MANAGERIAL ACCOUNTING RESEARCH 319
to more traditional perspectives of managerial accounting: can these various
perspectives be compared and contrasted or possibly blended, and a ‘‘champion’’
paradigm isolated? In exploring the structure of more general scientific revolutions,
Kuhn (1970) reasoned that because of fundamentally different philosophical
presumptions, it is impossible to employ the tenets of one paradigm to assist those
subscribing to a second paradigm to transition to understand the first paradigm.
But rather, the ‘‘leap’’ from one paradigm to another must be based on faith in
order to fully appreciate what a particular paradigm may offer for understanding
our existence. In this spirit, and more closely concerned with organizational
analysis, Morgan (1980) (see also Burrell and Morgan 1979; in accounting see
Dirsmithet al.1985) theorized that different paradigms both address different
sorts of problems and, where paradigms address common problems, portray them
in fundamentally different ways and thereby offer differing insights into their
nature. Thus, what is called for is not a blending of paradigms nor the isolation of
a particular paradigm as champion, but rather paradigmatic pluralism as a way
of enhancing our understanding of issues in the social sciences. Consequently,
we offer the various paradigms not as competing perspectives but in some sense
as alternative ways of understanding the multiple roles played by management
accounting in organizations and society.
Extending this theme and drawing upon Churchman’s (1971) characterization of
influencing systems, Mitroff and Mason (1982) offered a useful way to understand
the properties of more orthodox research approaches and one which calls for a
plurality of theories used in a dialectic fashion. The argument of Mitroff and Mason
(1982) also highlights fundamental differences in the types of problems which
may be addressed. Within the traditional approach to management accounting
research, one seeks regularity, consistency or consensus by two means. In the first,
one seeks patterns in specific sets of empirical data in a purely inductive mode.
‘‘Consensus’’ of data is in essence a guarantor of faithfully representing a concrete
reality. Any lack of regularity or consensus in the data (e.g., lowr^2 )servesto
question the validity of the pattern isolated or theory used. One seeks improved
understanding by refining the model’s specification. In the second approach, one
seeks internal consistency in a postulational system wherein realityisthe axiomatic
structure. In a deductively driven system, only the lack of internal consistency or
conflict in the propositional network can cause one to abandon it in favor of a
competing network.
The rational frame of reference importantly assumes that the phenomena under
investigation are either well specified and well known, or able to be well known
through some preliminary fieldwork (Keating 1995), or further refinement of the
model or propositional network, and hence, are eminently structurable. However,
because of its reliance on a set of fixed concrete data or variables expressed in
a fixed postulational structure, it is limited in its abilities to preserve or reflect
anomalies and uniqueness in phenomena and to capture the essence of ill struc-
tured problems. Its use is, therefore, relegated to examining well structured though
perhaps technically complex problems. Paradoxically, because of their very com-
prehensiveness, these traditional perspectives tend to suppress conflict, anomaly
and uniqueness. By contrast, a more interpretive or critical view emphasizes