12
CULTURES IN ORGANIZATIONS
approach to absorb or distort another. The next section of this chapter explains the
advantages of this framework and how it differs from more traditional approaches
to theory-building.
Introducing the Three Perspectives
This book examines the commonalities and fundamental conflicts in culture research
by distinguishing and then analyzing three social scientific perspectives:
the Integration, Differentiation, and Fragmentation views.3
Studiesconducted from an Integration perspective have three defining characteristics:
all cultural manifestations mentioned are interpreted as consistently
reinforcingthe same themes, all members of the organization are said to share in
an organization-wideconsensus, and the culture is described as a realm where all
is clear. Ambiguity is excluded.
In contrast, research conducted from a Differentiation perspective describes
cultural manifestations as sometimes inconsistent (for example, when managers
say one thing and do another). Consensus occurs only within the boundaries of
subcultures, which often conflict with each other. Ambiguity is channeled, so
that it does not intmde on the clarity which exists within these subcultural bound-
aries.
Studiesconducted from a Fragmentation perspective focus on ambiguity as the
essence of organizational culture. Consensus and dissensus are issue-specificand
constantly fluctuating. No stable organization-wide or subcultural consensus exists.
Clear consistencies and clear inconsistencies are rare.
These three social scientific perspectives are summarized in Table 1-1 and
represented (in partial and idiosyncratic ways) in the second of the two culture
arguments earlier.
Why These Perspectives Are Subjective
Many cultural researchers assume or assert that a particular perspective represents
an accurate reflection of an objectively observed reality, rather than subjectively
construed conceptual judgments. Thus some companies are said to have more
consistency, organization-wide consensus, and clarity than others. Or a company
is seen as passing, in stages, from having an Integrated culture to having a
Differentiatedor Fragmented culture. Furthermore, some researchers would argue
that when enough high quality cultural research has been done, it will be
possible to declare a winner in the war of the three perspectives, in the sense that
one of the three (or some variant or combination of the three) will be shown to be
the single most accurate way to describe the majority of organizational cultures.
These conclusions reify cultures as having an objective reality that can be accurately
assessed as fitting one of the social scientific perspectives more than the
others.
In contrast, this book argues that a social scienttj'icperspective is an interprerive
framework that is subjectively imposed on the process of collecting and
analyzing cilltural data. A social scientiJicperspective is not considered here to
SEEING CULTURES FROM DIFFERENT POINTS OF VIEW
Table 1-1 Defining characteristics of the three perspectives
Perspective Integration Differentiation Fragmentation
Orientation to Organization-wide Subculturalconsensus Multiplicity of views
consensus consensus (no consensus)
Relation among Consistency Inconsistency ~omblexit~(not
manifestations clearly consistent or
inconsistent)
Orientation to Exclude it Channel it outside Focus on it 1
ambiguity subcultures I
Metaphors Clearing in jungle, Islands of clarity in Web. jungle
monolith, hologram sea of ambiguity
Adapted from Martin and Meyerson (1988). Table 1; Meyerson and Martin (1987), Figure 3; and
Frost, Moore, Louis, Lundberg, and Martin (1991), Table 1.1.
be an objective description of empiricalfacts. This is not because researchersare
careless, dishonest, or otherwise inadequate social scientists. It is because different
researchers, studying the same cultural members and the same organizational
events with equal care, skill, and honesty may evaluate, recall, and interpret what
happens differently. This is so, in part, because who a researcher is, or is seen to
be, may affect what cultural members say and do. In addition, different researchers
have different preconceptions, sensitivities, and skills. For example, an
African-American researcher with personal experience of subtle forms of racial
prejudice, in contrast to a white researcher less conscious of race, may elicit or
offer a different interpretation of the meaning of an inter-racial interaction.
Because of these issues, it is essential to realize that when a study concludes,
for example, that all members of a culture share a particular value, this is a
subjective judgment. Whether that judgment is based on quantitative or qualitative
data, the measurement, collection, and interpretation of that data are affected
by subjective factors. The three social scientific perspectives, then, should be
thought of as subjectively perceived "ideal types," rather than objective descriptions
of particular cultural realities. Culture is not reified--out there-to be
accurately observed.4
Once these social scientific perspectives are conceptualized as subjective, it
becomes easier to see their shortcomings and imagine them changing. An example
from another area might clarify this contention. There is nothing "natural"
about categorizing people by the color of their skin. For instance, why do race
classifications focus on skin color, rather than the color of a person's eyes? In the
United States, why is color so often conceptualized as a dichotomy (black or
white)? Once race is seen as a subjective, socially constructed category, different
ways of seeing racial relations become possible. A "fact of nature" becomes a
social relationship of domination and oppression that might be viewed quite
differently and changed.5 In a similar way, once the three social scientific perspectives
are seen as subjective and socially constructed, it becomes possible to
envision other ways of conceptualizing what cultures are and how they change.
46
CULTURES I N ORGANIZATIONS
included in order to illustrate the fact that each of the three social scientific
perspectives is based on an extensive body of empirical literature. Explanatory
material that I have added to these quotations is marked with brackets.
All or even most studies congruent with a given perspective could not be
included in these references. There are simply too many. For example, the
Integration perspective has become the dominant view of organizational researchers
and practitioners in the United States. One review of a limited sample
of journals counted almost 200 recent articles, most of which were conducted
from this perspective.2 Limitations in manuscript length and reader patience also
make it impossible to discuss all of a cited text (or all of the writings of a given
researcher). Therefore, those who are unfamiliar with particular researchers and
texts quoted or referenced in this book are encouraged to read some of this work
in its entirety, so that the context from which a quote is taken can be seen and the
fairness and accuracy of descriptions can be assessed.
ORGANIZATION-WIDE CONSENSUS
The core of the Integration perspective is the lure of organization-wide consensus.
Clark's description of "organizational sagas" (a precursor of organizational
culture) states this point cogently:
An organizational saga is a powerful means of unity in the formal workplace, It makes
links across internal divisions and organizationalboundaries as internal and external
groups share their common beliefs. With deep emotional commitment believers define
themselves by their organizational affiliation and in their bond to other believers they
share an intense sense of the unique. (Clark, 1972, p. 183)
In Integration views of culture, people at all levels of an organizational hierarchy
are said to agree about potentially divisive issues. For example:3
Just as individualsprocess information,so also do groups and units of people. In doing
so they develop collective belief systems about social arrangements. . .. . They include
beliefs about, among other things, organizational purpose, criteria of performance,
the location of authority, legitimate bases of power, decision-making orientations,
style of leadership, compliance, evaluation, and motivation. (Quinn and
McGrath, 1985, p. 325)
Integration studies often describe organization-wide consensus in (harmonious)
familial terms, which merge the (supposedly separate) public and private
domains, so that organizations are seen as families and families of employees are
described as part of the organization. For example, some OZCO employees, such
as Denise and Stuart, quoted in Chapter 3, spoke of OZCO using a family
metaphor. Other examples of familial language come from Schein's study of
organizational founders and Ouchi and Jaegar's examination of "Theory Z
cultures:
IHt INTEGRATION PERSPECTIVE
47
The people who were comfortablein this environment and enjoyed the excitement of
building a successfulorganizationfound themselvesincreasingly feeling like members
of a family and were emotionallytreated as such. Strongbonds of mutual supportgrew
up at an interpersonal level, and Murphy [the founder] functioned symbolically as a
brilliant, demanding, but supportive father figure. (Schein, 1991a, p. 23)
The slownessof evaluationand the stabilityof membership promote a holistic conckm
forpeople, particularlyfrom superiorto subordinate.Thisholism includes the employee
and his or her family in an active manner. Family members regularly interact with other
organizationmembers and their familiesand feel an identificationwith the organization.
(Ouchi and Jaegar, 1978, p. 688)
Perhaps because of the Integration perspective's emphasis on interpersonal,
sometimes even familial closeness, organization-wide consensus is often described
in highly emotional terms. For example, a study of the Los Angeles
Olympic Organizing Committee concludes with a description of a party held for
employees:
A vice president who had been on the road for several days with the torch relay
. . . told about the runner going over a winding road in the hills of West Virginia
and encountering a man standing alone on the top of a hill with a trumpet playing
"America the Beautiful" as the torch passed. There was not a dry eyein the house. The
speakerhimself broke down,,overcomeby emotion, and could not continuefor several
minutes. The staff filedout to the strainsof ceremonialmusic, clutchingtheir commemorative
mugs and pins reading "Team 84" that were handed to them at the exit.
(McDonald, 1991, p. 37)
This emotional language leaves no room for dissent. For example, in the quotation
above, no "eye in the house" failed to cry, responded with skepticism, or
felt embarrassed about the chauvinism, fervor, or abundance of tears.
Consensus is another name for conformity. Some Integration research has
frankly responded to the harsh criticism that this perspective prescribes, as well as
describes, a corporate form of fascism or cult religion:
Those white-shirted,polite, competent,hard-working [IBM] employeesof twentyyears
agowere oftenregaged as corporate "fanatics," or even corporate "fascists," because
they appeared not to displayin superficial ways their "American individuality." White
shirtswere mistaken for laundered minds. The shirtsare now colored,but it appearsthat
their wearers are still politely service-oriented, highly competent, and hard working
(everythingmay changebut the beliefs). And in ourculture, any evidenceof a reduction
in obvious "individuality," which naturally accompanies increases in organizational
commitment, will produce criticism from those who overvalue individuality. (Pascale
and Athos, 1981, p. 186)
Even if pressures toward consensus sometimes make it necessary to over-ride the
desires of individuals, some Integration studies find this justified because unity
provides an antidote to the conflicts of interest that can divide and paralyze an
48
CULTURES I N ORGANIZATIONS
organization. Integration studies sometimes acknowledge conflict or deviance,
interpreting them as reasons for seeking a transcendent, more powerful unityone
that gains the consent, even the enthusiastic commitment, of the governed.
CONSISTENCY
Many Integration studies acknowledge that organization-wide consensus is not
easy to achieve. It must, researchers argue, be reinforced by a myriad of interconnected
cultural manifestations, each of which is consistent with the others. For
example:
Each ideal type [of culture]represents a set of interconnected parts, each dependent on
at least one other part. (Ouchi and Jaegar, 1978, p. 685)
An Integration study usually includes three kinds of consistency: action, symbolic,
and content. Examples of each are discussed in the following sections.
Action Consistency
Action consistency occurs when content themes are consistent with an organization's
formal and informal practices. For example, in the Integration view of
OZCO's culture, some employees described management's espoused value of
egalitarianism as consistent with a wide variety of formal and informal practices,
including company participation in the United Way charity, the company's stock
plan, profit-sharing, answering one's own telephone, "Management By Walking
Around," need-based distribution of "perks," "bottom-up" consensual decision-making,
and lateral promotions. In more abstract terms, action consistency
occurs when
the structural elements and organizational processes making up the design type are
strongly underpinned by provinces of meaning and interpretive schemes that bind them
together in an institutionally derived normative order. (Hinings and Greenwood, 1987,
p. 2, quoted in Greenwood and Hinings, 1988, p. 295)
For example, Barley's study of funeral work examined practices consistent with a
content theme emphasizing the denial of death:
The funeral director seeksto createthe appearanceof normality or naturalnesswhenever
the living are in the presence of the dead. This intention underlies strategies that
organize the execution of many different activities; for example, preparation of the
body, removal of the deceased from a home. (Barley, 1991, p. 44)
Symbolic Consistency
A second type of consistency is symbolic. It occurs when the symbolic meanings
of cultural forms, such as physical arrangements, stories, rituals, and jargon, are
THE INTEGRATION PERSPECTIVE 49
described as congruent with content themes. For example, the Integration view of
OZCO's culture offered egalitarian interpretations of Jim Hamilton's behavior at
company rituals, "retirement village" jargon, and the story about the failure of
corporate strategic planning. Physical arrangements, such as casual dress norms,
a single cafeteria for all employees, and open office spaces also were described.as
reinforcing egalitarian themes.
Symbolic consistency is also evident in Pettigrew's study of public school
headmaster^.^ According to this study, headmasters tried to introduce new values
in their schools by reinforcing desired changes with cultural forms. For example,
they created rituals or told organizational stories that expressed appreciation for
the types of behaviors they were seeking to encourage. Similarly, at IBM mles
were said to apply equally-to all employees. This espoused value was once ~ u t
to a severe test, according to an Integration interpretation of the "Rule Breaking"
story:
A twenty-two-year old bride weighing ninety pounds, whose husband had been sent
overseas and who, in consequence, had been given a job until his return. . . . The
young woman, LucilleBerger, was obliged to make certainthat people enteringsecurity
areas wore the correct clearance identification. Surrounded by his usual entourage of
white-shinedmen, Watson [thepresident] approached the doorway to an area where she
was on guard, wearing an orange badge acceptable elsewhere in the plant, but not a
green badge, which alone permitted entrance at her door. "I was trembling in my
uniform, which was far too big," she recalled. "It hid my shakes but not my voice.
,,I1rn sorry," I said to him. I knew who he was alright. "You cannot enter. Your
admittance is not recognized." That's what we were supposed to say. The men accompanying
Watson were stricken;themoment held unpredictable possibilities. "Don't you
know who he is?" someonehissed. Watson raised his hand for silence, while oneof the
party strodeoff and returned with the appropriatebadge. (Rodgers, 1969,pp. 153-154,
quoted in Martin, Feldman, Hatch, and Sitkin, 1983, p. 440)
Martin, Feldman, Hatch, and Sitkin found that versions of this and six other
"common" organizational stories were told in a wide range of large and small,
public and private,organizations.s When we analyzed the scripts (the common
elements) in various versions of these seven common stories, we assumed that
these stories had only one interpretation--one that was consistent with the espoused
values of top management. Such an assumption is a hallmark of an
Integration study. '
Other Integration studies focus on the ways leaders and enthusiastic employees
can foster the development of symbolic consistency through attention to dress
norms and the physical arrangements of the work space. For example:
The architecture and office layout of Action reflected Murphy's assumptions about
creativity and decision making. He insisted on open office landscaping;preferred cubicles
for engineers instead of offices with doors; encouraged individualism in dress and
behavior; and minimized the use of status symbols, such as private offices, special
dining rooms for executives, and personal parking spaces. Instead, there were many
conference rooms and attached kitchens to encourage people to interact comfortably.
(Schein, 1991a, pp. 21-22)
54 CULTURES I N ORGANIZATIONS
Table 4-1 Defining culture from an integration perspective
1. The pattern of shared beliefs and values that give the members of an institution meaning, and
provide them with the rules for behavior in their organization. (Davis, 1984, p. 1)
2. Organizational culture can be thought of as the glue that holds an organization together through a
sharing of patterns of meaning. The culture focuses on the values, beliefs, and expectations that
members come to share. (Siehl and Martin, 1984, p. 227)
3. Culture is the set of important understandings (often unstated) that members of a community share
in common. (Sathe, 1985, p. 6)
4. I distinguish among these elements by treating basic assumptions as the essence-what culture
really is--and by treating values and behaviors as observed manifestations of the cultural essence.
(Schein, 1985, p. 14)
5. An organization might then be studied as a culture by discovering and synthesizing its rules of
social interaction and interpretation, as revealed in the behavior they shape. Social interaction and
interpretation are communication activities, so it follows that the culture could be described by
articulating communication rules. (Schall, 1983, p. 559)
6. A standard definitionof culture would includethe system of values, symbols, and shared meanings
of a group including the embodiment of these values, symbols, and meanings into material objects
amd ritualized practices. . . . The "stuff' of culture includes customs and traditions, historical
accounts be they mythical or actual, tacit understandings, habits, norms and expectations, common
meanings associated with fixed objects and established rites, shared assumptions, and intersubjective
meanings. (Sergiovanni and Corbally, 1984, p. viii)
7. To analyze why members behave the way they do, we often look for the values that govern
behavior, which is the second level. . . . But as the values are hard to observe directly, it is often
necessary to infer them by interviewing key members of the organization or to content analyze
artifacts such as documents and charters. However, in identifying such values, we usually note that
they represent accurately only the manifest or espoused values of a culture. That is they focus on what
people say is the reason for their behavior, what they ideally would like those reasons to be, and what
are often their rationalizations for their behavior. Yet, the underlying reasons for their behavior
remain concealed or unconscious. To really understand a culture and to ascertain more completely the
group's values and overt behavior, it is imperativeto delve into the underlying assumptions, which are
typically unconscious but which actually determine how group members perceive, think, and feel.
(Schein, 1984, p. 3)
8. Organizational theorists often claim that culture is best understood as a set of assumptions or an
interpretive framework that undergirds daily life in an organization or occupation. However, despite
such theoretical pronouncements, few organizational researchers have actually bothered to study the
deep structure of a work setting. Instead, most have focused on symbolic phenomena that lie on the
surface of everyday life: stories, myths, logos, heros, and assorted other verbal or physical artifacts.
For this reason, cultural research typically belabors the obvious while failing to reveal the core of the
interpretive system that lends a culture its coherence. (Barley, 1991, p. 39)
THE INTEGRATION PERSPECTIVE
55
In contrast, Schein (definition 7) and Barley (definition 8) argue that a focus on
thesekinds of content themes or on other kinds of cultural manifestations(such as
formal and informal practices, stories, or rituals) is undesirable because these are
all relatively superficial cultural manifestations. Instead, Schein and Barley argue
that cultural studies should focus on deeper content themes, labeled basic assumptions,
because such themes are less subject to rationalization and self!
conscious manipulation. For example, Schein suggests the following basic
assumptions may serve as content themes for a cultural analysis: humanity's
relationship to nature (dominant, submissive, harmonizing, etc.); the nature of
reality and truth (what is real and what is not); the nature of human nature (good,
evil, or neutral); the nature of human activity (active, passive, self-developmental,
etc.); and the nature of human relationships (cooperative, competitive,
etc.)." One (somewhat tautological) rationale for this focus on deeply held
assumptions, offered in Integration studies, is that culture should be defined in
terms of those manifestations that are most likely to elicit organization-wide
consensus. For example:
Basic assumptions,in the sense in which I want to definethat concept,have become so
taken for granted that one finds little variation within a cultural unit. (Schein, 1985,
P. 18)
content themes, Kilmann "measured culture" by asking respondents to answer a
questionnaire with items, for example, that asked respondents to report whether
cultural members do or do not "encourage creativity" or "try to please the
organization."lO Such relatively superficial content themes may be espoused by
culturalmembers because of a desire to present oneself or one's organization in a
sociallydesirable manner, either as an intentional, impression management strategy
or as a less conscious distortion of perception or memory.
This debate about depth versus superficialitycan be illuminated by examining
the results of three Integration studies of the same organizational culture, referred
to here by a pseudonym--GEM Co. These three studies all work within an
Integrationdefinition of culture as that which is shared, but they all come to quite
different conclusions about the nature of GEM Co.'s culture because they place
different emphasis on the importance of depth. Martin, Anterasian, and Siehl, in
the most superficial of the three studies, content-analyzed the externally espoused
values in the annual reports of 100very large corporations.12In comparisonto the
other companies, GEM Co. expressed less concern about the well-being of its
employees, emphasizing instead bottom-line issues such as profitability and
products-appropriate concerns for an annual report.
A second study by Siehl and Martin focused on internally enacted content
themes at GEM Co.13 Four content themes emerged from a study of cultural
forms, such as stories, rituals, and jargon (symbolic consistency): "people are
our most important asset," "the company's products and people are unique,"
"the family of an employee is part of the company," and "never take a shortterm
perspective." At least three of these themes express a humanitarian concern
for employee well-being.
In a third, independently conducted research project, Dyer studied GEM Co.
using Schein's emphasis on internally enacted, fundamental assumptions.14The
themes which emerged from the observation of formal and informal practices
(action consistency) and cultural forms (symbolic consistency) included: "longterm
perspective on employees' careers," "egalitarianism," "truth through confrontation,''
and "protect women." A deeper look at the cultural manifestations
expressiveof these apparently humanitarian content themes revealed evidence of
wide range of studies claimed to show that under the right kind of vision-creating
leadership, organizations can develop cultures based on shared values, harmony,
and homogeneity. The Differentiation research to be discussed in the next chapter
challenges this conclusion. According to these studies, the Integration perspective
is a myth, created and perpetuated for the benefit of top management, to
cover up the contradictions and intergroup conflict that inevitably characterize
organizational cultures.
NOTE
1. Hatch (1990), Hoffman (1982).
The Differentiation Perspective:
Separation and Cmflict
According to the Differentiation perspective, the apparently seamless unities of
the Integration perspective mask a series of overlapping, nested organizational
subcultures. These subcultures co-exist, sometimes in harmony, sometimes in
conflict, and sometimes in indifference to each other. The Differentiation perspective
unveils the workings of power in organizations, acknowledges conflicts
of interest between groups, and attends to differences of opinion. This chapter
introduces the Differentiation perspective and extends it to incorporate insights
about environmental influences on the development of cultures in organizations.
Differentiation views of organizational culture have three definingcharacteristics.
First, interpretations of content themes, practices, and forms are often inconsistent.
Second, the Differentiation perspective is suspicious of claims of
organization-wide consensus. To the extent that consensus exists, it is seen as
located primarily within subcultural boundaries. Third, within subculturalboundaries,
clarity reigns, while ambiguity is relegated to the periphery. These defining
elements of the Differentiation perspective (inconsistency, subcultural consensus,
and the relegation of ambiguity to the periphery of subcultures) are
explored in more detail below, using quotations from studies that illustrate this
viewpoint.
INTRODUCTION TO DIFFERENCE
Integrationstudies focus on that which is similar, often moving to higher levelsof
abstraction that sidestep difference by encompassing it. Because abstractions
have a higher logical status, they are sometimes seen as having a firmer claim to
importance or even a moral priority. Thus, claims of difference are often seen as
"lower level, fragmenting, particularistic sorts of concerns," while similarities
are considered "higher level, integrative, and universalistic sorts of concerns."l
This process of valuing that which unifies, and devaluing that which differentiates,
can be observed at all levels of organizations. Managerialadvocatesof unity
may urge a recalcitrant labor force to accept a firmwidewage freeze "because we
96 CULTURES IN ORGANIZATIONS
opposition shaped by the consciousness of being on the bottom. . . . That these
activities have been obscured from traditional social scientists should come as no
surprise. Oppressed peoples may maintain hidden consciousness and may not reveal
their true selves for reasons of self-protection. (Collins, 1986, p. 23)
Other Differentiation studies focus on what Feldman labels acceptance based
on unconscious distortion. This distortion occurs because it can be difficult for
members of less powerful groups to see their subordination as unjust; a dominant
culture can silence dissent, making it difficult for the less powerful to articulate,
enact, or even see an alternative way of life. This is a description of "false
consciousness." Jermier offers two descriptions of a blue-collar worker at a
manufacturing plant.21 In the first description, the protagonist accepts management's
assurances about the safety of his work and occupies his mind with
consumerist fantasies. The only breaches in this false consciousness are vague
and fleeting glimpses of an alternative reality-a partially articulated unease that
is quickly dismissed. In the second half of this portrait, the same individual is
overtly alienated, disbelieving management's claims of safety, angry about his
exploitation, and well aware of the costs of joining the consumerist "rat race."
Although Jermier's study is unusual in its portrayal of two starkly different
psychic realities, Differentiation research generally assumes a self-concept that is
compartmentalized by conflicting subcultural demands. In summary, the Differentiation
view differs from an Integration perspective at the individual, subcultural,
and organizational levels of analysis.
DEFINING CULTURE FROM A DIFFERENTIATION PERSPECTIVE
Differentiation studies define culture in terms that are surprisingly similar, in
some ways, to the definitions used in Integration research: culture is defined as
that which is shared. Table 6-1 presents a range of these Differentiation definitions.22In
contrast to Integration definitions, however, Differentiation definitions
specify that it is a group, rather than an entire organization, that.is doing the
sharing. See, for example, Trice and Morand's definition (I), Louis's definitions
(3) and (lo), Gregory's definition (5), Smircich's definition (ll), and Van
Maanen and Barley's definition (12).
Some (but not all) Differentiation studies define culture as unique or distinctive
to a particular group (see, e.g., Louis's definition [lo], Smircich's definition
[I11, Van Maanen and Barley's definition [12], and Gregory's definition [13]).23
This theme of cultural uniqueness was also present in many Integration definitions
of culture. Because this issue of distinctness transcends the boundaries
between the Integration and Differentiation approaches to the study of culture,
and because there is reason to believe that this claim of cultural uniqueness may
not be well founded, this issue is discussed separately in the concluding sections
of this chapter.
The most important aspect of these definitions is that they vary in the extent to
which they admit the possibility that subcultures may co-exist with some form of
Table 6-1 Defining culture from a Differentiation perspective
1. Organizational subcultures may be defined as distinct clusters of understandings, behaviors, and
cultural forms that identify groups of people in the organization. They differ noticeably from the
common organizational culture in which they are embedded, either intensifying its understandings
and practices or deviating from them. (Trice and Morand, in press, p. I)
t
2. Culture's utility as a heuristic concept may be lost when the organizational level of analysis is
employed. Work organizations are indeed marked by social practices that can be said to be "cultural,"
but these practices may not span the organization as a whole. (Van Maanen and Barley, 1985,
pa 32)
3. Organizations are referred to as "culture-bearing milieus". . . . The [top of the] organization,
/
veaical and horizontal slices, and other formal unit designations [such as department] all represent
typical sites in and through which cultures may develop. . . . As such, they serve as breeding
grounds, if you will, for the emergence of local shared meanings. (Louis, 1985, pp. 75-79) '
4. [Studies] of culture most often portray organizational systems. . . . as working together in a
shared cohesive totality. The theoretical position expressed in this paper develops an alternative
stance4 perspective of organizational cultures that expects organizations to have subcultures and
allows for rival images and competing systems of meanings. (Riley, 1983, pp. 414415)
5. More researchers have emphasized the homogeneity of culture and its cohesive function than its
divisive potential. This paper suggests, however, that many organizations are most accurately viewed
as multicultural. Subgroups with different occupational, divisional, ethnic, or other cultures approach
organizational interactions with their own meanings and senses of priorities. (Gregory, 1983,p. 359)
6. From this perspective, internal conflict becomes a frequent feature of organizational cultures.
Subculturescan obviously clash over issues, programs, and missions. Also, they can exist sideby side
for long stretches of time without conflict, and clearly can be compatible. . . . In turn, the concept
of power comes into focus since it would logically be generated and differentially distributed in and
among subcultures. It follows that a political view of organizational behavior becomes relevant.
(Trice and Morand, in press, p. 8)
7. From the perspective we have elaborated, the study of cultural organization is therefore closely
bound to the study of organizational conflict. (Van Maanen and Barley, 1985, p. 48)
8. In a political-cultural approach to organizational analysis, conflict rather than being ruled out a
priori, is the ground from which interest groups collectively construct the figure of organizational
culture. (Lucas, 1987, p. 153)
9. Cultural arrangements, of which organizations are an essential segment, are seen as manifestations
of a process of ideational development located within a context of definitematerial conditions. It is a
contextof dominance (males over femalestowners over workers) but also of conflictand contradiction
in which class and gender, autonomous but overdetermined, are vital dynamics. Ideas and cultural
arrangements confront actors as a series of rules of behavior; rules that, in their contradictions, may
variously be enacted, followed or resisted. (Mills, 1988, p. 366)
10. A set of understandings or meanings shared by a goup of people. The meanings are largely tacit
among members, are clearly relevant to a particular group, and are distinctive to the group. (Louis,
1985, p. 74)
11. In a particular situation the set of meanings that evolves gives a groupits own ethos, ordistinctive
character, which is expressed in patterns of belief (ideology), activity (norms and rituals), language
and other symbolic forms through which organization members both create and sustain their view of
the world and image of themselves in the world. The development of a world view with its shared
understanding of group identity, purpose and direction are products of the unique history, personal
interactions and environmental circumstances of the group. (Srnircich, 1983a,p. 56, emphasis added)
98 CULTURES IN ORGANIZATIONS
Table 6-1 (cont.)
12. Only when members of a group assign similarmeanings to facets of their situation can collectives
devise, through interaction, unique responses to problems that later take on trappings of mle, ritud,
and value. (Van Maanen and Barley, 1985, p. 34, emphasis added)
13. A culture is conceptualized as a system of meanings that accompany the myriad of behaviors and
practices recognized as a distinct way of life. (Gregory, 1983, p. 364, emphasis added)
organization-wide consensus. At one extreme, Trice and Morand's definition (1)
assumes the existence of an overarching, "common" organization-wide culture,
in which subcultures are embedded. Louis's definition (3), like many other Differentiation
definitions, focuses on groups as the site where subcultural development
may begin, but does not explicitly exclude the possibility of organization-wide
consensus. Van Maanen and Barley's definition (2) argues that
organization-wide unity may not occur. At another extreme, some Differentiation
definitions set out to articulate an alternative to the Integration perspective by
describing subcultures as oriented toward each other, rather than emerging in
opposition to some "dominant" culture articulated by top management (see,
e.g., Riley's definition [4] and Gregory's definition [5]).24
The definitions also vary in the extent to which they emphasize power differences
and conflicts of interest between groups. Some Differentiation studies
define organizational cultures as hierarchically ordered clusters of subcultures,
putting conflict and power in the forefront of their analysis (see, e.g., Riley's
definition [4], Gregory's definition [5], and Lucas's definition [8]), while others
do not mention conflict or allow for the possibility that both organization-wide
consensus and subcultural conflict might co-exist (see, e.g., Trice and Morand's
definition [6] and Van Maanen and Barley's definition [7]).
To summarize, like Integration studies, Differentiation research defines culture
as that which is shared. Unlike Integration research, however, Differentiation
studies define the boundary of a culture at the group level of analysis,
focusing on consensus within subcultures. Some Differentiation definitions deliberately
set out to provide an alternative to the Integration approach, denying the
possibility of organization-wide consensus, while other studies allow that subcultures
might co-exist with some kind of organization-wide sharing. Differentiation
definitions vary in the extent to which they dejine culture in terms of uniqueness
and the extent to which they emphasize conflict.
As was the case with the Integration definitions of culture discussed in Chapter
4, these Differentiation definitions can be misleading. Although in some cases
these Differentiation definitions sound much like an Integration definition of
culture, the tone of a Differentiation study is quite different, even when evidence
of organization-wide consensus is being discussed. This difference is discussed
next.
THE POSSIBILITY OF ORGANIZATION-WIDE CONSENSUS
Some Differentiation research describes an organization-wide culture that coexists
with various subcultures. However, that organization-wide culture is interTHE
DIFFERENTIATION PERSPECTIVE 99
preted quite differently than it would be in an Integration study. For exampie, in a
study of the machinists on the shop floor, Young sees statements that apparently
assert organization-wide commonality as a place where subcultural differences
are subtly articulated:
In this company, the existenceof solidary values was explicitly marked by widely held
views concerning the firm, assertions of its unique qualities, and regular statements of
collective identity by shop floor groups. These were also the features of company
culture proclaimed by its managers. Yet a closerappraisal suggests that precisely these
statementsof collectivityalso constitutethe vehicles whereby different interests among
shop floor workers asserted superiority and celebrated sectional boundaries. Unity and
division existed in tandem. (Young, 1991, p. 91)
Brunsson also offers a vision of the co-existence of organization-wide unity
and subcultural development. He begins with the premise that there is sometimes
a conflict between the acceptance of inconsistency and the need for action:
Sothe problem for these organizations is how to produce both consistencyand inconsistency,
how to be both integrated and dissolved. (Brunsson, 1986, p. 174)
Brunsson argues that an organization can respond by decoupling over issues,
using some issues as a magnet for conflict (usually th&e where action is not
essential) and other issues as an arena for action. Organizations can also decouple
into subunit structures, some of which can produce conflictive talk and inconsistent
decisions, others which can produce coordinated action. Finally, organizations
can decouple over environments, responding to some segments of the
environment with evident inconsistency and internal conflict, and reacting to
other segments with clear and unified actions. Brunsson incorporates into his
description some evidence of organization-wide unity, although his emphasis on
subunit differentiation and inconsistency is more characteristic of a Differentiation
perspective. The three kinds of decoupling (regarding issues, subunits, and
environments) are described as occurring simultaneously, suggesting an unusually
complex configuration of difference.
It is important to distinguish Differentiation research from Integration studies
that acknowledge the existence of an occasional inconsistency or the rare (usually
enhancing) subculture. The cultural descriptions of Young and Brunsson, for
example, are riddled with inconsistencies. Subcultural differentiation is their
primary focus; unifying cultural elements are only a secondary consideration.
Differentiation reseaich does not see elements of organization-wide unity as
mediating or transcending the potential for chronic, deep conflicts of interest
between groups. Thus Differentiation studies do not argue that organizational
productivity is enhanced by an appropriate balance of integrating and differentiating
forces, as Blau, Chandler, and Lawrence and Lorsch have done.25
Other Differentiation studies describe managerial attempts to create organization-wide
consensus with evident skepticism and stress employeeresistance.
For example, Kunda examines employees' reactions to deliberate attempts to
create shared values through participation in rituals. To the extent that employees
132 CULTURES I N ORGANIZATIONS
These same connectionsare often ephemeral, deactivated instantly as other issues
and other individualsenter the foreground of attention. The speed of (dis)connection,
the plethora of information and problems to be solved, and the difficulty of
resolving any one issue for long-these complexities bring cognitive overload.
Demographic and international sources of diversity within organizations mean
that contacts among employees are mediated by ethnic, racial, social, religious,
and age differences, by geographical distance, and by other disjunctions of interest
and experience that are only incompletely understood. In addition, some
employees at the lowest levels of hierarchies (and their family and friends who
may be unemployed) experience a material suffering that is incommensurable
with the life experiences of higher-status executives. Taken together, these factors
create an organizational world characterized by distance rather than closeness,
obscurity rather than clarity, disorder rather than order, uncontrollability
rather than predictability.
The Fragmentation perspective brings these sources of ambiguity to the foreground
of a cultural description. Building on the complexities introduced by the
nexus approach to understanding culture, Fragmentation studies see the boundaries
of subcultures as permeable and fluctuating, in response to environmental
changes in feeder cultures. The salience of particular subcultural memberships
wax and wane, as issues surface, get resolved, or become forgotten in the flux of
events. In this context, the manifestations of a culture must be multifacetedtheir
meanings hard to decipher and necessarily open to multiple interpretations.
From the Fragmentation viewpoint, both the unity of Integration studies and the
clearly defined differences of the Differentiation perspective seem to be myths of
simplicity, order, and predictability, imposed on a socially constructed reality
that is characterized by complexity, multiplicity, and flux. When culture is
viewed from a Fragmentation viewpoint, the Integration and Differentiation perspectivesseem
to deepen confusion and misunderstanding by misrepresenting the
complexities of living in an inescapably ambiguous world.
Like the other two perspectives, the Fragmentation viewpoint is not just an
intellectual position. When researchers exclude ambiguity from the study of
culture, they are making a moral judgment about the ambiguous aspects of
contemporary life:
Many anthropologists have a kind of temperamental preference for the simplicity,
order, and predictability of less complicated societies, in which everyone knows what
everyone else is supposed to do, and in which there is a "design for living." If you
share that preference, then you can turn culture into an honorific term by denying it to
those social arrangements which donot "deserve" it, thereby making adisguisedmoral
judgment aboutthose ways of life. But that leaves a good part of modem life . . . out
of the culture sphere altogether. (Becker, 1982, p. 518)
From this point of view, if theory and research are to be relevant to problems of
contemporary organizationallife, the exclusion of ambiguity cannot be an option.
THE FRAGMENTATION PERSPECTIVE
BEYOND EXCLUSION AND CHANNELING:
THE CENTRALITY OF AMBIGUITY
Some organizational researchers, such as Weick, Daft, Starbuck, and March add
his colleagues, have long stressed the importance of ambiguity, although usualiy
without a particular focus on culture. Weick lists various sources of organizational
ambiguity that these researchers have explored:
They include things like high mobility of people among positions, faulty memories,
attempts to cope with overload by lowering the standards of acceptableperformance,
public compliance undercut by private deviation, sudden changes in authority or job
descriptions, merging of odd product lines, and the like. (Weick, 1985,p. 117)
Fragmentation studies follow in this tradition, bringing ambiguity to the foreground
of a cultural description. For example, textbook publishing has been
described as a business riddled with ambiguity. According to Levitt and Nass,
what makes a good textbook, why one succeeds and another fails, even what
makes an editor have a successful career-all these critical factors are unclear:
The editors consistentlydescribed their work in gamblingterms, such as "a lotterywith
bad odds," "an attemptto hedge one's bets," or "a crapshoot." . . . The sense of
confusion experienced by participants inhabiting this haphazard and unpredictableuniverse
is captured in the following comment from a sociology editor: "Editors can
become schizophrenic. You think a manuscript is good and it doesn't make money.
Thenyou get a manuscriptthat you think isbad, and it makes money-but not always."
(Levitt and Nass, 1989, pp. 191-192)
Social work has also been portrayed as an occupational culture permeated with
ambiguity. According to Meyerson, there is no clear definition of what constitutes
social work:
Boundariesseem unclearbecause the occupation of social work includesa wide rangeof
tasks and responsibilities, many of which are performed by members of other occupations.
In a hospital, social work can include everything from concrete discharge
planning-such as placing an individual in a nursing home-to less well-defined clinical
work with patients and families. Yet nurses also plan discharges; psychologists
counsel;and members of the clergy coordinatecommunityresources.Thus, insiders,as
1-tellas outsiders, hold diffuse ideas about what social work is and about who is and is
not a social worker. In addition, technologies seem ambiguousbecause what one does
as a social worker (e.g., talk to clients)seems loosely related to what results (e.g.,how
clients behave). (Meyerson, 1991a, p. 136)
These social workers used metaphors to express the pervasiveness of ambiguities
in their work lives. For example:
One day I was feeling real scattered and I was trying to get a good image of what that
meant for me. Tom said, "It sounds like you're trying to find a place to stand in the
134 CULTURES I N ORGANIZATIONS
middle of a kaleidoscope." And he just captured what I was feeling. (social worker,
quoted in Meyerson, 1991a, p. 137)
Similarly, Feldman sees ambiguity as a central attribute of the work of policy
analysts at the Department of Energy in Washington, D.C. In this cultural context,
as in social work, metaphors were used to describe the experience:
Another claimed that he had "never really perceived the department as a thing or an
entity. [It]is an amorphous collection of things-parts that fit together only in a rather
rudimentary way and without an obvious sense of order." (policy analyst, quoted in
Feldman, 1983b, p. 229)
According to these researchers, no adequate cultural description of textbook
publishing, social work, or policy analysis could exclude ambiguity. However, to
include ambiguity requires a new approach to thinking about culture.
DEFINING AMBIGUITY
Beforethis new approach can be described, some clear definitions of ambiguous
phenomena are needed. Ambiguity is subjectively perceived; its meaning is interpreted.2
Something is judged to be ambiguous because it seems to be unclear,
highly complex, or paradoxical. A lack of clarity occurs because something
seems obscure or indistinct, and therefore hard to decipher. Silences and absences
can also create a lack of clarity. Something is highly comaex because a
plethora of elements and relationships makes it difficult to comprehend in any
simple way. Both a lack of clarity and high complexity can sometimes be resolved
with more information or a fresh insight, making the ambiguity disappear.
Paradoxes are not so easily resolved. A paradox is an argument that apparently
derives contradictoryconclusions by valid deduction from acceptable premises.3
Ambiguity is perceived when a lack of clarity, high complexiry, or a paradox
makes multiple (rather than single or dichotomous) explanations plausible.
It is helpful to distinguish these three aspects of ambiguity (lack of clarity,
complexity, and paradox) from external sources of and internal reactions to
ambiguity. External sources of ambiguity stem from conditions external to the
person perceiving the ambiguity. For example, uncertainties in an organization's
environment,such as an unexpectedenvironmental "jolt," can cause the perception
of ambiguity.Structural variables can also be an external source of perceived
ambiguity, as when an organization's technical core is loosely coupled with the
facade it presents to the outside world.
Such external sources of ambiguity should be distinguished from the various
internal reactions people can have to ambiguity. These internal reactions range
from disgust and antipathy (because many individuals do not tolerate ambiguity
well) to joy and elation. Internal reactions to ambiguity, then, can be negative
(e.g., a reaction of debilitating confusion or action paralysis) or positive (e.g., a
reaction of feeling free of unhelpful constraints or able to innovate).
THE FRAGMENTATEON PERSPECTIVE 135
INTRODUCTION TO FRAGMENTATION: MODES OF THINKING
These definitions of ambiguity raise several questions: How is it, or is it, possible
to think about multiple interpretations in a manner that is different from the
oppositional modes of thinking characteristic of the Differentiation perspective?
Furthermore, if there is a different mode of thinking characteristic of the Fragmentation
perspective, what use is it? That is, what new kinds of understandings
can emerge from this mode of thinking that would be, in a sense, "unthinkable"
from the other two perspectives?
These questions must be answered in order to understand how Fragmentation
studies conceptualize variations on the themes of consistency and consensus.
Some of the more thoughtful answers come from postmodernism, an intellect'ual
movementthat was developed by philosophers, literary critics, feminist theorists,
and anthropologists, among others. Particularly in the United States, postmodernism
is controversial because it challenges the widely held assumption that
theory and research can increase knowledge by bringing us closer to understanding
objective truth. Most Fragmentation studies do not adopt a postmodern
framework, although those that do offer insights inaccessible to those who rely on
more traditional approaches.4Because the postmodem approach is fundamentally
incommensurable with the assumptions of most of the research reviewed in this
book, the implications of postmodernism will be discussed separately, in Chapter
10. Thefew postmodern ideas introduced in this chapter can be understood and
utilized without making postmodernism a necessary foundation for working
within the Fragmentation perspective.
Shortcomings of Oppositional Thinking
Before describing alternatives, it is important to understand why these alternatives
are needed. In other words, what is wrong with the oppositional modes of
thinking, such as dichotomies, that are used in Differentiation research? Studies
conducted from a Differentiation viewpoint use oppositional thinking to distinguish
subcultures, such as labor versus management, support versus professional
staff, men versus women, blacks versus whites. Usually, subcultures represent
ends of a dichotomy, and one of these dichotomous alternatives is viewed as
having a higher status than the other. This emphasis on superior-subordinate
relations may not be simply a reflection of how organizations are structured;some
scholars argue that any kind of dichotomous thinking is inevitably hierarchical:
Thus, whites rule Blacks, males dominate females, reason is touted as superior to
emotion in ascertaining truth, facts supersede opinion in evaluating knowledge, and
subjects rule objects. Dichotomousoppositional differences invariably imply relationships
of superiority and inferiority, hierarchical relationships that mesh with political
economies of domination and subordination. (Collins, 1986, p. 520)
One problem with dichotomous thinking is that it oversimplifiesand misrepresents
the attributes and viewpoints of members of lower status groups. When
140 CULTURES IN ORGANIZATIONS
The interplaybetween presence and absence that produces meaning is posited as one of
deferral: meaning is never truly present, but is only constructed through the potentially
endless process of referring to other, absent signifiers. The "next" signifier can in a
sense be said to give meaning to the "previous" one, and so on ad infmitum.Therecan
thus be no transcendentalsignifiedwhere the process of deferral somehow would come
to an end. (Moi, 1985, p. 106)
In this way, the postmodem approach challenges all claims to have found a single
objective "truth" that transcends all other interpretations.
If meanings lie in absences, and multiple interpretations can be made, how is a
researcher to proceed? Integration and Differentiation studies exclude or channel
ambiguity from the domain that is labeled culture, in effect excluding from
analysis all that which cannot be explained. Yet
how could we ever discover the nature of the ideology that surrounds us if it were
entirelyconsistent,withoutthe slightestcontradiction,gapor fissurethat might allowus
to perceive it in the first place? (Moi, 1985, p. 124)
Only a definition of culture 'as fragmented-marked by gaps, slides, and
silences-would enable cultural theory to explain how even the most homogeneous
cultures generate their own lacunae.9 Eagleton makes a similar point,
speaking of Macherey's view of textual analysis:
It is in the significant silences of a text, in its gaps and absences that the presence of
ideology can be most positively felt. . . . The text is, as it were, ideologically forbidden
to say certain things; in trying to tell the truth in his own way, for example, the
author findshimself forced to reveal the limits of the ideology within which he writes.
He is forced to reveal itsgapsand silences,what it is unable to articulate.Becausea text
contains these gaps and silences, it is always incomplete. Far from constituting a
rounded, coherent whole, it displays a conflict and contradiction of meanings; and the
significanceof the work lies in the difference rather than the unity between these
meanings. (Eagleton, 1976, pp. 34-35, quoted in Moi, 1985, p. 94).
The problem then is how to "read" gaps, slides, and silences.
An example may help address this problem and also clarify the difference
between a Differentiation and a Fragmentation analysis. A picture was engraved
on the Pioneer spacecraft. The image showed a nude man and woman, with the
man's arm raised in greeting. Anderson analyzed this image by offering two
dichotomous interpretations of the raised arm:it could mean goodbye or hello. In
a subsequent analysis of the same image, Owens offered other possible interpretations,
including
the same gesture could also mean "Halt!" or represent the taking of an oath, but if
Anderson's text does not consider these two alternatives that is because it is not concerned
with ambiguity,with multiplemeanings engenderedby a single sign;rather two
clearlydejined but mutually incompatiblereadings areengagedin blind confrontationin
such a way that it is impossible to choose between them. (Owens, 1983,p. 60)
Owens castigated himself for initially not noticing other interpretations:
THE FRAGMENTATION PERSPECTIVE 141
I had overlooked something-something that is so obvious, so "naturd" that it may at
the timehave seemed unworthy of comment. It doesnot seem that way to me today. For
this is, of course, an image of sexual difference, or, rather, of sexual differentiation.
. . . For in this . . . image, chosen to represent the inhabitantsof the Earth for
the extraterrestrial Other, it is the man who speaks, who represents mankind. The
woman is only represented;she is (as always) already spoken for. (Owens, 1983,pp.
60, 61)
The first interpretation of this image, by Anderson, offers two opposing interpretations
(hello and goodbye). This is the oppositional thinking characteristic of
a Differentiation viewpoint. In the second analysis, Owens draws attention to the
complexity of a seemingly simple image and delineates a variety of ways its
message could be interpreted. He also notes an absence: the woman doesnot raise
her hand in greeting, as the man does. In this analysis. Owens is working from a
Fragmentation perspective, seeing a cultural manifestation as ambiguous and
offering multiple interpretations of its meaning. Owens's analysis also illustrates
the dynamics of power working through silence (in this case, about gender).
Differance is not purely a cognitive strategy for analysis; it reveals the power
inequalities implicit in absences, the suppressed ideology that slips "between the
lines" of a text or drawing.
To summarize, a Fragmentation study uses analysis of d12erance to explore
multiple meanings, paying attention to absence as well as presence. This is a
mode of thinking that is particularly usefulfor understanding variation within
groups and revealing the ways contexts injluence interpretations. Although this
section of the Fragmentation chapter has relied primarily on gender and race as
examples, most Fragmentation studies do not focus on these particular group
identities. However, when within group variation, contextual determinants of
behavior and absences are seen as important sources of understanding, the
silenced voices of demographic minorities are more likely to be heard.
The next section of this chapter describes how Fragmentation studies utilize
versions of this mode of thinking about differance to bring ambiguity to the
forefront of a cultural analysis, reconceptualizing what is meant by the absenceof
consistency and consensus.
BEYOND CONSISTENCY AND INCONSISTENCY TO COMPLEXITY
Integration studies describe the relationship between one cultural manifestation
and another as consistent. Differentiation studies describe these relationships as
clearly inconsistent; the interpretation of one manifestation directly contradicts
the interpretation of another. Fragmentation studies move beyond clear consistencies
or clear inconsistencies to reconceptualize these relationships in multivalent
terms, as partially congruent, partially incongruent, and partially related by tangential,
perhaps random connections. In some Fragmentation studies, this reconceptualization
of (in)consistency includes a rudimentary exploration of differance.
For example, one of Meyerson's social workers saw her work in shades of
142
CULTURES I N ORGANIZATIONS
gray, rather than in the black and white oppositions favored by the Differentiation
perspective:
When they [other social workers]come to me for a simple, clear solution,I tell them:
"Life is grey. If you want black and white go to Macy's. Black and white are in this
year." (social worker, quoted in Meyerson, 1991a, p. 138)
This next section of this chapter traces patterns of differance across and down
matrices. Rather than seeing relationships among cultural manifestations as either
consistent or inconsistent, the relationships appear, from a Fragmentation perspective,
to be unclear and multivalent. Three kinds of relationships are examined
here. On a matrix these would appear horizontally between themes and
practices (action ambiguity), horizontally between themes and cultural forms
(symbolic ambiguity), and vertically among content themes (ideological am-
biguity).
Action Ambiguity I
Fragmentation studies often begin with an idealized vision of how things ought to
be. These themes are then shown to bear an unclear relationship to observed
practices. For example, some OZCO employees were confused about the relationship
between the company's espoused values regarding employee well-being
and its formal benefits policies. Sometimes this confusion arose simply because
particular employees were ignorant of the relevant policies, as when the personnel
director could not answer a question about spousal relocation benefits. And
sometimes confusion arose because the policies themselves were so complex.
Whether the relationship between espoused values and practices was unclear
because of complexity or lack of information, these OZCO employees reacted
with confusion and sometimes action paralysis: they simply did not know what to
do.
Action ambiguity was also observed at the Department of Energy:
Analysts are supposedto analyzerelatively well definedproblems and produce solutions
that can be implemented by politicians (Feldman, 1989). . . . [But] analyses do not
lead to positions that are promoted through politics. Analyses do not even support
positionschosenby politicians.Analysesarebeing produced, but it is not clearfor what
or for whom. (Feldman, 1991, pp. 154-155)
This description of analysts' decision making bears some resemblance to the
"garbage can" model of decision making. Cohen, March, and Olsen describe
this model in terms of streams of loosely coupled problems, decision makers,
choice opportunities, and "solutions" entering (and leaving) the organizational
context according to a temporal, rather than a rational-causal logic.1° Some
problems attract others, some choice opportunities never happen, some decision
makers enter or leave the decision making context, and some "solutions" get
attached to new problems, while others go unheeded.
Organizations where garbage can decision making is the norm have been
THE FRAGMENTATlON PERSPECTIVE 143
labeled organizedanarchies.l 1 In these organizations, confusion and paradox are
the rule, rather than the exception. Patterns of connection are diffuse, membership
and participation in decision making is fluid, and coordination is hard to
comeby. When it does occur, it does so often on one level (perhaps agreementon
a policy), but not on another (how to implement that policy or why that policy,is
desirable).12
The U.S. military has been described as an organized anarchy. The U.S.
militarybureaucracy includes the Department of Defense, the Officeof the Secretary
of Defense, the Joint Chiefs of Staff, and the various arms of the military.
Although the government appointees and military officers in charge do what they
can to see that decisions are made in a structured, "rational" manner,
information still becomes lost in the system, directed to the wrong people, or both.
Similarly, during a crisis, the wrong people may try to soIve a problem becauseof their
prowess at bureaucraticgamesmanship,orthe right people(becauseof mismanagement
or oversight) may be overlooked or sent elsewhere. (Sabrosky, Thompson, and
McPherson, 1982, p. 142)
When decision-making practices differ so profoundly from the rhetoric about
how such decisions are supposed to occur, the rhetoric provides no guide for
action and and so patterns of response have to be developed.
Weick provides a listing of these kinds of reactions to action ambiguity:
A loosely coupled system is a problem in causal inference. For actors and observers
alike, the prediction and activation of cause-effect relations is made more difficult
because relations are intermittent, lagged, dampened, slow, abrupt, and mediated.
Actors in a loosely coupled system rely on trust and presumptions, are often isolated,
find social comparison difficult, have no one to borrow from, seldom imitate, suffer
pluralisticignorance, maintain discretion, improvise,and have lesshubris becausethey
know the universe is not sufficiently connected to make widespread change possible.
(Weick, 1979a, p. 122)
It is worthwhile to contrast Weick's description, written from a Fragmentation
view, with the clear inconsistencies of Meyer and Rowan's discussion of loose
coupling in school systems. Where Weick sees ambiguity, Meyer and Rowan
describe clear inconsistencies between the school's formal practices (visible to
the external constituencies of the school) and the informal practices of the
teachers in the classrooms. Meyer and Rowan's teachers know what do do and
loose coupling simply buffers them from outside interference. In Weick's description,
actors are more confused about what they should do and what effect
their actions will have and loose coupling makes everything more difficult to
figureout.
A Fragmentation study seldom offers clear and comforting prescriptions for
action. When an organizational situation is ambiguous, it is difficult to know if
action is called for, which actions wolild be inappropriate, and what their consequences
might be. The frequent result is inaction.13
Fragmentation studies that focus on action ambiguity often assume that the
experience of ambiguity is noxious. This can be seen, for example, in the descriptions
of OZCO in Chapter 7 and in parts of the studies of the military and the
Department of Energy, described earlier. A negative reaction to ambiguity is also
evident in Weick's study of the factors that complicated decision making one
night at the Tenerife airport. The fog was exceptionally thick, one flight crew
(due to flight time regulations) was in a ~ s h ,and it was very difficult to turn
around large airplanes (like the two KLM and Pan American 747jets waiting for
instructions) on the small runways. In addition,
controllers at Tenerife were also under pressure because they were shorthanded, they
did not oftenhandle 747's, they had no ground radar, the centerlinelights on therunway
were not operating, they were working in English which was a less familiar second
language, and their normal routines for routing planes on a takeoff and landing were
disrupted because they had planes parked in areas they would normally use to execute
these routines. (Weick, 1991, p. 122)
The stress on both controllers and flight crews was severe that night, amplifying
the ambiguity they perceived:
As stress increases perception narrows, more contextual information is lost, and parameters
deteriorateto more extreme levels before they are noticed, all of which leads
to more puzzlement, less meaning, and more perceived complexity. (Weick, 1991,
p. 129)
The Fragmentation studies quoted in this section, including Weick's study of
Tenerife, echo Katz and Kahn's assumption that ambiguity is noxious and potentially
a threat to effective performance:
Such research began with the assumptionthat ambiguity frustrates the human need for
clarity and structure in the environment, accordingly regarded it as a stressor, and
sought evidenceof resulting strain and performance decrement. (Katzand Kahn, 1978,
p. 206, quoted in Meyerson, 1989, p. 2)
Ambiguity can also be reacted to with neutrality, as an unavoidable aspect of
organizational life that is tolerable, not dangerous, and only moderately stressful.
For example, the textbook editors and social workers studied by Levitt and Nass
were generally very aware of the ambiguity that pervaded their working lives, but
relatively few expressed strong negative reactions to it.
There is a final possible reaction that is virtually absent in these Fragmentation
accounts: ambiguity could be embraced with joy. A social worker came closest to
this perspective:
The socialworker is really likethe bastard who could go in anywhere. The socialworker
gets in between them all and can do it all. That's one of the advantages. It's certainly not
a limitation. It is a flexibility that is phenomenal. (social worker, quoted in Meyerson,
1991a,p. 142)
Although few of Feldman's policy analysts expressed the unmitigated enthusiasm
of the social worker quoted above, one analyst did acknowledge the possibility of
joy in ambiguity, describing the Department of Energy as
t
p THE FRAGMENTATION PERSPECTIVE 145
"a vehicle that can be ridden for awhile with great joy." He went on to say that the
vehiclecould alsojust "be endured. One would hope it's a vehiclethat can be steeredthough
a lotof ablepeople have dashed their hopeson that one." (policyanalyst,quoted
in Feldman, 1991, pp. 150-151)
In contrast to this analyst's half-hearted observations about joy, other sc'holars
seem to revel in ambiguity. For example, Barthes advocates7t&ing joy in am-
biguity:
Imagine someone . . . who abolisheswithin himself all barriers,all classes,all exclusions,
not by syncretismbut by simplediscardof that old specter:logical contradiction;
i who mixes every language, even those said to be incompatible; who silently accepts
i every charge of illogicality, of incongruity. (Barthes, 1975, p. 3)
I
Imagine a cultural description focused on paradoxes, grounded in the idea that
"the center does not hold," that things are neither consistent nor inconsistent,
that it is not even clear what consistency would mean in a world so confusing that
oppositions, congruencies, and orthogonalities cannot be deciphered. Few Fragmentation
studies of culture revel in ambiguity with this kind of enthusiasm, but
they could do so-perhaps capturing a previously unexplored aspect of life in
organizations.
/ Symbolic Ambiguity
' From a Fragmentation point of view, there are no clearly consistent or clearly
inconsistent relationships between themes and cultural forms (such as physical
arrangements, jokes, and organiiationai a t o ~ e s )At OZCO, for exarnpie, egalitarian
values seemed to bear no clear relationship to the physical arrangements of
the open office plan; some effects of these arrangements were apparently egalitarian,
others seemed inegalitarian, and still other effects were difficult to decipher.
Physical arrangements14 also seemed to bear an unclear relationship to espoused
egalitarian values in Tom's study of the Women's Bank in New York City:
I Of the trainers, only Elaine has an office of her own. . . . All other trainers and
trainees work in open areas. Some trainees are assigned desks of their own when their
jobs require that they have a permanent work place. Many trainees do not actually
belong in any one place at all but must sit where they can find a place close to the task
I
they are performing. All trainershave desks of their own, usually somewhatlarger than
the trainees' desks. In general, the assignment of space is confused and fluid. People
often lose things that they leave sitting in a space they had carved out for themselves
when someone else appropriates the space. People use others' desks when they are
vacant, and it is not uncommon forElaine to be forced to vacate her office. (Tom, 1986,
PP. 58, 62)
In this discussion, acknowledgment of ambiguity does not prevent the dynamics
of status differences from surfacing. However, differences between groups do not
appear as clear cut or as static as in a Differentiation account.
Some Fragmentation studies interpret cultural forms in terms that show aware-
CULTURES I N ORGANIZATIONS
ness of and ambivalence toward paradox. For example, according to Meyerson,
social workers found cynical humor a helpful response to unresolvable dilemmas:
Cynicism defused the felt ambiguitythat erupted from seeminglyunsolvableproblems
(e.g., when a patient refuses to cooperate with his rehabilitation),from irreconcilable
differences(e.g., when the attending physician ignores seemingly essential emotional ,
factors),and most frequently from situationsin which social workers lacked the clarity
or authority to take action (e.g., when they are faced with incomprehensible "red ,
tape"). By acknowledging and suspending the ambiguity with a cynical remark, the
cynic enabledthe conversation to proceed without prematureclosure: allowing unsolvable~,irreconcilables,and
untenablesto remain unresolved. (Meyerson, 1991a,p. 141)
There is a striking contrast between this cynical ambiguity acknowledgment and
the clear hostility between groups reflected in the "Kotex" and "egg in the hard
hat" jokes in Differentiation studies.
A similar cynicism was found in some of the self-deprecating jokes told by
traders working in the chaos of Wall Street investment banking firms:
What's the difference between a Wall Streeterand a pig?Not even a pig would stoop so
low. What's the differencebetween abond and abond trader?A bond matures.Did you
hear about the new Drexel bond?The maturity is twenty to life. (Abolafia, 1989,p. 14)
These jokes involve a play on difference that has echoes of differance. They are
funny because of the similarities (e.g., between a greedy Wall Street trader and
an omnivorous pig), as well as the differences between the two concepts in each
joke. The fact that this similarity is not mentioned sets up awareness of other
silences (e.g., A bond gets more worthwhile as it ages--does a trader? A greedy
pig makes better eating, etc.). These jokes are popular, in part, because their
silences reverberate; the cumulation of unspoken similarities and differences
between the two juxtaposed concepts ultimately leaves each intact and unclassifiable:
neither the same or different.
Organizational stories provide particularly rich data for this kind of analysis.
The Fragmentation view starts from the premise that a story has multiple mean-
ings:
We can also see how the very same surfacereality may embody many differentmeanings,someofwhich
may be complementaryand otherscontradictory,as when an action
signifyinggenuine friendship on one occasion may on another be hollow and perfunctory,
and on yet another, be used as a manipulative ploy. (Morgan, 1983, p. 388)
All meaning is contextual. . . . A text can be taken to have any number of contexts.
Inscribing a specific context for a text does not close orfixthe meaning of that text once
and for all: there is always the possibility of reinscribing it within othercontexts. (Moi,
1985, p. 155)
For example, when the "Rule Breaking" story recounted in Chapter 4 is regarded
from the Fragmentationperspective, the meanings of this story can vary, depending
on the context in which the story was told, who was telling it, the audience
THE FRAGMENTATION PERSPECTIVE
147
listening, and so on.I5 One person might conclude the boss in this story was
admirably obeying the rules; another might say he was self-consciousIysetting an
example. A third might wonder why he sent someone else back for the badge he
had forgotten rather than getting it himself.
The silences in this story are also eloquent. For example, the story is worth
telling only because of an unstated assumption that most top executives don't
obey company rules. In addition, the story is silent about gender differences. The
story recounts the shock experienced by the men observing the confrontation
between the low-ranking woman and her male boss. Why is the story silent about
the woman's reaction to the situation? Would the story have the same shock value
if a man, holding the subordinate position, challenged the boss? Why, in so many
versions of this story, is the woman described as short and lightweight? Whyz.is
the boss never a woman? The absence and presence of such "unnecessary"
details suggest a deeper, silenced level of meaning in this story.
A Fragmentation analysis can also illuminate reactions to the "Caesarean"
story (recounted in full in Chapter 3) told by the president of OZCO. When he
was asked what OZCO was doing to "help" women employees with children,
the president said:
We have a young woman'who is extraordinarily important to the launching of a major
new [product].We will be taking about it next Tuesdayin its first worldwide introduction.
She has arranged to have her Caesarean [operation] yesterday in order to be
prepared for this event. (Martin, 1990, p. 139)
A Differentiation analysis of this story would examine what the story says,
focusing on the inconsistency between the president's claims to be helping
women, in accord with OZCO's commitment to employee well-being, and the
fact that this woman is being praised for altering the timing of her Caesarean
operation to fit OZCO's product introduction schedule, rather than the baby's
maturational needs. Rather than "helping" the woman and her baby, a Differentiation
analysis would conclude that the company is helping itself at the woman's
and baby's expense.
A Fragmentation analysis of this story would not disagree with the points
raised by the Differentiation analysis above, but it would go deeper. A Fragmentation
analysis would focus on multiple interpretations of the story's language--
including what is not said. Martin used deconstruction (a postmodem analytic
strategy for systematically examining texts) to analyze the Caesarean story.16
This deconstruction focused on the connotations of metaphors and puns, as well
the implications of silences, revealing unstated assumptions and sexual taboos
implicit in the story's language. For example, the first phrase in the story, "we
have a young woman," could have been restated as "we employ a young
woman." The choice of the verb "have" in this sentence implies an extraordinary
degree of corporate control, in excess of the rights and duties inherentin the
usual employment contract. The phrase "having a young woman" also is a
sexual pun which has male heterosexual connotations that are repeated in other
sexual puns throughout the story.
1 4 0
L U L IUKES I N ORGANIZATIONS
In a subsequent part of this Fragmentation analysis, Martin revealed hidden
assumptions of the story by examining the effects of two small changes in the
text. Instead of a woman undergoing a Caesarean birth, the central character was
rewritten to be a man undergoinga coronary bypass operation. This small change
had massive ramifications. The story's structure, its use of metaphor, and the
nature of its sexual puns, no longer "made sense"; the hidden workings of
gender-biased ideology were revealed.
As these examples indicate, when the language in this story (its unstated
assumptions, silences, metaphors, puns, etc.) is deconstructed, we can read
"between the lines," finding traces of what has been suppressed by a dominant
ideology. The workings of power and the interests of members of relatively
powerless groups (in this story, a woman) can be exposed in this kind of Fragmentation
study.
As illustrated in this analysis of the Caesarean story, a Fragmentation study
offers a highly complex portrait of the relationship between espoused values and
the multiple meanings inherent in a cultural form. In some Fragmentation studies,
the relationship between espoused values and cultural forms is attenuated even
further, so that these two types of cultural manifestation become decoupled and
the forms loose their meaning. For example, Schultz describes how forms have
become
a hard and repetitious outward show without any underlying system of meaning. The
rituals, stories, metaphors, etc, of the culture appear isolated from the fragments of
meaning created by the members of the organization. . . . The same basic values are
repeated perpetually:In speechesand annualreports, advertisementsforrecruiting staff,
press releases and manuals for the personnel, on ceremonial occasions, speeches of
thanks and appointments. . . . The same storiesaretold perpetually:In the canteento
new members of the staff, in nostalgic moments while drinkingcoffee in the afternoon.
(Schultz, in press)
According to Schultz, these meaningless slogans, endlessly repeated stories, and
empty rituals are perpetuated, carrying a seductivepromise that is never fulfilled.
Fragmentation studies offer a complicated view of the relationship between
content themes and cultural forms. Sometimes this relationship is hard to decipher
because it is obscure or indistinct. In other instances, the relationship is
difficult to comprehend because of the complexity of relevant factors. In still
other cases, hidden traces of the suppressed interests of members of subordinated
groups complicate interpretation, or empty cultural forms are perpetuated even
though they have become decoupled from their meanings. From the Fragmentation
viewpoint, symbolic interpretation is not simple.
Ideological Ambiguity
Given the ambiguities discussed above, it should be no surprise that Fragmentation
studies also see ambiguity in the relationships among content themes. In
accord with this point of view, March has long argued that tastes, values, and
1 THE FRAGMENTATION PERSPECTIVE
1 preferences are seldom clearly consistent or inconsistent with each other. For
j example:
/ Choices are often made without respect to tastes. Human decision rnaLers routinely
ignore their own, fully conscious, preferences in making decisions.They follow $es,
traditions,hunches, and the adviceor actionsof others. Tastes changeover time in such
a way that predicting future tastes is often difficult. (March, 1978, p. 596)
If content themes operate like March's description of tastes or values, cultural
perceptions should not be organized in any way that would be recognized by the
consistency-seeking Integration study or the inconsistency-seeking Differentiation
study.
In a Fragmentation study of cultures in organizations, themes do not *rovi;e a
clarifying ideology. Instead, ideology is as ambiguous as other aspects of working
life. For example, the female management of the Women's Bankattempted to
provide benefits for women from all class backgrounds by instituting a bank teller
training program for mothers who had previously been receiving welfare payments.
The management staff who served as trainers in this program saw the
trainees as immature and ungrateful, with poor working habits and inadequate
mothering skills. These complications created unresolvable ideological tensions
for the trainers:
Theyspeakvaguely of equalopportunityforwomen, buildingwomen's financial "presence"
and sophistication,and offeringa "chance" to trainees. This vaguely-feltsense
of purpose offers the trainers little defensewhen faced with some of the contradictions
of their organization--such as explaining the Bank's refusal to grant joint checking
accounts to women and their husbands when someof them prefer such accounts and go
to another bank to get them. (Tom, 1986, p. 93)
I In Tom's study, both trainers and trainees encountered tensions that ideology did
not help them resolve.
Meyerson reported that social workers also experienced ideological ambiguity.
For example:
One social worker mentioned that being "the elbow in the system's side" was her
professional responsibility. Others viewed themselves as the patient's advocate. However,
because social workers work in organizations in which they have little formal
power, they must comply with and even become exemplars of the system to gain
legitimacy. Some admitted that their job was to uphold, even "grease" the system.
. . . Thus, although some social workers believed that their role was to change
orresist the status quo, they alsobelieved that to be effectivethey must work within and
therebyperpetuate the status quo. Social workers must simultaneouslyadvancereforms
and preclude them, critique the medical model and enforce it. (Meyerson, 1991a,p.
140)
This is a paradox-an ambiguity that is unlikely to be resolved, once and for all,
by a fresh insight or more information.
When a Fragmentation study moves beyond consistency and inconsistency it
sometimes encounters a type of paradox that is a logical "dead end." Such a
paradox (sometimes called an aporia) exposes a system riddled with irreconcilable
ideological difficulties. For example, organizational behavior professors
(like myself) are often horrified at the ways students demand simple, unambiguously
"correct" answers to the complex problems of management:
It was then suggested to the studentthat for managers to be decisive in the mannerbeing
advocated there had to be an assumption that the managers concerned actually knew
what to do. This was received with a cry from another student of, "Managers arepaid to
know what to do," followed by an aside from a further student of, "It's your job to
teach us." . . . This seemed to reflect a rather worrying conceptualisationof management
as, being-told-what-to-do. (Golding, 1987, p. 3)
Although as organizational behavior teachers, we often expect our students to
recognize and accept ambiguity, we do not hold ourselves to the same standards
when we do research on organizations. We do not often admit, when we write
about our research findings, what we do not or cannot know. l7
A few academics and more than a few managers have confronted this difficulty,
observing that organizational research does not capture context-specific
complexities of the problems managers face.18 This creates Starbuck's paradox:
Prescriptions for managing organizations have to be simple in order to be understandable.
When prescriptions describe methods and strategies which are easily translated
into actual behaviors, these prescriptions oversimplify, they ignore contingencies, and
they state half-truths. When prescriptions specify methods and strategies applying to
complete, complexsystems,these prescriptionsread likepoems that expressveritiesbut
that have obscure applications to actual behaviors. (Starbuck, Greve, and Hedberg,
1978, pp. 122-123)
Starbuck's paradox implies that helpful advice from academics (and others) is
always going to be hard to come by. The practitioner is left in the lurch, and
teachers and researchers face an aporia: they cannot do research or teach in a way
that is complex enough, and simple enough, to be both comprehensible and
useful.
In summary, the Fragmentation view offers no comfort to those--academics or
practitioners-who long for clarity. Truth claims are seen as invalid bids for
domination, ideologies become false dogmas that conceal their opposites, and
ambiguities multiply endlessly.
BEYOND CONSENSUS TO MULTIPLE,
FRAGMENTED INTERPRETATIONS
The Fragmentation perspective reconceptualizes consensus in a manner which
acknowledges that cultural members sometimes change their views from moment
to moment as new issues come into focus, different people and tasks become
THE FRAGMENTATION PERSPECTIVE 151
salient, and new information becomes available. Group identities (such as gender,
race, or job classification) do not form stable subcultures in a Fragmentation
study. Instead, multiple interpretations and reactions are always possible. For
example, when OZCO employees acknowledged ambiguities in Chapter 7, some
of them worried, others were apathetic, a few confessed ignorance, and many f9t
dismay. No clear organization-wide or subcultural consensus was evident in this
Fragmentation view of OZCO.
Multiplicity of meanings can have extremely important practical consequences.
For example, at Tenerife airport:
Five hundred and eighty-three people were killed when the KLM jet tried to
take off and hit the Pan Am airplane in its path. The Spanish government conducted
a postcrash investigation that, in part, analyzed the multiple interpretations
and misunderstandings implicit in the conversation reported above. In addition
to these cognitive confusions, the government investigators were sensitive to
the power dynamics implicit in gaps and silences. The report focused on the
silence that occurred after the pilot replied with an emphatic "yes":
Perhaps influenced by [thepilot's] great prestige, making it difficult to imaginean error
of this magnitude on the part of such an expert pilot, both the copilot and flightengineer
made no further objections. The impact took place about 13 seconds later. (Spanish
Ministry of Transport and Communication's report on the crash, p. 71, quoted in
Weick, 1991, p. 121)
truction of the Caesarean story,
an analysis of the ambiguities of silences brings the hidden dynamics of power
inequality into focus, illuminating why widespread consensus is unlikely.
Fragmentation studies often portray goals as unclear, making consensus very
difficult to achieve. For example, in the study of the military:
No consensus exists on either the proper means of assuring the common defense, or the
preferred intermediate goals one ought to pursue. Without agreement on goals and
general force types, there can be no agreement on programs. And without agreement on
the definitionof acceptablegoals and adequate forces, "progress" toward either cannot
be measured readily. (Sabrosky, Thompson, and McPherson, 1982, p. 142)
Lack of consensual understanding can also be seen in one of the few Fragmentation
studies to mix quantitative and qualitative methods. Krackhardt and
KiIduff used constructs generated by employees in open-ended interviews, so
After the KLM plane made the 180degree turn at the end of the takeoff runway,rather
than hold as instructed, they started moving and reported, "we are now at takeoff."
Neither the air trafficcontrollersnor the Pan Am crewwere certain what this ambiguous
phrase meant, but Pan Am restated to controllers that they would report when they were
clear of the takeoff runway, a communique heard inside the KLM cockpit. When the
pilot of the KLM flight was asked by the engineer, "Is he not clear then, that Pan
Am?", the pilot replied "yes" and there was no further conversation. The collision
occurred 13 seconds later at 5:06 P.M.(Weick, 1991, pp. 118-120)
their data are less superficial and have fewer of the researcher-generated constraints
of many other quantitative studies of culture. After a series of statistical
analyses, the authors conclude:
n e results demonstrate that individuals in a joint enterprise can construe the same
interpersonal reality in completely different ways on constructs that capture the main
dichotomies in the organization. The puzzle of how people with different perspectives
succeed in enacting roles toward each other in organizations has partly been explained
by the findingthat the diversity of attributions is patterned by the network of personal
friendships. (Krackhardt and Kilduff, 1990, pp. 18-19)
The language in the last sentence is important. The heterogeneity is explained
only in part by the friendship network and that network is more a matrix of pairwise
relationships than it is a set of subcultures.19 This is a quantitative image of
an absence of organization-wide consensus. From a Fragmentation viewpoint,
this study's absence of consensus does not indicate an absence of culture, but
rather the presence of a Fragmented culture. This conclusion is, in part, an effect
of how culture is being defined.
DEFINING CULTURE FROM A FRAGMENTATION PERSPECTIVE
The definitions of culture offered in Integration and Differentiation studies often
bear relatively little relationship to what is actually studied. In contrast, Fragmentation
studies often abstain from defining culture at all. The exceptions, however,
are informative and directly relevant to the conclusions of this kind of research.
A reviewer of one of my papers suggested that I seemed to be defining culture
in these terms:
Culture is a loosely structured and incompletely shared system that emerges dynamically
as culturalmembers experienceeach other, events, and the organization's contextual
features.(anonymousreviewer, 1987)
This definition does not exclude ambiguity from the domain of culture and it
captures the sense of constant flux that is implicit in many Fragmentation studies.
However, this definition leaves open a question that is particularly relevant for
organizationai researchers: Given this ambiguity, how do people engage in coordinated
action? How is this definition a socially constructed reality, rather than a
collection of unique individual views? Meyerson has one answer to this dilemma:
Members do not agree upon clear boundaries, cannot identify shared solutions, and do
not reconcile contradictorybeliefs and multiple identities. Yet, these members contend
they belong to a culture. They share a common orientation and overarching purpose,
face similar problems, and have comparable experiences. However, these shared orientations
and purposes accommodate different beliefs and incommensurabletechnologies,
these problems imply different solutions, and these experiences have multiple
meanings . . . Thus, for at least some cultures, to dismiss the ambiguitiesin favor of
strictly what is clear and shared is to exclude some of the most central aspects of
members' cultural experience and to ignore the essence of their cultural community.
(Meyerson, 1991a, pp. 131-132)
Feldman offers an issue-specific response to the same dilemma:
As others have noted (Martin and Meyerson, 1988; Van Maanen and Barley, i985)
culture does not necessarily imply a uniformityof values. Indeed quite differentvalues
may be displayedby people of the sameculture. In such an instance, what is it thatholds
together the members of the organization?I suggestthat we look to the existence of a
common frameof referenceor a sharedrecognition of relevant issues. There may not be
agreement about whether these issues should be relevant or about whether they are
positively or negatively valued. . . . They [may] array themselves differently with
respect to that issue, but whether positively or negatively they are all oriented to it.
(Feldman, 1991, p. 154)
This issue-specific focus is essential to understanding how the Fragmentation
perspective defines culture. A focus on specific issues addresses one of the
problems inherent in the Differentiation approach to understanding culture. In a
Differentiation study, clear dichotomies and clearly defined subcultures cut up
and constrain the ways we see a culture, making unintelligible all the issuespecific
complexities that "fall between the cracks" of this way of thinking. In
contrast, the Fragmentation perspective sees issues as connecting individuals in
temporary, issue-specific coalitions. Other individuals and other issues are linked
in different, overlapping, temporary patterns of connection.20
One metaphor for this approach to understanding culture is a web:
Individuals are nodes in the web, connected by shared concerns to some but not all the
surrounding nodes. When a particular issue becomes salient, one pattern of connections
becomes relevant. That pattern would include a unique array of agreements, disagreements,
and domains of ignorance. A different issue would draw attention to a different
pattern of connections-and different sources of confusion. Whenever a new issue
becomes salient to culturalmembers or researchers, a new pattern of connectionswould
become significant. (Martin and Meyerson, 1988, p. 117)
Sometimes these issues can be content themes. For example, the Fragrnentation
view of OZCO presented in Chapter 7 focused on three content themes:
egalitarianism, innovation, and employee well-being. Connecting concerns can
also be more specific-a particular problem or a shared set of tasks. From a
Fragmentationperspective, then, an organizational culture is a web of individuals,
sporadically and loosely connected by their changingpositions ona variev
of issues. Their involvement, their subcultural identities, and their individual
self-definitionsjuctuate, depending on which issues are activated at a given
moment.
A jungle metaphor for the Fragmentation view of culture captures some of the
complexity evident in the web metaphor and also retains more of an emphasis on
the unknown and the unknowable. In Chapter 4, culture was defined, from an
Integration perspective, as that which is clear, "an area of meaning cut out of a
vast mass of meaninglessness, a small clearing of lucidity in a formless, dark,
always ominous jungle."*' Rather than denying or channeling ambiguity, the
Fragmentation perspective accepts it and makes it the focus of attention. In this
view, culture is no longer the clearing in thejungle; it is thejungle itself. Table 1-
1 in Chapter 1contrasts the characteristics and metaphors used to define culture in
Fragmentation studies to those in Integration and Differentiation research.
FRAGMENTATION AT DIFFERENT LEVELS OF ANALYSIS
Some critics have argued that a Fragmentation view reduces culture to an individual
level of analysis because this perspective sees no stable organization-wide or
subcultural consensus. This is a misunderstanding. The Fragmentation perspective
offers a distinctive approach to the subcultural, individual, and organizational
levels of analysis, as described next.
Fragmentation at the Subcultural Level
At the subcultural level of analysis, a Fragmentation study avoids the clear group
boundaries and the clearly defined subcultural relationships (enhancing, conflicting,
orthogonal) of the Differentiation perspective. Instead, a Fragmentation
view of subcultures portrays boundaries as permeable, subcultural membership as
fluctuating, and relations among subcultures as multivalent. In other words:
Group alliances look like affinities or coalitions rather than identities, and they are
characterizedby fluidity, the abilityto mobilize and disperse. Sometheoristscall them
microresistances. (Gagnier, 1990, p. 23)
When subcultures are regarded from a Fragmentation rather than a Differentiation
perspective, an examination of differancereplaces an oppositional mode of
thinking. A Differentiation study of gender subcultures, for example, would start
with a men versus women dichotomy. In one such study, Millman and Kanter
observe that men often fail to see that gender-segregated groups exclude women:
When male sociologists(ormen in general)look at a meeting of a board of trustees and
see only men, they think they are observing a sexually neutral or sexless world rather
than a masculine world. (Millman and Kanter, 1975, p. xiv)
This is a Differentiation analysis because it makes a generalization about men in
tacit, binary opposition to women. It assumes that all male sociologists are
"essentially" the same, with unified interests that stand in opposition to the
interests of an equally homogeneous grouping of women.22 Thus, all male sociologists
would see an all-male board of trustees in sexually neutral terms, while
presumably all female sociologists would not.
In contrast, a Fragmentation study would offer a far more complex analysis of
this situation. Such an analysis would not deny the bias inherent in seeing a sexsegregated
world as sexually neutral. Both Differentiation and Fragmentation
studies work from the premise that people notice what they are comfortable
I Ht tKAtiMENTATION PERSPECTIVE 155
noticing; inequities from which they benefit will be ignored. However, a Fragmentation
study would go further, analyzing how Millman and Kanter's generalization
masks important differences between and within groups.
A Fragmentation analysis might begin by observing that Millman and Kanter
do not mention the race of the trustees and the sociologists. Are the trustees all
white? If so, why is this not mentioned by the male sociologists or by Kanter and
Millman?Are all these researchers white? Next, genderand race might be considered
simultaneously. Would black male sociologists tend to be sensitiveto racial
segregation among the trustees, but perhaps blind to sexual segregation?Would
black female sociologists tend to be sensitive to both?
Finally, the possibility of variation within all these groups might be considered.
Millman and Kanter imply that because the male sociologists are members,
of the dominant gender, they are blind to the gender segregation of the all-male
group of trustees. However, membership in a dominant group does not necessarily
mean that one is blind to the exclusion of subordinate group members.
Some male sociologists might notice the gender segregation of the trusteesgroup.
Furthermore,some female sociologists might be gender-blind, seeing the trustees
in sexually neutral terms. Similarly, some whites might be sensitive to racial
segregation, while some blacks might not focus on this issue.
The complications revealed by a Fragmentation analysis are not merely "error
variance" that can be easily dismissed as trivial compared with group averages.
For example, women have different ethnic, class, and racial backgrounds and
these sources of diversity mean that women can see things very differently, have
similar or conflicting interests, and benefit from diverse and possibly diverging
paths of social action. Similar diversity exists for men. In addition, denying
differences within a subordinated group can further the subordination of minority
members within that group. And denying the differences within a dominant group
can mean the disregard of potential allies for those who are subordinated. These
observations are as relevant for other groups (based, e.g., on occupation or
hierarchical level) as they are for demographic groupings. Generalizations based
on dichotomous statements about between group differences can often mask
differences that are critical for understanding inequality or working toward
change.
A FragmentedView of the Self
This approach to the study of subcultures also has implications for the individual
level of analysis. When differences within categories are acknowledged, the
boundaries of subcultures become diffuse, permeable, and fluctuating. Because
these boundaries are subjectively construed and socially constructed, one person
sees them differently than another. Subcultures overlap, they are nested within
each other, and they intersect in the individual:
One might plausibly keep dissecting levels of subculturesuntil one reaches the level of
the individual.Each individual does, in fact, live in at least a slightly different subculture
or intermeshing of subcultures. (Sternberg, 1985, p. 1116)