Theory Development

Constructing Mystery: Empirical Matters in Theory Development
Author(s): Mats Alvesson and Dan Krreman

Source: The Academy of Management Review, Vol. 32, No. 4 (Oct., 2007), pp. 1265-1281
Published by: Academy of Management
Stable URL: http://www.jstor.org/stable/20159366
Accessed: 18-05-2015 08:19 UTC
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/
info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content
in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact support@jstor.org.
Academy of Management is collaborating with JSTOR to digitize, preserve and extend access to The Academy of Management
Review.
http://www.jstor.org
This content downloaded from 139.179.82.188 on Mon, 18 May 2015 08:19:56 UTC
All use subject to JSTOR Terms and Conditions
?
Academy o?Management Review
2007,Vol. 32,No. 4, 1265-1281.
CONSTRUCTINGMYSTERY:EMPIRICAL
MATTERSINTHEORYDEVELOPMENT
MATS ALVESSON
DAN K?RREMAN
Lund University
We outline a research methodology developed around two basic elements: the active
discovery and/orcreation of mysteries and the subsequent solving of the mysteries. A
key element is the reflexive opening up of established theory and vocabulary through
a systematic search for deviations from what would be expected, given established
wisdom, in empirical contexts. "Data" are seen as an inspiration for critical dialogues
between theoretical frameworks and empirical work.
How do we develop theory? Broadly speaking, capable of showing the right route to theory or
we can rely on speculative thinking or empirical screening out good ideas frombad. Rather, empir
observation (followed by careful analysis). Some ical material is an artifact of interpretations and
have argued that empirical material has no sys the use of specific vocabularies. Data are inextri
tematic role to play in theory building. Popper cably fused with theory. Acknowledging this fu
(1963, 1972), for example, compared theory cre sion?which is broadly accepted in the philoso
ation with guesswork and
explicitly called un phy of science (Denzin & Lincoln, 2000; Gergen,
justified (or unrefuted) theories "conjectures." 1978; Hanson, 1958; Kuhn, 1962)?has major conse
Others have tended to rely heavily on and per quences forhow we consider the theory-empirical
haps overplay the importance of empirical ma material relationship.
terial?often viewed as data. We emphasize the potential of empirical ma
Typically, theory is claimed to be developed terial as a resource for developing theoretical
either through discovery?by sifting through da ideas through the active mobilization and prob
ta?or by the accumulation of verified (or cor lematization of existing frameworks. In particu
roborated) hypotheses. These views of social lar, we point to the ways empirical material can
science are in many ways different, but each be used to facilitate and encourage critical re
relies on data as the central elements in social flection: to enhance our ability to challenge, re
research. Theory is supposed to "fit" data? think, andillustrate theory. This approach rec
either by design, where misfit should lead to ognizes the constructed nature of empirical
rejections or revisions of theory (Fetterman, material and "proofs" (Astley, 1985; Shotter, 1993;
1989), or by default, where theory is understood Shotter & Gergen, 1994; Steier, 1991). It advocates
as emerging from data (Eisenhardt, 1989; Glaser a light or moderate version of constructionism?
& Strauss, 1967; Strauss & Corbin, 1994). assuming that something is going on out there
In this paper we adopt a different approach. In and there may be better or worse ways of ad
conventional terms, we focus on the discovery (or dressing things, but also that the frameworks,
creation) of theory, rather than its justification. Al preunderstandings, and vocabularies are cen
though we find novel approaches toward the re tral in producing versions of the
particular
finement and justification of theory valuable, we world. We propose a relaxation of the emphasis
aim for more creative ways of theorizing. Like on "data" and a greater interest in the contribu
many others, we claim that data?or, our pre tion of how data are constructed for the benefit
ferred term, empirical material?are simply not of theoretical
reasoning (cf. Sutton & Staw, 1995).
A key element here is the role of empirical
material in inspiring the problematization of
We are grateful to guest editor John Van Maanen, the
theoretical ideas and vocabularies. To prob
anonymous reviewers, Andy Van de Ven, and Karen Lee
Ashcroft for helpful and and the
lematize means to challenge the value of a the
challenging comments,
Vinnova research foundation for a research grant on devel ory and to explore itsweaknesses and problems
oping qualitative methodology. in relation to the phenomena it is supposed to
_1265_
Copyright of theAcademy o?Management, all rightsreserved.Contents may not be copied, emailed, posted toa listserv,or otherwise transmitted
without thecopyright
holder's express written permission. Users may print,download, or email articles forindividual use only.
1266 Academy of Management Review October
explicate. Itmeans to generally

open up and to We are
inspired by Asplund's (1970) stimulat
point out the need and possible directions for ing idea of social science as involving two ele
rethinking and developing the theory. We con ments: the discovery or creation of a breakdown
sequently attempt to develop a methodology for in understanding of theoretical interest (the con
theory development through encounters be struction of a mystery) and the recovery of un
tween theoretical assumptions and empirical derstanding (the resolution of the mystery).2 In a
impressions that involve breakdowns. It is the sense, our project also shows an affinity for
unanticipated and the unexpected?the things Poole and Van de Ven's (1989) suggestion to view
that puzzle the researcher?that are of particu paradoxes as resources for theorizing. However,
lar interest in the encounter. In this sense our in contrast to Asplund's and Poole and Van de
approach attempts to take systematic advan Ven's strong focus on armchair theorizing, we
tage of what Robert Merton calls "serendipity"? pay particular attention to the interplay be
that is, "the art of being curious at the opportune tween theory and empirical material, thus focus
but unexpected moment" (Merton & Barber, 2004: ing on how inconsistencies and breakdowns de
210). Accordingly, theory development is stimu rived from empirical observation, rather than
lated and facilitated through the selective inter (pure) theoretical speculation, may help us de
est of what does nof work in an existing theory, velop theory. Chiefly, our goal is to explore how
in the sense of encouraging interpretations that empirical material can be used to develop the
allow a productive and noncommonsensical un ory that is interesting rather than obvious, irrel
derstanding of ambiguous social reality. evant, or absurd (Davis, 1971).
The empirical material, carefully constructed, Theorization may be understood as disci
thus forms a strong impetus to rethink conven plined imagination (Mills, 1959; Weick, 1989).
tional wisdom. However, the ideal is nof, as in Empirical material can facilitate theorization
neopositivist work, to aim for an "intimate inter because it provides resources for both imagina
action with actual that "produces
evidence" the tion and Breakdowns
discipline. create spaces
ory which closely mirrors reality" (Eisenhardt, where imagination can be put to work. And al
1989: 547).1 The empirical material may be mo though empirical material never exists outside
bilized as a critical dialogue partner?not a perspectives and interpretative repertoires, it
judge or a mirror?that problematizes a signifi nevertheless creates a relative boundary for
cant form of understanding, thus encouraging imagination. Some constructions make more
problematization and theoretical insights (cf. sense than others. Empirical material anchors
Ragin, 1987: Chapter 9). The dialogue metaphor the process of theorization in specific claims
is not uncommon in contemporary qualitative about the object under study, thus prohibiting
research. Emphasizing the critical aspect of the arbitrary ideas from being put into play.
ory as well as data construction?involving Exploiting breakdowns is, of course, not new
careful consideration of alternative representa to social science. In particular, in ethnographic
tions?frames the enterprise somewhat differ work the initial difference between the tradi
ently from established views. We think it is im tions involved (the researcher's and the topic of
portant to draw attention to (the construction of) study) produces breakdowns in understanding:
friction (as a potentially productive force) rather "A breakdown is a lack of fit between one's
than harmony in the interplay among theory, encounter with a
tradition and the schema
researcher subjectivity, and empirical material. guided expectations by which one organizes ex
perience" (Agar, 1986: 21). The researcher re
solves this problem by trying to understand the
1 cultural elements causing the breakdown and
Neopositivism (or postpositivism) assumes the existence
then adjusting the research schema. Break
of a reality that can accurately but imperfectly and proba
be the observer and the observed downs continue to appear until the researcher
bilistically apprehended,
separated, and data and theory treated as separable, al fully understands the studied culture. This
though the theory ladenness of data is acknowledged. The
aim is to produce generalizable results (Lincoln & Guba,
2000). Most contemporary social research and 2Asplund (1970) developed two metaphors for creating
quantitative
research like grounded theory (although there novel understanding of social reality: the riddle and the
qualitative
are different versions of the latter; Charmaz, 2000) appear to crime mystery. In this paper we use a generalized version of
be based on neopositivist the mystery metaphor as a device for developing theory.
assumptions.
2007 Alvesson and K?rreman 1267
means that ethnography can be described "as a THE FICTION OF "FACTS" IN

process of coherently resolving breakdowns" ORGANIZATIONAL ANALYSIS
(Agar, 1986: 39). In this sense ethnography has a
In interpretive and reflexive research, scholars
built-in propensity toward the type of theory de
view data as constructions, created through inter
velopment we outline in this paper. action between the researcher and the group un
However, ethnography is far from the only
der study (Denzin & Lincoln, 2000; Rorty, 1979;
method that can take advantage of breakdowns
new theoretical ideas. An example Rosenau, 1992; Van Maanen, 1988). Since the met
fordeveloping
of quantitative studies producing a breakdown is aphor "data collection" tends to be understood far
too literally and, thus, is potentially misleading,
Lincoln and Kalleberg's (1985) piece on job satis
we prefer, as noted above, the expression "empir
faction and organizational commitment among
ical material" as a representation of what is con
U.S. and Japanese workers. The result showed
ventionally understood as data. The metaphorical
higher scores for the former,which certainly was
quality of "material" indicates that we, as re
surprising. The "mystery" can possibly be solved
searchers, must actively do something with it.
through seeing questionnaire responses as less
in
With this mind, we use the twomentioned terms
objective measurements of objective phenomena
interchangeably. A key assumption is that "in the
than clues to cultural norms for expressions and
social sciences there is only interpretation. Noth
the following of language rules (Alvesson & Deetz,
2000). Another example is the classical Hawthorne ing speaks for itself" (Denzin, 1994:500). Sensitivity
on how to language is vital. Most conventional research
studies, which started with experiments
ers assume that language operates as a kind of
light affects performance and ended with open
ended ethnographic research that explored radi medium, albeit imperfect one owing to noise,
an
cally new ideas on the dynamics of workplace distortion, ambiguity, which

and ideally mirror
social interactions?a shift clearly encouraged the world "out there." However, the linguistic turn
by
in social science has attacked this language as
empirical material that challenged the initial
frameworks of the researchers mirror perspective (cf.Alvesson & K?rreman, 2000;
(cf. Schwartzman,
1993). Deetz, 1992; Potter & Wetherell, 1987; Rorty, 1979),
Our objective in this paper is to suggest an pointing to the ways all observation and all data
to theory development that uses the are theory laden and embedded in language.
approach
ory and imagination to critically open up alter Thus, vocabularies simply don't mirror the
world. They produce and conceal as much as
native ways of framing empirical material. We
follow a large amount of work in methodology, they reveal. The language used in a study to a
of, for exam extent determines the results. Theories
including significant contributions large
can be understood as repertoires of lenses
ple, Mills (1959),Garfinkel (1967),Davis (1971),
Weick (1989),Becker
Gergen (1978),Peirce (1978), (Deetz, 1992), each providing and communicat
(1996), and many others in philosophy of science ing particular understandings. This metaphor
and interpretive social science. Critical reflec points out the productive and pragmatic charac
tion, theory-driven disclosure, and the specific teristics of language. Language is a human ar
of working with breakdowns and tifact that affects our vision?blurring, clarify
procedure
combine to create an overall method ing, magnifying, and diminishing the things we
mysteries
see through it. From our point of view, theories
ology. This process systematizes attempts to ex
do not express the underlying engines of gener
plore new terrain and develop novel ideas, thus
potentially overcoming the inherent conserva alized empirical patterns. Rather, they are in
tism in well-established frameworks. In this pa struments that provide illumination, insight,
per we focus on exploring a maximalist version and understanding. In this sense theories oper
of breakdown-induced theory development. ate as idealizations (Freese, 1980). Our concep
However, we also briefly address broader strat tion of theory may be looser than the mantra of
egies for taking advantage of breakdowns for explicitness, abstractness, discreteness, syste
theory development. Our ambition is not to try to maticity, and completeness. It is, however, more
colonize empirical research through a specific useful?as Shotter (1993: 113) points out, few if
design but, rather, to provide some overall any theories meet the criteria above.
guidelines and concepts potentially useful for From this perspective, empirical evidence is
novel theorizing. constructed within particular paradigmatic and
linguistic conventions and is typically less ro self-critique (cf.Mills, 1959;Weick, 1989). Reflexiv
bust when approached from any other angle ity enters the picture (Alvesson & Sk?ldberg, 2000;
(Gergen, 1978). Most interesting (complex) ideas Calas & Smircich, 1999;Hardy & Clegg, 1997),
cannot be easily "checked" against data, and pointing to the struggle to acquire an awareness
empirical measures are always contestable. Or of how paradigms, sociopolitical contexts, frame
ganizations, for example, are complex, dynamic, works, and vocabularies are involved in shaping
and difficult to observe.
Rigorous studies have the researcher's constructions of theworld at hand
their limits, and the researcher has to depend on and his or her moves in doing something with the
pictures, maps, and metaphors (Morgan, 1980; world.
Weick, 1989). Social changes?partly fueled by REFRAMING THE RESEARCH OF

social science itself?tend to render empirical
ORGANIZATIONAL PHENOMENA
findings obsolete over time (Cronbach, 1975).
What is an interesting research problem? As
Ideas about evidence, rea
empirical objectivity,
son, truth, coherence, measurement, and
we see it, an interesting research problem
validity,
fact no longer great comfort or direction. includes the high potential for an empirical re
provide
If such concepts are relative, not absolute, they sponse a novel
and insight that adds signifi
are always contestable in whatever form they to?or understand
cantly against?previous
appear?although this is not to say that such
ings. Contrary to conventional wisdom, we think
are rendered irrelevant or un
concepts thereby
that it is fruitless, even counterproductive, to
thinkable (Van Maanen, 1995: 15).
attempt tominimize the influence of theory and
Values other than verification become impor These should not be denied and
subjectivity.
tant for the assessment of the value of a theo hidden but should be reflexively and self
retical contribution: "Theories favor be
gain critically cultivated and mobilized, reinforcing
cause of their conceptual appeal, their logical the ability to discover research is
interesting
structure, or their psychological plausibility. In sues. As Weick puts it, "Whenever one reacts
ternal coherence, parsimony, formal elegance, with the feeling thafs interesting, that reaction
and so on prevail over empirical accuracy in is a clue that current experience has been tested
determining a theory's impact" (Astley, 1985: and the past under
against past experience,
503). Although we do not advocate solipsism, has been found (1989: 525).
standing inadequate"
relativism, or an exclusive focus on the rhetori In order to make this experience more valu
cal qualities of research texts and theories, we able and relevant, it must be abstracted and
think there are good reasons tomove from made more general. We address this through
such as sensitive construc
concepts reflexivity,
a strong focus on data to an interest in the
construction of empirical material;
tions, and interpretive repertoires. More specifi
a view of theory and data as separate to an cally, we suggest that theory-developing orga
acknowledgment of the "internal" relation nizational research is characterized by
ship between them?the theory impregna
research themes that can be in
tion of all data; and empirically
a on the procedures and vestigated?empirical material that carries
strong emphasis
some credibility, meaning that it is capable
techniques for "collecting" and analyzing
data to a greater interest in researcher re of offering clues for thinking and themaking
and
flexivity in dealing with the empirical ma
of claims and/or counterclaims,
terial?that is, how to interpret and reinter ideas that offer challenges to conventional
within an area, pointing at short
pret the material. thinking
comings or paradoxes; this requires an in
From this perspective, the acts of construction? tensive empirical material/theory interplay
is also
always guided by theory in some form?become
where theory used "negatively"?a
resource is (models, vo
central. The knowledge and the person doing significant theory
cabularies) that fails to be useful to account
knowledge work/development cannot be sepa for a phenomenon, which does not a
imply
rated (Calas & Smircich, 1992). The framework, the Popperian ideal of falsification but can be
researcher, and social reality?inescapably repre seen as a chance for problematization, a
sented through potentially contested representa vital element in theory development as we

see it.
tions?are thus always interrelated and provide
an interconnected net of potential insights and The inference mechanism that guides this
ideas, ideally cultivated through discipline and kind of theory development is usually labeled
abduction (Peirce, 1978). It consists of three challenged, surprised, bewildered, and con
steps: (1) the application of an established inter fused may take center stage in research.3 The
pretive rule (theory), (2) the observation of a sur researcher's preunderstanding, including his or
prising?in light of the interpretive rule? her academic framework(s), may be be used as a
empirical phenomenon, and
(3) the imaginative tool that opens up a dialogue with the empirical
articulation of a new interpretive rule (theory) material. The
dialogue needs to include the
that resolves the surprise. This approach in reader. The
researcher is normally a part of a
cludes an interest in problematizing and re broader we, which includes the research com
thinking dominating ideas theory, when
and munity (or communities) that the researcher be
empirical impressions encourage such need for longs to and which informs preunderstanding
novel thinking. The rationale for this is that "the and preferences. How this community is tar
contribution of social science does not lie in geted, convinced, and challenged are key issues
validated knowledge, but rather in the sugges in doing field work, interpreting empirical ma
tion of relationships and connections that had terial, and?even more
so?crafting
a text.
not previously been suspected, relationships Key elements in this project are
that change actions and perspectives" (Weick,
a flexible theoretical framework requiring
1989:524).
multiple readings of the talk, the behaviors,
This way of looking at empirical material the events, and the documents one faces in
means that its dialogic qualities are empha fieldwork, and
sized. The researcher must call upon or actively a reflexive approach to empirical material
that alternative constructions
try to reach empirical material that can produce, encourages
and the self-critical interpretations of one's
or inspire the construction of, a variety of alter
own theoretical,
paradigmatic, political,
native "stories." Thus, the process of engage and social
methodological, predispositions.
ment, in which the languages and theories of
are activated, is central. This Without the first element there is insufficient
the researcher
direction or an inability to produce sufficiently
view differs from a position aiming to passively
mirror reality?for open and challenging observations and inter
example, through collecting
pretations, which can then be picked up as op
data and coding, processing, and trying to "dis
portunities for breakdowns and problematiza
cover" the facts and meanings that are assumed
tion. Without the second element the empirical
to be already present. For instance, when con
statements of
research
material may not be dealt with in sufficiently
sidering subjects?
or through observa rich and varied ways to engage in a critical
whether in interviews
tion?we can see these not just as possibly dialogue with theory. Our point is that we do not
the meanings just encounter empirical material and see where
revealing of those studied (or
it leads us. Rather, we are always doing some
facts about their organizations) but as political
action, moral story telling, identity work, script thing with it?framing and constructing it. A
so forth (Alvesson, 2003). Rather careful consideration of alternative construc
application, and
tions is necessary in order to produce a dialogue
than assume that the subject is reporting au
we can see the subject as a that may be theoretically inspiring and innova
thentic experiences,
politically motivated producer of what are, for
him or her, favorable "truths," or as a person 3
We realize that there are many ways in which research
repeating institutionalized standard talk about ers of different camps and with various personal convictions
a specific theme. Thus, interview talk can be work. Some people, in associating themselves with
seen as useful for a study of political action grounded theory, would probably share Strauss and Cor
bin's (1990, 1994) beliefs that objectivity, and
or the circulation of discourse, rather than for a reproducibility,
unbiased data collection provide a robust base for theory
study of the experiences, meanings, and beliefs
building; others would open up more constructivist consid
of individuals. erations (Charmaz, 2000). A strict focus on coding would
The proposed view?sensitive construc probably, formost, mean a minimization of researcher sub
tions?is different from most conventional ap jectivity for the benefit of reliable procedure. One may, how
ever, work with coding in different ways, do multi
guided by a desire to order and con perhaps
proaches, based on rereadings and of one's
ple codings, reframings
trolwhat is studied. But the impulse to control?
position, take incoherences and contradictions seriously,
through measuring, codifying, checking, and so and generally try to open up experiences of productive
on?can be bracketed, and a desire to become breakdowns.
tive, transcending the received wisdom and pre key elements here are finding ways of encounter
ferred line of constructing. ing breakdowns and creating mysteries. Below,
An important question?and an exercise of re we outline a methodology for doing this.4 (See
flexivity?is to ask oneself, "Can I construct/make Figure 1.)
sense of this material in another way than sug
1. Familiarizing oneself with the setting under
gested by the preferred perspective/vocabulary?
study and making inquiries about themes
Can I let myself be surprised by this material? in a fairly open way: This is based on pre
Can itproductively and fairly be constructed in a liminary decisions on a field of interest
way that kicks back at my framework and how and an initial, fairly broad focus for the
see Rather than on nar
we?in my research community?typically investigation. focusing
row themes?for
and reconstructions example, "knowledge
interpret things?" Such
sharing," "teamwork," or "leadership"?one
should meet the criterion of being well supported can ask oneself, "What is going on here?" or
by the empirical material (assuming that this can "What do the natives think they are up to?"
support different constructions) and should be as Obviously, a study must have a degree of
sessed to have some theoretical potential. The se direction. The trick is to balance thiswith a
capacity to expose oneself to something un
rious consideration of alternative representations
expected, something that can't easily be
and interpretations thus becomes crucial towork
disciplined by the preferred vocabulary and
that encounters empirical reality. Reflexivity can framework and too narrow of a research
be encouraged by using various theoretical per question. One may, for example, start with,
but not stick to, an idea of
spectives and metaphors, listening to alternative necessarily
"knowledge" being "shared," workers hori
voices of the research subjects, imagining multi
zontally coordinating their work, or manag
ple reader groups, considering different political ers influencing their subordinates' meaning
interests and research purposes (emancipation, constructions, and then see what may turn
thick description, better management), trying to up?what one may produce?in terms of un
consider oneself in various identity positions (gen expected empirical material in that kind of
area, broadly defined. Reflexivity here
der, ethnicity, class), working with coresearchers could involve a critical awareness of the
from another background or with a different theo risks of imposing and sticking to a set of
retical framework, and thus increasing the chance favored themes and a willingness to invoke
to be challenged when encountering empirical alternative themes, vocabularies, and un
Issues around and
material. The dialogue among framework, re derstandings. politics
ethics may also enter here: Who may bene
searcher, and empirical material should be, fit from studying a specific set of phenom
whenever possible, multilingual. ena in a particular way?
Of course, all this leads to considerably more 2. Encountering/constructing breakdowns in
freedom, compared to an approach in which the understanding: Fieldwork should be theo

researcher tries to stay very close to data and sees retically informed but also varied and rich
the latter as providing the robust building blocks enough in the sense that it allows for the
existence and exploration of breakdowns. A
of theory. This does not mean that the researcher breakdown means that an
really interesting
has a licence to follow any creative hunch. Still, empirical "finding" can't easily be ac
the empirical material has a very important and counted for by available theory. The break
is not an outcome of the igno
as partner, providing con
critical role a down, thus,
dialogue or narrow-mindedness
rance, naivety, of the
siderable constraints on what can be done.
researcher. The surprise should be the reac
tion likely to be experienced by most mem
bers of the research who are
A METHODOLOGY OF SORTS FOR to be able
community,
to understand/explain
supposed
THEORIZING FROM EMPIRICAL MATERIAL the empirical observation/construction trig
The metaframework outlined above offers

guidelines and direction but, crucially, is not
4
"locked" into a narrow way of seeing that deter This is a of the ideas we are advocating.
full version We
assume close contact with and
mines the results a priori. This makes it possible here
the possibility of having
back and forth to the research site. As pointed out
towork with a methodology that stimulates a di going
above, breakdown-oriented research can be associated with
alogue between theory and empirical material, in the use of any kind of method and can also be used in more
which the preunderstandings, expectations, and moderate ways, but, for clarity and space, we concentrate
imaginations of the researcher are crucial. The here on one version.
FIGURE 1
The Research Process: Decision Tree forMystery-Focused Research
Breakdown
Not so interesting, Potentially interesting

difficult to grasp
Further theoretical
Existing literature explains it
and empirical study
Stop
Signs of mystery Signs of mystery

stand up to scrutiny, uncertain,
Stop weakly
formulation of supported
mystery
Stop
Mystery is Mystery is
not solved solved
Present mystery as Present mystery

contribution and solution as
contribution
gering the breakdown. Hence, it is not just Standing through the formulation of the
the individual researcher but also the col mystery. This phase includes the critical
lective theoretical and paradigmatic frame checking of whether a breakdown can lead
work and the knowledge shared within the to something new that is of potential theo
research community that are involved in ac retical relevance. Not all breakdowns allow
knowledging the breakdown. The re for the construction of a "real" mystery. In
searcher is wise to make certain that the deed, most do not. A breakdown may?in
surprise appears in the context of a sophis the context of this paper?be viewed as a
ticated position and is not an out candidate, and a mystery can be
partly mystery
come of poor scholarship. seen as a breakdown with a strong poten
3. Moving from breakdown to mystery: After tial to offer a theoretical contribution. A key
encountering an unexpected finding, the re distinction is that a breakdown ismainly of
searcher's next move is to formulate some local relevance and can sometimes be over
preliminary interpretations of a theoretical come through additional empirical work

contribution through showing (a) the (leading to deeper or broader empirical
broader relevance of an empirical finding, knowledge) and/or through consulting the
(b) the problems with the earlier theory or literature. A mystery, as we use the term
critique, and (c) some hints of a new under here, requires a novel theoretical contribu
tion. In other words, when asking more domain itmay cover are to estab
important
questions, hanging around (Dingwall 1997), lish. This is not just a matter of type of
and walking to the library and reading organization or organizational phenomenon
more books fails to be sufficient, a mystery but of time and history and the relative in
is at hand. Self-critique and reflexivity are terpretive value of a theoretical concept or
important elements here, as antidotes to the metaphor.
tendency tobe carried away by the prospect
of constructing a true mystery. Reflexivity This or stages
list of elements, in work, easily
may also mitigate the risk of being insuffi gives a or
too mechanicaloverly structured im
ciently careful in monitoring the empirical pression of this process. It is not intended as a
grounding and potential theoretical value manual or a model of how this kind of research
of the claim to mystery.
4. Engaging in more systematic work to de typically takes place, although we hope it can
a new understanding/theory, be used as a source of guidance and inspiration.
velop inspired
by a "negative finding" (breakdown in As Mills (1959) pointed out, research is a craft. It
additional resources, includ cannot be reduced to steps, manuals,
duced): Here, and mod
ing philosophy and social theory, are used. els. Rather, the list above should be seen as a
This work typically also involves further
rough description of the elements in research
empirical investigations, guided by devel
and interpretations processes that can bring the role of sophisti
oped understanding
supported by the use of additional theoret cated preunderstandings and the possibility of
ical and resources.
linguistic gradual development of theoretical understand
5. Solving or reformulating the mystery
ings more into focus in fieldwork. One can imag
through the development of a new idea that ine different modes of working with some over
offers a new interpretation of the phenome
non that inspired the mystery: This move lap from the framework. Work can be conducted
typically draws on the critical use of the cyclically?one may want to revisit and reframe
interplay between different theories being the field with a "preliminarily solved" mystery
problematized by the empirical input. One in order to develop the idea, metaphor, or theory.
can throw some novel light on the phenom
enon indicated by themystery by using new
It is also possible that a really challenging en
a new theoretical framework, or a counter triggers an excellent idea on the spot?
concepts,
new metaphor. This move can also involve making the breakdown/mystery distinction and
the formulation of new research tasks. The bypassing stage 3 and 4.
idea is also to transcend the empirically the research process in ways as
Structuring
specific and to produce something of
illustrated
broader relevance. where acts of cre by the model facilitates interplay
Again,
ativity are central, moments of reflexivity among theory, researcher subjectivity, and em
are important in enabling the rethinking of pirical options that can encourage theoretical
one's and vocabularies.
preferred positions development through problematizing existing
6. Developing the (re)solution of the mystery is a
theory. As stated, the framework presented
so that it gains a broader relevance for a
kind of full version associated with fieldwork
specific terrain and positioning it more
in to other theories:
research. The process may differ when working
clearly relationship
This means more considerations with breakdown/mystery ideas in other kinds of
systematic
of other, but not too diverse, terrains than research. What is important are the major orien
the one that "produced" or inspired the tations, not the details or the stages of the re
breakdown and subsequent mystery. This search process.
development may be about theoretical ab
straction, as well as considering where and
when this may encourage a productive un
THE CREATION AND RESOLUTION OF
derstanding. ?o theory is always wrong or
always right?all are more or less relevant BREAKDOWNS AND MYSTERIES
and helpful in different situations. And it is
Having outlined a mystery approach, we now
important to have a good idea ofwhen and
how be relevant. At the same time, indicate some key aspects of how breakdowns
they may
the approach suggested here is not so much and mysteries can be produced. Crucial in this
concerned with generalization and abstrac kind of work is an open affifude. Here, of course,
tion. It is more oriented to the specific and
it is important to avoid the naive idea of being
related empirical terrain that provides the
"nontheoretical" or blank as a means of being
empirical inspiration for the mystery?and
thus has a local touch. However, some ideas open, as implied by some views on grounded
about the nature of this locality and what theory (e.g., Eisenhardt, 1989; Glaser & Strauss,
1967). This simply means that cultural taken-for This is, of course, to a large extent a matter of
granted assumptions and other implicit theories creativity, but it is also a matter of wanting to
take precedence. Illiteracy does not lead to an achieve "anthropological" rather than familiar
open mind. Openness?the consideration of al or "technical-pragmatic" results. To some de
ternative routes of interpretation and analy gree it is a matter of using the critical strategy of
sis?is better accomplished through familiarity defamiliarization: "Disruption of common sense,
with an extensive repertoire of theories and vo doing the unexpected, placing familiar subjects
cabularies used reflexively (Rorty, 1989). In in unfamiliar, even shocking, contexts are the
terms of gender, for example, "openness" is not aims of this strategy to make the reader con
just a matter of making gender visible through scious of difference" (Marcus & Fischer, 1986:
observing sex differences ("body counting") or 137). Apart from general intellectual efforts to
through paying attention to the meanings and accomplish this, one can employ such tactics as
experiences of men and women. It involves using unconventional and varied literature,
questioning these two seemingly homogenous drawing from personal and research
experi
categories, paying attention to various forms of ences that are different from those salient in a
cultural masculinity and femininity, the possi previous study, and putting together a research
ble shifting character of these cultural mean team so that different viewpoints?and, thus,
ings in local contexts, and the ways they in different inclinations to see a variety of familiar
scribe a particular order on the world. It also and unfamiliar aspects?are represented.
means openness to how researchers may order What is needed, we believe, is a combination
the world through constructing it in terms of of theories that allows the researcher to see a
masculinity and femininity (Alvesson & Billing, multitude of perspectives and facilitate the de
1997; Ashcroft & Mumby, 2004; Calas & Smircich, velopment of results that may be frommore than
1999). one point of view. We label the set of perspec
Openness, thus, is not a matter of avoiding tives, concepts, and themes that a researcher
theory or postponing the use of it; rather, it in masters his or her interpretive repertoire (Alves
cludes broadening the repertoire of vocabular son & Sk?ldberg, 2000). Such a repertoire in
ies and theories that can be mobilized in order cludes the paradigmatic, theoretical, and meth
to consider more and less self-evident aspects. A odological qualifications and restrictions that
particular interpretive bias, following from a guide and constrain research work. The inter
closed theoretical/cultural/private orientation, pretive repertoire is made up of theories, basic
may be counteracted. Theory is often seen as assumptions, commitments, metaphors, vocabu
providing direction and control, but it can also laries, and knowledge. It indicates the "aca
be mobilized as a tool for disclosure. A theory demic" part of the researcher's preunderstand
can open up not only other theories and their ing and the whole spectrum of theoretical
lines of interpretation but also sensitive con resources that may be put into use when the
structions and interpretations of empirical ma researcher confronts empirical material. It
terial. marks the limits forwhat a researcher can do in
When studying relatively familiar phenom terms of making something out of certain empir
ena like organizations and management within ical material?material that in itself is produced
one's own country, the problem often is not only based on the interpretive inclinations of the re
or even primarily resolving breakdowns; there is searcher. It offers input to the struggles of, as
typically an element of creating them required. Becker puts it, "getting control over how we see
Ifwe accept the socially constructed nature of things, so that we are not simply the unknowing
social reality as well as research, this creative carriers of the conventional world's thoughts"
element is always involved. But more of an effort (1996:8).
is called for in organization studies than inmore The interpretive repertoire is made up of ele
unfamiliar settings, even though one occasion ments of relative degrees of depth and superfi
ally encounters original and exotic organiza ciality. Of course, few people master a broad
tions. The trick is to locate one's framework (cul spectrum of theories in depth. At one extreme
tural understanding) away from the cultural the researcher has a firm grasp of some theories
terrain being studied so that enough significant and discourses and can therefore skillfully use
material emerges to resolve the breakdown. them. At the other extreme the researcher has a
mere familiarity with other theories and dis breaking experience'" that challenges an estab
courses and can therefore only apply them in a lished position and encourages rethinking. One
crude and uncertian manner. We can refer to can imagine the same effect also through the
these end points as the deep (or scholarly) and use of less divergent approaches than those as
the shallow (or lay) elements in the repertoire. sociated with different paradigms. If this inter
The deep elements are central in the interpre theory challenge interacts nicely with the em
tive repertoire and easily activated, whereas the pirical material, the likelihood of a productive
shallow elements may be described as crude in breakdown in relationship to empirical material
terms of mastery and peripheral in terms of in increases. The combination of questioning in
terest and awareness. Typically, researchers empirical experience and intertheory confronta
have a strong tendency to use the deep elements tion gives the input to the rethinking of a partic
of their repertoire, since there is a likelihood ular understanding.
that they will lead to results, albeit in a rather
predictable way. AN ILLUSTRATION:A "FEMININE"
The shallow elements in the interpretive rep
ORGANIZATION DOMINATED BY MEN?
ertoire are only activated in research work if the
empirical material obviously appears to be in Below, we provide an example of how empir
line with these elements. This typically indi ical material can be used productively to rethink
cates that the empirical material is seen as im and develop theory. We want to stress that we
portant or interesting when framed in this way. use the example as an illustration. For a more
The researcher has three alternatives when he thorough discussion of the case, see Alvesson
or she thinks that the empirical material trig (1998). The empirical material stems from an eth
gers thinking activating the shallow/peripheral nography of an advertising agency (LAA). The
elements in the interpretive repertoire: (1) to study was initially fairly open, guided by a
drop the theme, (2) to refer to it briefly or mainly broad interest in organizational culture?
in empirical/low-abstract terms, or (3) to develop facilitated by the small size of the organization
the relevant parts of the interpretive repertoire (twenty-one people)?but soon we discovered a
and then do a more advanced investigation of somewhat extreme division of labor along with
this phenomenon. The third alternative means other interesting gender themes. All the men,
that the shallow part of the repertoire takes with one exception, occupied the professional
more center stage and the researcher develops positions, while all the women worked as assis
her or his skills in using it, thus moving it to the tants. In addition, the men were ten years older
deeper part of the repertoire. In such a case, than the women, who were typically twenty-five
empirical material typically has the chance to to thirty years old. The women were all attrac
make a real impact on the research outcome. tive and well dressed. LAA was an organization
The ambitious use of the idea of an interpre led by men, while the women managed routine
tive repertoire inspires a critical use of theory in jobs and the "domestic chores."
which empirical material and alternative theo There was no specific intention to focus on
ries are employed as elements in theory devel gender issues at the outset of the study, but this
opment. Carefully constructed empirical mate "discovery" was seen as a surprise. Why did it
rial to problematize
is used a targeted theory, emerge as
such? A gendered division of labor?
thus opening it up for reconsiderations and al including vertical division?is common, and
ternative understandings. In organization stud many students of gender may have constructed
ies the work ofMorgan (1980,1997) has been vital the case as a standard one, indicating broad
in this regard. Also, the literature advocating patterns. But the pattern here seemed extreme
multiparadigmatic studies is relevant here (e.g., and unexpected in this kind of work. We guess
Gioia & Pitre, 1990; Lewis & Grimes, 1999). One that most researchers, not interested in gender,
can debate the extent to which it is possible to would not have made much of this observation,
cross and master several paradigms (Burrell & but the researcher here had an interest and com
Morgan, 1979; Deetz, 1996; Hassard, 1991; Parker petence in gender and identity themes. Of
& McHugh, 1991), but we agree with Lewis and course, careful consideration here preceded the
Grimes (1999: 686) that "exploring 'foreign' para choice to explore this in depth. Other factors,
offers theorists a potentially 'frame including age, education, occupational back
digms
ground, and forth, were

so considered. One tionality as a core dimension at work, whereas
question concerned
raised whether the case "masculine" occupations and organizations typ
was an example of an overall gender division of ically do the opposite. We thus have interesting
labor that would not come as a surprise for an breakdowns of understandings based on theories
expert in gender theory. Or could there be some that men and masculinities go together and that
thing "local," associated with organizational feminine values are at odds with male-dominated
conditions that might inspire new ideas? When institutions.
the accounts of the men dominating the agency The interview accounts?and statements
were interpreted, the fieldwork revealed even noted during observations?were carefully con
more interesting and surprising results. They sidered in a multitude of ways before being
emphasized that the men were intuitive, emo seen as cultural constructions with a gender
tional, sensitive to interpersonal relationships, relevance. One may view the statements as
family oriented even at work, uninterested in purely factual?referring to the personalities of
careers and management, and so on. the people in the agency?or treat them as non
gendered. The construction of the constructions
Advertising people are normally very outgoing
and they are emotionally charged. Because feel of the male advertising people in feminine
ings and things like that are the basis of creativ terms was eventually viewed as a (1) a good
ity, so to speak. They are often very rich in ideas interpretation of the empirical material and (2)
and can associate with
associative, they quickly one that was part of the construction of a break
various They are rather
phenomena. normally down with potential mystery qualities.
difficult to steer and jump for joy when they be
come happy or hit the roof when they become Hence, we have a possible mystery: How can
mad. The amplitude of their reactions is much highly asymmetrical gender relations (with the
higher than for example people in companies' men dominating) coexist with "feminine" values
are
accounting departments. Advertising people and meanings? Or how come an organization
seldom very systematic or structured (male ad
that is dominated by men is constructed by them
vertising worker).
in feminine terms? Further consultations of the
They described themselves, their occupation, gender were
literature unhelpful in making
and their organization inways that were closely sense of this. Gender organization studies gen
in line with cultural views of femininity, at least erally emphasize how workplaces dominated by
on an overall and clich?-like level. One male men are constructed in masculine terms (e.g.,
used the metaphor of pregnancy to describe the Hall, 1993; Leidner, 1991;Mills, 1988). They do not,
work of developing an advertisment. Of course, on the whole, seem to be able to produce a good
these accounts are not just facts, or even authen understanding of an organization that is ex
ticmeanings, but, rather, constructions. As such tremely strongly hierarchically structured in
they are of considerable interest. Once again, it terms of gender, where men dominate and
is perhaps not surprising that advertising peo where the dominant understanding matches
ple construct themselves in these terms, but what a large body of literature sees as feminine
given the context of the gendered division of orientations and values.
labor, we encounter a second breakdown. It Literature reviews and additional empirical
seems reasonable to see that the statements work supported the case for a "mystery." The
show considerable alignment with the ideals of case may be uncommon but may still encourage
many feminists around the importance of emo us to revise some theoretical ideas around the
tion and the personal in terms of thinking, work tight connection of male domination and domi
ing, and organizing (Jaggar, 1989; Mumby & Put nation of masculine cultural constructions, mu
nam, 1992). Correspondingly, males are tually supporting each other, emphasized by the
conventionally constructed as nonemotional gender literature. Without denying that this the
(Hearn, 1993). Hollway writes that, "in our soci oretical idea often makes sense, perhaps the
ety, the judgment is a sexist one: expressing case can problematize the operations of gender
feelings is weak, feminine and in contradiction and and help us rethink constructions of mascu
tomen's rationality" (1984: 253). "Masculine" oc linities and femininities.
cupations require people "to be cool, impassive The case indicates
that the link between con
or stern" (Cockburn, 1991: 150). But the dis structionof the organization in feminine terms
courses of the advertising industry stress emo and women's positioning is not straightforward.
The ambiguities of the work situation, results, for their symbolic labor, despite the construction
and client relations of the advertising workers of themselves, their work, organization, and po
heighten identity problems. As in much other sition in client relationships as feminine.
professional service work, "the largely fluid To conclude, the study suggests the possibil
character of anything external to interactional ity of a loose coupling between male domina
accomplishments, provides for very active sym tion and the domination of masculinities (as
bolic labour" (Deetz, 1998: 157). In the present these are described in the literature and typi
case it complicates issues of gender. The con cally culturally defined). In particular, the pres
struction of the work and organization through ence/absence of specific linkages made by sub
the use of the emotionality-intuition-personal jects in organizations between what in the
chemistry-antibureaucracy vocabulary facili gender literature is viewed as masculine/
tates identity work. It indicates positive values, feminine properties and the two sexes is impor
coherence, and distinctiveness, for example, in tant for the fate of men and women. This is
relation to client's personnel and other conven partly a matter of power: explicitly labeling
tional people. These are constructed as the op what is generally, but not necessarily con
posite of the advertising people: as cautious, sciously, seen as culturally feminine may well
bureaucratic, and lacking the right intuition. upset gender orders. In the present case, a gen
What the gender literature identifies as femi dered division of labor would be more difficult
nine orientations?which it claims that men to reproduce if the constructions of work content,
avoid and downgrade?are used as symbolic client relations, and organizational practices ac
and discursive resources in the identity con knowledged the correspondence with what is
structions of the advertising people. But the fem broadly defined as culturally feminine. The case
inine undertone/low degree of masculinity presented here provided inspiration for a theory
makes this solution a mixed blessing. The ad of workplace gender relations that allows for a
vertising agency appears as subordinate and discrepancy between abstract ideas of mascu
feminine in relation to its clients?the relation line/feminine properties proposed by gender re
ship is often referred to as a marriage, and it is searchers and
local constructions of gender. It
clear that the agency assumes the female part. also provides a framework to understand gen
This discourse puts some strain on gender iden der stereotypes as resources in social processes,
tity. In sum, the precarious character of the oc thus illuminating the elastic and relative as
cupational identity has a clear gendered mean pects of gender relations that enable richer in
ing. From the other angle, one can say that the terpretations of their social effects.
gender identity of male advertising profession
als is only partially, and in some respects even
ALTERNATIVEWAYS OF TAKING
badly, supported by work, organization, and cli
ADVANTAGE OF BREAKDOWNS FOR THEORY
ent relations.
DEVELOPMENT
In LAA the weak symbolic support formascu
linity in the work content (connected to the low Although we think the approach developed in
degree of technical expertise) and client rela this paper is a fruitful and underutilized way of
tions is compensated for by highlighting work developing more novel interpretations of empir
place sexuality and perpetuating internal gen ical phenomena and innovative theoretical
der structures. Masculinities emerge in relation ideas?and most people probably agree that
to female personnel, subjected to what may be there is a shortage of such?some (self-reflexive)
referred to as "hyperfemininization." Gender be words of caution are necessary. The maximalist
comes structured so that male work/gender version sketched out above is not a low-risk
identities are supported. One aspect here is the strategy. Constructing and solving a mystery
location of men and women in the division of calls for a fortunate combination of inspiring
labor, where male power accounts for the re empirical material, access to a rich frameworks
cruitment of younger, sexually attractive, lower and resources for reflexivity about how to use
positioned women. Another is the heightened these, creative construction work, and, in the
state of gender interaction. These two mean that available literature, empty space for a theoreti
the men can place themselves in "masculine cal contribution. Many research projects have
subject positions," using gender as a resource other agendas and/or do not lead to the discov
ery/construction of great mysteries with theoret ies. The norm seems to be that the researcher
ical potential. is in control, producing a linear, coherent
However, more moderate uses of the break study, where research questions, framework,
down idea are also possible. Arguably, all re fieldwork, empirical results, and conclusions
search approaches confront (or have the poten follow a rational procedure. Even in some re
tial to construct) breakdowns, as long as we search drawing on Foucauldian and other
accept that social reality is not fully understood. poststructuralist ideas, the studies reported
It is possible to imagine variation in emphasis tend to produce conventional "depersonalized,
on elements of breakdowns and mysteries in third-person and apparently objective and au
research. We propose a spectrum that includes thoritative representations" (Wray-Bliss, 2002:
breakdown-focused, breakdown-sensitive, and 20; see also Richardson, 2000). This may say
breakdown-considering research, with varying more about the established standards for pre
degrees of interest in and attention paid to ex sentation in journals?despite decades of pos
ploring and exploiting breakdowns. itivism critique?than about how researchers
Breakdown-focused research means working actually work. Arguably, breakdowns and
fully in line with the ideas suggested here, aim projects following these are not so rare, but
ing for a full-scale mystery-scanning approach there may be a need to make them more legit
and being more than willing to explore and con imate and explicit.
struct breakdowns. Of course, sometimes this Which methods are most suitable for research
intention is not fulfilled, and the research working with breakdowns and mysteries? Here
project may be turned into something else. we have two answers. The first is that the more
Breakdown-sensi five research means a strong to a study is processual, emergent, open, and em
modest interest in potential mysteries. Itmay be pirically and rich, the more likely an in
varied
carriedout as part of a more conventional study, teresting mystery, via breakdowns, will be pro
which is guided by a specific research question duced and solved. Ethnographic studies
and a design for studying it. In this case the (Prasad, 1997;Wolcott, 1995) here have some ad
mystery approach operates as an
additional vantages. Other studies that are open to the
guiding principle. is
The researcher open to the views of the research subjects (perhaps viewing
possibilit?s of an unanticipated theme and keen them as coparticipants; Heron, 1981)?allowing
to follow it, even though this is not the initial or them to express unconstrained voices in the re
primary intent of the study. Possible outcomes search?may also increase the frequency with
could be refinement of theory or suggestions for which breakdowns will appear. Our second an
new lines of inquiry. The breakdown-consider swer?and this is our main point?is that all
ing researcher is less inclined to actively work kinds of research can lead to?or be used for?
with breakdowns and mysteries, unless he or the discovery or construction of breakdowns and
she bumps into something really interesting. mysteries. As our initial reference to Lincoln and
He or she has some awareness of the possibility Kalleberg (1985) and the Hawthorne studies in
of taking advantage of breakdowns but takes dicates, even questionnaire studies and experi
this road only when extraordinary opportunities ments may provide interesting breakdowns. The
emerge. For researchers and research projects Hawthorne studies are particularly illuminating
guided by this orientation, breakdowns only in this respect. The ideas discussed here are
occasionally play a significant role in account thus of potential broad relevance, even though
ing for results. When they do, self-critique and research that does not allow for the flexibility of
new research questions are more likely to result developing and exploring new ideas in the pro
than the formulation and solving of a mystery. cess of gathering additional empirical material
However, occasionally, the researcher who is not may have difficulties solving a mystery. Often,
intially not very breakdown oriented may encoun however, the formulation of a mystery can be a
ter breakdowns that trigger radical rethinking. great contribution:it can be a vital step in en
Presumably, most researchers have such a couraging reflexivity and new lines of inquiry.
breakdown-considering research orientation, Asking innovative questions can be as impor
although it is difficult to find examples of re tant as providing answers.
searchers actually espousing it explicitly, at In addition to being feasible in any kind of
least in management and organization stud research, breakdowns can, in principle, occur
at almost any point in the research process, questioning (Asplund, 1970). Solving here means
based on serendipity or conscious efforts to re that the mystery becomes more understandable:
flexively remain open to them. Working with it is less puzzling and less ambiguous, and we
empirical material in different phases is impor have concepts, a line of reasoning, a metaphor,
tant here. The trend to shift the emphasis from or other tools that give us a sense of what to
fieldwork to textwork (Geertz, 1988; Richardson, expect and how to intellectually understand the
2000; Van Maanen, 1988) has pointed to the im mystery.
portance of writing in crafting ideas and articu A mystery emerges as a combination of the
lating findings. Our approach does not neces researcher's preunderstanding, including ac
sarily imply a linear development. We indicated cess to theoretical framework(s) and vocabular
earlier the potentially cyclical nature of this ies, and the inspiration of empirical material.
kind of research. Breakdowns and mystery con The ratio of input from empirical experiences
struction may start with the writing process, and the intellectual-creative work necessary to
which then may lead the researcher to return to construct a mystery may vary. Since this is a
fieldnotes or other empirical
material (interview paper emphasizing empirical work and method
protocols, questionnaires), the literature, and ology, we have devoted much attention to the
even the field. The kind of curiosity and willing role of empirical studies in triggering a mystery,
ness to reconsider received wisdom that charac but, as mentioned previously, "pure" empirical
terizes the research methodology suggested impressions do not lead us far. In addition, cre
here is thus not limited to a specific phase in the ativity and concentrated work in supplementing
research project. and focusing theoretical work are necessary to
assess whether the mystery candidate is fruitful
for theoreticaldevelopment?that is, is not just a
CONCLUSION
breakdown for the researcher only and/or within
In this paper we have advocated the use of a narrow terrain. A mystery promising a theoret
empirical material as input for theorizing. In ical contribution meetmust

high criteria?it
short, our approach encourages researchers to can't be solved
through a literature search but,
actively work with, expand, and vary their inter rather, calls for innovative theoretical work. The
pretive repertoire by being open to and focusing successful solving of a mystery means that one
on breakdowns. Breakdowns, in most research, produces a theoretical understanding that (1)
are seen as a nuisance?they indicate that the illuminates the phenomenon leading to the
researcher is not in control and may obstruct the breakdown and subsequently mystery and (2)
research design and threaten the production of allows an abstracted set of ideas and concepts
predictable results. Students interested in "lead with broader bearing on how to make sense of
ership," for example, may face settings inwhich similar phenomena in other settings.
people do not seem concerned about or refrain Because the literature on the interplay be
from/fail to produce strong asymmetrical rela tween theory and empirical material is vast and
tions and coherent behaviors that fit a "leader varied, it is difficult to claim that our contribu
ship style" concept. Such experiences will typi tion is a great invention. Rather, we synthesize,
cally not make the student of leadership happy. expand, sharpen, and refine ideas that, to a de
From the approach suggested here, which is in gree, have already appeared in social science.
line with a long line of scholarship (e.g., Becker, We can distinguish among three elements in our
1996), breakdowns are potentially good news? contribution:
they may make space for theoretical reconcep
1. One contribution concerns the introduction
tualizations and development.
and, to some degree, development of a gen
Breakdowns offer a vital step in the produc eral framework for and an alternative con
a
tion of mystery. Establishing a mystery in it of the research process. The
ceptualization
self offers an interesting source of further think aim is not primarily to provide a blueprint
it encourages and for methodology but to offer inspiration
ing, since problematization
through a guiding set of generative ideas.
self-reflexivity. This may be an important con We have advocated a framework for think
tribution. But the formulation of mystery also
ing about empirical material and how it can
provides an impetus for solving it and, thus, be used in more creative and challenging
adding new knowledge beyond the critical ways than may be common. This means go
ing beyond recommending openness and and data and the minimization of re
following where data may lead us and, in searcher subjectivity.

stead, actively working with alternative
constructions. One aspect here is the en To a believer in conventional methodology,
of a to be sur
couragement willingness
including the most popular versions of qualita
prised in research and a willingness to re
tive methods, this may appear to be a dangerous
vise the frameworks and traditions from
and unreliable enterprise. But similar critique
which we originate. Not just encountering
but also trying toproduce breakdowns is a can be directed at hypothesis testing and induc
vital part of this approach, in which prob tive projects that frequently exhibit a mislead
lematizing existing ideas is crucial.
ing surface of rigor and robustness. Since the
2. A second, somewhat more contri new
specific, purpose is to generate ideas, it is important
bution concerns vocabulary. We can cap not to emphasize rigor too much and to allow
ture the advocated alternative conceptual
ization through metaphors. We have used space for the researcher's
imagination when
both critical and positive ones. Convention working with
empirical material. Still, we are
ally, data are seen as building blocks in not propagating an "anything goes" version or a
research, as unknown territory (of facts license for researchers to be creative and try to
and/or meanings) to be discovered and/or innovate for the sake of saying something novel.
judged in terms of what are true/valid and The researcher
needs to persuade the skeptical
false/nonacceptable claims to knowledge.
reader?building a convincing case involves il
We are skeptical of such metaphors as data
guiding or ultimately validating theory.Em luminating empirical material, using a well
pirical material is, in most cases of interest mastered interpretive repertoire, and demon
strating elements of reflexivity in the process, as
for organization studies, not robust but
and in various de
shaped reshaped ways, well as showing a careful and sophisticated un
pending on the language and perspectives
of the relevant literature. In the
used. We alternative derstanding
propose metaphors
and material end, this is not less demanding than building
conceptualizations. Empirical
is seen as a potential theory from data or validating and falsifying
dialogue partner,
leading to questioning, doubting, and prob hypotheses.
lematizing existing/dominant expectations
and frameworks. is viewed as a po
Theory
so are
tential tool for disclosure,
downs in understanding. We suggest
and break
the
REFERENCES
creation and solving ofmysteries?aided by Agar, M. 1986. Speaking of ethnography. Thousand Oaks,
breakdowns?as a root metaphor for the re CA: Sage.
search process. We also suggest that con
Alvesson, M. 1998. Gender relations and identity at work: A
cepts such as sensitive constructions, inter case study of masculinities and femininities at an ad
pretive repertoires, and reflexivity are
vertising agency. Human Relations, 51: 969-1005.
helpful in realizing the full generative po
Alvesson, M. 2003. Beyond romanticists, and
tential in breakdowns and mysteries. neopositivists,
localists: A reflexive approach to research interviews.
3. A third contribution concerns the specific
of Management Review, 28: 13-33.
methodology proposed for working with Academy
breakdowns and mysteries. We hope this is Alvesson, M? & Billing, Y. D. 1997. Understanding gender and
not read as a recipe, and we would argue organization. London: Sage.
that in an area of methodology where "pro Alvesson, M., & Deetz, S. 2000. Doing critical management
gressive" (e.g., constructivist) ideas fre
research. London: Sage.
quently are rather abstract and of uncertain
relevance for research a Alvesson, M., & K?rreman, D. 2000. Taking the linguistic turn
practice, outlining
in organizational research: Challenges, responses, con
research process taking these ideas seri
sequences. Journal of Applied Behavioral Science, 36:
ously may be supportive. There is a strong
134-156.
norm to present research results in a fairly
linear and rational way. Researchers have Alvesson, M., & Sk?ldberg, K. 2000. Reflexive methodology.
difficulty fully using constructivist ideas in London: Sage.
empirical studies and take the insight
Ashcroft, K., & Mumby, D. 2004. Reworking gender. Thousand
about the fusion of theory and empirical CA: Sage.
Oaks,
material seriously. We have formulated an
alternative to dominating and sometimes Asplund, J. 1970. Om undran inf?r samh?llet. Lund: Argos.
misleading notions of research as a mainly G. 1985. Administrative science as con

Astley, socially
rational process of planning, execution, and structed truth. Administrative Science Quarterly, 30:497
analysis based on a separation of theory 513.
Becker, H. 1996. Tricks of the trade. Chicago: University of Gioia, D., & Pitre, E. 1990. Multiparadigm perspectives on
Chicago Press. theory building. Academy of Management Review, 15:

584-602.
Burrell, G., & Morgan, G. 1979. Sociological paradigms and
organisational analysis. London: Heinemann. Glaser, B., & Strauss, A. 1967. The discovery of grounded
theory: Strategies for qualitative research. Chicago: Al

Calas, M., & Smircich, L.
1992. Re-writing gender into orga
dine.
nizational theorizing: Directions from feminist perspec
tives. In M. Reed & M. Hughes (Eds.), Re-thinking orga Hall, E. 1993. Smiling, deferring, and flirting: Doing gender
nization: New directions in organizational theory and by giving "good service." Work and Occupations, 20:
analysis: 54-62. London: 452-471.

Sage.
Calas, M., & Smircich, L. 1999. Past postmodernism? Reflec Hanson, N. 1958. Patterns of discovery: An enquiry into the
tions and tentative directions. ofManagement conceptual foundations of science. Cambridge: Cam
Academy
Review, 24: 649-671. bridge University Press.
Charmaz, K. 2000. Grounded and con Hardy, & Clegg,

C, S. 1997. Relativity without relativism:
theory: Objectivist
structivist methods. In N. Denzin & Y. Lincoln (Eds), Reflexivity in post-paradigm organization studies. Brit
Handbook of qualitative research (2nd ed.): 509-535. ish Journal of Management, 8: S5-S17.
Thousand Oaks, CA: Sage.

Hassard, J. 1991. Multiple paradigms and organizational
Cockburn, C. 1991. In the way of women. London: Macmillan. analysis: A case study. Organization Studies, 12: 275
299.
Cronbach, J. L. 1975. Beyond the two disciplines of scientific
American 30: 671-684. Hearn, J. 1993. Emotive subjects: Organizational men, orga
psychology. Psychologist,
nizational masculinities and the (de)construction of
Davis, M. S. 1971. That's interesting! Towards a phenomenol
"emotions." In S. Fineman (Ed.), Emotions in organiza
ogy of sociology and a sociology of phenomenology.
tions: 148-166. London: Sage.
Philosophy of the Social Sciences, 1: 309-344.
Heron, J. 1981. Experiential research methodology. In P. Rea
Deetz, S. 1992. Democracy in an
age of corporate coloniza
son & J.Rowan (Eds.), Human inquiry: A sourcebook for
tion: Developments in communication and the politics of new paradigm research: 53-81. Chichester, UK: Wiley.
everyday life. Albany: State University of New York
Press. Hollway, W. 1984. Gender difference and the production of
subjectivity. In J. Henriques, W. Hollway, C. Urwin,

Deetz, S. 1996. Describing differences in approaches to or
C. Venn, & V. Walkerdine (Eds.), Changing the subject:
ganizational science: Rethinking Burrell and Morgan
227-263. London: Methuen.
and their legacy. Organization Science, 7: 191-207.
Jaggar, A. M. 1989. Love and knowledge. Inquiry, 32: 151-176.
Deetz, S. 1998. Discursive formations, strategized subordina
Kuhn, T. 1962. The structure of scientific revolutions. Chicago:
tion, and self-surveillance: An empirical case. In A. Mc
University of Chicago Press.
Kinley & K. Starkey (Eds.), Foucault, management and
organizational theory: 151-172. London: Sage. Leidner, R. 1991. Serving hamburgers and selling insurance:
Denzin, N. 1994. The art and of In Gender, work, and identity in interactive service jobs.
politics interpretation.
Gender and Society, 5: 154-177.
N. Denzin & Y. Lincoln (Eds.), Handbook of qualitative
research: 500-515. Thousand Oaks, CA: Sage. Lewis, M., & Grimes,
A. 1999. Metatriangulation: Building
theory from multiple paradigms. Academy of Manage
Denzin, N., & Lincoln, Y. 2000. Introduction: The discipline
ment Review, 24: 673-690.
and
practice of qualitative research. In N. Denzin &
Y. Lincoln (Eds.), Handbook of qualitative research (2nd Lincoln, Y., & Guba, E. 2000. Paradigmatic controversies,
ed.): 1-45. Thousand Oaks, CA: Sage. contradictions, and emerging confluences. In N. Denzin
& Y. Lincoln (Eds.), Handbook of qualitative research
Dingwall, R. 1997. Accounts, interviews and observations. In
(2nd ed.): 163-188. Thousand Oaks, CA: Sage.
G. Miller & R. Dingwall (Eds.), Context and method in
qualitative research: 51-65. London: Sage. Lincoln, J.,& Kalleberg, A. 1985. Work organization and work
force commitment: A study of plants and employees in
Eisenhardt, K. 1989. Building theories from case study re
the US and Japan. American Sociological Review, 50:
search. Academy of Management Review, 14: 532-550.
738-760.
Fetterman, D. M. 1989. Ethnography: Step by step. Newbury
Marcus, G., & Fisher, M. 1986. Anthropology as cultural cri
Park, CA: Sage.
tique. Chicago: University of Chicago Press.
Freese, L. 1980. Formal theorizing. Annual Review of Sociol
Merton, R. K., & Barber, E. 2004. The travels and adventures of
ogy, 6: 187-212.
serendipity: A study in sociological semantics and the
Garfinkel, H. 1967. Studies in ethnomethodoiogy. Englewood sociology of science. Princeton, NJ: Princeton University
Cliffs, NI: Prentice-Hall. Press.
Geertz, C. 1988. Work and lives: The anthropologist as author. Mills, A. 1988. Organization, gender and culture. Organiza
Cambridge: Polity Press. tion Studies, 9: 351-370.
Gergen, K. 1978. Toward generative theory. Journal of Per Mills, C. W. 1959. The sociological imagination. New York:
sonality and Social Psychology, 36: 1344-1360. Oxford University Press.
Morgan, G. 1980. Paradigms, metaphors, and puzzle solving Rorty, R. 1989. Contingency, irony and solidarity. Cambridge:
in organization theory. Administrative Science Quar Cambridge University Press.
terly, 25: 605-622. Rosenau, P. M. 1992. Post-modernism and the social sciences:
Morgan, G. 1997. Images of organization. Thousand Oaks, Insights, inroads and intrusions. Princeton, NJ: Princeton
CA: Sage. University Press.
Mumby, D., & Putnam, L. 1992. The politics of emotion: A Schwartzman, H. B. 1993. Ethnography in organizations. New
feminist reading of bounded rationality. Academy of bury Park, CA: Sage.
Management Review, 17: 465-486.
Shotter, J. 1993. Conversational realities: The construction of
Parker, M., & McHugh, G. 1991. Five texts in search of an life through Park, CA:
language. Newbury Sage.
author: A response to lohn Hassard's "Multiple para
Shotter, J., & Gergen, K. 1994. Social construction: Knowl
digms and organizational analysis." Organization Stud
edge, self, others and continuing the conversation. Com
ies, 13: 451-456.
munication Yearbook, 17: 3-33.
Peirce, C. S.
1978. Pragmatism and abduction. In C. Hart
Steier, F. 1991. Reflexivity and methodology: An ecological
shorne & P. Weiss (Eds.), Collected papers, vol. V: 180
constructionism. In F. Steier (Ed.), Research and reflex
212. Cambridge, MA: Harvard University Press.
ivity: 163-185. London: Sage.
Poole, M. S., & Van de Ven, A. 1989. Using paradox to build
and theories. of Strauss, A., & Corbin, J. 1990. Basics of qualitative research.
management organization Academy
14: 562-578. Newbury Park, CA: Sage.
Management Review,
K. and refutations: The growth of Strauss, A., & Corbin, J. 1994. Grounded theory methodology.
Popper, 1963. Conjectures
In N. K. Denzin & Y. S. Lincoln (Eds.), Handbook of qual
knowledge. London: Routledge and Kegan Paul.
itative research: 273-285. Thousand Oaks, CA: Sage.
Popper, K. 1972. Objective knowledge: An evolutionary ap
Oxford: Clarendon Press. Sutton, R., & Staw, B. 1995. What theory is not. Administrative
proach.
Science Quarterly, 40: 371-384.
Potter, J.,& Wetherell, M. 1987. Discourse and social psychol
attitudes and behaviour. London: Van Maanen, J. 1988. Tales of the field: On writing ethnogra
ogy: Beyond Sage.
phy. Chicago: University of Chicago Press.
Prasad, P. 1997. Systems of meaning: Ethnography as a
for the study of information technologies.
Van Maanen, J. 1995. An end to innocence: The ethnography
methodology
& J. I. DeGross of ethnography. In J.Van Maanen (Ed.), Representation
In A. S. Lee, J. Liebenau, (Eds.), Informa
tion systems and qualitative research: 101-118. London: in ethnography: 1-35. Thousand Oaks, CA: Sage.
Chapman & Hall. K. 1989. construction as im

Weick, Theory disciplined
C. C. 1987. The method: agination. Academy of Management Review, 14: 516?
Ragin, comparative Moving beyond
and quantitative Uni 531.
qualitative strategies. Berkeley:
versity of California Press.
Wolcott, H. 1995. Making a study "more In J.
ethnographic."
Richardson, L. 2000. Writing: A method of inquiry. In N. Den Van Maanen (Ed.), Representation in ethnography: 79
zin & Y. Lincoln (Eds.), Handbook of qualitative research 111. Thousand Oaks, CA: Sage.
(2nd ed.): 923-948. Thousand Oaks, CA: Sage.

Wray-Bliss, E. 2002. Abstract ethics, embodied ethics: The
and the mirror of nature. of Foucault and in labor
Rorty, R. 1979. Philosophy Princeton, strange marriage positivism
NJ: Princeton University Press. process theory. Organization, 9: 5-39.
Mats Alvesson (Mats Alvesson@fek.lu.se) is professor at the School of Economics and
Management, Lund University, Sweden, and visting professor at University of Queens

land Business School, Australia. His research interests include knowledge work,
organizational culture, identity, critical theory, and qualitative methodology.
Dan K?rreman (Dan.Karreman@fek.lu.se) is associate professor at the School of Eco

nomics and Management, Lund University, Sweden. His research interests include
knowledge work, organizational control, identity, leadership, and research methods.

Theory Development

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Theory Development

Uploaded by

Copyright:

Available Formats

Constructing Mystery: Empirical Matters in Theory Development

Author(s): Mats Alvesson and Dan Krreman

discovery and/orcreation of mysteries and the subsequent solving of the mysteries. A

explicate. Itmeans to generally

means that ethnography can be described "as a THE FICTION OF "FACTS" IN

cally new ideas on the dynamics of workplace distortion, ambiguity, which

Weick, 1989). Social changes?partly fueled by REFRAMING THE RESEARCH OF

sented through potentially contested representa vital element in theory development as we

freedom, compared to an approach in which the understanding: Fieldwork should be theo

The metaframework outlined above offers

imaginations of the researcher are crucial. The here on one version.

Not so interesting, Potentially interesting

Signs of mystery Signs of mystery

Present mystery as Present mystery

preliminary interpretations of a theoretical come through additional empirical work

ground, and forth, were

empirical material as input for theorizing. In ical contribution meetmust

following where data may lead us and, in searcher subjectivity.

misleading notions of research as a mainly G. 1985. Administrative science as con

analysis based on a separation of theory 513.

Chicago Press. theory building. Academy of Management Review, 15:

theory: Strategies for qualitative research. Chicago: Al

analysis: 54-62. London: 452-471.

Charmaz, K. 2000. Grounded and con Hardy, & Clegg,

Thousand Oaks, CA: Sage.

subjectivity. In J. Henriques, W. Hollway, C. Urwin,

sonality and Social Psychology, 36: 1344-1360. Oxford University Press.

Chapman & Hall. K. 1989. construction as im

(2nd ed.): 923-948. Thousand Oaks, CA: Sage.

Mats Alvesson (Mats Alvesson@fek.lu.se) is professor at the School of Economics and

Management, Lund University, Sweden, and visting professor at University of Queens

Dan K?rreman (Dan.Karreman@fek.lu.se) is associate professor at the School of Eco

knowledge work, organizational control, identity, leadership, and research methods.

You might also like