Scandinavian Journal of Management 39 (2023) 101273

Contents lists available at ScienceDirect

Scandinavian Journal of Management

journal homepage: www.elsevier.com/locate/scajman

What theory is – A late reply to Sutton and Staw 1995

Peter Kesting 1
Aarhus University, Fuglesangs Allé 4, DK-8210 Aarhus V, Denmark

A R T I C L E I N F O A B S T R A C T

Keywords Theory plays a central role in research in management science, with theoretical contribution an essential
Theory measure for evaluating research. However, the many ambiguities regarding the use of the theory concept make
Explanation judgments imprecise and, to an extent, arbitrary. The discussion of the concept of theory in management science
Causality
has remarkably little anchoring in the findings of the philosophy of science. This article presents some of these
Hypotheses
Induction
findings and discusses the concepts of explanation and theory on this basis. In particular, Sutton and Staw’s
Science (1995, p. 385) notion that theory should provide a logical explanation of causal relationships formulated in
hypotheses is critically questioned.

1. Introduction in Google Scholar (as of August 6, 2022), but also that 25 of these ci­
tations have been added since July 7, 2022, in a mere month. This shows
Theory plays an important role in management science, and an how widely the essay is still used today. It is therefore no exaggeration to
essential criterion for evaluating research is its theoretical contribution call this essay a milestone in understanding theory in management sci­
(Colquitt & Zapata-Phelan, 2007; Rynes, 2005). In his editorial, Suddaby ence. It has largely shaped the understanding of theory in management
(2015, p. 2) states, “Knowledge accumulation simply cannot occur science and continues to do so today.
without a conceptual framework.” Hambrick (2007) even diagnoses a To be clear, Sutton and Staw are not to blame for the position of their
“devotion to theory.” In view of this central importance, however, sur­ paper, written as an editorial, now holds. However, views from this
prisingly little attention is devoted to the object itself, the theory, in essay were and are adopted with little critical questioning. The view of
management science, its understanding only weakly anchored in the theory as a logical explanation of causal relationships, i.e., that it pro­
findings of the philosophy of science. The philosophy of science has vides an explanation as to why certain causal relationships exist, is
developed the understanding of theory over centuries and gained particularly problematic. This view not only contradicts the funda­
valuable insights in the process. These insights usually receive little mental insights of the philosophy of science, but furthermore, it conveys
attention in management science, especially in influential contributions a problematic impression of the epistemological status of hypotheses.
dealing directly with theory or theoretical contributions, such as Corley This paper aims to initiate a debate that creates a stronger connection
and Gioia (2011), Hambrick (2007) and Whetten (1989). Consequently, to the philosophy of science and discusses the implications of its findings
Suddaby (2015, p. 1) diagnoses, “we appear to disagree as a profession for research in management science. To this end, I begin by showing the
about why we need theory and what role it should play in creating, importance of theory for research, particularly for explanation and
maintaining, and shaping what type of knowledge we value in the field.” prediction. Following, I specify the concept of theory based on an open
Similarly, Brunsson (2021, p. 1) sums up, there is "little agreement" definition and show similarities and differences between different po­
about "what theory is," and "many issues remain unsolved." A central sitions. Building on this, I explain the challenges that management
benchmark for assessing research thus remains unclear, and judgments research faces when developing theory (theory building and testing) and
become, to a degree, arbitrary. Management science sees too little effort the consequences that result. Against this background, I show why the
to counteract this shortcoming. requirement of Sutton and Staw (1995, p. 375), according to which it is
The essay "What Theory is not" by Robert I. Sutton and Barry M. the task of theory to “explain why variables or constructs come about or
Staw, published in 1995 in ASQ remains an essential reference point for why they are connected,” is so problematic.
the understanding of theory in management science. The importance of
this essay is demonstrated not only by the fact that it has 2795 citations

E-mail address: [email protected].

1
https://orcid.org/0000-0001-6780-8299.

https://doi.org/10.1016/j.scaman.2023.101273
Received 31 August 2022; Received in revised form 24 January 2023; Accepted 2 March 2023
Available online 6 March 2023
0956-5221/© 2023 Elsevier Ltd. All rights reserved.
P. Kesting Scandinavian Journal of Management 39 (2023) 101273

2. What is an explanation? things occurred in the past, and furthermore, allows a glimpse into the
future. We know much more now than that there will be cats in the
A basic idea of theory was already formulated by Plato in reply to future. We can foresee things (in a loose sense of the word) and know
Heraclitus. In Cratylus, 509a (Plato & Reeve, 1998, see also Keller, under what conditions they will occur, but more importantly, we can
2000) Plato attributed to Heraclitus the well-known assertion that now assess the consequences of actions.
everything is in constant motion (panta chorei), and contesting this, All of this is highly relevant for management science, which is
Plato replied that there are things not subject to change, especially strongly decision related. Here, alternatives are compared and the
so-called forms (Russell, 1946). According to Plato, forms are ideas and consequences of decisions shown (e.g., for different approaches to
as such are the non-physical essence of all things. It is particularly internationalization or different leadership styles, different strategic
important that these are stable, absolute, and universal. Plato gives the positioning on markets, etc.) Forecasts are particularly relevant for the
form of the cat as an example. This form is based on many individual practical application of management research in a specific company
characteristics (every cat is different), but the form itself is universal context, for managerial decision-making, or for management consulting.
(they are all cats; note that Plato, of course, had no concept of biological For such considerations and recommendations, logical conclusions in
evolution). The crucial point is that in this way we can classify phe­ the form of the covering-law model are indispensable. Therefore, I
nomena and furthermore, look into the future, because we know that would venture to claim that the thinking pattern of the covering-law
cats will continue to exist in the future. The cat’s form will continue, just model plays a major role in management science.
as the dodo’s form continues even past extinction. Much of Plato’s It is epistemologically of great importance that in the covering-law
theory of forms cannot be maintained in its absoluteness, but in stability model, statements of type (i) [the "natural law"] represent a funda­
it names an essential cornerstone of theory. mentally different class of statements as opposed to statements of types
The next logical step is the natural law formulated in its modern form (ii) [the initial condition] and (iii) [the state of affairs to be explained].
by René Descartes (Descartes, Veitch & Hoyt-O’Connor, 2008). Not only In Kant’s terminology, both represent synthetic judgments, as they
are forms stable, causal relationships are likewise stable. This allows combine a subject with a predicate that is not contained in the concept of
new access to the concept of understanding. We can now explain and the subject (Kant & Meiklejohn 2018). However, (i) is a matter of uni­
even predict things, but what exactly does that mean, and how does versally valid causal relationships (in Kant, synthetic judgments a pri­
theory factor into it? ori), and in (ii) and (iii), observable singular events (in Kant, synthetic
The covering-law model formulated by Hempel and Oppenheim in judgments a posteriori). They are therefore relatively easy to distin­
1948 provides a widely recognized specification of the structure of guish, and there is an epistemological watershed between them. This
explanation and prediction. An explanation therefore provides a reason distinction plays a central role in the philosophy of science.
for an empirical fact, i.e., why things happened. This reason consists of a It is also epistemologically significant that explanation and predic­
causal conclusion composed of two conditions (the explanandum) and a tion in the sense of the covering-law model cannot do without state­
conclusion (the explanans). Fig. 1 shows the structure of the covering- ments of type (i), i.e., without universally valid causal relationships.
law model and an explanation based on this model. These first establish the connection between the initial condition and the
The covering-law model establishes a connection between observed event to be explained. This brings us directly to the notion of theory.
initial conditions and an observed result on the basis of a generally valid
causal relationship. The water is boiling because it was heated to 71 3. What is theory?
degrees Celsius and because water on Mt. Everest boils at 71 degrees
Celsius. In this way, the covering-law model provides a logical expla­ Descartes’ conception of the law of nature, as well as more
nation for the occurrence of a specific event. contemporary concepts such as that of axiomatic theory (Bourbaki,
Hempel and Oppenheim postulate that prediction and explanation 1994) or the “hypothetico-deductive” model of Carnap et al. (2019),
are related symmetrically to the time axis, with explanation relating to Hempel and Fetzer (2001), suggest that theory, at its core, is about the
the past and prediction relating to the future. From this, Hempel and formulation of generally valid relationships of type (i) in the
Oppenheim (1948, p. 323) develop the demand, “It may be said, covering-law model. If that is the case, the definition of theory is quite
therefore, that an explanation is not fully adequate unless its explanans, unproblematic. I would like to use an “open definition” for this, which
if taken account of in time, could have served as a basis for predicting the distinguishes between essential characteristics that are largely shared
phenomena under consideration.” This requirement is fairly strict and and characteristics about which there is often no consensus.
not uncontroversial (Caldwell, 1982), but is at the same time very The essential characteristic of theory is then that theory consists of
instructive. the determination of conditionally universal connections between con­
The covering-law model allows for the possibility to explain why structs (at this point, I am already using the contemporary term). This is

Fig. 1. Schematic representation of the Covering Law model.

2
P. Kesting Scandinavian Journal of Management 39 (2023) 101273

where the decisive epistemological difference lies: theory is a matter of have often encountered this idea myself in various forms, and it seems to
general statements that include reference to the future, while observa­ have become deeply ingrained in our discipline.
tions, on the other hand, are singular statements that always relate to the Indeed, theory is presented in very different forms, with the most
past (arguably the present as well) but never the future; it is impossible developed being the so-called axiomatic theory (Bourbaki, 1994). It
to scientifically observe future events. The basic form of a theoretical contains a complete specification of all elements of a theory, the al­
statement is thus A—B, with “A” and “B” representing constructs. I have phabet, the formation rules, the axioms, and the rules of inference
denoted the conditionally universal connection between the constructs (Kesting & Vilks, 2005). Axiomatic theory consistently distinguishes
in this vagueness with the symbol "—" (not with the causal connection between first propositions, the axioms, which cannot be concluded from
→, to be elaborated upon later). Theories can be complex and extend the system, intermediate steps, the so-called lemmas, and theoretical
beyond the basic form A—B. They can include a variety of constructs, as statements which can be logically concluded from the system, the the­
well as mediators and moderators. The conditionally universal connec­ orems. In economic theory, Debreu (1959) comes quite close to this
tions are and will remain the decisive factors. This definition is quite form. In management science, however, there is hardly anything com­
close to that of Bacharach (1989, p. 496), according to whom theory parable and little to no fully developed theory in this sense. In addition,
offers “a statement of relations between concepts within a set of research in management science is characterized by little use of logical
boundary assumptions and constraints.” or mathematical conclusions drawn within theory, i.e., we seldom use
The conditionality of universality refers to the constructs, A and B. any formal models. This is not necessarily a weakness, and there is good
Theory does not make any statement outside of A and B, but rather is reason to refrain from fully formalizing theory (Kesting & Vilks, 2005).
limited to precisely these constructs. An assessment of the empirical Nevertheless, we should be aware of how far most research in man­
validity of the theoretical connection therefore requires that A and B are agement science is from a fully developed theory in the sense of an
well defined. If this is not the case, the validity of the theoretical axiomatic theory.
connection also becomes unclear. On the other hand, the basic form of theory remains, necessitating
Additionally, the conditionality of universality relates to the mod­ universal statements in the form A→B. Such sentences can offer valuable
erators and mediators that further limit and determine the validity of the gains in knowledge, even in isolation. As an example, Galinsky and
theoretical context (the famous "it depends"). The validity of the Mussweiler’s (2001) found that the first offer to negotiate has a strong
connection is thus limited by conditions or determined by third factors. influence on the final outcome. With its high empirical relevance (Orr &
Within these conditions, however, the connection is universal. In the Guthrie, 2005), this sentence offers valuable insights, even in isolation,
example of boiling water, the boiling point depends on factors such as for individuals faced with the decision of making the first offer or
sea level and salinity of the water. This conditions the relationship be­ waiting for the other party. As Weick (1995) rightly points out, vari­
tween temperature and boiling point, making it more complex. ables, diagrams, and hypotheses can also represent a valuable contri­
This definition of theory is certainly very broad, but at the same time, bution to theoretical work. In principle, theory can therefore also be
very specific, as it designates its own class of statements and is precisely very simple.
delimited. With this definition, theory is by no means in danger of In management science, theory is rather fragmented. Many studies
becoming “meaningless” (as Merton, 1967 fears, quoted in Sutton and only examine a handful of hypotheses (Sandberg and Alvesson, 2011),
Staw, 1995: 371). and these are often included informally (sometimes hardly at all) in the
Up to this point, there should be considerable agreement in our research context. Limitation to a few hypotheses is methodologically
discipline. Disagreement, then, primarily stems from two questions, unavoidable, especially in quantitative studies; however, a more
namely, what constitutes the substance of the connection between the comprehensive scope and bigger picture can aid our understanding of
constructs, and in which form the theory should be presented. causal relationships and help capture the complexity of decision-making
In his seminal work of 1748, David Hume showed that the notion of situations. When making decisions, is often impractical to concentrate
causality is problematic in that causality itself cannot be observed on individual aspects and ignore others. Holding to the example above:
(Hume & Beauchamp, 2000). Observations are limited exclusively to there is already value in knowing that there is a stable connection be­
sequences of events, and causality can only be inferred from these. There tween the first offer and the final outcome. To make a good decision,
have been many attempts to overcome this problem, but to my knowl­ however, it is helpful to know more, e.g., when and for what amount to
edge, none have succeeded. Consequently, causality is always a make the first offer, when it is better to refrain from the first offer, and
construct, an assumption. This applies not only to causal relationships, how all this fits into the overall negotiation context. Complex theories
but also to the concept of causality as such. It can be proven neither that such as behavioral decision theory or transaction cost theory also serve
causality is stable nor that it even exists. In response, in his quest to ban importantly as a theoretical lens for research. Additionally, complex
all metaphysics from science, Mach (1976) proposed replacing causality theories play an important role in management education. Conse­
with functional relationships. However, these should not be understood quently, there is an obvious need for complex theories, and the concept
as independent statements (in the sense that such functional relation­ of the full-blown theory can fit here. But how is this term to be under­
ships actually exists), but merely as an aid to organizing data. However, stood conceptually? What constitutes a full-blown theory, and how is it
it was soon recognized that this view does not allow the use of func­ to be presented? What demands are to be made of a full-blown theory?
tionality for explanation and prediction and thus deprives science of Does this have to be summarized in one comprehensive, logically
essential content. As a result of this and other discussions, causality has consistent building (in the sense of a “classic achievement” according to
asserted itself, as it has ultimately proved indispensable for scientific Schumpeter (1994)), or is a partially inconsistent body of literature
work. This specifies the conditionally universal connection between the sufficient? At this point, I would like to emphasize the value of con­
constructs, and the new definition is that theory consists of the estab­ ceptual studies that establish such overarching connections, as my
lishment of conditionally universal causal connections, in the notation of impression is that the value of such studies is insufficiently recognized.
Hilbert and Ackermann (1950), A→B. However, causality remains a
construct, and a certain skepticism about causality is therefore neces­ 4. The development of theory
sary. This is particularly important for theory building.
Moreover, there is disagreement regarding different ideas as to 4.1. Theory building
which form theory must suffice. Such ideas are found implicitly in
Sutton and Staw (1995) and their statement that variables, diagrams, The main difficulty in constructing theory stems from a lack of direct
and hypotheses are not about theory. However, they can also be found in connection between empirical and theoretical knowledge. Not only that
Weick and his idea of the “fully blown theory” (Weick, 1995, p. 385). I causality, as presented above, cannot be observed. In addition, there is

3
P. Kesting Scandinavian Journal of Management 39 (2023) 101273

no inductive inference; it is not possible to draw conclusions about “Different theories tend to reflect different perspectives, issues and
universal connections from singular observations. Hume recognized this problems worthy of study, and are generally based upon a whole set
as early as 1748. In its quest to remove all metaphysics from science, of assumptions which reflect a particular view of the nature of the
logical positivism has done everything it can to overcome Hume subject under investigation.”
(Caldwell, 1982) and failed gloriously. Gloriously, because the re­
But this, in turn, means that theory is always a choice and focusing on
searchers involved were their own biggest critics, discontent with simple
one aspect means that other aspects are pushed aside, neglected, or
solutions. This failure has found its expression in the renaming of logical
ignored entirely. Seidl (2007, p. 16) describes this as the "dark" side of
positivism as logical empiricism. I am convinced that the claim of
knowledge:
positivism to build science on reliable knowledge and refrain from any
speculation or metaphysics has failed, at least for the moment. But with “Knowledge thus means selection; and selection implies contingency
that, the platonic idea that it is possible for the prisoner to leave his cave – one could have selected differently. The selectivity of knowledge,
through true insight has also burst. This characterizes current science however, remains latent. That, and what knowledge excludes, is not
like hardly any other finding. included in the knowledge. Knowledge, thus, inevitably implies
Consequently, theory cannot be derived from observation or other nonknowledge as its other, or “dark,” side.”
evident knowledge but is always constructed. Until Hume is overcome,
this is one of the fundamental principles of research. Statements such as,
“Most qualitative papers advance theory by building it inductively” 4.2. Theory testing
(Bansal & Corley, 2012: 509) are counterproductive in their suggestion
that such a connection could exist after all. My research experience, In my opinion, the most important contribution of Popper (1992,
along with the fact that constructivism is still understood as a philo­ 2020) is his skeptical approach to confirmation. In his time, Popper was
sophical option, casts doubt that this insight has really penetrated our particularly bothered by discussions with adherents of Marxism and
discipline. I take issue with the concept of constructivism because there psychoanalysis who provided extensive confirmative evidence for their
is no counterpart to it – we are all constructivists when we formulate theories but immunized their theories, thus evading critical discourse.
theory. To be very clear at this point, it is not possible to induce theory This revealed the inconsequential value of confirmative evidence in
from any data. Caldwell (1982: 51) formulated the consequences of this evaluating theory. As exemplified by conspiracy theories, confirmatory
in all his radicality: “Again, it is well known that for any set of data, an evidence can be found for even the most absurd theory. Critical
infinite number of theories can be developed to explain them.” discourse is therefore crucial for assessing theory.
The construction of theory thus becomes a creative process. How­ Kuhn (1996), however, has shown that Popper’s idea of falsification
ever, as early as 1908, well before Popper, Schumpeter pointed out that in connection with theory is unrealistic, both from a sociological and
this process is not arbitrary, but that the construction is carried out with systematic point of view. Empirical findings can "prove" neither the
regard to an explanatory goal. accuracy nor the fallacy of theory, or as Harré (1985, p. 44) put it, "there
are no brute facts." This leaves critical scientific discourse as the only
“The purely static economy is nothing else than an abstract picture of
way to assess theory. Critical scientific discourse consists of arguments
certain economical facts, a schema that is supposed to serve for the
based on empirical findings and logical conclusions. Kuhn (1996) has
description of the same. It is based on certain assumptions and in­
shown that this is a sociological process. Popper (1992) rightly
sofar a creature of our arbitrariness, just the same as that every other
emphasized the importance of research questions by which research and
one is exact science. So if the historian says that our theory is a
research discussion must be measured. This still seems to be accepted
figment of our imagination then he is correct in a sense. Surely, in the
today. In Sutton and Staw (1995), this discourse can be the only measure
world of phenomena itself there are neither our ‘assumptions,’ nor
for distinguishing between weak and strong theory. It is therefore a
our ‘laws.’ But an objection against the same does not follow from it
subjective distinction based on the sociology of science rather than a
yet; because this does not prevent that they suit the facts. Where does
strict methodological distinction. Statements such as "few of them take
this now come from? Simply from there that we have indeed pro­
the form of strong theory" (Weick, 1995, p. 385) or "Most products that
ceeded arbitrarily but rationally during the construction of our
are labeled theories actually approximate theory" (Weick, 1995, p. 385)
schema; we have simply constructed the same with regards to the
may be justified by good arguments, but ultimately represent subjective
facts.” (Schumpeter, 2010, p. 386)
judgments. This assessment is influenced by habits and ideas about what
One difficulty is the complexity of socio-economic phenomena, is important and what is unimportant. We should be aware of this when
which is further increased by human decisions and historical processes. evaluating theory. Too strong a focus on confirmative evidence and
It is often not possible to scientifically grasp real phenomena in all their confirmation of hypotheses in the discourse on the assessment of theory
complexity, and the construction of theory is essentially characterized ("theory testing") is not unproblematic, especially in light of Popper’s
by reduction. In this context, Weber (2013, p. 124) coined the term justified objections. A critical perspective is imperative.
“ideal type”: The evaluation of theory and thus the progress of science rests
entirely on discourse, and its importance therefore cannot be under­
“It [the ideal type] presents us with an ideal image of what goes on in
estimated. This discourse must be open, objective, and follow the pri­
a market for goods when society is organized as an exchange econ­
macy of argument. Logical inconsistencies are not to be ignored. Kuhn
omy, competition is free, and action is strictly rational. This mental
(1996) has shown that this cannot be avoided, but science must always
image brings together certain relationships and events of historical
struggle against cliques and power. Science is not just about gaining
life to form an internally consistent cosmos of imagined in­
knowledge; it is also a sociological phenomenon.
terrelations. The substance of this construct has the character of a
utopia obtained by the theoretical accentuation of certain elements of
5. Theory as an explanation of causal connections?
reality.”
Theories are not accurate representations of reality but instead Against the background of the findings of the philosophy of science,
accentuate and are guided by the research interests of the observer. the specification of theory in Sutton and Staw (1995, p. 375) seems
Theory thus becomes an interpretation and structuring of reality. particularly problematic. They demand, “A theory must also explain
Different research interests can lead to different theories. Against this why variables or constructs come about or why they are connected.”
background, Burrell and Morgan (1979, p. 10) argue that theory always Sutton and Staw (1995, p. 372) illustrate their requirement with an
embodies a certain perspective: example:

4
P. Kesting Scandinavian Journal of Management 39 (2023) 101273

“To illustrate, this sentence from Sutton’s (1991: 262) article on bill Is that what Sutton and Staw have in mind? The problem, however, is
collectors contains three references but no theory: ‘This pattern is that every micro-foundation itself is a construct. It consists of condi­
consistent with findings that aggression provokes the “fight” tionally universal causal connections of the same form as the original
response (Frijda, 1986) and that anger is a contagious emotion statement (it must consist of these to establish a causal connection), only
(Baron, 1977; Schacter & Singer, 1962).’ This sentence lists publi­ now A→R, R→S, and S→B. All of these statements are constructed again.
cations that contain conceptual arguments (and some findings). But Do these statements in turn require an explanation of their respective
there is no theory because no logic is presented to explain why causal connection? That would lead to an infinite regress. Do they not
aggression provokes ‘fight’ or why anger is contagious.” require an explanation? What then differentiates A→B on one hand and
A→R on the other? In addition, micro-foundations in management sci­
The conditioned universal causal relationship A[aggression]→B
ence are usually informal and thus produce only informal connections.
[fight response] in itself does not constitute a theory, as theory consists
In doing so, they fail to meet the requirement of Sutton and Staw (1995,
only in the presentation of a logic that shows why this connection exists.
p. 372) to provide a “logical” explanation of the causal relationship
For this reason, hypotheses are not theories. So Sutton and Staw.
A→B.
This way of thinking has become entrenched in management science
Micro-foundations can be helpful to break down and better under­
and must be challenged and changed. This perspective not only funda­
stand processes, “to unpack some of macro-management’s preferred
mentally contradicts the findings of the theory of science, but further,
aggregate concepts (e.g., ‘capabilities,’ ‘absorptive capacity,’ ‘routines,’
sets a standard that is counterproductive for research and cannot be met.
and ‘institutions’) in terms of individual action and interaction” (De
First, it should be noted that a statement in the form A→B is
Massis & Foss, 2018: 387). However, they do not provide any explana­
completely sufficient for an explanation and prediction in the covering-
tions of causal connections but are merely foundations.
law model. What Sutton and Staw (1995) overlook is that the statement
If Sutton and Staw (1995) do have a micro-foundation in mind, from
A→B itself provides an element for an explanation. Water begins to boil
what understanding of theory does this requirement arise? How is this
on Everest because it has been heated to 71 degrees and because this is the
demand justified? Why is A→R→S→B theory, but A→B not? If it is not
boiling point there. I don’t have to understand why the connection be­
micro-foundations they have in mind, then what? How else can a logical
tween boiling temperature and air pressure exists, only that it exists. An
explanation of the causal relationship A→B be provided?
“explanation” (whatever this means in this context) of the conditionally
There is also a practical problem. Let us assume that the statement A
universal causal connection A→B itself is therefore necessary neither to
[aggression]→B[“fight” response] is the result of an empirical study in
understand developments in the past nor to assess the consequences of
which data supports the proposed hypothesis. However, this data cannot
certain decisions in the future. It is enough to know that A→B holds.
sufficiently support all the necessary theoretical statements for a micro-
Second, no explanation can be given for the conditionally universal
foundation. What else can offer an “explanation” of the hypothesis, i.e.,
causal connection in the form A→B, simply because there is no such
a logic that explains its causality? Ultimately, this can only be based on
explanation. The central insight of the philosophy of science since Hume
plausibility, a review of the literature, considerations, and speculation.
is that such statements are—must be—constructed, as there is no
This is exactly what the practice in management science resembles. Is
inductive inference, and causality cannot be directly observed. As
that what Sutton and Staw envision?
Schumpeter (2010) pointed out, statements in the form A→B are crea­
In contrast, I advocate focusing the theoretical work on the formu­
tures of researchers’ arbitrariness. Sutton and Staw’s (1995) require­
lation of conditionally universal causal connections in the form A→B
ment at this point cannot, in a strict sense, be fulfilled. This is
itself. As previously demonstrated, theoretical propositions are always
problematic because it misjudges the epistemological status of such
constructed and created by researchers. However, they are not arbitrary
statements and creates the impression that such a justification could
but are formulated with a goal of explaining real facts. The formulation
exist. At this point, we should remember Schlick et al. (1979) and his
of theoretical statements and their introduction into the scientific
complaint that demands of precisely this kind hinder the progress of
discourse should therefore stand on good reasons, which often lie pri­
science because they cannot be fulfilled and thus lead to fruitless
marily in data showing their empirical relevance. However, the context
discussions.
of justification should extend beyond pure empiricism to the existing
In a less strict sense, research indeed can offer “explanations” for
theory.
universal causal connections in the form A→B. In the justification
Thus understood, the demands of Sutton and Staw (1995) make
context of theorems in axiomatic systems, a causal chain is established
perfect sense to me: It can be helpful to place the theoretical statements
from which A→B follows. However, management science does barely
made in the context of existing research and make them plausible in this
practice axiomatic theory, and furthermore, the causal chain always
way. Which existing findings support the statement A→B and make it
refers to axioms within the system. The only claim, therefore, is that the
seem plausible? What are its implications? In what points does it
causal connection A→B can be concluded from certain assumptions, a
contradict existing views? How do they contribute to our understanding
very weak form of explanation. It is doubtful that Sutton and Staw
of the issue? This may well include a presentation of the reasons that
(1995) have that in mind either.
suggest a causal relationship A→B. However, this is not a logical justi­
Still, there is the justification context, which is carried out in a micro-
fication of the causal connection A→B itself but is instead a justification
foundation. A micro-foundation examines which microprocesses are
for bringing the hypothesis A→B into the scientific discourse. The focus
subject to a macro-relationship (Felin et al., 2012). Schematically, the
then is on the hypotheses, with the “explanation” serving as support.
structure of a micro-foundation is represented by Coleman’s (1990)
In this context, Styhre’s (2022) article is worth mentioning, which
bathtub model (Fig. 2).
formulates a “theory of theorizing,” i.e., examines how researchers
“render an empirical material meaningful on basis of a description that
oftentimes includes abstract, yet precise analytical terms.” This way, an
explanation for A→B could in fact be provided. However, this does not
consist of showing the logic of the causal connection, but rather the
reasons that led the researcher to construct this connection in the
context of management studies.

6. Outlook

Fig. 2. Schematic representation of Coleman’s (1990) bathtub model. Theory is indispensable for research, providing its aim is to explain

5
P. Kesting Scandinavian Journal of Management 39 (2023) 101273

or predict. Our discipline’s current focus on theory can only be dispense with empiricism. Matsuo Basho, a 17th-century Japanese haiku
welcomed against this background. Without theory, we cannot assess master, once said, “To learn about pine trees, go to the pine tree; to learn
the consequences of actions and therefore cannot give any recommen­ of the bamboo, study bamboo.” I’m sure Popper and Schumpeter would
dations for action. In our theoretical work, however, we should be aware wholeheartedly agree. According to Popper and Schumpeter, re­
of the findings of the philosophy of science: searchers should dig deep into the field, observe and collect data, study
the literature, and discuss different aspects. However, the leap to theory
(i) Theory describes stable causal relationships. It essentially con­ is based not on induction, but on a creative, thoroughly intuitive syn­
sists of conditionally universal causal statements in the form thesis of the knowledge gained, which finds its form in the bold
A→B. conjecture or vision. The systematic collection of data subsequently
(ii) Theory is indispensable for the explanation and prediction of follows to critically test the formulated new hypotheses or models.
empirical facts according to the covering-law model. But what about management science? A study by Sandberg and
(iii) Because causality cannot be observed and because there is no Alvesson (2011) provides interesting insights. This study shows “that the
inductive inference, theory is always constructed; for every most common way of producing research questions is to spot various
empirical fact, there is an infinite number of theories that can gaps in existing literature, such as an overlooked area, and based on that
explain it. to formulate specific research questions” (p. 24). Sandberg and Alvesson
(iv) For this reason, conditionally universal causal statements in the also refer to this modus operandi as "gap-spotting." As a result, research
form A→B themselves cannot be explained; nor can constructs be “is more likely to reinforce or moderately revise, rather than challenge,
explained, as they, like theories, are constructed. already influential theories” (p. 25). This approach is quite distant from
(v) There are no brute facts; theories can neither be proven nor that of Popper and Schumpeter. Sandberg and Alvesson (2011) identify
explained, but can only be decided on in a scientific discourse a number of factors that convey this development and lead to a path
based on arguments. dependency of science. These include: the individual striving for secu­
rity and simplicity, the integration of research into traditions, a required
It is right and important that theory is an essential aspect of the recognition by established researchers, but also the institutional practice
scientific discourse; however, if that is the case, it is also important to of funding committees and journals.
clarify and reflect on the concept of theory. It must be clear what The excessive adherence to empiricism in theory-building studies
research is aiming for and by what criteria it should be judged. If this is and the repeatedly increased demand for inductive derivation (or at
not the case, research becomes arbitrary. “Theoretical contribution” least empirical foundation) of new theory seems to be an essential
then becomes synonymous with “I like it.” In this context, it is important expression, if not even the driver, of this practice. Cornelissen (2017, p.
to drop ideas that contradict the above findings of the theory of science, 368) points in exactly this direction, saying, “In recent years qualitative
above all, that there is a logical explanation for causal relationships and papers are increasingly being fashioned in the image of quantitative
that theory can be induced from data. Just recently, I read in the review research, so much so that papers adopt ’factor-analytic’ styles of theo­
report of a leading journal, “I admit that I didn’t fully follow your claim rizing that have typically been the preserve of quantitative methods.”
that we cannot induce theory from qualitative data”—and this claim was This is exactly what Popper and Schumpeter warned against, that
not related to weak induction. Such ideas are problematic in leading to research focuses too closely on the path and loses sight of the goal. But
claims that cannot be fulfilled; in reaction, researchers perform an “in­ how should new theory be motivated if not by data? At this point,
duction theater” (I use this term in reference to the term “innovation Burrell and Morgan’s (1979) understanding of theory as a perspective
theater” as found in Blank (2019)) and pretend to derive their propo­ can be helpful. The introduction of a new theory should focus on the new
sitions from qualitative data (many even believe they do). This distorts perspectives that arise from it, the new relevant aspects or perspectives
the evaluation of research and impedes scientific discourse. that can be revealed with it. Does it make sense to dare a little more
My suggestion in this study is to base theory on the universality of Popper and Schumpeter in management science? Discussing this ques­
causal connections, which can be simple hypotheses, but also complex tion could also help better understand what makes a theory novel.
models. However, theory always involves statements of type (1) in the A central point for scientific discourse and theory testing is that
covering law model, which are necessary for explanations, forecasts and theory is not only constructed but also reduced, and this reduction is
recommendations for action in this form. In doing so, I propose to subject to prioritization. This means, however, that acceptance of one
logically distinguish theory from explanation, i.e., theory is an element theory does not necessarily invalidate another, even if the two are
of explanation, but not the explanation itself. This distinction seems contradictory. Burrell and Morgan (1979) take a strongly instrumental
integral to me, but it stands in sharp contrast to Sutton and Staw’s (1995, position here, viewing theories as a resource for developing and testing
p. 374) specification that "theory explains why empirical patterns were new perspectives. Theory can thus guide observations and in­
observed or are expected to be observed." I consider such an equation of terpretations and might also be used to create new opportunities. Theory
theory and explanation problematic in its implication that theory would is then discussed relative to specific objectives.
also contain empirical statements (statements of type (2) in the covering In management science, it is indeed often the case that several the­
law model). Such a concept of theory is very ambiguous and would ories can exist simultaneously. Just think of the different strategic ap­
complicate defining the boundaries of theory. But that is only one view; proaches such as the market-based view, resource-based view, or
it makes perfect sense to critically examine this view and discuss dynamic capabilities view. In his book, The Sense of Dissonance, Stark
competing perspectives. The aim is not necessarily to agree on a single (2009) argues that this ambiguity should be understood not as a defect,
concept of theory, but rather to gain a better understanding of the nature but as a source of knowledge. Though Stark was primarily referring to
and structure of the concept of theory. organizations, his findings can also apply to science. Research should
With regard to theory building, the main question is on what basis therefore not strive to eliminate differences, but to understand them and
new theories should be introduced if they cannot be induced from use them to gain knowledge. Stark explains (p. 17), “Whether we refer to
qualitative data. Popper (1992) suggests that theory building should the process as research, innovation, exploration, or inquiry, the kind of
detach itself from data and start with a bold conjecture. “On these search that works through interpretation rather than simply managing
grounds, Popper rejects the confirmationist goal of discovering theories information requires reflective cognition.” This recognition of diversity
which have high inductive probabilities” (Caldwell, 1982, p. 43). stands in sharp contrast to the empiricists, whose attitude Suddaby
Similarly, Schumpeter (1994, p. 33) sees the starting point of theory in a (2015, p. 2) describes as follows: “When a single theory fails to emerge
“vision”, a “pre-analytic cognitive act that supplies the raw material for (as is inevitable), empiricists tend to reject the value of theory entirely
the analytic effort.” This does not mean, however, that we completely and focus energy exclusively on the collection of data.” One downside of

6
P. Kesting Scandinavian Journal of Management 39 (2023) 101273

diversity, however, is that it comes at the expense of clarity. To coun­ simple and unproblematic. Ultimately, this consists only in convention,
teract this, it is necessary to specify diversity and mutually relate in a determination of an object of knowledge. Much more problematic is
different perspectives; the individual perspectives should be internally the formulation of theory, and then, above all, dealing with theoretical
consistent. How should such a scientific discourse look in detail? Further statements. Much ambiguity remains in management science, as well as
clarification is still needed. some real misunderstandings. A look at the theory of science can help
Now, however, the question arises anew as to what theory actually is; sharpen the view here.
not how it is defined, but how it should be understood and used. What
insight do theories provide if they are not right or wrong in a strict sense, Data Availability
but ideal types, simplifying models constructed to serve an explanatory
goal, emphasizing some aspects and abstracting from others? At this No data was used for the research described in the article.
point, it may be useful to consider the findings of the instrumentalism-
realism debate, the positions of which are aptly specified by Caldwell
References
(1982, p. 26):
“Realists claim that theoretical terms must refer to real entities and Bacharach, S. B. (1989). Organizational theories: Some criteria for evaluation [Article].
Academy of Management Review, 14(4), 496–515. https://doi.org/10.5465/
that theories which do not are false. Instrumentalists are agnostic on AMR.1989.4308374
the point, for they insist that theories are only instruments, that as Bansal, P., & Corley, K. (2012). Publishing in AMJ -Part 7: What’s Different about
such it is incorrect to speak of theories as being either true or false, Qualitative Research? [Editorial]. Academy of Management Journal, 55(3), 509–513.
https://doi.org/10.5465/amj.2012.4003
and that the only relevant questions that can be asked regarding Beyer, J. M. (1982). Introduction. Administrative Science Quarterly, 27(4), 588–590.
theories concern their adequacy.” Blank, S. (2019). Why companies do "innovation theater" instead of actual innovation.
Harvard Business Review Digital Articles, 2–5.
The advantage of the instrumentalist view is that it frees itself from Bourbaki, N. (1994). Elements of the history of mathematics. Springer-Verlag.
the problems of the theory of science, but this comes at the cost of giving Brunsson, K. (2021). The use and usefulness of theory. Scandinavian Journal of
Management, 37(2), 1–3. https://doi.org/10.1016/j.scaman.2021.101155
up the search for truth. Furthermore, theory immunizes itself and eludes
Burrell, G., & Morgan, G. (1979). Sociological paradigms and organisational analysis:
critical scrutiny. Elements of the sociology of corporate life. Heinemann.
Caldwell, B. (1982). Beyond positivism: Economic methodology in the twentieth century.
“By avoiding the rigors of critical, potentially falsifying tests, in­ Allen & Unwin.
strumentalists can always preserve their theories. In cases of failure, Carnap, R., Carus, A. W., Friedman, M., Kienzler, W., Richardson, A. W., Schlotter, S.,
they can claim that the theory should not have been applied to the Carnap, R., & Carnap, R. (2019). Early writings (First edition). Oxford University
Press.
situation in question” (Caldwell, 1982, p. 52). Coleman, J. S. (1990). Foundations of social theory. Belknap Press of Harvard University
Press.
Instrumentalism can then become an excuse to disregard reality. Colquitt, J. A., & Zapata-Phelan, C. P. (2007). Trends in theory building and theory
Against this background, radical instrumentalism seems very problem­ testing: A five-decade study of the academy of management journal. Academy of
atic, whereas a strict realism, on the other hand, is always confronted Management Journal, 50(6), 1281–1303. https://doi.org/10.5465/
AMJ.2007.28165855
with the problem of induction. However, it should be emphasized that a Corley, K. G., & Gioia, D. A. (2011). Building theory about theory building: What
perspective as such is not an instrument, but a point of view and constitutes a theoretical contribution. Academy of Management Review, 36(1), 12–32.
therefore connected to reality. So how to understand theory, how to deal https://doi.org/10.5465/amr.2009.0486
Cornelissen, J. P. (2017). Preserving theoretical divergence in management research:
with it? This is perhaps where the greatest need for discussion in man­ Why the explanatory potential of qualitative research should be harnessed rather
agement science lies. than suppressed. Journal of Management Studies, 54(3), 368–383. https://doi.org/
This is the context in which the debate about “rigor” and “relevance” 10.1111/joms.12210
De Massis, A., & Foss, N. J. (2018). Advancing Family Business Research: The Promise of
in management science should also be examined. Beginning in the
Microfoundations [Article]. Family Business Review, 31(4), 386–396. https://doi.org/
1980s, there has been increasing criticism that research in the field of 10.1177/0894486518803422
management science “has had little effect on life in organizations” Debreu, G. (1959). Theory of value; An axiomatic analysis of economic equilibrium. Wiley.
Descartes, R., Veitch, J., & Hoyt-O′ Connor, P. (2008). Principles of philosophy. Barnes &
(Beyer, 1982, p. 588). “Critics question the relevance of what we teach
Noble.
business students as well as the meaning of the research performed at Eriksson-Zetterquist, U., Hansson, M., & Nilsson, F. (2021). On the use and usefulness of
business schools” (Gulati, 2007, p. 776). The debate about formalization theories and perspectives: A reply to Brunsson [Article]. Scandinavian Journal of
in economics (Kesting & Vilks, 2005) not only revealed a trade-off, but Management, 37(4). https://doi.org/10.1016/j.scaman.2021.101178
Felin, T., Foss, N. J., Heimeriks, K. H., & Madsen, T. L. (2012). Microfoundations of
also a tendency to lose oneself in the logic of theory and methods. routines and capabilities: Individuals, processes, and structure. Journal of
Despite Gulati’s (2007) unhelpful dismissal of this debate as a “socially Management Studies, 49(8), 1351–1374. https://doi.org/10.1111/j.1467-
constructed” struggle between “tribes,” the danger seems real to me, and 6486.2012.01052.x
Frijda, N. H. (1986). The emotions. Cambridge University Press.
it seems sensible to track down and critically question such de­ Galinsky, A. D., & Mussweiler, T. (2001). First offers as anchors: The role of perspective-
velopments time and time again. This debate is not socially constructed taking and negotiator focus. Journal of Personality & Social Psychology, 81(4),
but must be conducted again and again to prevent researchers from 657–669. https://doi.org/10.1037/0022-3514.81.4.657
Gulati, R. (2007). Tent poles, tribalism, and boundary spanning: The rigor-relevance
losing their grounding in reality. debate in management research. Academy of Management Journal, 50, 775–782.
Another recent debate is how theory should be applied in manage­ https://doi.org/10.5465/AMJ.2007.26279170
ment education. In this debate, Brunsson (2021) questions the role of Hambrick, D. C. (2007). The field of management’s devotion to theory: Too much of a
good Thing? [Article]. Academy of Management Journal, 50(6), 1346–1352. https://
theory in management education, stating, "It is suggested that
doi.org/10.5465/AMJ.2007.28166119
full-fledged theories obstruct the usefulness of theory, which should be Harré, R.. (1985). The philosophies of science (2nd ed.). Oxford University Press.
experience and intuition based and allow for discoveries, theorizing and Hempel, C., & Oppenheim, P. (1948). Studies in the logic of explanation. Philosophy of
Science, 15, 135–175.
new concepts” (Brunsson, 2021, p. 1). In their reply, Eriksson-Zetter­
Hempel, C. G., & Fetzer, J. H. (2001). The philosophy of Carl G. Hempel: Studies in science,
quist et al. (2021) point out the benefits of theory for management ed­ explanation, and rationality. Oxford University Press.
ucation to describe and analyze phenomena. Weighing these two effects Hilbert, D., & Ackermann, W. (1950). Principles of mathematical logic. Chelsea Pub. Co.
against one another begs the empirical question of to what extent theory Hume, D., & Beauchamp, T. L. (2000). An enquiry concerning human understanding: A
critical edition. Clarendon Press; Oxford University Press.
sharpens the view and to what extent it impedes creative discovery. Kant, I., & Meiklejohn, J. M. D. (2018). Critique of pure reason (Dover thrift editions. ed.).
Dover Publications, Inc.
7. Conclusion Keller, S. (2000). An interpretation of Plato’s Cratylus. Phronesis, 45(4), 284–305.
Kesting, P., & Vilks, A. (Eds.). (2005). Formalism. Edward Elgar.
Kuhn, T. S. (1996). The structure of scientific revolutions (3rd ed.). University of Chicago
As a result, it can be stated that the definition of theory is surprisingly Press.

7
P. Kesting Scandinavian Journal of Management 39 (2023) 101273

Mach, E. (1976). Knowledge and error: Sketches on the psychology of enquiry. D. Reidel Pub. Seidl, D. (2007). The dark side of knowledge. Emergence: Complexity & Organization, 9(3),
Co. 16–29.
Merton, R. K. (1967). On theoretical sociology; five essays old and new. Free Press. Stark, D. (2009). The sense of dissonance: Accounts of worth in economic life. Princeton
Orr, D., & Guthrie, C. (2005). Anchoring, information, expertise, and negotiation: New University Press.
insights from meta-analysis. Ohio State Journal on Dispute Resolution, 21(597–628). Styhre, A. (2022). Theorizing as scholarly meaning-making practice: The value of a
Plato, & Reeve, C. D. C. (1998). Cratylus. Hackett Pub. Co. pragmatist theory of theorizing. Scandinavian Journal of Management, 38(3). https://
Popper, K. R. (1992). The logic of scientific discovery. Routledge. doi.org/10.1016/j.scaman.2022.101215
Popper, K. R. (2020). The open society and its enemies. Princeton University Press. Suddaby, R. (2015). Editor’s comments: Why theory? Academy of Management Review,
Russell, B. (1946). History of western philosophy and its connection with political and social 4015, 1–5. https://doi.org/10.5465/amr.2014.0252.test
circumstances from the earliest times to the present day. G. Allen and Unwin ltd. Sutton, R. I. (1991). Maintaining Norms about Expressed Emotions: The Case of Bill
Rynes, S. L. (2005). Taking stock and looking ahead [Editorial]. Academy of Management Collectors [Article]. Administrative Science Quarterly, 36(2), 245–268. https://doi.
Journal, 48(1), 9–15. https://doi.org/10.5465/AMJ.2005.15993108 org/10.2307/2393355
Sandberg, J., & Alvesson, M. (2011). Ways of constructing research questions: Gap- Sutton, R. I., & Staw, B. M. (1995). What theory is not. Administrative Science Quarterly,
spotting or problematization. Organization, 18(1), 23–44. https://doi.org/10.1177/ 40(3), 371–384. https://doi.org/10.2307/2393788
1350508410372151 Weber, M.. (2013). The ’objectivity’ of knowledge in social science and social policy. In
Schlick, M., Mulder, H. L., & Velde-Schlick, B. F. B. v. d. (1979). Philosophical papers. D. H. H. Bruun & S. Whimster (Eds.), Max Weber: Collected methodological writings (pp.
Reidel Pub. Co. 100–138). Routledge.
Schumpeter, J. A. (1994). History of economic analysis. Oxford University Press. Weick, K. E. (1995). What theory is not, theorizing is. Administrative Science Quarterly, 40
Schumpeter, J. A. (2010). The nature and essence of economic theory (English ed.). (3), 385–390. https://doi.org/10.2307/2393789
Transaction Publishers. Whetten, D. A. (1989). What constitutes a theoretical contribution. Academy of
Management Review, 14(4), 490–495. https://doi.org/10.5465/AMR.1989.4308371