Pragmatism vs interpretivism in qualitative
information systems research
Göran Goldkuhl
Linköping University Post Print
N.B.: When citing this work, cite the original article.
This is the authors’ version of the original publication:
Göran Goldkuhl, Pragmatism vs interpretivism in qualitative information systems research,
2012, European Journal of Information Systems, (21), 2, 135-146.
http://dx.doi.org/10.1057/ejis.2011.54
Copyright: Palgrave Macmillan
http://www.palgrave-journals.com/pal/index.html
Postprint available at: Linköping University Electronic Press
http://urn.kb.se/resolve?urn=urn:nbn:se:liu:diva-76528
1
Pragmatism vs. interpretivism in qualitative information systems
research
Göran Goldkuhl, {goran.goldkuhl@liu.se},
Department of Management and Engineering, Linköping University, Sweden
Abstract
Qualitative research is often associated with interpretivism, but alternatives do exist. Besides
critical research and sometimes positivism, qualitative research in information systems can be
performed following a paradigm of pragmatism. This paradigm is associated with action,
intervention and constructive knowledge. This paper has picked out interpretivism and
pragmatism as two possible and important research paradigms for qualitative research in
information systems. It clarifies each paradigm in an ideal-typical fashion and then conducts a
comparison revealing commonalities and differences. The possibilities of combining
pragmatism and interpretivism in qualitative research in information systems are analysed. A
research case that combines interpretivism and pragmatism is used for illustration. The paper
thus contributes to a discussion about different paradigms and methods for qualitative
research in information systems.
Key words: Qualitative research, interpretivism, pragmatism, paradigm, information systems
Introduction
Background
The interest in qualitative research in information systems (QRIS) has grown for several years.
Many scholars acknowledge the difficulties in reducing the complex social and technical
phenomena in the IS-field to quantitative figures. There is a need for more open and nuanced
ways to study and analyse IS complexities. Historically, significant compilations of articles
discussing and presenting qualitative IS research have been made, such as Mumford et al
(1985), Nissen et al (1991), Lee et al (1997), Trauth (2001) Myers & Avison (2002a) and
Kock (2007). There have also been special issues in journals containing papers on qualitative
research or certain methods within such a tradition; cf. e.g. Myers & Walsham (1998), Kock
& Lau (2001), Baskerville & Myers (2004).
One important discussion concerning QRIS is whether qualitative research is equal to
interpretive as this has sometimes been considered the case. Trauth (2001b, p 7) states that
“interpretivism is the lens most frequently influencing the choice of qualitative methods”,
There are however some reservations to make against such views. Myers & Avison (2002b, p
5) write “It should be clear from above that the word ‘qualitative’ is not synonym for
‘interpretive’. Qualitative research may or not be interpretive, depending on the underlying
philosophical assumptions of the researcher.” They mention three possible epistemologies
(interpretive, positivist, critical) following Orlikowski & Baroudi (1991) and Chua (1986).
The question of positivism vs. interpretivism in IS has been discussed by several scholars.
Some attempts have been made to reconcile the differences and propose integrated views (e.g.
Lee, 1991; Weber, 2004; Fitzgerald & Howcroft 1998). Other scholars claim and insist that
the differences between these two paradigms are great and irreconcilable (e.g. Orlikowski &
2
Baroudi, 1991; Walsham, 1993, 1995). It seems actually that much of the discussions and
comparisons concerning interpretivism vs. positivism have had the character of interpretivists
claiming the differences and positivists disregarding the differences. If one wants to discuss
the differences between positivism and interpretivism in connection with qualitative research,
it is obvious that interpretivism is an established, elaborated and adapted research paradigm
for this type of research. Even if positivism can be applied to qualitative studies (e.g.
Benbasat et al, 1987), ideal-typically it seems to have been adapted for use within quantitative
studies.
Is it so that interpretivism should be seen as the dominant research paradigm for qualitative
research? Are there, then, no real competitors? Alternative research paradigms which can be
compared and evaluated together with interpretivism do exist. Critical research is one such
paradigm according to a division made by Chua (1986) and Orlikowski & Baroudi (1991),
although there are scholars (e.g. Butler, 1998) who prefer to see this paradigm as a variant
within interpretivism. On the basis of these discussions and comparisons I do not see an
urgent need to proceed with making comparative reviews of interpretivism and critical
research.
As stated, a major part of the meta-scientific debate has concerned the two rivals
interpretivism and positivism. In a paradigm analysis within business ethics, Wicks &
Freeman (1998) have added pragmatism as a third alternative besides interpretivism and
positivism. A similar stand has been taken by Fishman (1999) in psychology. Inspired by
Wicks & Freeman (1998), Goles & Hirschheim (2000) also argue that also the IS research
paradigm debate should include pragmatism.
Pragmatist thinking has influenced IS research to a great extent, although the paradigmatic
foundations have not been fully acknowledged. When introducing the MIS Quarterly special
issue on action research (AR), Baskerville & Myers (2004) claim that paradigmatic
foundations for this research approach should be found in pragmatism. Actually, they
explicitly refer to the classical pragmatist philosophers (Pierce, James, Dewey and Mead)
when making this statement. Far from everyone applying AR makes such a paradigmatic
reference to pragmatism. Another evolving research approach within IS, design research (DR),
can also be located within pragmatic ground. Lee & Nickerson (2010) state that pragmatism is
a more adequate research paradigm for design research than positivism.
Pragmatism is concerned with action and change and the interplay between knowledge and
action. This makes it appropriate as a basis for research approaches intervening into the world
and not merely observing the world. This would be the case if the intervention is
organisational change (as in action research) or building of artefacts (as in design research).
The growing interest in action research and design research (e.g. Cole et al, 2005; Järvinen,
2005; Iivari & Venable, 2009) makes it important to investigate pragmatism as one possible
paradigmatic base for QRIS.
Braa & Vidgen (1999) have presented a research-methodological framework consisting of
three epistemological orientations: Research 1) aiming for explanation and prediction, 2)
aiming for interpretation and understanding and 3) aiming for intervention and change. The
first approach is of course located within positivism and the second in interpretivism. For the
third they do not give any clear reference to a corresponding school of thought. They refer to
inventionary research and action research as variants of research for this epistemological
3
orientation. Braa & Vidgen (1999) is a typical example of researchers who talk about action
and change oriented research without explicitly locating it within a pragmatist paradigm.
Braa & Vidgen (1999) propose a research method, action case research, which combines
interpretive and interventionary research. There are other scholars who also have identified an
affinity between change and interpretation in research (e.g. Baskerville, 1999). Action
research (ibid) and specialities as action case research (e.g. Braa & Vidgen, 1999), grounded
action research (Baskerville & Pries-Heje, 1999) and dialogical action research (Mårtensson
& Lee, 2004) all seem to comprise qualitative, interpretive and pragmatist research
orientations.
Purpose and procedure
Certain scholars advise against blending interpretivism and positivism; instead recommending
that they should be kept apart as separate research paradigms. How should one view
pragmatism and interpretivism as paradigms? Should they be kept apart or could they be
blended? Some hybrid forms have already been alluded to above. Do we understand the
grounds for mixing pragmatism and interpretivism in QRIS sufficiently? Are there reasons for
not adding pragmatist thinking to interpretive studies or vice versa?
If one follows the quest for pragmatism in IS research by Goles & Hirschheim (2000), and
other scholars (e.g. Goldkuhl, 2004; 2008b; Marshall et al, 2005), there seems to be a need for
more comparative evaluations between research paradigms within IS that include pragmatism.
Is pragmatism to be seen as suitable paradigm for qualitative research? If so, how is it related
to interpretivism? What similarities and differences can be found? These research questions
constitute the core of the current inquiry. Goles & Hirschheim (2000) have taken an important
first step here, comparing positivism, interpretivism and pragmatism. A more thorough
investigation is, however, seriously required.
The purpose of the paper is thus to clarify characteristics of interpretivism and pragmatism as
possible research paradigms for qualitative research within information systems. The purpose
is to make a comparative review of these two research paradigms. Similarities and differences
are sought for. The clarification and comparison will be made with the aid of paradigmatic
constituents such as assumptions concerning ontology, epistemology, methodology and
researcher – practice relations (Orlikowski & Baroudi, 1991; Goles & Hirschheim, 2000;
Iivari, 2007). As a first step, the possible divergences need to be clarified. To do this I will
conduct an ideal-typical approach in order to achieve clarification of each research paradigm.
In this sense I will follow similar approaches which compare different research paradigms as
ideal-types (e.g Orlikowski & Baroudi, 1991). In their analysis of the three research
orientations Braa & Vidgen (1999) describe both the distinct research orientations and hybrids
as mixtures between the “genuine forms”. Initially in my own analysis I will try to elaborate
ideal-typical genuine forms and avoid hybrid forms. A second step will be to investigate
similarities and clarify the possibilities to combine the two research paradigms in practice. In
connection with this I will also use an account of an empirical research project which
comprises both interpretive and pragmatist elements.
For this paradigm comparison between interpretivism and pragmatism, what can be learnt
from the debate concerning interpretivism vs. positivism? There are purist arguments
claiming that paradigms should not be mixed; they should be kept apart as distinct approaches.
There are, on the contrary, opponents against ideal-typically discerning of differences. To
4
contrast research paradigms is seen as a hindrance of blending different approaches in practice.
Goles & Hirschheim (2000) even state that the introduction of pragmatism “undercuts the
traditional dichotomistic warfare between conflicting paradigms by providing a philosophical
basis grounded in pluralism”. I do not think that researchers firmly rooted within one research
paradigm (positivism or interpretivism) agree to this radical proposal. We have not yet come
to an end of paradigm history in IS.
There are differences between research paradigms and I cannot see that such differences
should be blurred. The identification of such differences contributes to our paradigmatic
awareness. This is also a pre-condition for an informed mixing of views and elements from
different research paradigms in practical research. There are arguments for discerning
differences and similarities but also for investigating possibilities to blend and combine. My
aim is to bring more clarity to the choice of qualitative research methods in IS: I want to
reduce uncertainty among IS scholars as to whether it is possible to combine interpretive and
pragmatist approaches in QRIS. This is especially important in regard to the growing interest
in action research and design research in IS.
There are several reasons for bringing pragmatism into a comparative review of research
paradigms for QRIS. Pragmatism may contribute with the broadening of possible research
alternatives for a qualitative researcher; to see that interpretivism is not the main viable option.
The bringing in of pragmatism may also contribute with clarifications of pure and hybrid
forms of interpretivism and pragmatism in QRIS. One additional reason is that there are
qualitative researchers that apply action research and/or design research who may wish to
subscribe to a clear paradigmatic basis for their work.
Interpretivism in qualitative research
Main characteristics
Interpretivism is not a unified and unequivocal tradition. There are many forms of
interpretivism. Butler (1998) identifies several different variants such as conservative,
constructivist, critical and deconstructionist. The deconstructionist approach seems equivalent
with postmodern structuralism and this approach does not appear to be central in the
interpretive IS tradition. As indicated above, a critical tradition can be seen as a viable
separate tradition within IS (cf. e.g. Orlikowski & Baroudi, 1991) and this approach is
therefore also left out from the current study. In my analysis I will mainly focus on the
constructivist tradition and partially on the conservative (such as classical hermeneutics). This
means that the analysis here will focus on hermeneutic and phenomenological traditions. My
attempt is also, as mentioned, to make an ideal-typical account of interpretivism in IS.
The aim of understanding the subjective meanings of persons in studied domains is essential
in the interpretive paradigm. This was a central claim in the Verstehen sociology of Max
Weber (1978): the postulate of subjective interpretation. Alfred Schutz (1970) brought the
Verstehen sociology further with inspiration from phenomenology. He claimed that scientific
knowledge (concerning social life) was of second-order character. It must be based on the
meanings and knowledge of the studied actors. “The constructs involved on common-sense
experience of the intersubjective world in daily life…are the first-level constructs upon which
the second-level constructs of the social sciences have to be erected” (ibid p 274). Silverman
(1970) describes the difference between natural scientists and social scientists as being that
they work with different realms. The natural world of matter is meaningless until the scientist
