You are on page 1of 15

GeoJournal 47: 395–409, 1999.

© 1999 Kluwer Academic Publishers. Printed in the Netherlands.


395

Grid-group cultural theory: an introduction

Virginie Mamadouh
Amsterdam Study Centre for the Metropolitan Environment, University of Amsterdam, Nieuwe Prinsengracht 130, 1018 VZ
Amsterdam, The Netherlands
(E-mail: v.d.mamadouh@frw.uva.nl)

Received 24 June 1999; accepted in revised form 7 July 1999

Key words: cultural bias, cultural theory, grid-group analysis, grid-group cultural theory, myths of nature, theory of
sociocultural viability

Abstract
This article offers an introduction to grid-group cultural theory (also known as grid-group analysis, Cultural Theory or
theory of socio-cultural viability), an approach that has been developed over the past thirty years in the work of the British
anthropologists Mary Douglas and Michael Thompson, the American political scientist Aaron Wildavsky, and many others.
This assessment begins with a presentation of the main claims of the theory, distinguishing two characteristic breads of grid-
group cultural theory, in the one it is conceived as a heuristic device, in the other it is seen as a full explanatory theory. This
brief is followed by a discussion of the typology generated by the theory. This includes a presentation of the two dimensions
of sociality it posits, the cultural map they produce, as well as the four (or five) cultural types derived from them and their
designations. The article proceeds with a discussion of key issues including the incorporation of other typologies (such as
the one developed to analyse myths of nature), the relations between cultures or rationalities and several methodological
issues. Finally the article introduces the contributions to this special issue of the GeoJournal.

This special issue of GeoJournal is organised around a political violence and political cultures, and they cover a
theory. The theory in question is known under several names wide range of cases on three continents.
(see below), among which we have chosen here the label The purpose of this article is to introduce grid-group cul-
grid-group cultural theory. This diversity is symptomatic of tural theory. This introduction is meant as an assessment of
the variety of interpretations and applications that one en- the main features of the theory as well as the issues at stake
counters in the literature linked in one way or the other to when it comes to the coherence of the theory, its method-
this theory. ological implications, its relevance and its application to
It is not surprising that such variety is found, as the diverse fields of social science. As the grid-group map is
approach emerged thirty years ago (Douglas, 1970, 1973, the most famous exponent of theory it will be discussed at
1975, 1978). During three decades, insights evolve. Besides length. The article goes on with a short review of the main
the ongoing changes in mainstream social science theory and pitfalls regarding the application of the theory and it ends
methodology did also contribute to this diversification. But with a presentation of the contributions to this special issue.
diversity is mostly the result of the diverse backgrounds of
the many authors involved. It is perhaps odd that a theoret-
ical framework designed to deal with cultural diversity in What’s in a name?
remote places by an author interested in rituals, symbols,
witchcraft, food and drinking habits, can be relevant in so The theory is introduced here as the grid-group cultural the-
many fields of social science at home. But the challenging ory, but this choice needs to be justified, as it is not the
character of the theory has aroused people from various aca- most common usage. The theory was first referred to as grid-
demic disciplines and they have tried to apply the theory in a group analysis (Douglas, 1978; also in the volume of essays
broad range of research fields to generate insights about the edited by Mary Douglas in 1982; in Douglas, 1982a; Gross
most diverse topics. and Rayner, 1985) but this label has been abandoned as the
The purpose of this special issue is to show what the the- awareness grew that the approach was a theory. By the end of
ory has to offer to geographers. Although most contributors the eighties the term cultural theory had become established
are not geographers, they all deal with topics most rele- (especially Wildavsky, 1987; in Thompson et al., 1990; Dou-
vant to geographers and those interested in environmental glas, 1992) as a way to emphasise that the theory puts culture
sciences, such as water management, water pollution, post- at the centre of the explanation of social life. Nevertheless
industrialisation, international relations, urban democracy, the label is most confusing as this specific theory is not the
only cultural theory in town. Because of that, authors often
396

used capital letters to emphasise that they mean the Cultural cultures. She discusses specifically the problems of British
Theory. Of course, this in turn often irritates readers, es- anthropologists trained with theories based on field work in
pecially those interested in other cultural theories, as it is Africa who lack tools to analyse their findings during field
understood as an arrogant gesture. Neophytes are puzzled work in Melanesia. From the start she makes a few provoca-
and labelling loses its power.1 tive thoughts: we need a framework that is able to deal
Cultural theorists, also less respectfully called grid- with culture everywhere and we should start with cultural
groupies, have been discussing a possible way out of this diversity at home (in the western societies).
terminological fog, but it seems that they cannot get rid of Mary Douglas came up with a radical proposal: that of
the CT label. To this day, the most accurate description put identifying two basic dimensions of sociality, arguing that all
forward to clarify the content of the theory is the following: cultures can always be assessed and classified according to
the theory of sociocultural viability (Thompson et al., 1990) these two dimensions. Subsequently she takes four cosmo-
but this adequate denomination is so unattractive that the logical types and shows how they are linked to four extreme
authors of that volume propagate the use of Cultural Theory types of social context, as ‘for purposes of demonstrating the
as a shortcut – in fact this is also the plain title of their book. present argument, it will be clearest if only extreme cases
I have chosen to join an emerging convention (Ward, are considered’ (Douglas, 1978, p. 19). The typology was
1998; Hood, 1998) in which the two most used denomina- constructed.
tions grid-group analysis and Cultural Theory are merged The approach has been subsequently developed in many
into one. It emphasises that it is a cultural theory and at the essays by Mary Douglas, especially Risk Acceptability ac-
same time it makes clear which kind of cultural theory it cording to the Social Sciences (1985), How Institutions
is.2 However there is no reason to believe that this will the Think (1986), Risk and Blame (1992), Thought Styles
conclusive solution to the labelling problem and as we will (1996), as well as the very influential Risk and Culture in
see below the terminological fog is not limited to the name collaboration with Aaron Wildavksy (1982) and recently
of the theory. Missing Persons with Steven Ney (1998). Other core pub-
lications are by Steve Rayner, Michael Thompson, Aaron
Wildavsky and Richard Ellis such as Measuring Culture
The theory behind the typology (Gross and Rayner, 1985), Divided we stand (Schwarz and
Thompson, 1990), Cultural Theory (Thompson et al., 1990),
Grid-group cultural theory is often confused with the cul-
Culture Matters (Ellis and M. Thompson, 1997) and Human
tural map it has produced. The terms grid and group refer to
Choice and Climate Change (Rayner and Malone, 1998).
two dimensions of sociality which structure that map. Based
Several edited volumes have bundled applications of the
on these two dimensions, ideal typical social positions are
theory, especially Essays in the Sociology of Perception
constructed which are claimed to be able to account for cul-
(Douglas, 1982b); Politics, Policy and Culture (Coyle and
tural diversity in the most parsimonious way possible. The
Ellis, 1994), Kultur som Levemåte (Grendstad and Selle,
typology is catchy and has often attracted all the attention of
1996) and Cultural Theory as Political Science (Thompson
readers, at the expense of the theory on which it is grounded.
et al., 1999).4
In this section I would like to highlight the theory, the ty-
The main features of the theory, soft and hard, can be
pology and its many variants will be discussed in the next
summarised in a few claims.
one.
The main claim of grid-group cultural theory is that cul-
Grid-group cultural theory exists in two breeds, and all
ture matters. Preferences and justifications shape the world
intermediary shades. In the soft version, it is conceived as
of social relations. Everything human beings do or want is
a theoretical framework, a heuristic device, a classification
culturally biased. Therefore this is a cultural theory.
scheme. The hard variant is a full explanatory theory.
The second claim is that it is possible to distinguish a
The original presentation (grid-group analysis) belongs
limited number of cultural types. That can be done by con-
to the soft breed. The grid group approach has been devel-
structing a typology of cultures. This typology includes vi-
oped as a tool to deal with cultural diversity. In Cultural Bias
able combinations of patterns of social relations and patterns
published in 1978 (reprinted in Douglas, 1982a, pp. 183–
of cultural biases (or cosmologies). These combinations are
254)3 Mary Douglas addresses directly the failure of an-
often called (sub)cultures, ways of life or rationalities, some-
thropologists to theorise cultural diversity and to compare
times ways of organising, social orders, solidarities, political
1 For example, the expression is useless to search for relevant literature cultures, or simply types. Combinations are viable when so-
in a library catalogue.
2 Another alternative is the expression grid-group theory which has control others to the left, and from zero to increasingly controlled by other
been used by some authors in this volume. people’s pressure to the right (Douglas, 1973, p. 84). Nonetheless the basic
3 In my view, the earlier presentation of grid and group in Natural elements of grid-group analysis are there: the concordance between natural
Symbols (Douglas, 1970, 1973) was a very embryonic one. It is mostly symbols and social experience, and the idea of four distinctive systems of
focused on classifications and codes, and is still very close to the work of natural symbols.
Basil Bernstein (1971). In fact the diagram of grid and group presented 4 For introductions see a recent review article (Verweij, 1997), an in-
there (Douglas, 1973, p. 84) is reproduced in a subsequent publication as a troduction to a cultural theory of risk (Tansey and O’Riordan, 1999), to
scheme of classification and control, as adapted from Bernstein’s work (in Cultural Theory and international relations (Verweij, 1995; Ward, 1998)
Douglas, 1975b, p. 218). Grid runs from a private system of classification and to Cultural Theory and new institutionalism (Grendstad and Selle,
to a system of shared classification. Group stands for the relation between 1995). For a benevolent critique, see Selle (1991) and for a less benevolent
ego and the others: it runs from zero to increasingly exerting pressure that critique limited to risk analysis see Boholm (1996).
397

cial relations and cultural biases reinforce each other, that is: bunch of preferences) and the second that the two patterns
the cultural bias justifies the social relations which confirm reinforce each other. Second the impossibility theorem (p. 3,
the expectations raised by the cultural bias. In other words, also Wildavsky, 1987, p. 5) states that there are five and
when people behave rationally. The very fact that rationality only five7 ways of life (the ones deduced from the grid and
is plural and that several types of rationality can be estab- group dimension) that are viable combinations of bias and
lished is an important contribution of the theory to social relations. Third the requisite variety condition (p. 4) alleges
science in which the rational choice paradigm is hegemonic. that ways of life need each other to be viable: because of its
As such it can be seen either as an alternative to it or an specific blind spots, each cultural bias leads to catastrophe
expansion of it (see also Wildavsky, 1994). if it is not ‘corrected’ by the other (curvilinearity). In addi-
The viable combinations can be ‘discovered’ by assess- tion, each way of life needs its rivals to define itself against.
ing two main dimensions of sociality. We could use more, There is an interdependence, ‘conflict among cultures is a
but to be parsimonious we stick to two dimensions. It is cru- precondition of cultural identity’ (Wildavsky, 1987, p. 7).
cial to retain more than one dimension to allow for change in As a result, the authors assume the five ways of life to be
different directions. Any unidimensional typology suffers a present in any society at any time,8 be it in various strength
serious flaw: moving away from one pole means necessarily and in various patterns of interaction. The competition be-
moving towards the other pole.5 Instead of opposing individ- tween the ways of life is a state of constant disequilibrium
ualism and collectivism as two poles of the same dimension (Thompson et al, p. 4, pp. 83–100): adherents are constantly
(a common practice in social sciences), they are conceived moving from one way of life to the other (the authors use
as two dimensions: one of individuation and one of social the metaphor of a flock of starlings, p. 83). Finally, change
incorporation (Douglas, 1978, p. 7 or 1982a, p. 190). These is explained by the theory of surprise (pp. 69–81). Ways of
two dimensions of sociality have been named group and life are resistant to change, and events that do not fit the
grid. Group stands for incorporation into a bounded group expectations raised by a way of life, are explained away. But
(it is strong when the individual is a member of one corpo- the cumulative impact of successive anomalies or surprises
rate group, weak when individuals do not belong to such a (major, painful accidents) provoke a change of paradigm.
group) while grid stands for ‘the cross-hatch of rule to which The content of the theory will be clarified in the fol-
individuals are subject in the course of their interaction’ lowing by a discussion of its main components (the two
(Douglas, 1978, p. 8, or 1982a, p. 192). dimensions of sociality, the cultural map, the analytical lev-
The third claim is that the typology of viable combi- els and the cultural types), followed by a discussion of a few
nations is universal. It can be applied anywhere anytime of the main problematic or controversial issues.
because the two dimensions of sociality grasp the funda-
mental nature of the social being. Grid-group cultural theory
postulates that people derive a great many of their prefer- The two dimensions of sociality
ences, perceptions, opinions, values and norms from their
adherence to a certain way of organising social relations, These two dimensions of sociality have been named group
which is revealed by their preferences with regard to the two and grid. The exact nature of the two dimensions have been
basic dimensions of social life: incorporation or bounded- the topic of many discussions, especially related to method-
ness (group) and regulation or prescription (grid). In a way, ological consequences.9 The two dimensions should be seen
institutions do the thinking for us (Douglas, 1986). Grid- as polythetic scales (Douglas, 1978, p. 15; Hampton, 1982;
group analysts can therefore deduce preferences, attitudes Gross and Rayner, 1985, Chapter 4). That is: they include
and behaviours regarding all kind of topics for each ideal a series of aspects, but those are not necessarily present in
type.6 each case observed.
These three claims together are already quite a program, For grid (perceived as more difficult to grasp than the
but the hard version of the theory goes a few steps further. group dimension), Mary Douglas distinguishes in Cultural
It has been presented systematically by Michael Thompson, Bias four elements: insulation, autonomy, control and com-
Richard Ellis and Aaron Wildavsky in a book published in petition (Douglas, 1978, p. 16). In his study of work crime,
1990: Cultural Theory. They posit the following series of Gerald Mars develops a classification of occupations, he lists
propositions. First the compatibility condition (p. 2) asserts four elements to assess the strength of grid dimension: au-
that social relations (patterns of impersonal relations) and tonomy, insulation, reciprocity and competition; whereas the
cultural bias (shared values and beliefs) can not be combined four elements of group strength include: frequency (of inter-
contrary to each other: they must be mutually supportive. It action), mutuality, scope (of activities) and (group) boundary
implies both consistence and coherence: the first refers to (Mars, 1982, pp. 24–28).
the fact that both are patterned (a cultural bias is a consistent 7 Four or five is however a matter of discussion, more later on.
8 If the ways of life are conceived as preferences (as in Verweij, 1995)
5 On one and two-dimensional models, see also Ostrander (1982).
it is much easy to accept this condition than if ways of life are seen as
6 In Cultural Bias Mary Douglas (1978) works out views on nature,
directly determined by social environments. In that case, the requisite va-
time, human nature, and social behaviour. In Divided we stand, there are riety condition can only expected to be met in complex organisations and
many more: preferred way of organising, view of resources, scope of knowl- societies.
edge, learning style, ideal scale, idea of fairness, preferred economic theory, 9 Are the dimensions independent from each other? Do they intersect
perception of time, preferred form of governance, model of consent, etc. at right angles? These are important matters for those who want to test the
(Schwarz and Thompson, 1990, pp. 66–67). theory by statistical means.
398

A most detailed discussion of the dimensions can be


found in Measuring Culture in which Steve Rayner and
Jonathan Gross present a mathematical model for grid-group
analysis (Gross and Rayner, 1985). Their book is an at-
tempt to formalise and operationalise the two dimensions
and it includes a computer program to ‘calculate the val-
ues of grid/group predicates from experimental data’ (Gross
and Rayner, 1985, p. 117ff). Their model is however not
meant as the ultimate selection and weighting of predicates,
because the relevance of each predicate depends on context.
Gross and Rayner include five group predicates in their
Figure 1. The cultural map: two dimensions, four types (Douglas, 1978,
model (1985, pp. 73ff): proximity (closeness of members to p. 7).
each other), transitivity (of relations), frequency (proportion
of the allocable time of members during which they inter-
act with other members of the unit), scope (activities with bother to talk about the dimensions and just present the re-
members as proportion of all activities inside and outside the sulting types, what Michael Thompson provocatively calls
relevant social unit), and impermeability (which measures Cultural Theory without grid and group. But before turning
how easily eligible non-members who want to join actually to the cultural types themselves, I would like to discuss the
join the unit). They also mention other predicates such as a cultural map and its graphic representation.
commonality of experience and life-support dependency.
Likewise, they have four grid predicates (pp. 79ff): spe-
cialisation (operationalised as the proportion of all possible The typology portrayed: the cultural map
roles that a member assumes), asymmetry (lack of symmetry
in role exchange), entitlement (operationalised as the pro- Mary Douglas illustrates her essay Cultural Bias with a
portion of all roles that are ascribed to certain categories of diagram on which grid and group are portrayed as two di-
persons, as opposed to roles open to all that can be attained mensions generating four types of social environment. The
by achievement), and accountability (operationalised as the visual presentation of the typology has been very success-
proportion of all roles that are accountability interactions, ful and is found in many presentations of the theory (but
which are interactions in which there is one dominant role not all, especially not in studies focused on one or two cul-
and one subordinate role, the former can put sanctions on tural types). Most striking for someone who glances through
the later). The final step in the model is to aggregate the the literature is the capricious shape of the diagram. The
predicate scores by (weighted) average. main variants will be discussed here because they give away
In Measuring Culture, the model is applied to a fictitious some of the problematic aspects of the theory and its many
case-study about the location of a nuclear power plant in an varieties. A recurrent issue concerns the nature of the di-
imaginary small New England coastal town and this quan- mensions: are they continuous or dichotomous? Another
titative approach is contrasted with a qualitative analysis of discussion is about the number of positions: are there four
the same (fictitious) case. The model developed by Gross or five types to be taken into account? Both questions are re-
and Rayner did not come into use because of the obvious lated to each other as the fifth position is generated by certain
difficulty to gather the kind of detailed data needed to feed conceptualisations of the dimensions and not by others.
their computer program. The study remains however rec- As can be seen in Figure 1 (based on Douglas, 1978,
ommended reading to grasp the complexity of the grid and p. 7), the diagram originally published in Cultural Bias is
group dimensions and as such it may help understand why a map on which specific positions are located. Each quad-
presentations are so diverse in different publications. rant of the diagram is characterised by its score on the two
The two dimensions are generally acknowledged as such, dimensions. The diagram has a major flaw: the dimensions
but their presentations may vary. Aaron Wildavksy for ex- are continuous but the boxes that materialise the four types
ample generally introduces the two dimensions as grasping suggest discrete positions. It is not clear what happens when
the answer to two fundamental questions for each human one moves across the map. There are several ways to deal
being: (1) who am I? and (2) how should I behave? (for with this problem.
example, Wildavsky, 1987, p. 6). Many authors focus on One way has been to drop the boxes and represent only
specific grid and group features relevant to their topics of the two continuous dimensions as in Figure 2 (based on Dou-
investigation, therefore they select one specific predicate of glas, 1992, p. 178). The four types are generally pinpointed
grid and group: for example fettered/unfettered competition on the map,10 but sometimes the map is left empty (Douglas
(instead of group) and symmetrical/asymmetrical transac- and Isherwood, 1979, p. 23). In these cases ‘any conceivable
tions (instead of grid) (Thompson, 1997a, p. 142, and in combination of values of grid and group corresponds to a
this volume), group strength and number and variety of pre- point in a one-by-one square diagram’ (Gross and Rayner,
scriptions (Webber and Wildavsky, 1986, p. 25) or preferred 1985, p. ix). Such mapping can be found in a comparison
amount of group loyalty and preferred amount of prescrip- 10 Douglas, 1992b, p. 473; Ellis and Coyle, 1994, p. 4; Coyle 1994,
tions (Verweij, 1995, p. 92). Sometimes, authors do not p. 221; Verweij, 1995, p. 92.
399

Figure 2. Two continuous dimensions but no boxes (Douglas, 1992,


p. 178). Figure 3. Types as extreme regions of the continuous grid/group diagram
(Gross and Rayner, 1985, p. 113).

Figure 4. Two dichotomous dimensions, four boxes (Douglas, 1982, p. 4). Figure 5. Four boxes, no dimensions (Douglas, 1993, p. 45).

between Dinka, Nuer and Mandari by Mary Douglas (1973, Thompson et al, 1990, pp. 16–17, n. 23), which intersects
p. 129) or a comparison between eight social categories in with the other two at that same zero-point. The regular cul-
Boyle and Coughlin (1994, p. 204). tural map reveals only the bird’s-eye view. Thompson’s three
An alternative is to emphasise that the types are extreme dimensional model is portrayed in Figure 8 in three slices
cases, pushing them to the edges of the map as in Figure 3 from left to right: two positions score negatively on the ma-
(based on Gross and Rayner, 1985, p. 113), a presentation nipulation dimension; two positions score positively, and the
much used by Mary Douglas in the essays collected in Risk fifth type scores null: this individual does not manipulate
and Blame (1992, p. 106, 205, 264, 287; also in Boyle and and is not manipulated by others. Thompson’s demonstra-
Coughlin, 1994, p. 195). tion uses the properties of a morphogenetic field which has
A third way to deal with the problem is to represent just five flat areas which correspond to the five stabilisable
the two dimensions as dichotomous ones, that is not run- combinations (Thompson et al., 1990, p. 17; Thompson,
ning from 0 to infinite, but distinguishing between weak 1982, pp. 31–63).
and strong, low and high. Such dichotomous dimensions The third dimension is concomitant of the other two: no
produce four boxes, as in Figure 4.11 At times the dimen- new combinations emerge. Therefore it can be ignored in a
sions are not even mentioned, and the drawing only portrays parsimonious model, even if it is useful to comprehend the
four boxes as in Figure 512 while the map may even been fifth position. Related cultural maps show the grid and group
replaced by a table as in Figure 6.13 dimensions as crossing each other at the centre of the map.
Finally another option has been developed by Michael They contain four or five positions, according to whether or
Thompson, who moves the zero-point of both dimensions to
the centre of the map, where a fifth type emerges (Figure 7,
Thompson 1982, p. 36). To clarify what this fifth position
is, he also argues for the existence of a third dimension
called manipulation, or grip to complete the alliteration (see
11 Figure 4 is based on Douglas, 1982c, p. 4; see also in Bloor, 1982;
Ostrander, 1982; Mars, 1982, p. 29; Gross and Rayner, 1985, p. 7; Rayner,
1984, p. 162; Wildavsky, 1986, p. 335; Webber and Wildavsky, 1986, p. 25;
Coughlin and Lockhart, 1998, p. 36; Engbersen, 1993, p. 36; Ward, 1998,
p. 215; Tansey and O’Riordan, 1999, p. 78.
12 Figure 5 is based on Douglas, 1993; p. 45; see also: Douglas, 1996,
p. 43; Douglas and Ney, 1998, p. 101.
13 Figure 6 is based on Ellis, 1991, p. 311; see also Wildavsky, 1984,
p. 193, 1987, p. 6; Selle 1992, p. 102; Grendstad and Selle, 1991, p. 11,
Grendstad, 1995, p. 224; Hood, 1998, p. 9; and most presentations in this Figure 6. A table, instead of a map (Ellis, 1991, p. 311).
volume.
400

terpersonal relations, (2) cultural biases, and (3) behavioural


strategies. It underlines the freedom of each individual to
choose a strategic behaviour that fits the social environment,
the cultural bias or both, or to choose a strategic behaviour
that disrupts the social environment, discredits the cultural
bias or do both.
Although the cultural map portrays dimensions of the so-
cial environment, descriptions of the cultural types generally
blur the analytical distinction between patterns of interper-
sonal relations, cultural biases and behavioural strategies.
This is also reflected in the labels used to name the types
(more later on this particular problem).
Figure 7. A fifth type in the centre (Thompson, 1982, p. 36).
The type located in the lower left quadrant (in Figures 1
to 10) is characterised by weak group incorporation and
not they depict the fifth type, as in Figure 9, respectively weak regulation or role prescriptions. The individual is free
Figure 10 (as in Thompson and Wildavsky, 1986, p. 176; to enter transactions with others as (s)he wishes, as on a
Thompson et al., 1990, p. 8; Coyle, 1994, p. 223; Dake market. Boundaries are provisional. They are subject to ne-
and Thompson, in this volume). When the fifth position is gotiation. Individuals are relatively free of control by others
not shown (Schwarz and Thompson, 1990, p. 7; Thomp- but their ability to control others is a measure of their posi-
son 1997a, p. 142), the presentations are quite similar to tion in the network. This pattern is justified by the pursuit
those obtained with two dichotomous dimensions, especially of personal rewards in a competitive environment. Fairness
when the name of the dimensions are not mentioned, as in consists of equality of opportunity. Blame is put on personal
Figure 5 (for example Davy, 1997, p. 327). failure (or lack of competition). This way of organising is
Sometimes graphic representations are confusing be- vulnerable to the lack of co-operation.
cause they include contradictory elements, for example in The upper right quadrant is a social environment char-
Webber and Wildavsky (1986, p. 25; or Wildavsky, 1986, acterised by strong group boundaries and binding prescrip-
p. 335) where the dichotomous dimensions are represented tions. These prescriptions are justified by the importance of
by arrows crossing at the centre of the scheme. And at times the whole over the parts, the collective over the individuals.
things are really mixed up, such as the figure in which Unni Consequently division of labour, differentiated roles, hier-
Edvardsen (1997, p. 218) portrays both two continuous grid archical social relations are typical of these nested groups.
and group dimensions and two dimensions crossing at the Fairness consists of equality before the law. Blame is put
centre of the map or the one by Michael Shackleton (1993, on deviants who do not endorse the established procedures.
p. 581) who uses a table but still puts the fifth position in the This way of organising is vulnerable to misplaced trust in
middle. authority and expertise.
The graphic presentation of the typology is not only a The upper left quadrant is characterised by binding pre-
means to attract attention that as such should deserve care of scriptions in combination with weak group incorporation.
the authors. It also has methodological implications. Those Isolation can be imposed or by choice (a possibility not al-
who choose continuous grid and group dimensions are likely ways acknowledged). These individuals have little to say
to develop predicates for grid and group to measure the about the ways they live their life, it is organised from the
grid and group scores of a (series of) case(s). Those who outside. They live at the margins of organised patterns dis-
choose dichotomous dimensions, regardless of where they cussed above such as hierarchies or networks. Fairness is not
posit their intersection, are likely to focus on the cultural on this earth. Blame is put on fate (bad luck). This way of
types themselves and to select indicators for each of the four organising is vulnerable to unwillingness to plan ahead and
(or five) types. The first group may conclude that X is ‘up inertia.
grid’ and ‘down group’ as compared to an ideal-typical po- The lower right quadrant is characterised by strong group
sition D, whereas the second group are likely to speak of boundaries coupled with few regulations. The group is main-
mixed types or hybrids between two cultural types. As the tained through intensive relations between group members.
second option is more popular than the first – contributions Internal role differentiation is minimal. Fairness is equality
to this issue of the journal are no exception – it is now time of result. Shared opposition to the outside world keeps such
to present the cultural types. an egalitarian bounded group together. Blame is put on the
‘system’. This way of organising is vulnerable to deadlocks
because no authority is accepted to resolve internal conflicts.
The cultural types Finally, the fifth type is (when acknowledged) one of
withdrawal from coercive and manipulative relations: this
Cultural types (ways of life, cultures, rationalities or soli- individual is not controlled by others and is not controlling
darities) have been defined as viable combinations of social others (see Figures 8 and 10). In other presentations, it is
patterns and cultural patterns, but it is now common to conceived as an extreme form of the low grid positions, not
distinguish three levels of analysis instead of two: (1) in-
401

Figure 8. Michael Thompson’s three dimensional model (Thompson, 1982a).

Figure 9. A map with the zero point in the centre: Two dimensions of Figure 10. A map with the zero point in the centre: The five vignettes
sociality and four rationalities (Schwarz and Thompson, 1990, p. 7). mapped onto the dimensions of sociality (Thompson et al., 1990, p. 8).

a specific type. Mary Douglas did discuss autonomy but as hierarchist(s), hierarchy, hierarchies or hierarchical and
conceived it as ‘off the map’ (1978, p. 18). closely related terms as (hierarchical) collectivism.
For the other two types, divergences are more conse-
quential because the alternative terms have different conno-
The importance of labels: catchy or foggy? tations. At present, the most common label for the B position
is fatalism, fatalist(s) or fatalistic15 which seems to be a
In the original diagram Mary Douglas uses four letters: A, B,
Wildasvkian idea. Mary Douglas uses most often the term
C and D to describe the cultural types, starting in the lower
isolate(s), insulated (1978, p. 7) or atomised subordination.
left quadrant (low grid, low group) and moving clockwise
Still she uses fatalists once (Douglas, 1992, p. 264). Posi-
through the diagram (see Figure 1). From the start, the four
tion D is often referred to as egalitarianism or egalitarian(s)
letters were coupled with labels that became more popular
which seems to originate from Steve Rayner’s (1979, 1982,
than the letters, because it is for us easier to use them than to
1984, 1988, 1992) work on egalitarian groups. By con-
remember which was A, which was B, etc.14 The labelling
trast, Mary Douglas uses factionalism, sect/sectarian, and
unfortunately also produces a lot of confusion because labels
enclave/enclavist, but also dissenting groups, communard(s)
stimulate our imagination so readers tend to forget about grid
(Douglas, 1992, p. 264; also in Davy, 1997; Thompson et al.,
and group and fill up the quadrant with the connotations of
1990) and she uses once egalitarian enclaves (Douglas and
the labels, but also because of the inconsistencies in label use
Ney, 1998, p. 101).
(as can be seen in Figures 1 to 10). In addition, the same la-
The fifth type, whose existence and position on the
bels can be used with different diagrams, and different labels
map are much disputed, is labelled either the hermit or au-
with the same diagram.
tonomous/autonomy/autonomist or both16 although it should
There is a fair degree of consensus for the A and C posi-
be noted that Mary Douglas discusses this position in a chap-
tions, the two manipulative cultures (see Figure 8). It is no
ter called individualism in Cultural Bias (1978, p. 41ff.)!
coincidence, these are the two types widely acknowledged
This enumeration (although limited to the core publica-
in social science, most commonly as the opposite poles
tions of the key authors Douglas, Wildavksy and Thompson)
of one continuum: individualism and collectivism. Position
establishes a relative lack of standardisation. It is messy and
A is often referred to as individualist(s), individualism or
it may confuse some readers, but does it really matter? I
individualistic, or slightly more remote terms as competi-
tion, entrepreneurs, markets. Position C is often referred to 15 Schwarz and Thompson, 1990; Thompson et al., 1990, 1992; Ellis
and M. Thompson, 1997; Ellis and Coyle, 1994, p. 4; Davy, 1997, etc.
14 Once Mary Douglas tried to match labels and letters using: Active 16 Thompson, 1982; Thompson and Wildavsky, 1986; Thompson et al.,
individualism, Backwater isolation, Conservative hierarchy and Dissident 1990; Coyle, 1994, p. 223.
enclave (1996, p. 43).
402

think so. Labels matter because they transmit different con-


notations. As such they greatly determine how the results of
such analysis are perceived. It may be more acceptable to be
labelled an isolate than a fatalist, or an egalitarian than a sec-
tarian. It is (by definition) impossible to find unbiased labels.
Different labels convey different interpretations of the the-
ory, so discussing them is important, even if the contributors
to this volume spontaneously follow the most conventional
use: individualism, fatalism, hierarchy, egalitarianism.
The labels used for the cultural types do implicitly em- Figure 11. Myths of nature (Schwarz and Thompson, 1990, p. 5; Thomp-
phasise one analytical level rather than the other. Terms as sonet al., 1990, p. 27).
isolates, networks, hierarchies, clans, sects, point at pat-
terns of social relations, e.g. social environments. Terms
as individualism, fatalism, collectivism, egalitarianism, sec- forests, fisheries and grasslands.18 Michael Thompson and
tarianism, point at patterns of beliefs, e.g., cultural bias. his co-authors (Schwarz and Thompson, 1990; Thompson
Finally terms as individualist, fatalist, hierarchist, enclavist et al., 1990; see also Rayner and Thompson, 1998) have
point at persons holding certain views and choosing certain underlined the close links between each myth of nature and
strategies, the same applies to the series of attitudes towards the cultural types. These four myths of nature are nature
poverty (the entrepreneur, the ineffectual, the hierarchist, the benign (corresponding to the A position), and following
sectist and the autonomist) established by Thompson and clockwise nature capricious, nature perverse/tolerant and
Wildavsky (1986, p. 176). Other classifications are neutral, nature ephemeral (with nature resilient for the hermit in
the original A, B, C and D (Douglas, 1978, 1982a; Gross Thompson et al., 1990, p. 26).
and Rayner, 1985), and yet others explicitly metaphorical, Each myth of nature offers a partial representation of
such as the animal typology presented by Gerald Mars in his reality. It captures some experience, as such it is true. But
classification of occupations (hawks, donkeys, wolves and it can be ‘surprised’ by events, as such it is wrong. The
vultures, Mars, 1982) and as such they can easily address all myths are represented graphically by a ball in a landscape,
levels at the same time, but they remain biased and as such the shape of the landscape revealing the expected interaction
immediately repulsive or attractive to each of us. between ball and landscape (between life and the world) as
The very fact that some authors use labels pertaining in Figure 11.19 When nature is benign, it is robust, it can
to different levels in the same account of the typology17 take experimentation, trial and error, as the shape of the
reveals some confusion between these levels of analysis. In- hill always brings the ball back to the best position. This
deed the distinction is not so important if one believes that myth corresponds to the individualist viewpoint. When na-
only the matching combinations are to be found empirically. ture is capricious, it is a lottery, one does not know what
Still this fogginess weakens the presentation, application and to expect, one can not learn from experience. This myth
validation of the theory. is compatible with the fatalistic viewpoint. When nature is
perverse/tolerant it is robust, but to a certain point, the shape
of the hill brings the ball back to the best position, at least
An imperialist theory? if it remains in the safe zone, if it goes too far, things go
wrong. Therefore this myth justifies the power given to ex-
A dangerous temptation for those convinced of the potentials perts as they are the ones who can evaluate the safety zone,
of the grid-group typology is to incorporate all other 4-types it corresponds to the hierarchic viewpoint. When nature is
typologies they came across. A re-interpretation of existing ephemeral, it is fragile and the ‘ball’ should be as quiet as
schemes is often seen as just another proof that the theory possible, any movement could lead to a fall from the top
is a hit, but it generally irritates those not impressed in the of the hill. That is the egalitarian position. Finally nature re-
first place, confirming in their view that you don’t need grid silient, the myth of nature corresponding to the fifth position,
and group to do the job. Anyhow there is a real danger of autonomy, is that the four myths are true at the same time,
sterile re-labelling of the old, especially when the theory is it is a myth too, because the myths are contradictory, nature
used as a classifying device and not as an explanatory the- can not be both at the same time regarding the same object
ory. Still, incorporation may be useful to specify the content and the same conditions (Thompson et al., 1990, p. 29–33).
of the types in a specific context. Empirical typologies and Human nature is constructed accordingly, Thompson
classifications and a theoretical typology can strengthen each et al. (1990, p. 33–37) argue. For individualists, human
other: the first validating the later, the second grounding the nature is extraordinarily stable and self-seeking. For fatal-
former in an explanatory framework. ists, human nature is unpredictable, for hierarchists humans
One of the most successful cases of incorporation has
18 Especially Holling 1986 (Schwarz and Thompson, 1990, p. 4;
been the typology of myths of nature found by ecolo-
Tansey and O’Riordan, 1999, p. 80).
gists who studied institutions managing ecosystems such as 19 It is worth noting that here too, diagrams prompt discussion. In this

17 For example: individualist and isolated in the original presentation case, it is mostly the description of the C position (nature perverse/tolerant)
that is difficult, as the exact drawing may suggest a varied range of
(Douglas, 1978, p. 7), individualism and egalitarianism but isolates and
robustness within limits (putting it closer to A or closer to D).
hierarchy (Tansey and O’Riordan, 1999, p. 78).
403

are ‘born sinful but can be redeemed by good institutions’ alliance between two cultures. His four hybrid regimes are
(p. 35) whereas egalitarians see exact the opposite ‘man much entrenched in history (labels again!) as he calls them
is basically good but his nature is highly susceptible to social democracy (C + D), American individualism (D + A)
institutional influences’ (p. 36). state capitalism (A + B) and totalitarianism (B + C) (for
The two typologies fit so well into each other that the example in Webber and Wildavsky, 1986, p. 25; Wildavsky,
myths of nature have become an integral part of grid- 1986, p. 335). Surprisingly he neglected the diagonals as po-
group cultural theory. They have been often mentioned in tential regimes (although ‘the establishment’ is mentioned as
studies dealing with perceptions and views on the environ- a ninth regime in Wildavsky, 1986, p. 335). Discussing Cul-
ment (e.g., Thompson, 1997a; Selle and Grendstad, 1997; tural Theory and its shortcomings, Per Selle introduced the
Davy, 1997). Mary Douglas (1992, pp. 263–64) has also hermit as the ninth regime (Selle, 1991, p. 102) but dropped
adopted them also adapting them by slightly altering the it again a few years later (Grendstad and Selle, 1995, p. 11).
labels and applying them to human nature (1996, pp. 88– Virginie Mamadouh pursues a similar investigation by ex-
89, also in Douglas and Ney, 1998, p. 109): nature/person ploring other ideal types of conversation between cultures in
robust, nature/person unpredictable, nature/person robust national political cultures (Mamadouh, 1997, 1999). Con-
within limits, nature/person under duress20 (after a try as sidering all these patterns of interaction between cultures,
nature fragile). the comparison with chemistry has been made by Michael
Another example of a happy fit is that of Ouchi’s typol- Thompson: the grid-group typology as the table of basic
ogy of organisations (1980): markets, hierachies and clans, elements.
corresponding to the organised patterns of social relations When dealing with relations between cultures, it is im-
(A, C and D). Schmutzer (1994) uses Ouchi’s terminology portant to distinguish between active and passive cultures.
and re-constructs the fourth type as the zero-matrix (the A, C and D (individualism, hierarchism and egalitarianism)
isolates). are the active cultures, because their adherents propagate
their way of life: they tend to recruit new adherents and to
impose their cultures on others because it is the right way to
Relations between cultures do things. By contrast, isolation and autonomy are passive
cultures.
One of the postulates of the hard version of grid-group cul- Another fundamental point is that synthesis is not achiev-
tural theory is the requisite variety condition which posits able. The cultures are antagonist, they are inherently adver-
the interdependence between cultural types. As soon as sarial (Douglas and Ney, 1998, p. 104–106). It is not possible
this condition is stated, the relations between the differ- to merge them and to obtain the best of them all at once. It is
ent cultures or rationalities emerge as a topic of major however postulated by many Cultural Theorists that cultural
importance. plurality is an asset: a society or an organisation in which
At first sight, one may think that cultures sharing the the different cultures are present and acknowledged will be
same score on one dimension may have some mutual under- less prone to surprise, it has a broad repertoire of strategies to
standing: e.g., the individualist and the fatalist, the fatalist deal with a broad range of problems, and it is better equipped
and the hierarchist, the hierarchist and the sectarian, the to learn. Grid-group cultural theory in its strongest form is
sectarian and the individualist. Instead, cultural theorists un- thus a normative theory.21
derline the connivance between positions situated along the
same diagonal: one is called the positive diagonal (A–C),
the other (B–D) the negative diagonal. The diagonals link Applying the theory
cultural types that share the same position on the third di-
mension, manipulation or grip (see above): it is why one The above presentation of the theory has over-emphasised
speaks of the diagonal of affirmation and the diagonal of variety, but at this point the reader hopefully has a good
withdrawal (Douglas, 1992, p. 45). The positive diagonal is idea of the different breeds of grid-group cultural theory.
also called the establishment (Wildavsky, 1986; Webber and Before introducing the contributions to this issue, the main
Wildavsky, 1986; Thompson et al., 1990). Another metaphor discussions regarding the application, the development and
to highlight the diagonals is that of the centre (A + C) versus the validation of the theory will be presented in brief.
the border (B + D) used by Douglas and Wildavsky (1982) The first issue of importance is that of the distinction be-
in Risk and Culture. tween levels of analysis. The distinction of the (two or) three
Dealing with larger social environments (e.g. states) analytical levels (relations, biases, behaviours) have been
Aaron Wildavsky was interested in coalitions or alliances surprisingly poor in most grid-group cultural theory writ-
between culture, or regimes. He investigated systematically ings, probably because the cultural types over-emphasise
regimes, assessing four pure regimes (regimes dominated by the coherent combinations. The fact that the three levels in-
one culture) and four hybrid regimes characterised by the fluence each other in a double cycle does not mean that a
20 The application to persons is visualised by drawings by Pat Novy 21 On cultural pluralism as a definition of democracy see Wildavsky
in which the ‘ball’ is replaced by a person (Douglas, 1996, p. 89; Douglas and Thompson (1987, 1994), Edvardsen (1997), Molenaers and Thompson
and Ney, 1998, p. 109). Strangely enough, the person under duress does not (1999), Ney and Thompson (1999), on pluralism and learning see Hendriks
stand in a precarious equilibrium on the top of a hill, but is uncomfortably (1994, 1996, this volume).
imprisoned under a bell jar.
404

particular application of grid-group cultural theory can not But grid-group cultural theory has been applied at very
focus on a specific analytical level or a specific stage of different scales. In fact the ideal types of social environment
the cycle: for example on the question how cultural bias (networks, isolates, hierarchies and clans) can be regarded
affects strategic behaviour related to making ends meet (see as ideal types of relations between elements in any system:
Thompson et al., 1990). Quite the reverse is true. A careful these relations can be symmetrical or not, fettered or not. As
clarification of which level is put at the core of a specific such they can thus be applied to any system: with elements
investigation is necessary to make choices in the interpre- much larger or much smaller than human beings. For exam-
tation and the application of the theory intelligible. And it ple it could be applied to the individual person as a set of
may even be indispensable to avoid circular arguments in an roles, comparing relations between the social roles fulfilled
analysis. by a person to the ideal typical patterns. Or it could be ap-
The issue of coherence has been raised as the most im- plied to the international system as a set of states. So we
portant issue over the past ten years. Whether or not there have to conclude that the typology is applicable to a much
is coherence between the three analytical levels (that is broader range of phenomena than the theory itself.
when relations, bias, and behaviour reinforce each other) Michael Thompson (1997b) argues that the cultural types
is a question for empirical research. Inconsistent combina- can be operationalised at any scale, from village to world, no
tions may be viable after all, in the sense that they occur doubt a challenge for geographers. A complicating effect of
in the real world, even if they are probably not enviable scale is that scale can be different for the three analytical
because they are not comfortable. Such inconsistent com- levels. This happens for example when biases concerning
binations may be temporary, as predicted by the theory, in individual behaviours and patterns of interactions between
the sense that they lead to change in relational patterns, bias individuals are studied at the collective level. In that process
or behaviour to meet the hypothesised coherence, but at any of moving from individual to collective voices, the zoom
moment, incoherent combinations could still be more com- effect also occurs: an organisation can be assessed to be
monly observable than the coherent combinations posited as voicing a cultural bias in interactions with other organi-
ideal types by the theory, a possible finding that would be sations, but when we zoom in and study the organisation
consistent with the notion of a constant disequilibrium (see itself as an arena, nuances appear and different biases are
above). present inside the organisation, probably voiced by different
Thompson, Ellis and Wildavsky (1990, 1992) discuss at departments. And when we zoom again on a certain depart-
length so-called ‘stolen rhetoric’: the justification of certain ment, we’ll find the four biases again, inside that department,
behaviour by a rhetoric corresponding to another cultural voiced by different persons. This flexibility can be seen as an
bias. In addition we could speak of stolen strategy when asset for the theory, but also as a liability, because a theory
means corresponding to one cultural bias are used to achieve too widely applicable is often felt to have weak predictive
aims belonging to another bias. In both cases, stolen ele- power.
ments are dangerous as they transform the original social The third issue of prime importance is that of individu-
environment towards the phantom bias. Moreover, surprise als as carriers of cultural bias. Mary Douglas has proposed
(meaning: discrepancies between expected and observed re- a rigid interpretation in which cultural bias is a permanent
ality) is important to explain change. Therefore it becomes characteristic of a person: it may change over time (she often
more important to distinguish the different analytical levels discuss examples referring to changes coupled with transi-
and to put consistency (of the cultural bias) and coherence tion from one life phase to the other) but it is hegemonic, it
(across levels) to the test. It could well be that the cultural applies to all domains in life. The opposed viewpoint, voiced
types are not only extreme positions on the cultural map (see by Michael Thompson and Aaron Wildavsky is that indi-
Figure 3) but also extreme cases of coherence. viduals may adhere to different biases in different contexts,
A second issue of importance is the scale at which the the most obvious example being home and workplace. This
theory is meant to be applied. Originally Mary Douglas implies a lack of coherence of the individual in her/his daily
assumed the proper scale to be that of the relevant social routines and it is a major flaw for the utility of the theory.
environment, that is ‘the social accounting level, the level If we believe the rigid version, it makes sense to investigate
of justification and explanation’ (Douglas, 1978, p. 15). It is the cultural biases of individuals because it will predict their
often concretised as the scale of face to face contacts. The attitudes and their behaviours in a wide range of environ-
units of analysis should be the individuals in their social ments, paradoxically it weakens the explanatory power of
environments. From a methodological point of view, it is the theory because it still has to explain which environment
however difficult to determine the extent of the social unit to determined the hegemonic bias and why. If we believe the
be studied empirically (Gross and Rayner, 1985, p. 62–66). flexible version, cultural bias can only be audited for a par-
Of course, the problem is more acute in a low group context ticular context at a particular moment, the cultural bias of
than for strong groups with, by definition, strong (external) individuals is so volatile that it does not make sense to map
boundaries. Delimitation depends in the end on the research it.
agenda.22 Eero Olli (1995, 1999) has done pioneering work on
investigating the validity of the rigid and the flexible assump-
22 The delimitation of the justification context is not easy, an individual
may justify the same behaviour quite differently for different audiences, even referring to different biases, which brings us back to the issue of
coherence.
405

tions. Using Norwegian survey data, he measures coherence to keep culture/preferences/attitudes out of the picture and
at the individual level by allowing for adherence to differ- those who want to study culture but despise typologies as
ent biases at the same time and by investigating rejection23 (too) reductionist.
systematically. It means that he differentiates between ‘ad- Claiming that grid-group cultural theory is useful as a
herence to A and indifference to B’ against both ‘adherence tool for geographers and other social scientists does not
to A and rejection of B’, and ‘adherence to A and B’. He mean neglecting its pitfalls and problems. They are many,
posits three types of individuals: coherent, sequential and because both theory and typology are so malleable. Such an
synthetic. Coherent individuals support one cultural bias and observation invites authors to spell things out and specify
reject others, sequential individuals support one cultural bias how they use the theory. Inevitably, the lack of space gen-
but are open to the others (they may adhere to another bias erally prompts them to present only that inch of grid-group
in a different context) and synthetic individuals support two cultural theory they need for their specific analysis. This may
or more biases at the same time. Unfortunately, this exciting impede the further and coherent development of the theory,
and highly sophisticated application is bringing the typology but it does not make it less challenging.
away from the parsimonious model it promises to be.
Much effort has been put in establishing the theory in so-
cial and political sciences and making it testable according to An introduction to this special issue
the conventional standards. One of the major developments
has been the elaboration of survey measures to tap the grid Mary Douglas presents a short historical account of the evo-
and group dimensions or directly the cultural types through lution of the theory as a parsimonious model, a view from
regular survey questionnaires, an adventure in which two inside. She justifies the selection of four cultural types, deals
contributors of this volume, the late Karl Dake and Gun- with the original criticisms, and the responses to them. The
nar Grendstad, have done pioneering work (Dake, 1990, emphasis on personal autonomy in adopting a culture was
1991, 1992; Dake and Wildavsky, 1990; Dake and Thomp- meant to meet the criticism that the model was determin-
son, 1993; Grendstad, 1990, 1995; Grendstad and Selle, istic. Much attention was given to change in an attempt to
1997, 1999) the first by developing new survey questions, dynamise a model perceived as static. Finally she discuss
the second by re-interpreting available survey results. efforts to demonstrate the validity of the theory through
Such survey research mainly focuses on risk analysis conventional methodologies (surveys).
and environmental issues (recently: Ellis and F. Thompson, In a contribution with the late Karl Dake, Michael
1997; Coughlin and Lockhart, 1998; Marris et al., 1998 to Thompson demonstrates the very importance of plural ra-
name a few). It is assessed to have performed rather poorly tionality in the field of consumption and sustainability. By
(Boholm, 1996; Milton, 1996; Sjöberg, 1997; Tansey and contrast, most conventional models are based, at least im-
O’Riordan, 1999) partly due to weaknesses of the individual plicitly, on the assumption that there is an average household
research design,24 partly due to the fundamental inadequacy with an average consumptive pattern for the whole pop-
of survey questionnaires to investigate cultural biases (as ulation. This average behaviour is that affluence prompts
discussed above) and partly due to the lack of rigour of the more consumption. Such thinking is summarised as the I-
theory and its applications (Boholm 1996). PAT equation in which population (P), affluence (A) and
Grid-group cultural theory has proven much more pow- technology (T) are multiplied to calculate the environmental
erful in anthropological studies such as Mars’s (1982) study impact (I). Grid-group cultural theory is a tool to generate
of work fiddles and in many fields of political science. In alternative consumption styles as there are different ways to
policy analysis and public administration, it is a mighty make ends meet. To each of the five types of social beings
tool to identify competing arguments and conflicting strate- and their preferred form of social solidarity, corresponds a
gies (see for example the imposing study of waste siting consumption style. To show the existence of these consump-
and justice by Benjamin Davy, 1997), to generate compet- tion styles, Thompson and Dake discuss the results of a study
ing scenarios to assess problems and design policies (such they carried out in London, Lancashire and Merseyside in
as Van Asselt and Rotmans, 1996), to explain and predict England. Their combination of survey questionnaires and
failures (such as Hood, 1996, 1998), or to assess learning anthropological interviews provides data about consumption
capacities of institutions (such as Hendriks, 1994, 1996, patterns and social interactions. Obviously, if such diver-
1999). These applications are nevertheless vulnerable to il- sity exists, policy makers should take it into account when
lustrative examples and bird spotting,25 especially when the they design policies to influence the environmental impact
theory is used as a heuristic device without trying to explain of human behaviour.
why which actors voice which bias. They often lie under The importance of cultural pluralism is also the topic
cross fires being criticised by both those who would prefer of the next article. Frank Hendriks is addressing the post-
23 If adherence is measured through agreement with certain state-
industrialisation of the urban environment in the West. He
presents four influential political responses to these changes
ments, rejection is measured through disagreement with these statements
(as opposed to a neutral stand). in the city labelled liberal productivism, communitarian-
24 It appears to be extremely difficult to design survey questions that ism, institutional perfectionism and participatory democ-
would tap the cultural types satisfactorily. racy. Hendriks uses grid-group cultural theory to assess
25 Spotting representatives of each cultural bias, for the sake of spotting
the strength and the weakness of these four approaches.
them.
406

He presents a cultural approach to policy oriented learn- cultural theory provides a tool to make sense of interna-
ing, emphasising the importance of cultural pluralism (the tional relations before and after the 1987 watershed because
combination and the cross-fertilisation of the four biases it distinguishes different ways to organise (which applies to
distinguished by grid-group cultural theory) for so-called international relations as to everything else) . This watershed
double-loop learning and pleading for institutional settings was prompted by the major fire at the Sandoz warehouse in
that stimulate policy oriented learning. Basle in December 1986, a major ‘surprise’ in grid-group
Far from institutional perfectionism and the post- cultural theory terms. As a result, the hierarchical bias with
industrial city, Paul Richards applies grid-group cultural its preference for binding treaties which made agreement
theory to make sense of the irrational: the extreme violence very difficult to reach was discredited and replaced by a
of youth rebels in the civil war in Sierra Leone. His case more individualist way of co-operating internationally, with
study reveals the rationality of the ways these youngsters informal and non-binding agreements limited to goal setting
organise and the role of violence in it. Characterising newly and implemented pragmatically at the lowest level possible.
independent African states as dominated by two compet- This boosted the restoration of the river basin.
ing hierarchist biases: (traditional) patrimonial hierarchy and Both Gyawali and Verweij deal with the management
(modern) bureaucratic hierarchy, Richards shows how both of big rivers, the Ganga and the Rhine, still the two ap-
failed the youth, as young people are socially excluded. He proaches are very different. The point is of course not that
describes the mechanism of forced recruitment in the RUF the first is concerned with the management of water flow
and how captives adopt the enclavist or sectarian thought and the second with the reduction of water pollution, but
style that corresponds to their isolation in the forests of the rather that they use grid-group cultural theory differently.
Liberia-Sierra Leone border. Finally he discusses sectarian While Gyawali applies the theory to the ways of organising
violence and the need to account for world views in attempts water management, Verweij applies it to ways of organising
to conflict resolution, assessing grid-group cultural theory as international relations. They demonstrate how much the re-
an alternative to a-social theory of conciliation. searcher’s interest determines his/her way to employ insights
Dipak Gyawali is addressing the water conflict in the from grid-group cultural theory. It is therefore tempting to
Ganga Plains and its many proponents. The main dispute guess what they would have to say about each other’s case.
involves the state of Bihar and the Union government in In Gyawali’s perspective, nothing much changed before and
Delhi, with the neighbouring Indian states (Uttar Pradesh after 1986 in the ways Rhine pollution is managed (engi-
upstream and downstream West Bengal) and the neighbour- neering is still the answer and it is working pretty well in
ing sovereign states of Nepal (upstream of most tributaries that context). In Verweij’s perspective, relations between In-
of the Ganga) and Bangladesh (downstream) involved in dian states, between states and the federal state, respectively
diverse water projects (dams) with the Indian government between the Indian Union and its neighbouring states, are
with severe consequences for Bihar’s water rights. A sec- frozen. It seems that the recurrent inundations, the politi-
ond conflict opposes political active inhabitants of the Ganga cal mobilisations in Bihar and the democratisation of Nepali
basin to the state policies regarding the embankments of the politics are not ‘surprises’ big enough to trigger changes in
Ganga and its main tributaries and the technical agency that the way of organising inter-states relations in the region.
implement them. Gyawali uses grid-group cultural theory The last three articles bring us back to Europe.
to reveal the cultural bias between the institutional style of Gunnar Grendstad offers a political-cultural map of Eu-
the Bihar government, respectively the grassroots social and rope. He applies grid-group cultural theory to explore dif-
environmental activists. He shows how the hierarchic soli- ferences between the public of twelve European countries.
darity pervades the concerns of the Bihar government at the He uses the results of the European Values Survey of 1981
national level but its blindness for the water problems at the and 1990. Assuming that each respondent can be pinpointed
local level, namely the unintended effects of embarkments. as the adherent of one of the four cultures distinguished by
By contrast, grassroots actions are driven by the egalitarian grid-group cultural theory, he uses a principal component
solidarity, while the obvious failure of the water manage- analysis to construct the profile of these countries in 1981
ment eases recruitment among the fatalist masses. Because and in 1990. To do so he both uses gravity point (the av-
the Bihar government shares the same solidarity as the Delhi erage score of the country sample on the grid and group
government, it is easily pacified and it neglects completely components) and the proportions of adherents to the four
the activist critique of the past and present programs of dam cultures. In addition, he compares grid-group cultural the-
construction. ory to a more established model of characterising public
Marco Verweij is addressing international relations and attitudes: the materialism/post-materialism shift analysed by
more specifically collaboration between states to protect the Ronald Inglehart during the past three decades. The com-
environment. He uses grid-group cultural theory in a case parison includes a test of the predicting value of individual
study of the recent history of collaboration between the grid and group scores for (post-)materialist priorities and of
riparian countries of the Rhine. Before 1987, relations be- the predicting values of all three measures for environmental
tween states were ineffective (even if they did implement concern.
national programs) but since 1987 co-operation is extremely Virginie Mamadouh offers another political-cultural map
successful. Unlike established theories in International Re- of Europe. Here national political cultures are conceived as
lations (both neo-liberalism and neo-realism), grid-group norms and practices in the political arena, which is differ-
407

ent from the aggregated pattern of individual preferences how political mandates are held or the atmosphere and the
obtained in surveys. Each national political culture is char- debating style at party conferences. Faucher argues that this
acterised by a dominant ‘conversation’ between the rational- occurs because the two parties are confronted with differ-
ities distinguished by grid-group cultural theory. Mamadouh ent political opportunity structures and political cultures in
addresses the question of the origins of differences between France and in Great-Britain.
national political cultures. She discusses traditional family Although Faucher is rather vague about what the dif-
structures as a possible explanation, elaborating on the work ferences between the two national political cultures really
of the French historian and demographer Emmanuel Todd. are, the connection to the previous articles is obvious. If
Family structures are expected to have eased or hampered environmental concern is really connected to egalitarianism,
the weight of certain rationalities in the political domain. the green agenda is appealing to a much larger proportion
The two political-cultural maps differ in many respects. of the population in Germany, than in both Great-Britain
The two authors use different notions of political cul- and France (see Figure 2 and appendix in Grendstad’s ar-
tures. Their application of grid-group cultural theory differs: ticle). Regarding party organisation, differences in style
Grendstad focuses on the proportion of adherents to the dif- between French and British Greens echo the gap between the
ferent rationalities in the population whereas Mamadouh is obstructive/clientelistic political culture in France and the
more interested in the conversation between the different ra- entrepreneurial one in Britain (see Table 3, in Mamadouh).
tionalities in the political arena. Remarkably the clustering Surely a full series of such studies (including of course dif-
of countries on Grendstad’s Figure 3 fits well the character- ferent types of parties) is needed to substantiate the portrait
isation of national political cultures in Mamadouh’s article. of the national political cultures sketched by Mamadouh.
There are two clusters: one in the lower right quadrant fea- To complete this issue, two key publications published
turing the country with a moralistic political culture, and in 1998 have been reviewed. The first book review, by Paul
one in the upper left quadrant (with the gravity point at least Richards, deals with a tremendous enterprise, an American
once in the fatalist quadrant itself) for other national political initiative to tackle the issue of the social science of climate
cultures. But the distribution of the population among the change. The resulting series consists of four volumes edited
four rationalities (see Figure 2 and appendix in Grendstad’s by Steve Rayner and Elizabeth Malone under the title Hu-
article) varies much between countries characterised by the man Choice and Climate Change. The second book review,
same family type(s) (see Table 3 in Mamadouh’s article). by Robert Hoppe, tackles the issue of public management
The two measures of cultural context substantiate each other, and its failures as addressed by Christopher Hood in The Art
for example high proportions of fatalists and individualists of the State.
are found in countries where the egalitarian nuclear family
is predominant, as expected. Nevertheless certain inconsis-
tencies are revealed (especially regarding the importance of Acknowledgements
egalitarianism in Denmark and the Netherlands, as well as
the support for individualism in Germany). This is to be I would like to acknowledge the importance of lively ex-
explained through additional intervening factors such as re- changes during workshops such as the ECPR workshop in
ligion, industrialisation and literacy and the evolution of the Bern, March 1997, the authors’ meeting of Cultural Theory
political arena itself (see Mamadouh’s article), the latter be- as Political Science in Leiden, November 1997, the ESSHC
ing also the likely explanation for the changes in the makeup sessions in Amsterdam, March 1998 and the ECPR work-
of the cultural context in public opinion between 1981 and shop in Mannheim, March 1999, as well as the discussions
1990. on the electronic gridgroup-list created by Mary Douglas
Florence Faucher presents the results of a political an- and run first by Matthias Karmasin and now by James
thropological study of green activism in Great-Britain and Tansey. I am particularly grateful for stimulating discussions
in France, and in Oxford and Aix-en-Provence in partic- and enlightening insights from the many participants to these
ular (see also Faucher 1999). In this article she addresses exchanges, too many to name individually, but I make an ex-
the issue of party organisation and democracy. Her starting ception to thank Gunnar Grendstad who was so kind to give
point is the analysis presented in Risk and Culture by Mary me access to his extended Cultural Theory bibliography. Of
Douglas and Aaron Wildavsky (1982) in which great con- course, they do not necessarily endorse my view of the state
cern with environmental matters is said to correspond to the of affairs in grid-group cultural theory. All responsibility
egalitarian cultural bias. Faucher first assesses that French for mistakes, omissions and erroneous conclusions remains
and British greens are trying to organise political participa- mine.
tion according to these egalitarian preferences. The common
‘green culture’ is characterised by a strong emphasis on po-
litical participation for all, therefore aimed at facilitating References
face to face contacts between militants, at preventing the
emergence of political elites and at promoting participative Bernstein B., 1971: Class, Codes and Control, volume 1, Theoretical
Studies towards a Sociology of Language, London, Routledge & Kegan
democracy inside the new party. Nevertheless the two party Paul.
organisations are not as similar as one might expect, for Boholm Å., 1996: Risk perception and social anthropology: a critique of
example regarding the ways party officials are elected and cultural theory. Ethnos 61(1–2): 64–84.
408

Boyle R.P. and Coughlin R.M., 1994: Conceptualizing and operationalizing Douglas M. and Wildavsky A., 1982: Risk and Culture, An Essay on the
cultural theory. In: Coyle D.J. and Ellis R.J. (eds), Politics, Policy and Selection of Technological and Environmental Dangers. University of
Culture. Westview, Boulder, pp. 191–218. California Press, Berkeley.
Chai S.-Ki and Swedlow B. (eds), 1998: Culture and Social Theory: Aaron Edvardsen U., 1997: A cultural approach to understanding modes of
Wildavsky. Transaction, New Brunswick, NJ. transition to democracy. Journal of Theoretical Politics, 9(1): 211–234.
Coughlin R.M. and Lockhart C., 1998: Grid-group theory and political Ellis R.J., 1991: Explaining the occurrence of charismatic leadership in
ideology, A consideration of their relative strengths and weakness for organizations. Journal of Theoretical Politics, 3(3): 305–320.
explaining the structure of mass belief systems, Journal of Theoretical Ellis R.J. and Thompson M. (eds), 1997: Culture Matters, Essays in Honor
Politics, 10(1): 33–58. of Aaron Wildasky, Westview, Boulder.
Coyle D.J., 1994: The theory that would be king. In: Coyle D.J. and Ellis Ellis R.J. and Thompson F., 1997: Culture and the environment in the
R. (eds), Politics, Policy and Culture. Westview, Boulder, pp. 219–239. Pacific Northwest. American Political Science Review, 91(4): 885–897.
Coyle D.J. and Ellis R.J., 1994: Introduction. In: Coyle D.J. and Ellis R.J. Ellis R.J. and Wildavsky A., 1989: Dilemmas of Presidential Leadership:
(eds), Politics, Policy and Culture. Westview, Boulder, pp. 1–14. From Washington to Lincoln, Transaction, New Brunswick, NJ.
Coyle D.J. and Ellis R.J. (eds), 1994: Politics, Policy and Culture. Engbersen G., Schuyt K., Timmer J. and van Waarden F., 1993: Cultures of
Westview, Boulder. Unemployment: a Comparative Look at Long-Term Unemployment and
Dake K., 1990: Technology on Trial: Orienting Dispositions Toward En- Urban Poverty. Westview, Boulder.
vironmental and Health Hazards. PhD thesis, University of California, Grendstad G., 1990: Europe by Cultures, An Exploration in Grid/Group
Berkeley. Analysis. graduate Thesis, University of Bergen.
Dake K., 1991: Orienting dispositions in the perception of risk: an Grendstad G., 1995: Party followership and leadership in Norway: a
analysis of contemporary world views and cultural biases. Journal of political culture approach, Party Politics, 1(2): 221–243.
Cross-Cultural Psychology, 22(1): 60–81. Grendstad G. and Selle P., 1995: Cultural Theory and the new institutional-
Dake K., 1992: Myths of nature: cultural and social construction of risk. ism. Journal of Theoretical Politics, 7(1): 5–27.
Journal of Social Issues, 48(4): 21–37. Grendstad G. and Selle P. (eds), 1996: Kultur som Levemåte. Samlaget,
Dake K. and Thompson M., 1993: The meanings of sustainable develop- Oslo.
ment: household strategies for managing needs and resources. In: Wright Grendstad G. and Selle P., 1997: Cultural theory, postmaterialism and en-
S.D., Dietz T., Borden R., Young G. and Guagnanp G. (eds), Human vironmental attitudes. In: Ellis R.J. and Thompson M. (eds), Culture
Ecology: Crossing Boundaries. The Society for Human Ecology, Fort Matters, pp. 151–168.
Collins, Colorado. Grendstad G. and Selle P., 1999: The formation and transformation of pref-
Dake K. and Wildavsky A., 1991: Theories of risk perception: who erences, cultural theory and postmaterialism compared. In: Thompson
fears what and why? In: Wildavsky A. (ed.), The Rise of Radical et al. (eds), Cultural Theory as Political Science. Routledge, London,
Egalitarianism, pp. 133–150. pp. 43–58.
Davy B., 1997: Essential Injustice, When Legal Institutions Cannot Resolve Gross J.L. and Rayner S., 1985: Measuring Culture: A Paradigm for the
Environmental and Land Use Disputes, Springer-Verlag, Wien. Analysis of Social Organisation. Columbia University Press, New York.
Douglas M., 1970: Natural Symbols, Explorations in Cosmology. Barrie & Hendriks F., 1994: Cars and cultures in Birmingham and Munich. In Coyle
Rockliff, London. D. and Ellis R.J. (eds), Politics, Policy and Culture, pp. 51–69.
Douglas M., 1973: Natural Symbols, Explorations in Cosmology. First Hendriks F., 1996: Beleid, cultuur en instituties, het verhaal van twee
Vintage Books edition, Vintage Books, New York. steden. DSWO Press, Leiden.
Douglas M., 1975a: Implicit Meanings: Essays in Anthropology. Routledge Hendriks F., 1999: Political Institutions and Public Policy: A Tale of Two
& Kegan Paul, London. Cities. Edward Elgar (forthcoming).
Douglas M., 1975b: In the nature of things, Inaugural Lecture given at Holling C.S., 1986: The resilience of territorial ecosystems. In: Clark
University College, London, 1971. In: M. Douglas, Implicit Meanings, W.C. and Munn R. (eds), Sustainable Development of the Biosphere.
pp. 210–229. Cambridge University Press, Cambridge, pp. 292–388.
Douglas M., 1978: Cultural Bias. Royal Anthropological Institute Occa- Hood C., 1996: Control over bureaucracy: cultural theory and institutional
sional Paper No. 35, Royal Anthropological Institute, London (reprinted variety. Journal of Public Policy, 15(3): 207–230.
in: M. Douglas, 1982a, In the Active Voice, Routledge and Kegan Paul, Hood C., 1998: The Art of the State, Culture, Rhetoric, and Public
London, pp. 183–254). Management. Clarendon Press, Oxford.
Douglas M., 1982a: In the Active Voice. Routledge and Kegan Paul, Hoppe R. and Peterse A., 1994: Handling Frozen Fire: Political Culture
London. and Risk Management. Westview, Boulder.
Douglas M. (ed.), 1982b: Essays in the Sociology of Perception. Routledge Karmasin H. and Karmasin M., 1997: Cultural Theory: ein neuer Ansatz
and Kegan Paul, London. für Kommunikation, Marketing und Management. Linde Verlag, Wien.
Douglas M., 1982c: Introduction to grid/group analysis. In: Douglas M. Mamadouh V., 1997: Political culture: a typology grounded on cultural
(ed.), Essays in the Sociology of Perception. Routledge & Kegan Paul, theory. Geojournal, 43(1): 17–25.
London, pp. 1–8. Mamadouh V., 1999: National political cultures in the European Union.
Douglas M., 1985: Risk Acceptability according to the Social Sciences, In: Thompson M., Grendstad G. and Selle P. (eds), Cultural Theory as
Russell Sage Foundation, New York. Political Science. Routledge, London, pp. 138–153.
Douglas M., 1986: How Institutions Think. Syracuse University Press, Marris C., Langford I. and O’Riordan T., 1998: A quantitative test of the
Syracuse NY. cultural theory of risk perceptions, Comparisons with the psychometric
Douglas M., 1990: Risk as a forensic resource. Daedalus, 119 (4): 1–16. paradigm. Risk Analysis, 18(5): 635–647.
Douglas M., 1992: Risk and Blame: Essays in Cultural Theory. Routledge, Mars G., 1982: Cheats at Work, an Anthropology of Workplace Crime.
London. George Allen and Unwin, London.
Douglas M., 1992b: Governability: a question of culture. Millennium: Milton K., 1996: Environmentalism and Cultural Theory. Routledge,
Journal of International Studies, 22 (3): 463–481. London.
Douglas M., 1993: In the Wilderness: the doctrine of defilement in the Book Molenaers N. and Thompson, M., 1999: The cultural conditions for democ-
of Numbers, JSOT Press, Sheffield (Journal for the Study of the Old racy and their implications for transitional societies. In: Thompson M.,
Testament. Supplement series, No. 158). Grendstad G. and Selle P. (eds), Cultural Theory as Political Science.
Douglas M., 1996: Thought Styles: Critical Essays on Good Taste. Sage, Routledge, London, pp. 186–205.
London. Ney S. and Thompson M., 1999: Consulting the frogs, the normative im-
Douglas M., 1997: The depolitisation of risk. In: Ellis R.J. and Thompson plications of cultural theory. In: Thompson M. (eds), Cultural Theory as
M. (eds), Culture Matters. Westview, Boulder, pp. 121–132. Political Science. Routledge, London, pp. 206–223.
Douglas M. and Isherwood B., 1978: The World of Goods: Toward an Olli E., 1995: Cultural Theory Specified – The Coherent, Sequential and
Anthropology of Consumption. Penguin Books, Harmondsworth. Synthetic Individual Approach. Graduate Thesis, University of Bergen.
Douglas M. and Ney S., 1998: Missing Persons, A Critique of Personhood
in the Social Sciences, University of California Press, Berkeley.
409

Olli E., 1999: Rejection of cultural biases and effects on party preference. Thompson M. and Rayner S., 1998b: Risk and governance, Part I: The
In: Thompson M. et al. (eds), Cultural Theory as Political Science. discourses of climate change. Government and Opposition, 35(2): 139–
Routledge, London, pp. 59–74. 166.
Ostander D., 1982: One- and two-dimensional models of the distribution Thompson M. and Wildavsky A., 1986: A poverty distinction: from eco-
of beliefs. In: Douglas M. (ed.), Essays in the Sociology of Perception, nomic homogeneity to cultural heterogeneity in the classification of poor
pp. 14–30. people. Policy Sciences, 19: 163–199.
Ouchi W., 1980: Markets, bureaucracies and clans. Administrative Science Thompson M., 1997b: Security and solidarity: an anti-reductionist frame-
Quarterly, 25: 129–141. work for thinking about the relationship between us and the rest of
Rayner S., 1979: The classification and dynamics of sectarian forms of nature, The Geographical Journal, 163(2): 141–149.
organisations: grid/group perspectives on the far left in Britain, PhD Thompson M., Ellis R. and Wildavsky A., 1990: Cultural Theory. Westview
dissertation, University of London. Press, Boulder, Co.
Rayner S., 1982: The perception of time and space in egalitarian sects, a Thompson M., Ellis R. and Wildavsky A., 1992: Political cultures. In:
millenarian cosmology. In: Douglas M. (ed.), Essays in the Sociology of Hawkesworth M. and Hogan M. (eds), Encyclopedia of Government and
Perception, pp. 247–276. Politics. Vol. 1, Routledge, London, pp. 507–520.
Rayner S., 1984: Disagreeing about risk: the institutional cultures of risk Thompson M., Grendstad G. and Selle P. (eds), 1999: Cultural Theory as
management and planning for future generations. In: Hadden S.G. (ed.), Political Science. Routledge, London.
Risk Analysis, Institutions and Public Policy. Associated Faculty Press, Thompson M. Rayner S. and Ney S., 1998: Risk and governance, Part
New York, pp. 150–169. II: Policy in a complex and plurally perceived world. Government and
Rayner S., 1986: Management of radiation hazards in hospitals, plural ra- Opposition, 33(3): 330–354.
tionalities in a single institution. Social Studies of Science, 16: 573–591. Van Asselt M.B.A. and Rotmans J., 1996: Uncertainty in perspective.
Rayner S., 1988: The rules that keep us equal. In: Flanagan J.G. and Global Environmental Change, 6(2): 121–157.
Rayner S. (eds), Rules, Decisions and Inequality. Avebury, Brookfield, Verweij M., 1995: Cultural theory and the study of International relations.
Vermont/Aldershot. Millenium, 24(1): 87–111.
Rayner S., 1992: Cultural theory and risk analysis. In: Krimsky S. and Verweij M., 1997: Review article: developing the cultural analysis of Mary
Golding D. (eds), Social Theories of Risk, Praeger, Westport, pp. 83– Douglas. Government and Opposition, 32(3): 421–429.
115. Verweij M., 1998: Cultures and institutions in transboundary relations: the
Rayner S., 1993: Risk perception, technology acceptance, and institutional environmental protection of the Rhine and the Great Lakes. PhD thesis
culture: case studies of some new definitions. In: Ruck B. (ed.), Risk is European University Institute, Florence.
a Construct. Knesebeck, München, pp. 197–220. Ward V., 1998: Towards a theory of state – non-state actors: a grid-group
Rayner S. and Malone E.L. (eds), 1998: Human Choice and Climate cultural approach. In: Jacquin-Berdal D., Oros A. and Verweij M. (eds),
Change (four volumes), Battelle Press, Columbus, Ohio. Culture in World Politics. MacMillan, London/St Martin’s, New York
Schmutzer M., 1994: Ingenium und Individuum: eine sozialwis- (Millenium, Journal of International Studies) pp. 206–244.
senschaftlichr Theorie der Wissenschaft und Technik, Springer-Verlag, Webber C. and Wildavsky A., 1986. A History of Taxation and Expenditure
Wien. in the Western World. Simon and Schuster, New York.
Schwarz M. and Thompson M., 1990: Divided We Stand: Redefining Wildavsky A., 1984: The Nursing Father: Moses as a Political Leader, The
Politics, Technology and Social Choice. Harvester Wheatsheaf, Hemel University of Alabama Press, Alabama.
Hempstead. Wildavsky A., 1985. Change in political culture. Politics, The Journal of
Selle P., 1991: Culture and the study of politics. Scandinavian Political the Australasian Political Studies Association, 20(2): 95–102.
Studies, 14(2): 97–124. Wildavsky A., 1986: Budgeting, A Comparative Theory of Bud-
Shackleton M., 1991: The European Community between three ways of getary Processes. Second, Revised Edition. Transaction Books, New
life. Journal of Common Market Studies, 29(6): 575–601. Brunswick/Oxford.
Sjöberg L., 1997: Explaining risk perception: an empirical evaluation of Wildavsky A., 1987: Choosing preferences by constructing institutions:
cultural theory. Risk, Decision and Policy, 2(2): 113–130. a cultural theory of preference formation. American Political Science
Tansey J. and O’Riordan T., 1999: Cultural theory and risk: a review. Review, 81(1): 3–21.
Health, Risk, Society, 1(1): 71–90. Wildavsky A. (ed.), 1991: The Rise of Radical Egalitarianism. American
Thompson M., 1979: Rubbish Theory: The Creation and Destruction of University Press, Washington, DC.
Value. Oxford University Press, Oxford. Wildavsky A., 1991: What other theory would be expected to answer such
Thompson M., 1982a: A three-dimensional model. In: Douglas M. (ed.), profound questions? A reply to Per Selle’s critique of cultural theory.
Essays in the Sociology of Perception, pp. 31–63. Scandinavian Political Studies, 14(4): 355–360.
Thompson M., 1982b: The problem of the centre: an autonomous cos- Wildavsky A., 1994: Democracy as a coalition of cultures. Society, 31:
mology. In: Douglas M. (ed.), Essays in the Sociology of Perception, 80–83.
pp. 302–327. Wildavsky A., 1994: Why self interest means less outside a social con-
Thompson M., 1997a: Cultural theory and integrated assessment. Environ- text, Cultural contributions to a theory of rational choice. Journal of
mental Modeling and Assesment, 2: 139–150. Theoretical Politics, 6(2): 131–159.
Thompson M. and Rayner S., 1998a: Cultural discourses. In: Rayner S. Wildavksy A. and Dake K., 1990: Theories of risk perception: who fears
and Malone E. (eds), Human Choices and Climate Change. Vol. 1. The what and why? Daedalus, 119: 41–60.
Societal Framework, Batelle, Columbus, pp. 139–166.

You might also like