Description
This paper suggests how the contemporary theory of causality, based on the notions of
counterfactuality and contrastive thinking, offers helpful direction on how to generate
plausible causal arguments in interpretive research. For an interpretive researcher, this
opens a route from rich emic accounts to thick explanations; however, only if he/she so
wishes and the research question so requires. Perhaps with some surprise, causality can
be included in interpretive research framings without compromising the unique features
of such research – actually by even building on some of its strongholds
Exploring the possibilities for causal explanation in interpretive
research
Kari Lukka
?
Turku School of Economics, University of Turku, Finland
a b s t r a c t
This paper suggests how the contemporary theory of causality, based on the notions of
counterfactuality and contrastive thinking, offers helpful direction on how to generate
plausible causal arguments in interpretive research. For an interpretive researcher, this
opens a route from rich emic accounts to thick explanations; however, only if he/she so
wishes and the research question so requires. Perhaps with some surprise, causality can
be included in interpretive research framings without compromising the unique features
of such research – actually by even building on some of its strongholds. Examples from
interpretive management accounting research will illustrate the message of the paper.
Ó 2014 Elsevier Ltd. All rights reserved.
Introduction
The main thesis of this paper is that causality has the
potential to be a useful part of interpretive research. It is
suggested that it is possible to ?nd a meaningful role for
causality in interpretive studies and that this can also be
achieved while conducting interpretive research on its
own terms, without compromising its unique characteris-
tics. In fact, it will be argued that linking causal explana-
tions to interpretive research places some of its central
strengths at the forefront. The paper elaborates the value
of this thesis, its support and speci?c implications. The
analysis is illustrated by examples from interpretive man-
agement accounting research.
Understanding meanings of often unique phenomena,
considering people as taking ‘something as something’
(e.g. Meretoja, 2012), is the starting point and characteristic
feature of interpretive research (e.g. Chua, 1986). It tends to
focus on the emic perspective, an examination on how the
research subject his/herself develops his/her meanings,
rather than the etic one, whereby the issue is the research-
er’s interpretation and theorization on the studied
phenomena (Denzin, 1983; Headland, 1990; Pike, 1954).
The key features of interpretive research have often been
perceived as implying a stark contrast to causal explana-
tion, hence even eliminating the latter from such research
(e.g. von Wright, 1970; Winch, 1958; Burrell, 1979; cf.
Kakkuri-Knuuttila, 2006). The paper investigates this
alleged tension between interpretive research and causal
explanation in its attempt to explore the possibilities of
causality in interpretive research. Without denying the
importance of such interpretive research that seeks out-
comes other than discovering causality, the purpose of this
paper is to offer insight on potential avenues that open for
an interpretive researcher regarding causal explanation,
based on a notable amount of the recent literature on cau-
sality and explanation in the wider social sciences; it is also
informed by recent advances in the philosophy of science.
Indeed, the analysis of this paper is motivated, in partic-
ular, by recent increasing interest in the wider social sci-
ences to understand better how researchers conduct, or
might conduct, their work regarding causal inferences
(e.g. Tetlock & Belkin, 1996; Morgan, 2007; Hoerl,
McCormack, & Beck, 2011). While it certainly is not the
sole task of scholarly work to produce explanations, they
tend to play a central role in much that is sought to be
achieved (cf. Sobel, 1995, 1996; Morgan & Winship,http://dx.doi.org/10.1016/j.aos.2014.06.002
0361-3682/Ó 2014 Elsevier Ltd. All rights reserved.
?
Tel.: +358 2 3339315; fax: +358 2 3338900.
E-mail address: kari.lukka@utu.?
Accounting, Organizations and Society 39 (2014) 559–566
Contents lists available at ScienceDirect
Accounting, Organizations and Society
j our nal homepage: www. el sevi er. com/ l ocat e/ aos
2007). Further, although not the only option, the purpose
of producing explanations based on causality appears to
be important across many areas of research, including
business studies (e.g. DiMaggio, 1995; Sutton & Staw,
1995; Weick, 1995).
The paper starts with an outline of the signi?cant,
although not yet widely noticed, variation in how causality
is understood in the philosophy of science. In particular,
this analysis introduces the recent emphasis on the coun-
terfactual account of causality, suggested to be of great
importance with regard to generating causal explanations
in interpretive research. Causality and causal explanations
are then examined in the context of interpretive research,
focusing on knowledgeable agents and how their meanings
develop and operate in the surroundings where they act.
After outlining the ways in which interpretive framings
can be employed in management accounting research, con-
trastive thinking is introduced as a helpful method to focus
causal exploration, thereby making the counterfactual
analysis more effective. Subsequently, abductive reasoning
is depicted as an approach through which interpretive
researchers can apply the central resources of causal analy-
sis and also integrate the emic with the etic, thereby linking
an individual piece of interpretive research with the extant
body of knowledge on the focal ?eld. Thereafter the mes-
sage of the paper is illustrated by carefully opening up the
embedded causal storyline of one in?uential representative
of interpretive management accounting research, high-
lighting howthe conceptual weapons of causal exploration,
outlined in this paper, can be employed.
Finally, conclusions are drawn.
The regularity perspective versus the counterfactual
account of causality
Much of the ambiguity between causality and interpre-
tive research is probably driven by a limited and outdated
perception of causality based only on regularity. This roots
back to the Humean notion of causality as nothing but reg-
ular coincidence of phenomena, the rest being, Hume
argued, useless metaphysics, and the idea of a covering
law, according to which ‘explaining’ means the ability to
?t a phenomenon under a general law, such as falling
objects and the law of gravity (e.g. Raatikainen, 2011; cf.
von Wright, 1970).
Over recent decades, the positivistically tuned regular-
ity perspective on causality has been challenged by and
complemented with the counterfactual account of causality
(e.g. Morgan & Winship, 2007) that states: ‘‘An event Y
depends causally on a distinct event X if and only if both
X and Y occur, and if X had not occurred, then Y would
not have occurred either’’ (Lewis, 1973, p. 9).
1
This essen-
tially argues that explanatory factors are things that made
a difference to the thing being explained and that causality
refers to such dependency relations (Ruben, 1990;
Woodward, 2003).
The counterfactual approach to causality builds on a
systematic analysis of ‘what-if’ questions that compare fac-
tual observation with a counterfactual conditional, typi-
cally negating the actual observation, to test whether or
not the putative causal relationship holds.
2
In this notion
of causality, while not intending to argue against the corre-
lation of events being a cornerstone of causality, a constant
conjunction of events is perceived neither as a suf?cient
nor a necessary condition for its manifestations (e.g.
Kakkuri-Knuuttila, 2006). Hence, in contrast to Hume and
his followers, this notion of causality does not require regu-
larity and is, in principle, applicable wherever – also when
exploring unique situations in the ?eld, for instance such
that are typical of interpretive research in management
accounting. The conceptual move from the regularity per-
spective to the counterfactual account of causality is there-
fore a most signi?cant step with regard to how we
consider causality and its potential in the context of inter-
pretive research.
However, it is crucial to bear in mind that while the
counterfactual de?nition of causality is analytically strong,
it also implies that as the counterfactual in the putative
causal dependence naturally never simultaneously occurs
with the observable factual – strictly speaking, not even
in the case of controlled experiments – no causal argument
is ever de?nitive. This is illustrated in Fig. 1 in the case of a
controlled experiment.
Accordingly, with regard to the test group, we can only
observe the outcome in the case of treatment X, while the
counterfactual (i.e. the outcome of the test group in the
case of non-X) remains inevitably unobservable. On the
other hand, with regard to the control group, we can only
observe the outcome in the case of non-X, the counterfac-
tual (i.e. the outcome in the control group in the case of X)
remaining unobservable through necessity. As illustrated
in Fig. 1, even in the case of the strongest possible empiri-
cal research designs, causal arguments are, at least partly,
based on theorization and thought experiments (Holland,
1986; cf. Morgan & Winship, 2007).
3
An illustrative stylized example of counterfactuality
and thought experiments would be an attempt to explain
why a house had been razed by ?re. Suf?cient parallel
explanatory factors might include the house being struck
by lightning, a short circuit having occurred or the result
of pyromania. In a given situation, a lightning strike or
short circuit might possibly be eliminated and the thought
experiment, which mobilizes counterfactuals, focuses on
the potential action of a pyromaniac and takes into consid-
eration the evidence found. However, a necessary causal
1
Please note that the condition ‘other things equal’ is embedded in this
de?nition and hence, for instance, a now controlled Z can also be a relevant
explanatory factor. Therefore, X ? Y and Z ? Y can, at least in principle,
exist in parallel, indicating that while counterfactual analysis can identify
necessary or suf?cient causal factors (i.e. conditions) there can also be other
such factors. An outstanding account of the complex relations between
necessary and suf?cient conditions in causality, formulated as the so-called
INUS-conditions, is put forward by Mackie (1974).
2
The counterfactual account of causality is consistent with the so-called
manipulation theory of causality, according to which X is an explanatory
factor of Y if by producing X, other things equal, we can make Y happen.
This is a most natural method for distinguishing causality from mere
correlation (Woodward, 2003; cf. von Wright, 1970).
3
Mandel (2011) provides a profound analytical mapping on the land-
scape of necessary thought experiments, which he terms ‘‘mental simula-
tions’’, relating to causal explanations.
560 K. Lukka / Accounting, Organizations and Society 39 (2014) 559–566
factor for a ?re is always oxygen, although as that so self-
evidently exists and therefore makes no difference, it
would not normally even be mentioned. In this example,
the burnt-down house requires a combination of at least
one of the suf?cient factors and the necessary factor to
form a plausible explanation. Nevertheless, the explicated
explanation would probably only mention the alleged suf-
?cient factor (e.g. pyromania) as the suggested explanation
(cf. Mackie, 1974).
The counterfactual account of causality and interpretive
research
The contemporary perspectives on causality, outlined
above, open an avenue for arguing, in contrast to estab-
lished perceptions, that causal explanations are possible
also in the context of interpretive research; for instance,
in the area of management accounting. The main argument
of this paper is that in addition to ?nding a meaningful role
for causal thinking in interpretive research, it can also sup-
port interpretive research on its own terms without com-
promising its unique characteristics. In fact, as will be
explained below in this paper, linking causal explanations
to interpretive research places some of its central strengths
at the forefront.
While the term ‘explanation’ at times surfaces in the
context of the regularity notion of causality, often termed
the ‘Hempel-Oppenheim model’ or ‘the subsumption
model’, such explanations are ‘thin’ as causality is, in the
Humean spirit, eventually perceived as nothing but a con-
stant conjunction of events (cf. Hempel & Oppenheim,
1948). Hence, the processes and mechanisms through
which causality works need not be focal interests in this
vein of causal thought.
In contrast, the more recently established counterfac-
tual account of causality has a natural and serious linkage
to explanation. In addition, and importantly from the per-
spective of interpretive research, based on Wittgenstein’s
later philosophy in particular, the contemporary philoso-
phy of causality tends to closely link explanation with
interpretation and understanding. For Wittgenstein, in his
later ‘ordinary language philosophy’, understanding was
the correlate of explanation; to explain is to create under-
standing. Moreover, the attribution of understanding
means the ability to employ acquired information in action
and reasoning in ways sanctioned by the relevant commu-
nity (cf. Wittgenstein, 1953, §§ 148–159), thereby linking it
with a social aspect (cf. Chua, 1986). Accordingly, explana-
tion does not notably differ from understanding because
interpreting action correctly, that is, identifying the factors
(i.e. reasons) on which the action depended, simply means
explaining it (e.g. Henderson, 1993). From this perspective,
when interpretive studies emphasize the role of subjective
experiences in the construction of social reality, they can-
not eventually omit the explanatory aspect of the endeav-
our (Kakkuri-Knuuttila & Lukka, 2008).
But how does the idea of tracking down causality and
generating causal explanations relate to the tendency of
interpretive research to address the meanings of knowl-
edgeable agents in their life-worlds? This will be explored
in the next section.
Knowledgeable agents, subjective meanings and
causality
As interpretive research is interested in meanings
developed by subjects, it is closely related to the idea that
reality is, at least partially, socially constructed (e.g. Berger
& Luckmann, 1966; Hacking, 1999). Berger and
Luckmann’s (1966) core claim is that any adequate theo-
retical understanding on society necessarily requires an
integration of subjectivist and objectivist accounts. Hence,
society is both an objective fact and constituted by activi-
ties, which express subjective meanings. Through pro-
cesses of externalization, the latter have the potential to
become objecti?ed as facts that then re-affect the individ-
uals and their formation of new meanings. In the account-
ing context, Hines (1988) emphasizes that the social
construction of reality does not start from nothing; indi-
vidual action and interaction are embedded in social struc-
tures that pre-date the individual. New inter-subjective
constructions emerging in communication arise in the con-
text of particular social structures, which become continu-
ously sustained or reshaped through that communication.
Knowledgeable agents are often able to conceptualize
and re?ect upon the practices they entertain, leading to
the issue of double hermeneutics (Giddens, 1984). Hence,
while agents can trigger or hinder the actualization of par-
ticular causal mechanisms to implement their agenda, they
probably cannot, at least in the short term, change the cau-
sal mechanisms per se (Durand & Vaara, 2009). However, as
the entire notion of social constructionism is based on the
idea that agents often have the possibility of acting other-
wise (Hacking, 1999), the fact that knowledgeable agents’
intentions and actions are changeable and, over time,
new social structures and mechanisms can emerge, imply
that the stability of causal relationships in social sciences
is always at risk of being fragile (Alt, 2009; Mitchell, 2009).
Knowledge on peoples’ meanings (i.e. their understand-
ing of ‘something as something’) plays a central role when
we aim to explain human action and interaction. Seem-
ingly similar behavior can have widely different meanings
Test group Control group
Treatment X Outcome of variable Y = Y (X)
(observed)
Counterfactual (unobservable)
No treatment X Counterfactual (unobservable) Outcome of variable Y = Y (non-X)
(observed)
Fig. 1. Counterfactuality in a controlled experiment.
K. Lukka / Accounting, Organizations and Society 39 (2014) 559–566 561
and also causal effects in different contexts, both as
intended and as interpreted, and these do not necessarily
coincide; a classic example being the wink of an eye in
social interaction. Agents’ beliefs regarding existing causal
relations in the contexts where they operate have notable
bearing when they act; hence, both the beliefs and their
referents are relevant to understanding peoples’ meanings
and explaining their actions. As noted by Wittgenstein
(1953), peoples’ meanings are often not merely their pri-
vate, inner ostensions, but also include their abilities to
communicate and interact with other subjects (cf. Chua,
1986; Geertz, 1973). This perspective not only emphasizes
the social aspect of meanings but also suggests that
interpretive research does not need to be limited only to
exploring the emic understandings of single individuals
(Kakkuri-Knuuttila et al., 2008).
While, in line with the core agenda of the interpretive
paradigm, interpretive researchers certainly try to under-
stand how people receive, develop and send meanings in
social instances, they can and do tend to employ them
when attempting to explain the phenomena to which
these meanings relate. In fact, following Wittgenstein’s
(1953) line of thought, an argument can be developed that
not even the most subjectivist position of interpretive
research can entirely distance itself from explanation.
4
Accordingly, to be able to cope with living among other peo-
ple, we need the ability to understand what kind of causal
beliefs, meanings and intentions drive the actions of others
while simultaneously considering how our actions, funda-
mentally driven by similar factors, are read by others. Peo-
ples’ meanings, which are realized in their actions, become
part of the world’s causal linkages and can (and tend to)
have a lot of causal power (cf. Kakkuri-Knuuttila et al.,
2008; Lukka & Modell, 2010).
It is crucial to note that including causality in the anal-
ysis by no means implies that the world would be
assumed as deterministic. For interpretive research, it
rather means the incorporation of the idea that inten-
tional actors with a certain amount of agency can and
tend to take relevant causal relations within the context
where they operate into consideration and that their
actions can and tend to have causal consequences. These
features hold even though the web of causalities is always
inherently fragile in the social world (cf. Melan, 2013, pp.
117–119).
Causality in interpretive management accounting
research
Reviewing the literature of interpretive management
accounting research suggests that while causality is typi-
cally played down in what is explicitly stated, deconstruct-
ing the studies’ contents often discovers storylines that
include causal explanations. Indeed, the bulk of interpre-
tive studies in management accounting do not quite ?t into
the narrow window of the extreme subjectivist interpre-
tive position suggested by Winch (1958) or Burrell and
Morgan (1979). While interpretive studies indeed offer
thick descriptions (Geertz, 1973) or rich insights (Ahrens
& Dent, 1998; Dyer & Wilkins, 1991) and explore peoples’
subjective meanings in line with the interpretive para-
digm, they also tend to include objecti?ed ‘facts’ and
develop causal explanations, although most often only
implicitly (Ahrens & Chapman, 2006; Kakkuri-Knuuttila
et al., 2008; Lukka & Modell, 2010). It is therefore relieving
that the contemporary theory of causal explanation, based
on the notion of counterfactuality, supports conducting
case research and generating local explanations. Accord-
ingly, to discuss causal explanations, we do not need to
seek law-like generalizations and apply the covering law
model. Instead, we are encouraged to seek causal explana-
tions that capture (difference making) dependency rela-
tions between things in the world in any locality
(Kakkuri-Knuuttila et al., 2008; cf. Durand & Vaara,
2009). Clearly, the need always remains for causal claims
to sustain analysis with counterfactual conditionals (cf.
Mandel, 2011).
One of the strongholds of interpretive research, the ten-
dency to focus on responding to ‘how’ type of questions,
speci?cally addresses the question concerning how the
alleged explanatory factors work in producing outcomes
(Weick, 1989; Morgan & Winship, 2007; Tsang &
Ellsaesser, 2011). This is a notable difference from many
other kinds of study; for instance, surveys and archival
research in which analysis of processes is typically dif?cult,
if not impossible. Interpretive studies can progress over
and above merely describing associations between vari-
ables to exploring the processes and mechanisms by which
they are generated; that is, towards looking profoundly at
the linkages (i.e. the ‘arrows’) between the elements of the
explanatory scheme (cf. Ahrens & Chapman, 2006; Durand
& Vaara, 2009). This means, based on careful and profound
?eld work, opening up how processes of interest, including
causal linkages, work out from conditions to outcomes
regarding the explored research question, thereby possibly
revealing particular mechanisms that are at work in the
subject area (cf. Dent, 1991). This is precisely what inter-
pretive researchers in management accounting often seek
to achieve in their typical endeavours to ‘make sense’ of
what they have observed in their ?eld work. Due to this,
it can be argued that the explanations produced by inter-
pretive studies are particularly informative and strong.
Hence, interpretive studies can provide thick explana-
tions that are deeply rooted in the life-worlds of the people
being studied and, hence, incorporate peoples’ meanings,
which is the main focus of interest in interpretive research,
and also the processes and mechanisms that tie together
the elements of the developed explanation (cf. Lukka &
Modell, 2010). Consequently, when causal explanations
form part of the argumentation structure of an interpretive
study, there is no need for them to be hidden. On the con-
trary, they can be made even more explicit, which would
tend to sharpen and strengthen the arguments developed
while maintaining the interpretive framing of the study
and remaining faithful to the core agenda of interpretive
research (Kakkuri-Knuuttila et al., 2008; cf. Hammersley
& Atkinson, 2009, p. 185).
4
In addition to this epistemological grounding, the audiences of
interpretive research are likely to expect some theoretical (i.e. etic)
structure that overarches mere accounts of emic understandings.
562 K. Lukka / Accounting, Organizations and Society 39 (2014) 559–566
Contrastive thinking as a resource for causal analysis
An important aspect of contemporary thinking on cau-
sality is the acknowledgement that, in developing causal
claims, selection is an inherent part of scholarly work. It
is worth bearing in mind that causes typically have several
different kinds of effect while, in turn, effects tend to have
numerous possible and different kinds of cause.
5
Typically,
there is considerable room for choice regarding the screen-
ing of relevant explanatory factors and as such screening is
never a neutral, contextual or interest-free activity (van
Fraassen, 1977), the process of generating explanations
practically always needs to be carefully focused by the
researcher him/herself. It is also crucial to realize that, typ-
ically, we do not explain events or phenomena per se, but
only particular aspects of them (Hempel, 1965; cf. Tsang &
Ellsaesser, 2011).
An epistemologically sound approach is to let our
selected theoretical focus drive the choice of which possi-
ble explanatory factors are of true interest to us (e.g.
DiMaggio, 1995). The sharpening of explanations in terms
of contrastive thinking is of notable help in this regard
(Lipton, 1990; Tsang & Ellsaesser, 2011; Ylikoski, 2007).
Consider an innocent looking example observation, which
could certainly be part of an interpretive management
accounting study on budgetary biasing (cf. e.g. Lukka,
1988):
Divisional managers underestimate their sales forecasts
when submitting their budgets to the headquarters.
How can this example of budgetary biasing be
explained? Contrastive thinking advises us to ?rst sharpen
the target of the explanation (i.e. the explanandum). In so
doing, we get, for instance, the following four more speci?c
sub-questions, or so-called allomorphs (see Tsang &
Ellsaesser, 2011): Why divisional managers, not some other
party? Why underestimate rather than, say, overestimate?
Why sales forecasts, not some other item? Why when deal-
ing with the headquarters, not with some other party? All
of these questions lead to somewhat different explanation
focuses and processes of counterfactual analysis (cf. Lukka
& Modell, 2010).
The central lesson here is that we can easily endeavour
to explain overly broad sets of aspects of phenomena,
which tend to lead to weak and ambiguous argumentation,
potentially limiting the possibility of truly increasing our
understanding. Thoughtful application of contrastive
thinking makes analysis with counterfactual conditionals
more effective and also contributes to avoiding so-called
over-generation of causes (e.g. Hoerl et al., 2011). In fact,
there is reason to argue that careless disregard of the con-
trastive thinking aspect makes it very dif?cult to develop
causal claims that can be rendered meaningful and plausi-
ble (cf. Lukka & Modell, 2010).
The approach to causal exploration introduced in this
paper, revolving around the counterfactual notion of
causality, contrastive thinking and the power of thought
experiments suggests focusing on relatively narrow slices
of the world to enable contribution to the clarity and
strength of the generated explanation. However, the chal-
lenge of keeping the holistic nature of interpretive research
on the agenda remains and, as such, the abductive mode of
reasoning partially comes to the rescue as it, among other
things, tends to contextualize theoretically the core empir-
ical ?ndings and thereby helps to link the study with the
extant body of knowledge on the focal ?eld.
Abductive mode of reasoning and causal analysis
Interpretive research often functions through an abduc-
tive process (Dubois & Gadde, 2002; Peirce, 1903) instead
of the better known inductive and deductive modes of
reasoning.
6
Abductive reasoning typically starts from a
striking empirical observation that begs an explanation, trig-
gering the process of ‘making sense’ to begin. However, as
the observation is not perceived to exist in a theoretical vac-
uum, an abductive researcher then commences careful
development of theoretical explanations with the help of
everything he/she knows, both empirically and theoretically,
on the examined issue (Hanson, 1958, 1961). The process of
abductive reasoning can be paralleled to detective work as it
collects clues, focuses on those that are most promising and
attempts to put them into perspective with the conditions
and other evidence at hand. In brief, the abductive explana-
tory scheme can be formulated as follows (Peirce, 1960, p.
117):
Surprising phenomenon Y is observed.
But if proposition X would be true, then Y would be a
matter of course.
There is reason to assume X is true.
The outcome of an abductive research process is typi-
cally a causal explanation where X is concluded to plausi-
bly explain Y, representing so-called inference to the best
explanation. Rather than being a linear process of reason-
ing, an abductive process includes going back and forth
between empirical ?ndings and theoretical elements per-
ceived to be of relevance and interest (Lukka & Modell,
2010). Thought experiments relating to the application of
contrastive thinking and counterfactual analysis form an
integral part of the abductive reasoning process.
It is indeed worth emphasizing the important role of
contrastive thinking within the abductive reasoning pro-
cess; factors included in the various stages of the analysis
do not come ‘out of the blue’ but are selected, based on
particular grounds. In the case of interpretive research,
the endless spectrum of actors’ potential differing mean-
ings that might play a role in the research setting renders
5
Cf. Morgan and Winship (2007) and their discussion on the difference
between primarily seeking the causes of effects versus the effects of causes.
6
Abduction, developed and coined by C.S. Peirce approximately a
century ago, has much similarity with grounded theorization that was
introduced to the agenda of social studies in the 1960s by Glaser & Strauss
(1967) (cf. Strauss & Corbin, 1998; Suddaby, 2006). Hanson (1958)
criticized the dominant hypothetico-deductive and inductive models for
representing the logic of a completed research report rather than the actual
research process.
K. Lukka / Accounting, Organizations and Society 39 (2014) 559–566 563
their identi?cation an even more challenging matter. Con-
trastive thinking is crucial at the outset of the process as it
helps the researcher identify the interesting observation
perceived as requiring explanation – in other words, why
was precisely this observation picked up in contrast to
another one? Contrastive thinking is employed over the
entire abductive process as the researcher focuses his/her
attention on particular potential emerging explanatory fac-
tors instead of some others, collects further evidence based
on his/her theoretical contemplation as well as clues from
the ?eld and also runs thought experiments on expected
relationships between things (i.e. mobilizing counterfac-
tual conditionals). Contrastive thinking offers a focused
method with which to consider one’s own ideas and etic
preferences in relation to those of others.
To exemplify abduction in the context of interpretive
research, let us consider further the case of budgetary bias-
ing introduced in the previous section. Actors’ subjective
meanings relate particularly to their intentions concerning
their biasing attempts and different intentions can have
different likely consequences. In this regard, Lukka (1988)
identi?ed two main types of intention for a manager to
manipulate his/her budget: resource intention (i.e. aiming
to get additional resources) and performance evaluation
intention (i.e. aiming to receive more positive evaluations).
While a functionalist researcher would focus on trying to
measure bias and its quanti?able determinants, an inter-
pretive researcher pays attention to the process of biasing,
including its subjective elements. The contrasts are
brought in partly from extant theory; however, the
researcher still has to decide what he/she wishes to con-
trast and against what. In line with abduction, an interpre-
tive researcher can attempt to explain a particular
observed instance of budgetary biasing (e.g. contrasting
biasing efforts towards pessimistic and optimistic direc-
tions) and develop an explanation for it based on the anal-
ysis of the subjective biasing intentions of the focal actor
(cf. Lukka, 1988).
7
As abductive reasoning inherently combines empirical
and theoretical elements to develop a plausible argument
or set of arguments, it helps an interpretive researcher to
maintain the holistic feature of interpretive research,
despite conducting parts of his/her primarily empirically
grounded analysis in a focused and selective manner, based
on contrastive thinking. Abductive reasoning is inclined to
tie together the components of the study and has the poten-
tial of contextualizing the empirical ?ndings, possibly
including causal elements, to form a theoretically more
encompassing and consistent set of arguments. Through
abduction, emic-level empirical analysis at the core of inter-
pretive research is connected to etic-level analysis and
knowledge. This move makes the individual piece of inter-
pretive researchnotably more theoretical as it becomes part
of the knowledge set on the focal ?eld. This is a way by
whichtheoretical diversity anddevelopment ininterpretive
research across studies might be understood as an exercise
in complementarity rather than fragmentation.
Examples selected from published interpretive manage-
ment accounting research such as Covaleski and Dirsmith
(1986), Dent (1991), Ahrens (1996) and Vaivio (2006) give
other illustrations regarding how interpretive scholars
effectively apply abduction as their mode of reasoning. In
all of these studies, the authors offer implicit clues to hav-
ing employed counterfactual analysis and contrastive
thinking in making theoretical sense of their ?ndings and
delineating plausible explanations to accounting-related
phenomena (cf. Lukka & Modell, 2010). In all of these stud-
ies, the emic and the etic interact fruitfully, resulting in
interpretive studies producing more than merely rich emic
accounts. They offer theoretical arguments including
explanations and, as these are deeply rooted in the life-
worlds of the people being studied, we can term these
explanations thick. In the following section, one of the
above-mentioned studies will be analyzed in more detail
to more speci?cally illustrate how causal arguments and
a storyline can be generated by interpretive research,
although in an embedded manner.
Developing a causal explanatory storyline in
interpretive research: An example
Dent’s (1991) widely cited article ‘‘Accounting and
organizational cultures: A ?eld study of the emergence of
a new organizational reality’’ is a longitudinal interpretive
case study on a profound cultural change in an organiza-
tion, Euro Rail (ER), focusing on the roles played by new
accounting practices along the dynamic process of the cul-
tural transformation in the ?rm.
Dent’s study traces the process of major change in a
state owned railway company (i.e. ER), starting with the
introduction of new kinds of manager (i.e. ‘‘business man-
agers’’) under governmental pressure, culminating in a
profound cultural change whereby the old ‘‘railway
culture’’ became gradually replaced by a new ‘‘business
culture’’. The central interpretive content of Dent’s analy-
sis, in which he mobilized Geertz (1973) in particular,
encompasses that he carefully examined the process and
the ways by which the business managers gradually and
systematically acquired in?uence by transforming the
structures of meaning in the organization through, among
other things, various actions with considerable symbolic
signi?cance. Dent argued that the transformation of a
few accounting practices played a major role in this pro-
cess by changing the way people conceptualized key parts
of the organization and also the very role of the organiza-
tion in relation to the larger community. He identi?ed
and described a particular process and pattern in the busi-
ness managers’ actions whereby they gradually broadened
their in?uence at ER to implement their change agenda.
When causality is understood in the sense of referring to
dependencies, which do not need to be regularities and can
be contextual, Dent’s piece includes numerous explana-
tions (cf. Kakkuri-Knuuttila et al., 2008). The delineated
7
The researcher can also start the analysis from observed subjective
intentions and attempt to predict the likely biasing possibilities. It is worth
noting that the precise subjective biasing intention of the actor can make a
notable concrete difference. Resource intention, in contrast to performance
evaluation intention, tends to systematically relate to an attempt to
negotiate for more resources, leading to overestimations in the cost side of
the budget (cf. Lukka, 1988).
564 K. Lukka / Accounting, Organizations and Society 39 (2014) 559–566
linkages (i.e. the ‘arrows’ between various things) that tie
the isolated descriptions of events into a coherent whole
can be regarded as forming a set of causal explanations,
rendering the causal reading of the piece quite natural.
According to Dent, the accounting-related changes at ER
were important causal explanatory factors for the cultural
change, without which there would probably have been
no such change, thereby offering examples of counterfac-
tual conditionals that are frequently embedded throughout
his paper.
When Dent identi?es phenomena begging for explana-
tion (i.e. explanandums) at various levels of his analysis,
examples of contrastive thinking turn up in plenty, imply-
ing that he had selected speci?c focus issues to drive his
examination, forming the key elements of his storyline.
In their application, counterfactual thought experimental
analysis is present, although typically only implicitly. To
illustrate this, the process of cultural change began with
the appointment of new kinds of manager at ER; however,
why were they needed rather than, say, new kinds of loco-
motive engine? Because ER’s environment had changed in
such a way that its top management especially felt more
managerial resources would be required. The kind of man-
agers appointed were business managers; however, why
precisely this kind of managers rather than, say, hiring
more operational managers? Because ER’s top manage-
ment especially lacked marketing and business expertise
and a bottom line mentality. Had the government not
turned ‘hostile’, the need for new kinds of managerial
resources and practices would probably have not emerged.
Also, had the nature of the government’s hostility been dif-
ferent (e.g. less focused on ?nancial issues), the type of
newly recruited managers could have been different, too
(cf. Kakkuri-Knuuttila et al., 2008).
Overall, Dent’s (1991) analysis relies strongly on the
‘power of example’ (Flyvbjerg, 2001), whereby particular
critical events from ER’s trajectory of change are identi?ed
for careful and profound scrutiny and are presented as key
drivers for further changes. The analysis is formed by
numerous partial episodes and narratives, bound together
with the help of the theoretical resources introduced at the
beginning of the paper, hence providing a well-developed
example of how abductive reasoning can be usefully mobi-
lized in interpretive research. Therefore the outcome of
Dent’s (1991) analysis is inherently holistic. The paper also
offers a solid example of thick explanations as it presents a
careful emic account of the evolving change in the mean-
ing structures of the actors at ER and also provides an
explanation on the change from an etic perspective (cf.
Lukka & Modell, 2010).
Conclusions
This paper suggests how the contemporary theory of
causality, based on the notions of counterfactuality and
contrastive thinking, offers helpful direction on how to
generate plausible causal arguments in interpretive stud-
ies. For the interpretive researcher, if he/she so wishes
and it is required by the research question, they suggest
a route from rich emic-based accounts at the core of
interpretive research to thick explanations. Perhaps with
some surprise, causality can be included in interpretive
research framings without compromising the unique fea-
tures of such research, even by building on some of its
strongholds. In particular, this is due to the tendency of
interpretive studies to focus on responding to ‘how’ ques-
tions. Typically, interpretive research profoundly investi-
gates how associations between phenomena occur and
explores the processes and mechanisms that generate out-
comes from particular conditions, thereby shedding light
on the ‘arrows’ between variables.
The paper emphasizes and illustrates the role of con-
trastive thinking as a most useful tool with which to focus
on the development of causal explanations, rendering
analysis with counterfactual conditionals, including
thought experiments, more effective. The abductive mode
of reasoning, referring to the iterative movement between
empirical materials and theory elements and contributing
to making the set of explanations more holistic and theo-
retically contextualized, forms the principal way to gener-
ate plausible explanations in interpretive research. While
abduction works through all of the key tools of causal
exploration introduced in this paper, the role of contrastive
thinking is particularly notable as it guides the researcher
in his/her choices of analytical focus from the beginning
of the abductive process.
Explicating causal explanation in interpretive research,
for instance in the ?eld of management accounting, has
the potential to add to the clarity and strength of the argu-
ments developed, while also leading to the need to bear in
mind their inherent boundary conditions, such as the lim-
its of counterfactual analysis and those of abductive rea-
soning and also the fact that the stability of social
structures and relationships is always an open question.
Despite these caveats, it is suggested that the potential of
applying causal thinking more explicitly in interpretive
research is notable.
Acknowledgements
This paper is based on the plenary presentation at
GMARS conference in Copenhagen in June 2012. I would
like to thank Chris Chapman, Markus Granlund, Olli Koisti-
nen, Mikael Melan, Hanna Meretoja, Sven Modell, Susan
O’Leary, Jan P?ster and the anonymous reviewers for their
helpful comments along the process of preparing this
paper. Mike Power is thanked for his highly signi?cant
encouragement.
References
Ahrens, T. (1996). Styles of accountability. Accounting, Organizations and
Society, 21(2 & 3), 139–173.
Ahrens, T., & Chapman, C. (2006). Doing qualitative ?eld research in
management accounting: Positioning data to contribute to theory.
Accounting, Organizations and Society, 31(8), 819–841.
Ahrens, T., & Dent, J. F. (1998). Accounting and organisations: Realizing
the richness of ?eld research. Journal of Management Accounting
Research, 10, 1–39.
Alt, J. (2009). Comment: Conditional knowledge: An oxymoron? In C.
Mantzavinos (Ed.), Philosophy of the social sciences. Philosophical theory
and scienti?c practice (pp. 146–153). Cambridge University Press.
Berger, P., & Luckmann, T. (1966). The social construction of reality. Anchor
Books.
K. Lukka / Accounting, Organizations and Society 39 (2014) 559–566 565
Burrell, G., & Morgan, G. (1979). Sociological paradigms and organizational
analysis: Elements of the sociology of corporate life. Heinemann.
Chua, W. F. (1986). Radical developments in accounting thought. The
Accounting Review, 61(4), 601–632.
Covaleski, M. A., & Dirsmith, M. W. (1986). The budgetary process of
power and politics. Accounting, Organizations and Society, 11(3),
193–214.
Dent, J. F. (1991). Accounting and organizational cultures: A ?eld study of
the emergence of a new organizational reality. Accounting,
Organizations and Society, 16(8), 705–732.
Denzin, N. K. (1983). Interpretive interactionism. Sage.
DiMaggio, P. J. (1995). Comments on ‘‘What theory is not’’. Administrative
Science Quarterly, 40(3), 391–397.
Dubois, A., & Gadde, L.-E. (2002). Systematic combining: An abductive
approach to case research. Journal of Business Research, 55(7),
553–560.
Durand, R., & Vaara, E. (2009). Causation, counterfactuals and competitive
advantage. Strategic Management Journal, 30(12), 1245–1264.
Dyer, W. G., & Wilkins, A. L. (1991). Better stories, not better constructs, to
generate better theory: A rejoinder to Eisenhardt. Academy of
Management Review, 16, 613–619.
Flyvbjerg, B. (2001). Making social science matter: Why social inquiry fails
and how it can succeed again. Cambridge University Press.
van Fraassen, B. (1977). The pragmatics of explanation. American
Philosophical Quarterly, 14, 143–150.
Geertz, C. (1973). The interpretation of cultures. New York: Basic Books.
Giddens, A. (1984). The constitution of society. Outline of the theory of
structuration. Polity Press.
Glaser, B. G., & Strauss, A. L. (1967). The discovery of grounded theory:
Strategies for qualitative research. New York (NY): Aldine de Gruyter.
Hacking, I. (1999). The social construction of what? Harvard University
Press.
Hammersley, M., & Atkinson, P. (2009). Ethnography principles in practice
(3rd ed.). Routledge.
Hanson, N. R. (1958). Patterns of discovery. Cambridge: University Press.
Hanson, N. R. (1961). Is there a logic of scienti?c discovery? In H. Fiegl &
M. Grover (Eds.), Current issues in the philosophy of science. New York:
Holt, Rinehart & Winston.
Headland, T. N. (1990). Introduction: A dialogue between Kenneth Pike
and Marvin Harris. In T. N. Headland, K. L. Pike, & M Harris (Eds.),
Emics and etics: The insider/outsider debate (pp. 13–27). Sage.
Hempel, C. (1965). Aspects of scienti?c explanation. Free Press.
Hempel, C. G., & Oppenheim, P. (1948). Studies in the logic of explanation.
Philosophy of Science, 15(2), 135–175.
Henderson, D. (1993). Interpretation and explanation in the human sciences.
State University of New York Press.
Hines, R. (1988). Financial accounting: In communicating reality, we
construct reality. Accounting, Organizations and Society, 13(3),
251–261.
Hoerl, C., McCormack, T., & Beck, S. R. (2011). Introduction:
Understanding counterfactuals and causation. In C. Hoerl, T.
McCormack, & S. R. Beck (Eds.), Understanding counterfactuals,
understanding causation (pp. 1–15). Oxford University Press.
Holland, P. W. (1986). Statistics and causal inference. Journal of the
American Statistical Association, 81(9), 45–70.
Kakkuri-Knuuttila, M.-L. (2006). Kausaalisuhteet ja selittäminen
tulkitsevassa tutkimuksessa (Causal relations and explanation in
interpretive research). In K. Rolin, M.-L. Kakkuri-Knuuttila, & E.
Henttonen (Eds.), Soveltava yhteiskuntatiede ja ?loso?a (pp. 54–87).
Gaudeamus (Applied social science and philosophy).
Kakkuri-Knuuttila, M.-L., Lukka, K., & Kuorikoski, J. (2008). Straddling
between paradigms: A naturalistic philosophical case study on
interpretive research in management accounting. Accounting,
Organizations and Society, 33(2&3), 267–291.
Lewis, D. (1973). Counterfactuals. Harvard University Press.
Lipton, P. (1990). In D. Knowles (Ed.), Explanation and its limits
(pp. 247–266). Cambridge University Press.
Lukka, K. (1988). Budgetary biasing in organizations: Theoretical
framework and empirical evidence. Accounting, Organizations and
Society, 13(3), 281–301.
Lukka, K., & Modell, S. (2010). Validation in interpretive management
accounting research. Accounting, Organizations and Society, 35(4),
462–477.
Mackie, J. L. (1974). The cement of the universe: A study of causation. Oxford
University Press.
Mandel, D. R. (2011). Mental simulation and the nexus of causal and
counterfactual explanation. In C. Hoerl, T. McCormack, & S. R. Beck
(Eds.), Understanding counterfactuals, understanding causation
(pp. 147–170). Oxford University Press.
Melan, M. (2013). Will and mechanisms. Reports from the Department of
Philosophy. University of Turku.
Meretoja, H. (2012). Ymmärryskäsitteet ja toisen kohtaamisen ongelma
(concepts of understanding and the issue of encountering the other).
In V. Viljanen, H. Siipi, & M. Sintonen (Eds.), Ymmärrys (pp. 307–318).
Reports from the Department of Philosophy. University of Turku
(Understanding).
Mitchell, S. (2009). Complexity and explanation in social sciences. In C.
Mantzavinos (Ed.), Philosophy of the social sciences. Philosophical theory
and scienti?c practice (pp. 130–145). Cambridge University Press.
Morgan, S. L., & Winship, C. (2007). Counterfactuals and causal inference:
Methods and principles for social research. Cambridge University Press.
Peirce, C. S. (1903). The essential Pierce: Selected philosophical writings (Vol.
2). Bloomington: Indiana University Press.
Peirce, C. S (1960). In C. Hartshorne & P. Weiss (Eds.). Collected papers of
Charles sanders Peirce (Vols. V–VI). Belknap Press of Harvard
University Press.
Pike, K. L (1954). Emic and etic standpoints for the description of
behavior. In K. L. Pike (Ed.), Language in relation to a uni?ed theory of
the structure of human behavior (pp. 8–28). Summer Institute of
Linguistics.
Raatikainen, P. (2011). Teon teoria ja selittäminen (Theory of action and
explanation). Tiede ja edistys, 4, 291–300.
Ruben, D.-H. (1990). Explaining explanation. Routledge.
Sobel, M. E. (1995). Causal inference in the social and behavioral sciences.
In G. Armingger, C. C. Clogg, & M. E. Sobel (Eds.), Handbook of statistical
modeling for the social and behavioral sciences. Plenium Press.
Sobel, M. E. (1996). An introduction to causal inference. Sociological
Methods and Research, 24, 353–379.
Strauss, A., & Corbin, J. (1998). Basics of qualitative research: Techniques and
procedures for developing grounded theory. Thousand Oaks: Sage.
Suddaby, R. (2006). From the editors: What grounded theory is not.
Academy of Management Review, 49(4), 633–642.
Sutton, R. I., & Staw, B. M. (1995). What theory is not. Administrative
Science Quarterly, 40(3), 371–384.
Tetlock, P. E., & Belkin, A. (Eds.). (1996). Counterfactual thought
experiments in world politics. Princeton University Press.
Tsang, E. W. K., & Ellsaesser, F. (2011). How contrastive explanation
facilitates theory building. Academy of Management Review, 36(2),
404–419.
Vaivio, J. (2006). The accounting of ‘‘the meeting’’: Examining
calculability within a ‘‘?uid’’ local space. Accounting, Organizations
and Society, 31(8), 735–762.
Weick, K. (1989). Theory construction as disciplined imagination.
Academy of Management Review, 14(4), 516–531.
Weick, K. (1995). What theory is not, theorizing is. Administrative Science
Quarterly, 40(3), 385–390.
Winch, P. (1958). The idea of social science and its relation to philosophy.
Routledge & Kegan Paul.
Wittgenstein, L. (1953). Philosophical investigations. Basil Blackwell.
Woodward, J. (2003). Making things happen. Oxford University Press.
von Wright, G. H. (1970). Explanation and understanding. Cornell
University Press.
Ylikoski, P. (2007). The idea of contrastive explanandum. In J. Persson & P.
Ylikoski (Eds.), Rethinking explanation (pp. 27–42). Springer.
566 K. Lukka / Accounting, Organizations and Society 39 (2014) 559–566
doc_695864092.pdf
This paper suggests how the contemporary theory of causality, based on the notions of
counterfactuality and contrastive thinking, offers helpful direction on how to generate
plausible causal arguments in interpretive research. For an interpretive researcher, this
opens a route from rich emic accounts to thick explanations; however, only if he/she so
wishes and the research question so requires. Perhaps with some surprise, causality can
be included in interpretive research framings without compromising the unique features
of such research – actually by even building on some of its strongholds
Exploring the possibilities for causal explanation in interpretive
research
Kari Lukka
?
Turku School of Economics, University of Turku, Finland
a b s t r a c t
This paper suggests how the contemporary theory of causality, based on the notions of
counterfactuality and contrastive thinking, offers helpful direction on how to generate
plausible causal arguments in interpretive research. For an interpretive researcher, this
opens a route from rich emic accounts to thick explanations; however, only if he/she so
wishes and the research question so requires. Perhaps with some surprise, causality can
be included in interpretive research framings without compromising the unique features
of such research – actually by even building on some of its strongholds. Examples from
interpretive management accounting research will illustrate the message of the paper.
Ó 2014 Elsevier Ltd. All rights reserved.
Introduction
The main thesis of this paper is that causality has the
potential to be a useful part of interpretive research. It is
suggested that it is possible to ?nd a meaningful role for
causality in interpretive studies and that this can also be
achieved while conducting interpretive research on its
own terms, without compromising its unique characteris-
tics. In fact, it will be argued that linking causal explana-
tions to interpretive research places some of its central
strengths at the forefront. The paper elaborates the value
of this thesis, its support and speci?c implications. The
analysis is illustrated by examples from interpretive man-
agement accounting research.
Understanding meanings of often unique phenomena,
considering people as taking ‘something as something’
(e.g. Meretoja, 2012), is the starting point and characteristic
feature of interpretive research (e.g. Chua, 1986). It tends to
focus on the emic perspective, an examination on how the
research subject his/herself develops his/her meanings,
rather than the etic one, whereby the issue is the research-
er’s interpretation and theorization on the studied
phenomena (Denzin, 1983; Headland, 1990; Pike, 1954).
The key features of interpretive research have often been
perceived as implying a stark contrast to causal explana-
tion, hence even eliminating the latter from such research
(e.g. von Wright, 1970; Winch, 1958; Burrell, 1979; cf.
Kakkuri-Knuuttila, 2006). The paper investigates this
alleged tension between interpretive research and causal
explanation in its attempt to explore the possibilities of
causality in interpretive research. Without denying the
importance of such interpretive research that seeks out-
comes other than discovering causality, the purpose of this
paper is to offer insight on potential avenues that open for
an interpretive researcher regarding causal explanation,
based on a notable amount of the recent literature on cau-
sality and explanation in the wider social sciences; it is also
informed by recent advances in the philosophy of science.
Indeed, the analysis of this paper is motivated, in partic-
ular, by recent increasing interest in the wider social sci-
ences to understand better how researchers conduct, or
might conduct, their work regarding causal inferences
(e.g. Tetlock & Belkin, 1996; Morgan, 2007; Hoerl,
McCormack, & Beck, 2011). While it certainly is not the
sole task of scholarly work to produce explanations, they
tend to play a central role in much that is sought to be
achieved (cf. Sobel, 1995, 1996; Morgan & Winship,http://dx.doi.org/10.1016/j.aos.2014.06.002
0361-3682/Ó 2014 Elsevier Ltd. All rights reserved.
?
Tel.: +358 2 3339315; fax: +358 2 3338900.
E-mail address: kari.lukka@utu.?
Accounting, Organizations and Society 39 (2014) 559–566
Contents lists available at ScienceDirect
Accounting, Organizations and Society
j our nal homepage: www. el sevi er. com/ l ocat e/ aos
2007). Further, although not the only option, the purpose
of producing explanations based on causality appears to
be important across many areas of research, including
business studies (e.g. DiMaggio, 1995; Sutton & Staw,
1995; Weick, 1995).
The paper starts with an outline of the signi?cant,
although not yet widely noticed, variation in how causality
is understood in the philosophy of science. In particular,
this analysis introduces the recent emphasis on the coun-
terfactual account of causality, suggested to be of great
importance with regard to generating causal explanations
in interpretive research. Causality and causal explanations
are then examined in the context of interpretive research,
focusing on knowledgeable agents and how their meanings
develop and operate in the surroundings where they act.
After outlining the ways in which interpretive framings
can be employed in management accounting research, con-
trastive thinking is introduced as a helpful method to focus
causal exploration, thereby making the counterfactual
analysis more effective. Subsequently, abductive reasoning
is depicted as an approach through which interpretive
researchers can apply the central resources of causal analy-
sis and also integrate the emic with the etic, thereby linking
an individual piece of interpretive research with the extant
body of knowledge on the focal ?eld. Thereafter the mes-
sage of the paper is illustrated by carefully opening up the
embedded causal storyline of one in?uential representative
of interpretive management accounting research, high-
lighting howthe conceptual weapons of causal exploration,
outlined in this paper, can be employed.
Finally, conclusions are drawn.
The regularity perspective versus the counterfactual
account of causality
Much of the ambiguity between causality and interpre-
tive research is probably driven by a limited and outdated
perception of causality based only on regularity. This roots
back to the Humean notion of causality as nothing but reg-
ular coincidence of phenomena, the rest being, Hume
argued, useless metaphysics, and the idea of a covering
law, according to which ‘explaining’ means the ability to
?t a phenomenon under a general law, such as falling
objects and the law of gravity (e.g. Raatikainen, 2011; cf.
von Wright, 1970).
Over recent decades, the positivistically tuned regular-
ity perspective on causality has been challenged by and
complemented with the counterfactual account of causality
(e.g. Morgan & Winship, 2007) that states: ‘‘An event Y
depends causally on a distinct event X if and only if both
X and Y occur, and if X had not occurred, then Y would
not have occurred either’’ (Lewis, 1973, p. 9).
1
This essen-
tially argues that explanatory factors are things that made
a difference to the thing being explained and that causality
refers to such dependency relations (Ruben, 1990;
Woodward, 2003).
The counterfactual approach to causality builds on a
systematic analysis of ‘what-if’ questions that compare fac-
tual observation with a counterfactual conditional, typi-
cally negating the actual observation, to test whether or
not the putative causal relationship holds.
2
In this notion
of causality, while not intending to argue against the corre-
lation of events being a cornerstone of causality, a constant
conjunction of events is perceived neither as a suf?cient
nor a necessary condition for its manifestations (e.g.
Kakkuri-Knuuttila, 2006). Hence, in contrast to Hume and
his followers, this notion of causality does not require regu-
larity and is, in principle, applicable wherever – also when
exploring unique situations in the ?eld, for instance such
that are typical of interpretive research in management
accounting. The conceptual move from the regularity per-
spective to the counterfactual account of causality is there-
fore a most signi?cant step with regard to how we
consider causality and its potential in the context of inter-
pretive research.
However, it is crucial to bear in mind that while the
counterfactual de?nition of causality is analytically strong,
it also implies that as the counterfactual in the putative
causal dependence naturally never simultaneously occurs
with the observable factual – strictly speaking, not even
in the case of controlled experiments – no causal argument
is ever de?nitive. This is illustrated in Fig. 1 in the case of a
controlled experiment.
Accordingly, with regard to the test group, we can only
observe the outcome in the case of treatment X, while the
counterfactual (i.e. the outcome of the test group in the
case of non-X) remains inevitably unobservable. On the
other hand, with regard to the control group, we can only
observe the outcome in the case of non-X, the counterfac-
tual (i.e. the outcome in the control group in the case of X)
remaining unobservable through necessity. As illustrated
in Fig. 1, even in the case of the strongest possible empiri-
cal research designs, causal arguments are, at least partly,
based on theorization and thought experiments (Holland,
1986; cf. Morgan & Winship, 2007).
3
An illustrative stylized example of counterfactuality
and thought experiments would be an attempt to explain
why a house had been razed by ?re. Suf?cient parallel
explanatory factors might include the house being struck
by lightning, a short circuit having occurred or the result
of pyromania. In a given situation, a lightning strike or
short circuit might possibly be eliminated and the thought
experiment, which mobilizes counterfactuals, focuses on
the potential action of a pyromaniac and takes into consid-
eration the evidence found. However, a necessary causal
1
Please note that the condition ‘other things equal’ is embedded in this
de?nition and hence, for instance, a now controlled Z can also be a relevant
explanatory factor. Therefore, X ? Y and Z ? Y can, at least in principle,
exist in parallel, indicating that while counterfactual analysis can identify
necessary or suf?cient causal factors (i.e. conditions) there can also be other
such factors. An outstanding account of the complex relations between
necessary and suf?cient conditions in causality, formulated as the so-called
INUS-conditions, is put forward by Mackie (1974).
2
The counterfactual account of causality is consistent with the so-called
manipulation theory of causality, according to which X is an explanatory
factor of Y if by producing X, other things equal, we can make Y happen.
This is a most natural method for distinguishing causality from mere
correlation (Woodward, 2003; cf. von Wright, 1970).
3
Mandel (2011) provides a profound analytical mapping on the land-
scape of necessary thought experiments, which he terms ‘‘mental simula-
tions’’, relating to causal explanations.
560 K. Lukka / Accounting, Organizations and Society 39 (2014) 559–566
factor for a ?re is always oxygen, although as that so self-
evidently exists and therefore makes no difference, it
would not normally even be mentioned. In this example,
the burnt-down house requires a combination of at least
one of the suf?cient factors and the necessary factor to
form a plausible explanation. Nevertheless, the explicated
explanation would probably only mention the alleged suf-
?cient factor (e.g. pyromania) as the suggested explanation
(cf. Mackie, 1974).
The counterfactual account of causality and interpretive
research
The contemporary perspectives on causality, outlined
above, open an avenue for arguing, in contrast to estab-
lished perceptions, that causal explanations are possible
also in the context of interpretive research; for instance,
in the area of management accounting. The main argument
of this paper is that in addition to ?nding a meaningful role
for causal thinking in interpretive research, it can also sup-
port interpretive research on its own terms without com-
promising its unique characteristics. In fact, as will be
explained below in this paper, linking causal explanations
to interpretive research places some of its central strengths
at the forefront.
While the term ‘explanation’ at times surfaces in the
context of the regularity notion of causality, often termed
the ‘Hempel-Oppenheim model’ or ‘the subsumption
model’, such explanations are ‘thin’ as causality is, in the
Humean spirit, eventually perceived as nothing but a con-
stant conjunction of events (cf. Hempel & Oppenheim,
1948). Hence, the processes and mechanisms through
which causality works need not be focal interests in this
vein of causal thought.
In contrast, the more recently established counterfac-
tual account of causality has a natural and serious linkage
to explanation. In addition, and importantly from the per-
spective of interpretive research, based on Wittgenstein’s
later philosophy in particular, the contemporary philoso-
phy of causality tends to closely link explanation with
interpretation and understanding. For Wittgenstein, in his
later ‘ordinary language philosophy’, understanding was
the correlate of explanation; to explain is to create under-
standing. Moreover, the attribution of understanding
means the ability to employ acquired information in action
and reasoning in ways sanctioned by the relevant commu-
nity (cf. Wittgenstein, 1953, §§ 148–159), thereby linking it
with a social aspect (cf. Chua, 1986). Accordingly, explana-
tion does not notably differ from understanding because
interpreting action correctly, that is, identifying the factors
(i.e. reasons) on which the action depended, simply means
explaining it (e.g. Henderson, 1993). From this perspective,
when interpretive studies emphasize the role of subjective
experiences in the construction of social reality, they can-
not eventually omit the explanatory aspect of the endeav-
our (Kakkuri-Knuuttila & Lukka, 2008).
But how does the idea of tracking down causality and
generating causal explanations relate to the tendency of
interpretive research to address the meanings of knowl-
edgeable agents in their life-worlds? This will be explored
in the next section.
Knowledgeable agents, subjective meanings and
causality
As interpretive research is interested in meanings
developed by subjects, it is closely related to the idea that
reality is, at least partially, socially constructed (e.g. Berger
& Luckmann, 1966; Hacking, 1999). Berger and
Luckmann’s (1966) core claim is that any adequate theo-
retical understanding on society necessarily requires an
integration of subjectivist and objectivist accounts. Hence,
society is both an objective fact and constituted by activi-
ties, which express subjective meanings. Through pro-
cesses of externalization, the latter have the potential to
become objecti?ed as facts that then re-affect the individ-
uals and their formation of new meanings. In the account-
ing context, Hines (1988) emphasizes that the social
construction of reality does not start from nothing; indi-
vidual action and interaction are embedded in social struc-
tures that pre-date the individual. New inter-subjective
constructions emerging in communication arise in the con-
text of particular social structures, which become continu-
ously sustained or reshaped through that communication.
Knowledgeable agents are often able to conceptualize
and re?ect upon the practices they entertain, leading to
the issue of double hermeneutics (Giddens, 1984). Hence,
while agents can trigger or hinder the actualization of par-
ticular causal mechanisms to implement their agenda, they
probably cannot, at least in the short term, change the cau-
sal mechanisms per se (Durand & Vaara, 2009). However, as
the entire notion of social constructionism is based on the
idea that agents often have the possibility of acting other-
wise (Hacking, 1999), the fact that knowledgeable agents’
intentions and actions are changeable and, over time,
new social structures and mechanisms can emerge, imply
that the stability of causal relationships in social sciences
is always at risk of being fragile (Alt, 2009; Mitchell, 2009).
Knowledge on peoples’ meanings (i.e. their understand-
ing of ‘something as something’) plays a central role when
we aim to explain human action and interaction. Seem-
ingly similar behavior can have widely different meanings
Test group Control group
Treatment X Outcome of variable Y = Y (X)
(observed)
Counterfactual (unobservable)
No treatment X Counterfactual (unobservable) Outcome of variable Y = Y (non-X)
(observed)
Fig. 1. Counterfactuality in a controlled experiment.
K. Lukka / Accounting, Organizations and Society 39 (2014) 559–566 561
and also causal effects in different contexts, both as
intended and as interpreted, and these do not necessarily
coincide; a classic example being the wink of an eye in
social interaction. Agents’ beliefs regarding existing causal
relations in the contexts where they operate have notable
bearing when they act; hence, both the beliefs and their
referents are relevant to understanding peoples’ meanings
and explaining their actions. As noted by Wittgenstein
(1953), peoples’ meanings are often not merely their pri-
vate, inner ostensions, but also include their abilities to
communicate and interact with other subjects (cf. Chua,
1986; Geertz, 1973). This perspective not only emphasizes
the social aspect of meanings but also suggests that
interpretive research does not need to be limited only to
exploring the emic understandings of single individuals
(Kakkuri-Knuuttila et al., 2008).
While, in line with the core agenda of the interpretive
paradigm, interpretive researchers certainly try to under-
stand how people receive, develop and send meanings in
social instances, they can and do tend to employ them
when attempting to explain the phenomena to which
these meanings relate. In fact, following Wittgenstein’s
(1953) line of thought, an argument can be developed that
not even the most subjectivist position of interpretive
research can entirely distance itself from explanation.
4
Accordingly, to be able to cope with living among other peo-
ple, we need the ability to understand what kind of causal
beliefs, meanings and intentions drive the actions of others
while simultaneously considering how our actions, funda-
mentally driven by similar factors, are read by others. Peo-
ples’ meanings, which are realized in their actions, become
part of the world’s causal linkages and can (and tend to)
have a lot of causal power (cf. Kakkuri-Knuuttila et al.,
2008; Lukka & Modell, 2010).
It is crucial to note that including causality in the anal-
ysis by no means implies that the world would be
assumed as deterministic. For interpretive research, it
rather means the incorporation of the idea that inten-
tional actors with a certain amount of agency can and
tend to take relevant causal relations within the context
where they operate into consideration and that their
actions can and tend to have causal consequences. These
features hold even though the web of causalities is always
inherently fragile in the social world (cf. Melan, 2013, pp.
117–119).
Causality in interpretive management accounting
research
Reviewing the literature of interpretive management
accounting research suggests that while causality is typi-
cally played down in what is explicitly stated, deconstruct-
ing the studies’ contents often discovers storylines that
include causal explanations. Indeed, the bulk of interpre-
tive studies in management accounting do not quite ?t into
the narrow window of the extreme subjectivist interpre-
tive position suggested by Winch (1958) or Burrell and
Morgan (1979). While interpretive studies indeed offer
thick descriptions (Geertz, 1973) or rich insights (Ahrens
& Dent, 1998; Dyer & Wilkins, 1991) and explore peoples’
subjective meanings in line with the interpretive para-
digm, they also tend to include objecti?ed ‘facts’ and
develop causal explanations, although most often only
implicitly (Ahrens & Chapman, 2006; Kakkuri-Knuuttila
et al., 2008; Lukka & Modell, 2010). It is therefore relieving
that the contemporary theory of causal explanation, based
on the notion of counterfactuality, supports conducting
case research and generating local explanations. Accord-
ingly, to discuss causal explanations, we do not need to
seek law-like generalizations and apply the covering law
model. Instead, we are encouraged to seek causal explana-
tions that capture (difference making) dependency rela-
tions between things in the world in any locality
(Kakkuri-Knuuttila et al., 2008; cf. Durand & Vaara,
2009). Clearly, the need always remains for causal claims
to sustain analysis with counterfactual conditionals (cf.
Mandel, 2011).
One of the strongholds of interpretive research, the ten-
dency to focus on responding to ‘how’ type of questions,
speci?cally addresses the question concerning how the
alleged explanatory factors work in producing outcomes
(Weick, 1989; Morgan & Winship, 2007; Tsang &
Ellsaesser, 2011). This is a notable difference from many
other kinds of study; for instance, surveys and archival
research in which analysis of processes is typically dif?cult,
if not impossible. Interpretive studies can progress over
and above merely describing associations between vari-
ables to exploring the processes and mechanisms by which
they are generated; that is, towards looking profoundly at
the linkages (i.e. the ‘arrows’) between the elements of the
explanatory scheme (cf. Ahrens & Chapman, 2006; Durand
& Vaara, 2009). This means, based on careful and profound
?eld work, opening up how processes of interest, including
causal linkages, work out from conditions to outcomes
regarding the explored research question, thereby possibly
revealing particular mechanisms that are at work in the
subject area (cf. Dent, 1991). This is precisely what inter-
pretive researchers in management accounting often seek
to achieve in their typical endeavours to ‘make sense’ of
what they have observed in their ?eld work. Due to this,
it can be argued that the explanations produced by inter-
pretive studies are particularly informative and strong.
Hence, interpretive studies can provide thick explana-
tions that are deeply rooted in the life-worlds of the people
being studied and, hence, incorporate peoples’ meanings,
which is the main focus of interest in interpretive research,
and also the processes and mechanisms that tie together
the elements of the developed explanation (cf. Lukka &
Modell, 2010). Consequently, when causal explanations
form part of the argumentation structure of an interpretive
study, there is no need for them to be hidden. On the con-
trary, they can be made even more explicit, which would
tend to sharpen and strengthen the arguments developed
while maintaining the interpretive framing of the study
and remaining faithful to the core agenda of interpretive
research (Kakkuri-Knuuttila et al., 2008; cf. Hammersley
& Atkinson, 2009, p. 185).
4
In addition to this epistemological grounding, the audiences of
interpretive research are likely to expect some theoretical (i.e. etic)
structure that overarches mere accounts of emic understandings.
562 K. Lukka / Accounting, Organizations and Society 39 (2014) 559–566
Contrastive thinking as a resource for causal analysis
An important aspect of contemporary thinking on cau-
sality is the acknowledgement that, in developing causal
claims, selection is an inherent part of scholarly work. It
is worth bearing in mind that causes typically have several
different kinds of effect while, in turn, effects tend to have
numerous possible and different kinds of cause.
5
Typically,
there is considerable room for choice regarding the screen-
ing of relevant explanatory factors and as such screening is
never a neutral, contextual or interest-free activity (van
Fraassen, 1977), the process of generating explanations
practically always needs to be carefully focused by the
researcher him/herself. It is also crucial to realize that, typ-
ically, we do not explain events or phenomena per se, but
only particular aspects of them (Hempel, 1965; cf. Tsang &
Ellsaesser, 2011).
An epistemologically sound approach is to let our
selected theoretical focus drive the choice of which possi-
ble explanatory factors are of true interest to us (e.g.
DiMaggio, 1995). The sharpening of explanations in terms
of contrastive thinking is of notable help in this regard
(Lipton, 1990; Tsang & Ellsaesser, 2011; Ylikoski, 2007).
Consider an innocent looking example observation, which
could certainly be part of an interpretive management
accounting study on budgetary biasing (cf. e.g. Lukka,
1988):
Divisional managers underestimate their sales forecasts
when submitting their budgets to the headquarters.
How can this example of budgetary biasing be
explained? Contrastive thinking advises us to ?rst sharpen
the target of the explanation (i.e. the explanandum). In so
doing, we get, for instance, the following four more speci?c
sub-questions, or so-called allomorphs (see Tsang &
Ellsaesser, 2011): Why divisional managers, not some other
party? Why underestimate rather than, say, overestimate?
Why sales forecasts, not some other item? Why when deal-
ing with the headquarters, not with some other party? All
of these questions lead to somewhat different explanation
focuses and processes of counterfactual analysis (cf. Lukka
& Modell, 2010).
The central lesson here is that we can easily endeavour
to explain overly broad sets of aspects of phenomena,
which tend to lead to weak and ambiguous argumentation,
potentially limiting the possibility of truly increasing our
understanding. Thoughtful application of contrastive
thinking makes analysis with counterfactual conditionals
more effective and also contributes to avoiding so-called
over-generation of causes (e.g. Hoerl et al., 2011). In fact,
there is reason to argue that careless disregard of the con-
trastive thinking aspect makes it very dif?cult to develop
causal claims that can be rendered meaningful and plausi-
ble (cf. Lukka & Modell, 2010).
The approach to causal exploration introduced in this
paper, revolving around the counterfactual notion of
causality, contrastive thinking and the power of thought
experiments suggests focusing on relatively narrow slices
of the world to enable contribution to the clarity and
strength of the generated explanation. However, the chal-
lenge of keeping the holistic nature of interpretive research
on the agenda remains and, as such, the abductive mode of
reasoning partially comes to the rescue as it, among other
things, tends to contextualize theoretically the core empir-
ical ?ndings and thereby helps to link the study with the
extant body of knowledge on the focal ?eld.
Abductive mode of reasoning and causal analysis
Interpretive research often functions through an abduc-
tive process (Dubois & Gadde, 2002; Peirce, 1903) instead
of the better known inductive and deductive modes of
reasoning.
6
Abductive reasoning typically starts from a
striking empirical observation that begs an explanation, trig-
gering the process of ‘making sense’ to begin. However, as
the observation is not perceived to exist in a theoretical vac-
uum, an abductive researcher then commences careful
development of theoretical explanations with the help of
everything he/she knows, both empirically and theoretically,
on the examined issue (Hanson, 1958, 1961). The process of
abductive reasoning can be paralleled to detective work as it
collects clues, focuses on those that are most promising and
attempts to put them into perspective with the conditions
and other evidence at hand. In brief, the abductive explana-
tory scheme can be formulated as follows (Peirce, 1960, p.
117):
Surprising phenomenon Y is observed.
But if proposition X would be true, then Y would be a
matter of course.
There is reason to assume X is true.
The outcome of an abductive research process is typi-
cally a causal explanation where X is concluded to plausi-
bly explain Y, representing so-called inference to the best
explanation. Rather than being a linear process of reason-
ing, an abductive process includes going back and forth
between empirical ?ndings and theoretical elements per-
ceived to be of relevance and interest (Lukka & Modell,
2010). Thought experiments relating to the application of
contrastive thinking and counterfactual analysis form an
integral part of the abductive reasoning process.
It is indeed worth emphasizing the important role of
contrastive thinking within the abductive reasoning pro-
cess; factors included in the various stages of the analysis
do not come ‘out of the blue’ but are selected, based on
particular grounds. In the case of interpretive research,
the endless spectrum of actors’ potential differing mean-
ings that might play a role in the research setting renders
5
Cf. Morgan and Winship (2007) and their discussion on the difference
between primarily seeking the causes of effects versus the effects of causes.
6
Abduction, developed and coined by C.S. Peirce approximately a
century ago, has much similarity with grounded theorization that was
introduced to the agenda of social studies in the 1960s by Glaser & Strauss
(1967) (cf. Strauss & Corbin, 1998; Suddaby, 2006). Hanson (1958)
criticized the dominant hypothetico-deductive and inductive models for
representing the logic of a completed research report rather than the actual
research process.
K. Lukka / Accounting, Organizations and Society 39 (2014) 559–566 563
their identi?cation an even more challenging matter. Con-
trastive thinking is crucial at the outset of the process as it
helps the researcher identify the interesting observation
perceived as requiring explanation – in other words, why
was precisely this observation picked up in contrast to
another one? Contrastive thinking is employed over the
entire abductive process as the researcher focuses his/her
attention on particular potential emerging explanatory fac-
tors instead of some others, collects further evidence based
on his/her theoretical contemplation as well as clues from
the ?eld and also runs thought experiments on expected
relationships between things (i.e. mobilizing counterfac-
tual conditionals). Contrastive thinking offers a focused
method with which to consider one’s own ideas and etic
preferences in relation to those of others.
To exemplify abduction in the context of interpretive
research, let us consider further the case of budgetary bias-
ing introduced in the previous section. Actors’ subjective
meanings relate particularly to their intentions concerning
their biasing attempts and different intentions can have
different likely consequences. In this regard, Lukka (1988)
identi?ed two main types of intention for a manager to
manipulate his/her budget: resource intention (i.e. aiming
to get additional resources) and performance evaluation
intention (i.e. aiming to receive more positive evaluations).
While a functionalist researcher would focus on trying to
measure bias and its quanti?able determinants, an inter-
pretive researcher pays attention to the process of biasing,
including its subjective elements. The contrasts are
brought in partly from extant theory; however, the
researcher still has to decide what he/she wishes to con-
trast and against what. In line with abduction, an interpre-
tive researcher can attempt to explain a particular
observed instance of budgetary biasing (e.g. contrasting
biasing efforts towards pessimistic and optimistic direc-
tions) and develop an explanation for it based on the anal-
ysis of the subjective biasing intentions of the focal actor
(cf. Lukka, 1988).
7
As abductive reasoning inherently combines empirical
and theoretical elements to develop a plausible argument
or set of arguments, it helps an interpretive researcher to
maintain the holistic feature of interpretive research,
despite conducting parts of his/her primarily empirically
grounded analysis in a focused and selective manner, based
on contrastive thinking. Abductive reasoning is inclined to
tie together the components of the study and has the poten-
tial of contextualizing the empirical ?ndings, possibly
including causal elements, to form a theoretically more
encompassing and consistent set of arguments. Through
abduction, emic-level empirical analysis at the core of inter-
pretive research is connected to etic-level analysis and
knowledge. This move makes the individual piece of inter-
pretive researchnotably more theoretical as it becomes part
of the knowledge set on the focal ?eld. This is a way by
whichtheoretical diversity anddevelopment ininterpretive
research across studies might be understood as an exercise
in complementarity rather than fragmentation.
Examples selected from published interpretive manage-
ment accounting research such as Covaleski and Dirsmith
(1986), Dent (1991), Ahrens (1996) and Vaivio (2006) give
other illustrations regarding how interpretive scholars
effectively apply abduction as their mode of reasoning. In
all of these studies, the authors offer implicit clues to hav-
ing employed counterfactual analysis and contrastive
thinking in making theoretical sense of their ?ndings and
delineating plausible explanations to accounting-related
phenomena (cf. Lukka & Modell, 2010). In all of these stud-
ies, the emic and the etic interact fruitfully, resulting in
interpretive studies producing more than merely rich emic
accounts. They offer theoretical arguments including
explanations and, as these are deeply rooted in the life-
worlds of the people being studied, we can term these
explanations thick. In the following section, one of the
above-mentioned studies will be analyzed in more detail
to more speci?cally illustrate how causal arguments and
a storyline can be generated by interpretive research,
although in an embedded manner.
Developing a causal explanatory storyline in
interpretive research: An example
Dent’s (1991) widely cited article ‘‘Accounting and
organizational cultures: A ?eld study of the emergence of
a new organizational reality’’ is a longitudinal interpretive
case study on a profound cultural change in an organiza-
tion, Euro Rail (ER), focusing on the roles played by new
accounting practices along the dynamic process of the cul-
tural transformation in the ?rm.
Dent’s study traces the process of major change in a
state owned railway company (i.e. ER), starting with the
introduction of new kinds of manager (i.e. ‘‘business man-
agers’’) under governmental pressure, culminating in a
profound cultural change whereby the old ‘‘railway
culture’’ became gradually replaced by a new ‘‘business
culture’’. The central interpretive content of Dent’s analy-
sis, in which he mobilized Geertz (1973) in particular,
encompasses that he carefully examined the process and
the ways by which the business managers gradually and
systematically acquired in?uence by transforming the
structures of meaning in the organization through, among
other things, various actions with considerable symbolic
signi?cance. Dent argued that the transformation of a
few accounting practices played a major role in this pro-
cess by changing the way people conceptualized key parts
of the organization and also the very role of the organiza-
tion in relation to the larger community. He identi?ed
and described a particular process and pattern in the busi-
ness managers’ actions whereby they gradually broadened
their in?uence at ER to implement their change agenda.
When causality is understood in the sense of referring to
dependencies, which do not need to be regularities and can
be contextual, Dent’s piece includes numerous explana-
tions (cf. Kakkuri-Knuuttila et al., 2008). The delineated
7
The researcher can also start the analysis from observed subjective
intentions and attempt to predict the likely biasing possibilities. It is worth
noting that the precise subjective biasing intention of the actor can make a
notable concrete difference. Resource intention, in contrast to performance
evaluation intention, tends to systematically relate to an attempt to
negotiate for more resources, leading to overestimations in the cost side of
the budget (cf. Lukka, 1988).
564 K. Lukka / Accounting, Organizations and Society 39 (2014) 559–566
linkages (i.e. the ‘arrows’ between various things) that tie
the isolated descriptions of events into a coherent whole
can be regarded as forming a set of causal explanations,
rendering the causal reading of the piece quite natural.
According to Dent, the accounting-related changes at ER
were important causal explanatory factors for the cultural
change, without which there would probably have been
no such change, thereby offering examples of counterfac-
tual conditionals that are frequently embedded throughout
his paper.
When Dent identi?es phenomena begging for explana-
tion (i.e. explanandums) at various levels of his analysis,
examples of contrastive thinking turn up in plenty, imply-
ing that he had selected speci?c focus issues to drive his
examination, forming the key elements of his storyline.
In their application, counterfactual thought experimental
analysis is present, although typically only implicitly. To
illustrate this, the process of cultural change began with
the appointment of new kinds of manager at ER; however,
why were they needed rather than, say, new kinds of loco-
motive engine? Because ER’s environment had changed in
such a way that its top management especially felt more
managerial resources would be required. The kind of man-
agers appointed were business managers; however, why
precisely this kind of managers rather than, say, hiring
more operational managers? Because ER’s top manage-
ment especially lacked marketing and business expertise
and a bottom line mentality. Had the government not
turned ‘hostile’, the need for new kinds of managerial
resources and practices would probably have not emerged.
Also, had the nature of the government’s hostility been dif-
ferent (e.g. less focused on ?nancial issues), the type of
newly recruited managers could have been different, too
(cf. Kakkuri-Knuuttila et al., 2008).
Overall, Dent’s (1991) analysis relies strongly on the
‘power of example’ (Flyvbjerg, 2001), whereby particular
critical events from ER’s trajectory of change are identi?ed
for careful and profound scrutiny and are presented as key
drivers for further changes. The analysis is formed by
numerous partial episodes and narratives, bound together
with the help of the theoretical resources introduced at the
beginning of the paper, hence providing a well-developed
example of how abductive reasoning can be usefully mobi-
lized in interpretive research. Therefore the outcome of
Dent’s (1991) analysis is inherently holistic. The paper also
offers a solid example of thick explanations as it presents a
careful emic account of the evolving change in the mean-
ing structures of the actors at ER and also provides an
explanation on the change from an etic perspective (cf.
Lukka & Modell, 2010).
Conclusions
This paper suggests how the contemporary theory of
causality, based on the notions of counterfactuality and
contrastive thinking, offers helpful direction on how to
generate plausible causal arguments in interpretive stud-
ies. For the interpretive researcher, if he/she so wishes
and it is required by the research question, they suggest
a route from rich emic-based accounts at the core of
interpretive research to thick explanations. Perhaps with
some surprise, causality can be included in interpretive
research framings without compromising the unique fea-
tures of such research, even by building on some of its
strongholds. In particular, this is due to the tendency of
interpretive studies to focus on responding to ‘how’ ques-
tions. Typically, interpretive research profoundly investi-
gates how associations between phenomena occur and
explores the processes and mechanisms that generate out-
comes from particular conditions, thereby shedding light
on the ‘arrows’ between variables.
The paper emphasizes and illustrates the role of con-
trastive thinking as a most useful tool with which to focus
on the development of causal explanations, rendering
analysis with counterfactual conditionals, including
thought experiments, more effective. The abductive mode
of reasoning, referring to the iterative movement between
empirical materials and theory elements and contributing
to making the set of explanations more holistic and theo-
retically contextualized, forms the principal way to gener-
ate plausible explanations in interpretive research. While
abduction works through all of the key tools of causal
exploration introduced in this paper, the role of contrastive
thinking is particularly notable as it guides the researcher
in his/her choices of analytical focus from the beginning
of the abductive process.
Explicating causal explanation in interpretive research,
for instance in the ?eld of management accounting, has
the potential to add to the clarity and strength of the argu-
ments developed, while also leading to the need to bear in
mind their inherent boundary conditions, such as the lim-
its of counterfactual analysis and those of abductive rea-
soning and also the fact that the stability of social
structures and relationships is always an open question.
Despite these caveats, it is suggested that the potential of
applying causal thinking more explicitly in interpretive
research is notable.
Acknowledgements
This paper is based on the plenary presentation at
GMARS conference in Copenhagen in June 2012. I would
like to thank Chris Chapman, Markus Granlund, Olli Koisti-
nen, Mikael Melan, Hanna Meretoja, Sven Modell, Susan
O’Leary, Jan P?ster and the anonymous reviewers for their
helpful comments along the process of preparing this
paper. Mike Power is thanked for his highly signi?cant
encouragement.
References
Ahrens, T. (1996). Styles of accountability. Accounting, Organizations and
Society, 21(2 & 3), 139–173.
Ahrens, T., & Chapman, C. (2006). Doing qualitative ?eld research in
management accounting: Positioning data to contribute to theory.
Accounting, Organizations and Society, 31(8), 819–841.
Ahrens, T., & Dent, J. F. (1998). Accounting and organisations: Realizing
the richness of ?eld research. Journal of Management Accounting
Research, 10, 1–39.
Alt, J. (2009). Comment: Conditional knowledge: An oxymoron? In C.
Mantzavinos (Ed.), Philosophy of the social sciences. Philosophical theory
and scienti?c practice (pp. 146–153). Cambridge University Press.
Berger, P., & Luckmann, T. (1966). The social construction of reality. Anchor
Books.
K. Lukka / Accounting, Organizations and Society 39 (2014) 559–566 565
Burrell, G., & Morgan, G. (1979). Sociological paradigms and organizational
analysis: Elements of the sociology of corporate life. Heinemann.
Chua, W. F. (1986). Radical developments in accounting thought. The
Accounting Review, 61(4), 601–632.
Covaleski, M. A., & Dirsmith, M. W. (1986). The budgetary process of
power and politics. Accounting, Organizations and Society, 11(3),
193–214.
Dent, J. F. (1991). Accounting and organizational cultures: A ?eld study of
the emergence of a new organizational reality. Accounting,
Organizations and Society, 16(8), 705–732.
Denzin, N. K. (1983). Interpretive interactionism. Sage.
DiMaggio, P. J. (1995). Comments on ‘‘What theory is not’’. Administrative
Science Quarterly, 40(3), 391–397.
Dubois, A., & Gadde, L.-E. (2002). Systematic combining: An abductive
approach to case research. Journal of Business Research, 55(7),
553–560.
Durand, R., & Vaara, E. (2009). Causation, counterfactuals and competitive
advantage. Strategic Management Journal, 30(12), 1245–1264.
Dyer, W. G., & Wilkins, A. L. (1991). Better stories, not better constructs, to
generate better theory: A rejoinder to Eisenhardt. Academy of
Management Review, 16, 613–619.
Flyvbjerg, B. (2001). Making social science matter: Why social inquiry fails
and how it can succeed again. Cambridge University Press.
van Fraassen, B. (1977). The pragmatics of explanation. American
Philosophical Quarterly, 14, 143–150.
Geertz, C. (1973). The interpretation of cultures. New York: Basic Books.
Giddens, A. (1984). The constitution of society. Outline of the theory of
structuration. Polity Press.
Glaser, B. G., & Strauss, A. L. (1967). The discovery of grounded theory:
Strategies for qualitative research. New York (NY): Aldine de Gruyter.
Hacking, I. (1999). The social construction of what? Harvard University
Press.
Hammersley, M., & Atkinson, P. (2009). Ethnography principles in practice
(3rd ed.). Routledge.
Hanson, N. R. (1958). Patterns of discovery. Cambridge: University Press.
Hanson, N. R. (1961). Is there a logic of scienti?c discovery? In H. Fiegl &
M. Grover (Eds.), Current issues in the philosophy of science. New York:
Holt, Rinehart & Winston.
Headland, T. N. (1990). Introduction: A dialogue between Kenneth Pike
and Marvin Harris. In T. N. Headland, K. L. Pike, & M Harris (Eds.),
Emics and etics: The insider/outsider debate (pp. 13–27). Sage.
Hempel, C. (1965). Aspects of scienti?c explanation. Free Press.
Hempel, C. G., & Oppenheim, P. (1948). Studies in the logic of explanation.
Philosophy of Science, 15(2), 135–175.
Henderson, D. (1993). Interpretation and explanation in the human sciences.
State University of New York Press.
Hines, R. (1988). Financial accounting: In communicating reality, we
construct reality. Accounting, Organizations and Society, 13(3),
251–261.
Hoerl, C., McCormack, T., & Beck, S. R. (2011). Introduction:
Understanding counterfactuals and causation. In C. Hoerl, T.
McCormack, & S. R. Beck (Eds.), Understanding counterfactuals,
understanding causation (pp. 1–15). Oxford University Press.
Holland, P. W. (1986). Statistics and causal inference. Journal of the
American Statistical Association, 81(9), 45–70.
Kakkuri-Knuuttila, M.-L. (2006). Kausaalisuhteet ja selittäminen
tulkitsevassa tutkimuksessa (Causal relations and explanation in
interpretive research). In K. Rolin, M.-L. Kakkuri-Knuuttila, & E.
Henttonen (Eds.), Soveltava yhteiskuntatiede ja ?loso?a (pp. 54–87).
Gaudeamus (Applied social science and philosophy).
Kakkuri-Knuuttila, M.-L., Lukka, K., & Kuorikoski, J. (2008). Straddling
between paradigms: A naturalistic philosophical case study on
interpretive research in management accounting. Accounting,
Organizations and Society, 33(2&3), 267–291.
Lewis, D. (1973). Counterfactuals. Harvard University Press.
Lipton, P. (1990). In D. Knowles (Ed.), Explanation and its limits
(pp. 247–266). Cambridge University Press.
Lukka, K. (1988). Budgetary biasing in organizations: Theoretical
framework and empirical evidence. Accounting, Organizations and
Society, 13(3), 281–301.
Lukka, K., & Modell, S. (2010). Validation in interpretive management
accounting research. Accounting, Organizations and Society, 35(4),
462–477.
Mackie, J. L. (1974). The cement of the universe: A study of causation. Oxford
University Press.
Mandel, D. R. (2011). Mental simulation and the nexus of causal and
counterfactual explanation. In C. Hoerl, T. McCormack, & S. R. Beck
(Eds.), Understanding counterfactuals, understanding causation
(pp. 147–170). Oxford University Press.
Melan, M. (2013). Will and mechanisms. Reports from the Department of
Philosophy. University of Turku.
Meretoja, H. (2012). Ymmärryskäsitteet ja toisen kohtaamisen ongelma
(concepts of understanding and the issue of encountering the other).
In V. Viljanen, H. Siipi, & M. Sintonen (Eds.), Ymmärrys (pp. 307–318).
Reports from the Department of Philosophy. University of Turku
(Understanding).
Mitchell, S. (2009). Complexity and explanation in social sciences. In C.
Mantzavinos (Ed.), Philosophy of the social sciences. Philosophical theory
and scienti?c practice (pp. 130–145). Cambridge University Press.
Morgan, S. L., & Winship, C. (2007). Counterfactuals and causal inference:
Methods and principles for social research. Cambridge University Press.
Peirce, C. S. (1903). The essential Pierce: Selected philosophical writings (Vol.
2). Bloomington: Indiana University Press.
Peirce, C. S (1960). In C. Hartshorne & P. Weiss (Eds.). Collected papers of
Charles sanders Peirce (Vols. V–VI). Belknap Press of Harvard
University Press.
Pike, K. L (1954). Emic and etic standpoints for the description of
behavior. In K. L. Pike (Ed.), Language in relation to a uni?ed theory of
the structure of human behavior (pp. 8–28). Summer Institute of
Linguistics.
Raatikainen, P. (2011). Teon teoria ja selittäminen (Theory of action and
explanation). Tiede ja edistys, 4, 291–300.
Ruben, D.-H. (1990). Explaining explanation. Routledge.
Sobel, M. E. (1995). Causal inference in the social and behavioral sciences.
In G. Armingger, C. C. Clogg, & M. E. Sobel (Eds.), Handbook of statistical
modeling for the social and behavioral sciences. Plenium Press.
Sobel, M. E. (1996). An introduction to causal inference. Sociological
Methods and Research, 24, 353–379.
Strauss, A., & Corbin, J. (1998). Basics of qualitative research: Techniques and
procedures for developing grounded theory. Thousand Oaks: Sage.
Suddaby, R. (2006). From the editors: What grounded theory is not.
Academy of Management Review, 49(4), 633–642.
Sutton, R. I., & Staw, B. M. (1995). What theory is not. Administrative
Science Quarterly, 40(3), 371–384.
Tetlock, P. E., & Belkin, A. (Eds.). (1996). Counterfactual thought
experiments in world politics. Princeton University Press.
Tsang, E. W. K., & Ellsaesser, F. (2011). How contrastive explanation
facilitates theory building. Academy of Management Review, 36(2),
404–419.
Vaivio, J. (2006). The accounting of ‘‘the meeting’’: Examining
calculability within a ‘‘?uid’’ local space. Accounting, Organizations
and Society, 31(8), 735–762.
Weick, K. (1989). Theory construction as disciplined imagination.
Academy of Management Review, 14(4), 516–531.
Weick, K. (1995). What theory is not, theorizing is. Administrative Science
Quarterly, 40(3), 385–390.
Winch, P. (1958). The idea of social science and its relation to philosophy.
Routledge & Kegan Paul.
Wittgenstein, L. (1953). Philosophical investigations. Basil Blackwell.
Woodward, J. (2003). Making things happen. Oxford University Press.
von Wright, G. H. (1970). Explanation and understanding. Cornell
University Press.
Ylikoski, P. (2007). The idea of contrastive explanandum. In J. Persson & P.
Ylikoski (Eds.), Rethinking explanation (pp. 27–42). Springer.
566 K. Lukka / Accounting, Organizations and Society 39 (2014) 559–566
doc_695864092.pdf