form basic 'regimes'. In fact, one of the assumptions of the theory is that cultures never exist on their own. but that at least two cultures are always necessary to form a stable form of organization. Thus, the

combination of hierarchy and sectarianism is called 'social democracy', sectarianism plus individualism gives the specific culture which is assumed to prevail in the United States, i.e. American exceptionalism or 'American individualism', individualism and powerless fatalism lead to authoritarianism'(or 'state capitalism') and fatalism combined with hierarchy breeds totalitarianism. Finally, if one prefers to get by without coercing others or being coerced by others, there is a 'hermit' culture (Thompson 1982a, 1982b; Thompson et al. 1990). In Table 1 the two dimensions are combined to show where each culture and each regime fits into this scheme.

The interesting assumption underlying the theory is that 'people construct their culture in the process of decision making. When choices are not completely controlled by context, people discover their preferences by deciding whether they will reaffirm, modify, or abandon their way of life' (Wildavsky 1984c, 7). The important decisions individuals make are simultaneously choices of social order or, if you prefer, culture. People discover their preferences by continuously constructing and reconstructing their culture through decision-making. The theory does not accept that values and preferences are external to human decision-making, but wants to bring together values and facts, human preferences and human observations.

It is very important to understand the structure of the theory: Value

patterns (= shared values) legitimize and reinforce behavioral patterns (= social practices). Together they form 'social orders', 'ways of life' or 'cultures'. It does not make sense to ask which one come first, because the one cannot exist without the other. Together they are supposed to predict political behavior (politics), institutions (polity) and outcomes (policies) (cf. Figure 1). We do not, however, get any explanation of how people choose their culture in the first place. Those choices remain, for the time being, a priori assumptions.

Applications of the Theory

Let us now look at the kind of explanations at which the theory so far has aimed. The theory predicts that hierarchies seek to maintain social differences, sects to diminish them, and markets to allow them to grow large provided that competition remains possible. Thus regardless of whatever is to be decided, people who identify with a particular 'culture' or 'way of life' guess whether the decision will increase or decrease social distinctions and decide accordingly. These guesses are informed by what like-minded people think, but usually members of a market or a sect are able to figure out on their own whether they oppose or approve of a certain policy, e.g. progressive income tax, etc. This means that each culture has its own definition of equality. Hierarchs believe in procedural equality (everyone gets what his status entitles him to), market people in equality of opportunity can take part in market exchanges), and egalitarians in equality of result (everyone is treated equally).

The theory has also been used to explain organizational behavior, e.g. the handling of information in different forms of organizations and the evaluation of policies. Hierarchies believe in equality before the law so as to adjudicate statuses. Who has the right to do something matters at least as much as what is done. Therefore, hierarchies are concerned about preand data on the legality of expenditures, while data on the results of activities are secondary. Data on the differential effects of activities on various sub-units are ignored or even suppressed because they could threaten the legitimacy of obediency to the center. We are told that markets, on the other hand, 'could not care less about proper procedures". All they care about is the 'bottom line' - profit, cost-effectiveness, or popularity. They are in perpetual audit: 'What have you done for the bottom line lately?" Sects, which want to equalize differences, ignore information favorable to increasing social differences. They focus instead on information against central authority and inequality.

In the same way the theory attempts to explain why certain cultures rely
more on policy analysis than others, or why the United States, compared

with Europe, relies so much more on independent policy advice. The answer is by now obvious: the US is more market-like than for example the European countries, hence conflicting evidence is not looked upon as a threat to established hierarchies, but as a welcome addition to the ongoing competition of ideas. Graduate Schools of Public Policy only thrive in the US because other countries cannot really use all this conflicting evidence in their orderly hierarchies.

Finally, the theory has been used to explain why people are only concerned certain kinds of risks and ignore others. Why is the American public much more concerned about chemicals in food, while they at the same time accept the much higher risk in driving an automobile? People care about different kinds of risks, because they 'defend their way of life'. Thus sectarians care about the risk of industrial products because they oppose big industry. They believe that 'corporate capitalism causes cancer' because this belief embodies all their other values, i.e. distrust of corrupt worldliness, social distinctions, division of labor, material values, big technology, industry, big organizations, and so on. Sectarians believe that the 'established society is incorrigibly evil, being both coercive and hierarchical' (Douglas & Wildavsky 1982, 122), and therefore they care most for those risks which are caused by the establishment. At the same time their expectations of the future are negative: 'The sectarian cosmology expects life in the future to undergo a radical change for the worse. It is not confident that the disaster can be averted. There may be no time left. But it knows how the disaster has been caused' (Douglas & Wildavsky 1982, 127).

On the other hand, the 'establishment cultures', markets and hierarchies, are much less concerned about those kinds of risks. The market culture incorporates calculated risks as legitimate costs and tries to spread the possibility of risk as widely as possible. The concept is that of a short-term outlook with insurance. If the possibility of risk is small, why worry about it now? Nobody knows about all the forms of future interactions and their outcomes, the future will take care of itself. Hierarchies care about longterm but because of the preoccupation with the right procedures and the right to act, they have blind spots concerning the future. They cannot see the long-term dangers, and anyway the future will not be very different from the past. Risk and its perception are routinized through established procedures. Phenomena which do not fit these procedures are ignored for as long as possible. In other words, all cultures have their own 'cultural bias'.

A Critical but Benevolent Analysis

In order to improve (or ultimately reject) a theory it is necessary to spot

and analyze its weak points. While the approach is highly fascinating. it seems that the theory, even as it is formulated in the most recent book (Thompson et al. 1990), has serious flaws and that certain aspects must be clarified. In examining the theoretical and methodological problems of cultural analysis, the same steps as in the first section will be followed, i.e. after starting with the problems of the two dimensions, the four cultures and various regimes will be considered before concluding with the predictionsand of the theory.

The Dimensions

There is no comprehensive explanation as to why only two, and in particular the dimensions group/grid, have been selected to explain all kinds of social interaction and phenomena and, ultimately, all kinds of evil in the world. The main reason provided is that there must be forms of social organizations other than markets and hierarchies, and there are voluntary sects. One may fully agree, but that does not automatically entail grid-group analysis. It is surprising, moreover, that political science is supposed to have ignored voluntary organizations until now. Voluntary organizations play a veryimportant both in theories of mass-society and pluralism, two of the dominant perspectives in sociology and political science after World War 11. So what is new? This is the first time a single framework tries to integrate the main forms of human organization into a single typology. Such a typology, if it can be found, would indeed be very useful.

The theory has been inspired by cultural anthropology and rational choice perspective and tries to improve both. But one has to remember that at least the latter kind of theorizing starts out from a very special kind of theory of behavior: that of economics and the part of the organizational theory most strongly influenced by economics. In this tradition the ordering of assumptions does not have to be empirically or theoretically "proven". their only test is that they are theoretically meaningful and clear enough to give us a deeper understanding of similarities and differences of cultural forms, in other words, to allow us to systematize real world phenomena. Thus the question 'why these two?' can only be answered in retrospect: 'because they make sense". But this gives rise to the next question: Is it clear what these two dimensions mean? Unfortunately the answer here is negative. Definitions and operationalizations are unclear and even change. Let us look at each dimension in turn.

Grid is usually defined as "many vs. few" rules which govern behavior and limit choice. Thus markets and sects are characterized by few rules. But in reading Risk and Culture (Douglas & Wildavsky 1982) it becomes obvious that this cannot be the whole story. In this book two environmental groups, the Clamshell Alliance (CA), called an 'action group", and the

Environmental Coalition on Nuclear Power (ECNP), called a 'local intervenergroup', contrasted. The 'action group', CA, is supposed to be more sect-like, because its members do not recognize leaders and officers, while the other is more informal:

. . .whereas action groups (like CA) have strict rules about how decisions are to be made and how the organization is to be administered, these rules are designed to prevent hierarchical type leaders from taking control and to provide for the resolution of disputes in the absence of authority. In contrast, local intervener groups (like ECNP) have a much less formal view of their internal structures (Douglas & Wildavsky 1982, 147).

Obviously, the 'action group' has many more prescriptions. But still, the
conclusion reads:

While the ECNP also rejects formal leadership in the sense of elected officers, it lacks rules which are deliberately designed to prevent the emergence of leaders. Leaders of intervener groups tend to be entrepreneurial individuals who take on on the leading roles in promoting local opposition to nuclear plants. Hence the ECNP rates higher on the grid dimension that the strictly egalitarian Clamshell Alliance (Douglas & Wildavsky 1982, 148-149).

This does not make sense. Either the CA has more rules or it hasn't. In this case, the group with more rules has less hierarchy, so why should it be placed lower on the grid dimension? Three different explanations are possible: (1) the dimension has changed its meaning in the course of the argument such that it now measures hierarchical or egalitarian values or more abstract, social differentiation; (2) it is actually supposed to measure strong vs. weak rules; or (3) it measures internal group rules.

None of these alternatives is very convincing. If 'grid' meant acceptance or rejection of social differentiation, then markets and sects should be on the same side on issues like progressive income tax, free education, etc. Obviously they are not. In any case, social differentiation (the like or dislike of hierarchies) is what the grid dimension is supposed to explain.

'Grid' cannot be about strong and weak rules either. Low grid can mean very strong rules. The market and the sect may have few rules, but the ones they have are extraordinarily strong. Any reading about such sects as the Amish, Puritans, or Quakers shows that the life of their members is, or was, full of rules (one is not supposed to dance, gamble, use automobiles, and so on). To explain those rules by egalitarian values is not in accordance with the theory. The explanation should run the other way round: Low grid/high group is supposed to explain egalitarian values.

So perhaps those rules are only necessary to keep the group together? But then the distinction between grid and group disappears. This brings up the next question. What does the group dimension really define? What does it mean that 'group boundaries' are high? At least in Mary Douglas's writing, group has something to do with face-to-face contact. Scoring high on the group dimension would then imply preferring small face-to-face

groups in contrast to larger impersonal ones. But that would leave all people who belong to groups which are large, but which have strong boundaries (i.e. members of large hierarchies, like the Catholic Church), with a low score on the group dimension. If group boundaries are kept up by social pressure in face-to-face groups, what about people who are members of many groups? According to this definition, modern societies would be defined as 'low group' because they are characterized by constant cross pressures. If someone is a member of many groups, he must be low on group. Douglas (1982a, 201) argues thus. But why are members of groups concerned with environmental questions to be looked upon as 'high group'?

Does 'establishing high group boundaries', moreover, mean that one accepts the group's decision on different aspects of one's life? If so. is that not the same as 'grid'? If strong group boundaries need strong rules, 'grid' would be a function of 'group'. A group with very strong boundaries (say a caste in India) prescribes all kinds of behavior for its members. Since all rules can be interpreted as group rules, the grid dimension is by definition empty. This is obviously not what the theory wants to tell us. But especially in the discussion of sects one gets the impression that grid becomes low because group regulation is strong. This seems to be necessary to keep the low grid assumption. The fundamental problem, in short, is why a sect, whatever the type, has to be low on grid. If this argument holds true, it would mean that the two dimensions are not independent of each other.

There is a strong possibility that 'grid' and 'group" are in fact partly the same, i.e. that both dimensions measure partly the same thing or. just as disturbing, the same thing twice. If 'grid' is the social regulation of behavior per se, what is left for 'group'? Obviously, the grid-group dimensions are not quite clear. Unclear dimensions have severe consequences, because the location of social phenomena becomes somewhat arbitrary. This means that a conclusion/deduction based on one definition too easily starts to be valid in another area. Thus, our first reservations concern the underlying dimensions. They have to be theoretically refined and operationalized to make them useful for empirical analysis. But even having settled these problems, there are other aspects of the theory which can be discussed and which are questionable independently of whether "grid" and 'group' can be sufficiently defined. Let us look at some of these other problems.

Four Cultures

In the following, it shall be argued that the basis unit of analysis in this form of cultural analysis is not the two dimensions discussed so far. but the four cultures deduced from them. While the theory promises the least number of variables with the highest predictive consequences, it actually

delivers four descriptive typologies with innumerable connotations and variables. Do all these 'shared values legitimizing social practices' really follow from our two main dimensions and can we actually find them in reality? Does 'low grid/high group' logically lead to 'sects', and why is 'high grid/high group' always a hierarchy? Indeed, what is really meant by the terms, 'sects' and 'hierarchies'?

There are two sets of problems here. One concerns the logical coherence of the theory: What kind of auxiliary assumptions are necessary to keep this theory together? What kind of social practices belong necessarily to one type of culture? The other concerns its empirical usefulness: Can we find hierarchies, markets and sects which display the characteristics which the theory postulates?

We may begin by asking how one gets from two dimensions to four cultures? Cultural theory postulates that the four main cultures follow necessarily from the 'internal requirements' of groups of people who belong to one culture. It wants to 'account for behavior on the basis of the internal requirements of groups of people who choose to live a particular way of life' (Wildavsky 1984a, 36). Thus, if one chooses to live with high group boundaries and few rules, not only does it follow (according to the theory) that one would prefer equality of results and abhor authority, but the cultural choice also explains one's distrust of big business and big government preference for handwoven clothes and meals without a main course.

This kind of reasoning appears to rely on at least three clusters of assumptions which are questionable: (1) preferences are not external but forced on one by organizational requirements; (2) people act coherently within their chosen way of life; and (3) people actually choose their way of life and can leave it at their will. Each of these assumptions may be examined in turn.

External Preferences?

Cultural theory does not accept that preferences are external, but tries to show that they follow necessarily from your chosen way of life: 'People discover their preferences by continuously constructing and reconstructing their culture through decision making' (Wildavsky, 1984c, 8-9). The argument as follows: Egalitarian values and practices follow from a commitment totally voluntary organizations which in turn follow from a commitment of low grid/high group. I 'discover' the interconnectedness of all my values and actions while acting. Put differently: our preferences are something our organization does to us. I may not like equality, but I am forced into it because I like strong group boundaries and few rules.

Does this make sense? There are two obvious problems. First, why do I

like strong group/weak grid? Are these not external preferences? How are they explained? This theory cannot stand in thin air. In this explanation the turtle which supports the elephant which supports the world has to stand on something. Second, why would I ever join a sect or an egalitarian organization, say the Norwegian Bellona? Is it because I am upset about certain environmental hazards and feel that the existing parties tend to ignore them, or is it because I prefer few rules and strong groups? Do I only care for poor people and trees because I am strongly committed to a strictly voluntary way of life? Which explanation makes most sense?

Coherent Actors?

Is it possible to be a vegetarian corporate executive? What about a hierarchical or an autocratic leader of a sect? The theory has two other strong requirements which are rather doubtful. In order continuously to reconstruct my 'way of life' (1) I have to find out what supports and what endangers my way of life, and (2) I have to act accordingly, i.e. coherently.

How do I know what supports my way of life? In true market fashion, the theory tells me, I use trial and error to find out. A capitalist, we are told, would not extract a monopoly rent over a longer period of time, because that would be culturally irrational, it would endanger his way of life. But this argument would mean that people know about the unintended and unanticipated consequences of their actions. I would have to know, and care about, long and complicated causal chains in order to find out what supports my way of life and what goes against it. This is implausible if not impossible. If I only care about the most imminent and obvious consequences of my actions, furthermore, how can I then protect my way of life? There must be an invisible hand of cultural theory which guides our actions, otherwise all four cultures would be very unstable.

But even if I knew what was good for my way of life and what was not. would I act accordingly? Do people hold coherent views of life and do they act in accordance with them? To put it differently, do all members of a hierarchy have to share hierarchical values? Do all members of egalitarian sects have to share sectarian values? Obviously not. There must be more than one culture in each organization. But what about each person? In Mary Douglas's writing, the entrepreneurial businessman can become the tyrannical hierarch at home, and a lot of studies from psychology question cognitive consistency. If this is so, how can we then predict behavior?

The stress upon cultural forms as being autonomous has other consequencesas This is related to the unusual use of the term "voluntary organization. Tocqueville, who was among the first to connect egalitarianismwith organizations, claimed that increasing egalitarianismproduces

tarianismproducesa power vacuum that is filled with voluntary organizations. Douglas and Wildavsky argue that it is the other way round. They start with the power vacuum which creates the problem of organization,which partly solved by the principle of equality. Egalitarianism as an idea is in essence reduced to an internal organizational problem.

To make the rise of new and powerful groups of environmentalists, antinuclear and feminists into a problem of the internal coherence of organizations is too great a step to take. Furthermore, if the coherence argument is correct and people's values are reinforced all the time, how can it then be that such groups have to exclude people all the time? Is this just a problem of organization and unobtainable values, or does it have something to do with political differences with regard to acceptable solutions, which means inconsistency in each organization, i.e. more than one culture in each organization. These problems have to be dealt with by the theory, but can only be integrated if the assumption of autonomous and coherent cultures is given up. Wildavsky's latest theoretical work moves in that direction (Wildavsky 1987; Thompson et al. 1990, especially the last chapter). However, such a move in the direction of increased realism increases, as we will see later on, measurement problems, or even makes them unsolvable.

Free Choice?

Finally, in trying to get from the two dimensions to the four cultures and connecting shared values with social practices, there is the problem of choice. Why do I choose a certain culture in the first place? In true 'rational choice' manner, the theory declares that 'culture is a contract', it depends on its ability to provide satisfaction to its adherents. If I am not satisfied, I can exit. But this assumption cannot explain why there are fatalists and slaves.

Where does dominance and coercion fit into this theory? Is it only in the upper left-hand corner, among the alienated fatalists and unhappy slaves, that we find people who are forced to do things they would rather not do? Or does 'low grid' enlarge my options? Or perhaps 'low group'? That would mean that only the competitive entrepreneurs can really choose. If I cannot choose 'my culture', why would I repeatedly reconstruct it through my actions? If I can choose, why would anybody choose to be a slave? We seem to have a two-step model. First you choose your culture, then the social organization reinforces your values and beliefs. Choice means that the cost of exit and entry into a new culture is not too high. But choice is never placed in a historical context. Any discussion of increased choice needs to stress qualitative elements, the choice must be meaningful (Robertson1966). can never be a question of only more or less. If it

were, everybody would of course choose 'low grid'. But this would be a
one dimensional way of approaching choice, supporting a kind of liberal
philosophy which really is supposed to be overcome.

The reason why this kind of quantitative approach to choice is insufficient is that any real discussion of power and domination is lacking. The relationship freedom and choice is an important one, and one that cultural theory has to take seriously. If people choose their culture, even though some cultures offer more choices than others, where do the preferences for a certain culture come from? If after each choice the social context changes and reinforces beliefs, values and behavioral patterns, why do people become members of a certain culture in the first place?

The main reason for the coherence assumption is that the perspective has been transferred from anthropology, from studies of tribes and smaller societies, into modern, much larger societies. To what extent this is problematic never really discussed. The theory is built on the very interesting assumption of small differences between 'simpler' and more 'complex' societies and basic similarities of organizational forms. Scale and technology are not looked upon as determining social organization. In Cultural Bias, Mary Douglas (1982a) discusses the insufficiency of traditional dichotomies like mechanical/organic solidarity, status/contract society, and rightly stresses the emphasis of these concepts on the breakdown of corporate groups,

. . .as if the increase in individual freedom could only be traded against decrease in group
strength. I am trying to present a less impoverished view of social change (Douglas 1982a.
191).

Nevertheless, change (or development) in cultural analysis means decreasing strength and more than anything else a downward move on grid. Lower grid then means more choice options, which in turn implies greater individual freedom. Hence the theory, when used, is in line with traditional functionalist and liberal theory.

It would seem that the tension between the ahistorical classification and an underlying development scheme is the reason for the somewhat strange use of the terms 'hierarchy' and 'sect. After studying voluntary organizations some time we have never met real organizations as described in this perspective. Not even environmentalist organizations have 100 percent participation and commitment. These people also have to work, start a family, and relax. The impression we get from the theory is that participation in voluntary organizations is almost equal to living in a commune (Kanter 1972). Again, it is the coherence argument that forces unrealistic assumptions.

Since most environmentalist and other contemporary 'sectarians' do not
live as they are supposed to, but are 'members" of many and different

institutions and groups, we need a qualitative framework grasping which kinds of 'socialization' or 'membership' are the most important. We need this, because, according to Mary Douglas's own discussion in Cultural Bias (1982a), in modern societies most people, since they are members in many groups, become low on group. Furthermore, her 'group' dimension is explicitly based on face-to-face contacts. How can one make such assumptionsand the same time use such unrealistic definitions of voluntary organizations?

Even logically it is not readily evident why low grid and strong group has to be a sect or why strong groups with strong grid should be hierarchies. The 'hierarchy' of cultural theory is one without competition, where everything seems more or less fixed. This is far away from modern bureaucracies in the Weberian sense. It would appear that this has to do with the a priori assumption that 'strong group/strong grid' fits both the old 'status' society (like feudalism) and the modern bureaucracy (for example in a large firm, a political party or a governmental bureaucracy). Yet why is it that bureaucratic organizations cannot logically stress individualistic values, why is it that strong group and strong competition cannot go together?

Why the dimensions pose problems and why they do not always have to produce the cultures they are supposed to may be illustrated by a wellknown from the anthropological literature. The British anthropologist Barnes, writing mainly in opposition to Redfield's description of the so-called 'folk society', has studied the Norwegian county of Bremnes over a considerable period of time (Barnes 1954, 1971, 1978). His description of the development from 1950 to 1970 could be summarized as follows:

In 1950

• a very egalitarian society with strong and fixed geographical boundaries;
• very strong religious sects (cultural domination) with strong regulation
of social life and with unusual emphasis on egalitarian values; and

• very limited choice opportunities.

In 1970

• a much less egalitarian society;

• much weaker religious sects (no cultural domination); and
• strong increase in choice opportunities.

Going back to the dimensions, this implies a move down both group and grid. Individual choice opportunities have increased at the same time as group integration has decreased. This is a development towards the 'market-culture' and should not, as such, produce serious trouble for the model. The problem is, where are we coming from? Since group was strong and grid was strong, we must have left the upper right-hand corner (cf. Table 1), where hierarchies are supposed to rule.

However, highly regulated social life and strong group boundaries do not necessarily have anything to do with hierarchies. We have here a combination of strong regulation and strong egalitarianism. a combination that is not all that uncommon. (Think of the social science concept of 'community'.) As we understand the dimensions, there is no logical reason why egalitarianism and strong group/strong grid cannot go together.

Two Other Questions

For the time being let us assume that people act and think coherently, that they can choose their culture and that preferences follow from the internal requirements of each culture. Two important questions still remain, one logical and one empirical: (1) Do the 'shared values and legitimized practices' which are identified in the theory really belong to each culture? Is competition, for example, only a part of the market culture and is voluntarism necessarily confined to sects? (2) Can we actually find those cultures, as defined by the theory, in reality? With these questions in mind, each culture may be briefly considered.

Individualistic Sects?

The sectarian way of life is supposedly characterized by strong groups and 'low grid. Yet there are some definitional problems here. If we look at traditional religious sects, it becomes very questionable that they can be characterized by few or even weak prescriptions. Sectarian life is guided by innumerable prescriptions: 'don't play cards, don't dance, do not engage in sexual practices'. If all these prescriptions are defined as "group enforcing rules' it becomes very hard to distinguish between the two dimensions. It is then basically impossible to falsify the theory. If all rules imposed by the group (or organization) in order to strengthen its boundaries are looked upon as non-grid, then the hierarchical culture becomes empty. It is at least impossible to distinguish between sects and hierarchies, unless "grid" means 'social differentiation', but then the dimension has changed its meaning.

Let us try a different way of reasoning: Does a strong commitment to a completely voluntaristic way of life in cultural analysis guarantee personal freedom? Obviously not. In fact, it seems to restrict personal choice quite severely. So why would one choose to stay in this culture? How does this culture become reinforced?

Voluntary Hierarchies?

It is possible to extend this argument by looking at different kinds of
voluntary organizations. There are organizations which can be characterizedas

terizedasvoluntary hierarchies. A voluntary hierarchy is an organization which one may (or may not) join and which one can leave whenever one wishes to. These organizations have a strong internal structure, with, let's say, ordinary members, a general assembly, perhaps different levels of steering committees and a whole range of officers. The important aspects of those organizations are: (1) they have strong membership boundaries, i.e. you have to be accepted, you pay membership fees, and it is expected that you at least partially take part in the activities of the organization; (2) the organization is democratically controlled, i.e. all officers are elected, there is no co-optation, the power of the officers is limited, and so on; and (3) membership is voluntary and typically does not prescribe behavior outside the organization.

These kinds of organizations are very common for European voluntary organizations. All kinds of organizations are organized this way: Soccer clubs, 'Gesangvereine', fishing clubs, even political parties. Perhaps political best demonstrate the differences between European and American voluntary organizations. In Europe you become a member of a political party by signing an application, existing members decide whether you can join, you pay monthly dues, you are expected to take part in regular meetings, you elect all officials of the party at your level, and each level elects delegates who control and elect the next level of leadership. This is obviously a highly structured and at the same time very democratically controlled organization. Co-optation and interference from non-members are rejected.

So where do these organizations fit into this theory? Are they hierarchies or sects? They possess an elaborate internal structure and strong group boundaries, i.e., one could argue, high group, high grid. But there can also be no doubt that these organizations are voluntary. The point here is that strong preferences for voluntarism can be neither logically nor empirically linked to egalitarian sects alone. Hierarchies and voluntarism can go together well, as long as hierarchies are democratically controlled. But this in turn means that competition is not confined to the market culture, it thrives well in a high group, high grid environment. Thus the connection between competition and the 'market culture' also cannot be upheld. It is very easy to construct a hierarchy with high group boundaries and many prescriptions which is perfectly competitive. Just think of the ideal model of a European socialist political party. Each organizational unit elects the delegates to the next level and members compete fiercely for these important but each unit accepts the policy directions of the elected leaders.

Hierarchical Markets?

Let us look at another example. If an American wants to get some exercise

he or she will join what is called a 'Health Club', which is essentially a profit-making company. The person pays his or her membership fees and is allowed to use the facilities, and that is all the influence there is. In Europe, you could join a 'Sports Club' which would really be a club, i.e. a voluntary organization. There would be a general meeting of all members once a year, officials would be elected, policies decided, and so on. Thus, where Europeans would rely on a voluntary, hierarchical, but democraticallycontrolled Americans would rely on the market.

But are these market organizations less hierarchical and more democratic than voluntary hierarchies? Of course, you can always go to another 'Health Club' until you find one that provides exactly what you want (dependent on where you live), so in theory the market should work fine. But the argument here is on another level: What 'cultural analysis' considers typical market organizations can be looked upon as hierarchies, and those identified sects, are really market organizations.

Let us think of a huge 'Health Club' with lots of 'paying members', i.e. customers, and perhaps 100 employees. It seems obvious that this is also a strongly regulated organization, probably a hierarchy. The people who work in the organization do not have much influence on its rules and goals, only the customers can influence it through their 'money votes' or ultimately throught 'exit'.

Markets are really characterized by hierarchies of the involuntary kind, i.e. domination and centralization. If you want to earn a living, and most people in market societies must, you have to join one of the existing hierarchies. You join voluntarily, but what kind of choice is there? Basically, the choice between different kinds of hierarchies, but for most people they are all involuntary hierarchies. Of course, you can also become 'self-employed', but is that really a viable alternative for most people? Most people in market societies will spend most of their lives in involuntary hierarchies in which they have little control over the organization. In Bendix's words:

Ideologies of management are attempts by leaders of enterprises to justify the privilege of voluntary action and association for themselves, while imposing upon all subordinates the duty of obedience and of service at the best of their ability. In the ideal world of laissezfaire the right to combine is theirs alone based on the right of property ownership (1974. p. XXIII).

In this respect ideology' had to be separated from 'reality", a separation
which is thus far missing, or at least diffuse, in the treatment of the market
culture in cultural analysis.

Dominant Fatalism?

So what kind of organization is typical of a market? Most common, it would

seem, is a leadership (the owners, or better, the management, uncontrolled by the members of the organization) on top of an involuntary hierarchy combined with many independent customers. This leaves the leaders of these organizations in the low group, low grid culture, but not the members or the customers. Both must be characterized by low group, high grid; they have no direct control over what they have to do and what they are offered. Obviously, most people in a market culture will not be members of the entrepreneurial group, but will be 'lowerarchs' in some hierarchies.

Thus, from the dimensions as they have been defined so far, and from the same type of illustrative use of empirical data as is, for instance, found in Risk and Culture (Douglas & Wildavsky 1982), one could just as well argue that modern America is best characterized by the upper left-hand culture, the alienated fatalism. The name may not be very suitable or comfortable, but one could argue that the location follows logically from the theory. People do not belong (and do not want to belong?) to strong groups, but they are confronted by innumerable prescriptions. Anyway, why does the theory have so little to say about this upper left-hand corner? It is strange that the theory so far mostly ignores one of its archetypes. 'Fatalism' or whatever one wants to call this culture, has been treated as a kind of left-over category, the place which is neither market, nor hierarchy nor sect. It must be treated very seriously, and it is increasingly so in the latest book, Cultural Theory (Thompson et al. 1990).

Blame Yourself or the System?

Finally, let us look at how cultural theory assigns blame, because it can teach us something about how the theory is used. The theory assumes that people in different cultures react differently when something goes wrong, when their preferred culture does not work as intended and expected. Thus, sectarians blame the system, hierarchs blame deviants, and market individualists blame themselves. This seems to be a highly arbitrary interpretation the explanations used in different cultures, and if one goes wrong here, this implies important consequences for the interpretation of empirical data. Let us look more closely at the possible arguments.

The ruling assumption can be stated quite bluntly, sectarians assume that the voluntary group is without (immanent) fault, it cannot do wrong; hierarchs assume that the hierarchy is without fault; and market believers assume that the market can do no wrong. Whenever the favorite culture does not work, blame is put on the other cultures: sectarians assume that the market and uncontrolled bureaucracies ruin nature and human beings, hierarchs believe that sectarians and uncontrolled markets ruin stable societies; and market believers know that markets would work just fine if

it were not for government intervention and culturally dangerous ideologies
of equality, etc.

There are other possible explanations. One can always blame other human beings. Thus sectarians believe that otherwise perfect voluntary groups are ruined by power-hungry individuals who even strike deals with the enemy; hierarchs believe hierarchies have problems because people do not know and do not want to accept their place, and market believers know that problems only arise because people are either careless (industrial injuries) or not quite fit for the marketplace.

Conspiracy is a way of thinking which is widely used in all the three 'active' cultures. If the market does not work, there are trusts and cartels (usually hidden), or dangerous ideologies, which disturb it. People in hierarchies are suspected of wanting to build up another hierarchy or perhaps of being employed by another one, and sects always fears that their members strike deals with the devil.

Thus self-blame is usually only an explanation of the last resort. If I want to stay in a culture which otherwise treats me badly, and if I furthermore believe there is no fault in the culture and no 'foreign¹ intervention, selfcriticism the way to justify my actions. But that is well-known in all three cultures. Thus, the main reason why Wildavsky uses a different logic in explaining blame in the market culture is his strange use of the concept 'market', making it for instance impossible to separate utilitarian individualism expressive individualism (Bellah 1985). It is not clear whether business corporations should be seen mainly as market, or as hierarchy, or both. This also explains why the fatalist culture is not really integrated in the analysis. Thus far one finds no real discussion of how people's structural positions in big organizations differ. Wildavsky's 'market culture' could be looked upon as an attempt to resuscitate the myth of the lonesome rider brilliantly portrayed by Allan Ladd as Shane. Here, however, Wildavsky commits an ideological fallacy devastating to the perspective, because even Shane helped the small farmer. Not even he really believed in self-blame, but in social responsibility.

Eight Regimes?

The discussion so far leads us directly to the next step in the theory. Modern societies are not made up of single cultures but are combinations of different cultures. In fact, we are told that no society can consist of only one culture, each culture needs others to exist:

inegalitarian market and an inequitable hierarchy - to criticize (Wildavsky 1984 c. 7).

The point is well taken. It has indeed just been argued that markets really

need fatalists in order to operate. But this step in the theory engenders even more problems. Again they can be divided into empirical and logical ones. An empirical problem is whether it is possible to place existing societies in one of the eight (or nine) hybrid cultures, which are called regimes. Space limitations prohibit us from looking at all possible regimes and their historic examples, but we may look at one prominent example in the theory - American Individualism. The central question here is where should the United States be located in the theory, i.e. which of the regimes is exemplified? Moreover, if real-life regimes are made up of combinations of cultures, there is also a logical problem of aggregation. What is the unit of analysis in this theory? Are we talking about societies made up of different cultures, or organizations with different cultures? Is the theory really about different personality types (entrepreneurs, fatalists, hierarchs and sectarians), or does each person consist of different cultures?

Where Does the US Fit in?

In the theory the US is characterized as a society with strong markets, moderate sects and weak hierarchies, i.e. low on the grid dimension and somewhat unclear on the group side. This is what 'American Individualism' (or exceptionalism) is all about. Again, however, doubts may be raised. Let us look at the grid/rules side first. There are again two possible arguments. The US has fewer written rules than, for example, good old hierarchical Europe, or it has fewer unwritten rules.

Both possibilities are highly doubtful. Wildavsky himself realizes that the US in many areas relies more heavily on written rules than Europe, for example in the field of environmental protection. His explanation centers on the influence of sectarian groups in the US. Since they want to protect nature or make life hard for the market, they have to rely on written rules. Written rules are indeed in many areas at least as important in the US as in Europe, but for different reasons. American 'red tape' can be as annoying as that in Europe. Anyone who has had to deal with American banks or the Immigration and Naturalization Service knows that. But those are not exactly sectarian institutions. Why is it so much harder to deal with a bank in the US than in Europe? Why does a company (one of those new, entrepreneurial, low-grid ones) have a book of more than 50 pages with rules for its employees? Why are there more written rules on American ships than on British ones (Richardson 1956)? Why do American manufacturing organizations have more formal rules, but are nevertheless more informal and unbureaucratized and open than are British ones (Jamieson 1980)? Why do college students in the US have to sign a long list of rules before they are admitted to college?

Another explanation comes in handy here. The US needs more written

rules because there are so few unwritten ones. In hierarchical, class-ridden Europe, everyone knows his place, but not so in low-grid, low-group America. Yet here again one may disagree. Any European observer cannot fail to recognize the amazing number of conventions which govern the US. One could mention the area of attitudes towards sex, but that may be a special case, as Douglas assumes. Even so, the many rules governing sexual behavior in the US are simply amazing to a European observer.

A more pertinent example is Congress. There can hardly be any doubt that Congress scores lower on the group-dimension than most continental parliaments (little party coherence), but at the same time it scores high on rules and hierarchies. Congress is governed by an amazing number of unwritten rules and at the same time gives its 'hierarchs', i.e. the chairmen of committees and subcommittees, an amazing amount of power. Authority in Congress is not assigned competitively as in the 'hierarchical' parliaments of Europe (party leaders compete, especially within each party), but according to the traditional rule of seniority. Thus there is much less competition between members in 'market-like' Congress than in the hierarchical' of Europe. This has also to do with the concept of leadership in the US. The 'winner-take-all' system leads to the phenomenon that elected (or very often appointed) leaders have much more power than they would have in a European setting; they are much less controlled by 'group'.

This interpretation would suggest that the US and Europe do not necessarily on the 'rules' dimension, but on the 'group' side. Let us look at a rather typical American phenomenon, the so-called 'mail order sects", such as the Moral Majority, Ralph Nader, and so on. and also environmental such as Friends of the Earth (FOE) (compare the description in Risk and Culture). Wildavsky has problems in locating them within his mode. Some - one suspects the ones he just dislikes - like the Moral Majority or the National Rifle Association, he calls 'small hierarchies', while others - the ones he really can't stand - like Nader's Raiders. Friends of the Earth, etc., are identified as 'sects'.

It may be argued, however, that all are typical 'market organizations'. They have a highly active staff (sometimes volunteers) and a highly inactive 'membership', which basically consists of customers who can be persuaded to 'buy' the public goods the organizations promise to deliver. The typical organizational form of these groups is not the hierarchy (or the pyramid) or an active membership as in a sect. These organizations are characterized by extremely low group-boundaries and a highly independent leadership. The leadership' can do what it wants as long as it gets the votes (dollars, or whatever), and leaders may even resort to the most modern forms of manipulation through mail campaigns or television. Perhaps the member/ customers are independent low-grid/low-group entrepreneurs, but they

appear to be more like low-group/high-grid people, who want their rules extended to everybody else (school prayer, abortion, etc.). This hypothesis leaves room for a different interpretation of current developments in the American system. It is not, as assumed by Wildavsky, a dangerous movementtowards groups in a low-grid culture, i.e. towards sects, but a movement towards more and more strict rules in a low-group culture.

Unit of Analysis?

These are, of course, empirical problems. But they may be solved only if
we can agree on how this cultural theory can be tested. Here again one
runs into logical problems. Let us try to sort through the contradictions.

Nations, like the United States, are made up of different cultures. But unless we specify the exact mixture, the theory cannot be tested. If we are just told that the US has elements of markets, hierarchies and sects, then anything can be explained. Prediction and retrodiction only work if it can be shown that the mixture changes. An increase in the number of people concerned with their environment does not, of course, demonstrate that sectarians have become stronger. That has to be demonstrated by an independent measure. But what kind of measure?

Can organizations be the unit of analysis? Thus, if we have more sects, does the cultural mixture change? If so, we again have problems, however. The connection between organizational form and members is obviously not simple. If there are different cultures in each organization, then there must be market individualists in hierarchies, sectarians in markets, but also hierarchs in sects. The point is, why should sects be purer than other organizations? But then how could we aggregate?

Cultural Theory - An Assessment

This leads us to a very important question. What does cultural theory really explain? If we start off from 'grid/group', there will be no problems. But as soon as the four basic cultural regimes are used, the problem becomes evident. We have 'social practices' on each side of our equation. Egalitarians value patterns legitimizing equality. Therefore they dislike social differences. This is a tautology, and common sense. Nobody would claim it to be otherwise. In fact, as soon as the four cultural types are used as descriptive typologies, prediction and retrodiction become highly problematic.

It would seem that cultural theory, when used empirically, is thus so far
not really about different cultures or different organizations, but about
different personality types, i.e. a micro-level theory. There is a tension

here between 'theory' and 'practice'. The theory as it stands now, is not consistent and can 'travel' to explain everything, i.e. nothing. This inconsistency is related not only to unclear definitions of the dimensions, and of the four main cultures, but to the lack of a real discussion of the problem of scale and unit of analysis. This is a fundamental problem, because if 'group' means social interaction, then 'group' is a function of scale, and will change as scale changes. That means that 'group' at one level of analysis becomes 'grid' at another level. This implies that not only are the dimensions not independent, but that they also relate to different levels. Cultural theory has to decide what it wants to measure and explain. Typologies of regimes, organizations and personalities have to become clearer in order to test the theory.

Wildavsky's move from political economy to political anthropology is a move away from technological rationality. However, the attack on rational choice theory is by no means only coming from the new cultural theory (Black 1984; March 1978; March & Olsen 1989). Cultural theory then can be seen as part of a more general move away from technological rationality towards an increasing interest in preferences and conflicting interests (Pfeffer 1981, 11). An interesting consequence that follows logically from the perspective as such is that symbols and rituals, myths and ceremonies are seen not only as 'expressive', but as 'instrumental' for the more general 'interest' of different cultures. The new cultural theory never leaves the assumption of rationality, but turns it into a cultural one. Cultural practices reinforce cultural values.

Cultural theory with its emphasis on rationality and congruence can be looked upon not only as a way of combining functionalism and rationality, but as a serious attempt to get both on their feet again. In contrast to traditional functionalism, however, the theory does not imply that society as such possesses essential functions, but 'that social action has different consequences for different types of culture", and actually tries to identify them (Wildavsky 1984c, 30). The new functionalism is not at the level of the 'society', but concerns equilibrium and coherence at a "cultural" level, a level that opens up for more than one culture in any one society.

It is this congruence, or new functionalism, if one wishes, that makes this perspective different from the 'new institutionalism" in political science/ organizational theory (March & Olsen 1984, 1989). Both approaches underlinethe in predicting behavior in an institutional context from how people behave in a non-institutional context (or the other way around). Institutions reinforce behavioral patterns. The new cultural theory will furthermore follow the new institutionalism in its attack on reductionalism. pure instrumentalism and the belief in historical efficiency (traditional functionalism). However, the view of goals is different. Cultural theory does not de-emphasize the cognitive elements of politics and the role of

goals. The instrumental aspects of politics remain as important as before, but now in a wider and different context. Cultural theory, because of the typology, de-emphasizes the relativistic elements that are built into the 'new institutionalism', because what the typology is trying to grasp is what is possible and what is not. In this sense it is possible to view cultural theory as a promising typological version of the new institutionalism, rooted as it is in underscoring the importance of social organizations and institutions.

Within the perspective of 'bounded rationality' (Simon 1985), one is not able to specify how we would expect 'bounded rationality' to work in different contexts and under different circumstances. Wildavsky's perspective exactly about how we should expect 'bounded rationality' to work under different circumstances, depending on different types of cultures or regimes. Cultural theory, because of an explicit explanatory typology of different 'ways of life', serves to take the theory of 'bounded rationality' a long step forward. Closure is no longer mainly a question of psychological capacity, time or resources, but of what kinds of information, taken from which sources, are acceptable and satisfactory and which are not. The concept of 'cultural rationality' is giving us all our rationality back.

Conclusion

Any criterion will organize data - will order items in classes - but only some classification
will be scientifically useful. That is why we insist that typologies must be based on dimensions
that form categories, not on categories by themselves (Thompson et al. 1990, 261).

This article has questioned several of the most important assumptions upon which the new theory is based. However, even though we have concentrated on what are considered to be important logical and empirical problems, this does not mean that these problems are more devastating than alternative 'cultural' approaches. The new cultural theory is a general theory, useful for approaching any political or organizational problem where there are important differences in view, i.e. conflict. Adding functions to cultures, i.e. to types of social relations within a society, cultural functionalism explicitly addresses the question of conflict. And it is a genuine comparative approach, since no culture can be understood except in relation to other cultures.

Cultural theory is a theory of what goes with what, which argues that shared beliefs and values, though varied, are not free to float anywhere. They are closely tied to a social relation which they help sustain and render meaningful. Through a typology of 'cultural bias' the theory tries to find a meaningful medium between infinite diversity and total unity, i.e. constraineddiversity. is an absolute necessity if we really want to grasp cultural change. The perspective has a potential for producing much more

than interesting typologies supported by an illustrative use of empirical
data.

REFERENCES

Almond, G. 1980. 'The Intellectual History of the Civic Culture Concept", in Almond, G. &
Verba, S., eds., The Civic Culture Revisited. Boston: Little, Brovvn.

Almond, G. & Powell. G. B. 1978. Political Culture'. in Almond. G. & Powell. G. 8., eds.,
Comparaiive Politics. 2nd ed. Boston: Little. Brown.

Almond, G. & Verba, S. 1963. The Civic Culture. Princeton: Princeton University Press.

Barnes. J. 1971. 'The Righthand and Lefthand Kingdom of God", in Beidelman. T. O. ed..
The Translation of Culture. London: Tavistock Publications.

Barnes, J. 1978. "Neither Peasant norTownsmen: a Critique of a Segment of the Folk-Urban
Continuum", in Barth. F.. ed.. Scale and Social Organization. Oslo: Universitetsforlaget.

Bellah, R. N., ed., 1985. Habits of the Heart. Berkeley: University of California Press.

Burt. R. S. 1979. Towards a Structural Theory of Action. New York: Academic Press.

Douglas, M. 1982a. 'Cultural Bias'. ch. 9 in~Douglas. M.. ed.. //; the Active Voice. London:
Routledge & Kegan Paul.

Douglas, M. 1982b. 'Introduction to Grid/group Theory". in Douglas. M.. ed.. Essays in the
Sociology of Perception. London: Routledge & Kegan Paul.

Douglas. M. 1986. How Institutions Think. Syracuse: Syracuse University Press.

Douglas. M. & Wildavsky. A. 1982. Risk and Culture. Berkeley: University of California
Press.

Inglehart. R. 1977. The Silent Revolution. Princeton: Princeton University Press.

Kanter. R. M. 1972. Commitment and Community: Communes and Utopias in Sociological
Perspective. Cambridge: Harvard University Press.

March. J. G. 1978. 'Bounded Rationality. Ambiguity, and the Engineering of Choice'. Bt'll
Journal of Economics 9, 587-607.

March, J. G. & Olsen. J. P. 1984. The New Institutionalism: Organizational Factors in
Political Life". American Political Science Review 78. 734-749.

March, J. G. & Olsen, J. P. 1989. Rediscovering Institutions. The Organizational Basis of
Politics. New York: The Free Press.

Pfeffer, J. 1981. Power in Organizations. Marchfield: Pitman Publishing.

Richardson, S. A. 1956. 'Organizational Contrasts on British and American Ships'. Administrative
Quarterly 1. 189-207.

Robertson. D. 8.. ed.. 1966. Voluntarx Associations. A Studx of Groups in Free Societies.
Richmond: John Knox Press.

Simon. H. A. 1985. -Human Nature in Politics: The Dialogue of Psychology with Political
Science". American Political Science Review 79. 293-304.

Thompson, M. 1982a. 'The Problem of the Centre: An Autonomous Cosmology', in Douglas.
M., ed.. Essays in the Sociology of Perception. London: Routledge & Kegan Paul.

Thompson, M. 1982b. 'A Three-dimensional Model', in Douglas, M., ed., Essays in the
Sociology of Perception. London: Routledge & Kegan Paul.

Thompson, R., Ellis, R. & Wildavsky, A. 1990. Cultural Theory. Boulder: Westview Press.

Thompson, M. & Wildavsky. A. 1986. 'A Poverty of Distinction: From Economic
Homogeneity to Cultural Heterogeneity in the Ciassification of Poor People', Policy Sciences
19, 163-193.

Wellman, B. 1983. 'Network Analysis: Some Basic Principles", Sociological Theory 1983.
London: Jossey-Bass.

Wildavsky, A. 1982. 'The Three Cultures: Explaining Anomalies in the American Welfare
State', The Public Interest, no. 69.

Wildavsky. A. 1983. 'Models of Political Regimes or Pluralism means more than one Culture
in one Country at one Time". Berkeley: Survey Research Center. Mimeo.

Wildavsky, A. 1984a. 'Doing More and Using Less: Utilization of Research as a Result of
Regime", Berkeley: Survey Research Center. Mimeo.

Wildavsky, A. 1984b. The Nursing Father. Moses as a Political Leader. Alabama: University
of Alabama Press.

Wildavsky, A. 1984c. 'From Political Economy to Political Culture or Rational People Defend
Their Way of Life", Paper, American Political Science Association, Annual Meeting, 30
August-2 September.

Wildavsky, A. 1985. "The Logic of Public Sector Growth', in Lane, J. E., ed., State and
Market. London: Sage.

Wildavsky, A. 1987. 'Choosing Preferences by Constructing Institutions: A Cultural Theory
of Preference Formation", American Political Science Reuiew 81, 3-21.

Wuthnow, R., Wuthnow, R., Hunter, J. D., Bergeren, A. & Kurzwert, E. 1984. Cultural
Analysis. London: Routledge & Kegan Paul.