Accounting as Social Science:
Some Implications for Teaching and Research
Tony Lowe and Tony Puxty
It is now commonplace to suggest that accounting has moved on from a
discussion of "techniques and procedures" (Hopwood, 1974) to a contextual
basis: that is, the debate has moved from accounting as a closed-system which
largely ignored the purposes, behavioural consequences, and behavioural
antecedents of accounting information to a set of approaches which do recognise
that accounting information is only of importance in a given context. This is
of course entirely praiseworthy, and context cannot be overemphasised. At the
same time, opening up accounting to the real world-of-affairs, recognising that
it is social science and not a collection of abstract mathematical manipulations,
also brings into play the very considerable methodological problems which have
bedevilled social sciences generally and the parallel problems of professional
property-in-knowledge, boundary-guarding and closed-mindedness which are
unfortunately part of the social science scene. For there is a tendency in social
science (and for all we know in other fields) for boundaries to be tightly drawn
around disciplines. This manifests itself in two ways: first, in that the workers
in one field ignore the findings of those in other fields, even when they might
be of considerable importance to them and would appear to the outsider to be
closely linked: and second, that fellow researchers from other disciplines who do
attempt to cross the boundaries (there are some) are seen as poachers, and
The contextual approach to accounting has emphasised that it is not just linking
with the social sciences: accounting is itself a social science, and thus it is
interdisciplinary, breaking down the boundaries which convention has set up.
But, claiming that accounting is a social science is perhaps surprising and
deserves some expansion. We would suggest that there are two kinds (or levels)
of social science. First, there is the kind which tries to understand some aspect
This paper is an edited version of a paper presented at the Research Seminar Series at
the Department of Accounting and Financial Management, University of the South
Pacific on 11 March, 1990.

of society: this is exemplified by sociology, which takes as its purview all
aspects of society, but particularly in the widest sense class structures and
institution-societal relations: also by economics, which restricts its interest to
those aspects of society concerned with productions and exchange of goods.
Pyschology, anthropology and politics are other examples. Second, there are
social sciences which are concerned with the relations between people in society
in a "secondary" sense. Accounting is about this: it is concerned with how
certain technical matters (accounting principles and reports) affect people and the
relations between them. The inputs to accounting are human actions, and the
output of accounting information is likewise a human action. At every turn, in
the contextual nature of accounting, one comes across human nature.
In this paper we shall first consider the problems the accounting researcher faces
in his interdisciplinary researches; and then focus on one particular social
science, sociology, and discuss a particular problem within that area and its
implications for the accounting researcher.
Social Science or Social Sciences?
Knowledge is knowledge is knowledge. It has been deemed sensible to break
down into convenient boxes knowledge marked "natural science", "social
science" and so on, and then create smaller boxes within each. In part, the way
in which the division has been made is historical. For example it might be
thought that psychology and psychiatry have a great deal in common - the one
dealing with normal human behaviour and the other with deviant behaviour -
but because the early psychologists were philosophers, and the early
psychiatrists doctors, the two have developed in different ways and are inevitably
organised into different "large boxes" so that interaction is very small. Also
there has developed a very different ethos in regarding certain subjects, which
have caused the divisions between them to be emphasised, with the result that
"disciplines" grow further apart.
All this implies that the interdisciplinary researcher (and lets agree that such a
creature is very useful) encounters three major problems:
1 antipathy of certain members of disciplines to discipline cross-fertilisation,
2 the different ethos regarding attitude - political, methodological and so on -
towards the subject being studied, and

3 the different languages which have built up for different specialisations.
The first has been discussed already: let us consider the second and third in more
The "ethos" problem is in part historical, but is also grounded in the nature of
the particular social science itself. As a result of the historical context of a
science, traditions of research methods, and attitudes, etc. are passed from teacher
to student, and hence become part of the saturation in the subject. In the end,
this results in the practical impossibility of bridge building. A further aspect is
that the very nature of the field conditions the approach scientists take towards
it. For example, there is a strong behaviouralist trend in psychological research
which in part, we suggest, is the result of the ease, in the laboratory context, of
reducing research hypotheses to simple stimulus-response (S-R) experiments.
Since the S-R technology is there (mazes and the like) there is a resulting
tendency to see S-R as the approach to the science. Anthropologists, on the
other hand, have been far less able to manipulate the variables they are
concerned with, and in their methodologies are less behavioural. Similarly, the
proximity of the professional accountant to the stock market has led to the
belief among many accounting researchers that there is something of special
importance in financial markets (as opposed to the factor markets for other
participant groups such as workers - that is, labour markets) and hence has led
to the considerable research into share price movements, financial gearing and
dividend policies and related activities. The ease of availability of the data has
led to the manipulation of it, without questioning its significance to the overall
context of the organisation.
The third problem is that of "different languages". By this we do not just refer
to the use of specialised jargon. The problem goes much deeper, as was noted
by Kuhn (1970) and applies both when there are different "schools" within one
discipline, and when there are different disciplines approaching a problem.
Each approach builds up a language. This language does in part consist not
only of newly coined technical terms to describe the particular events noted by

the researchers but also of day-to-day usage which, however, takes on a
particular "slant". This can be extremely complex, and it is often unconscious.
To take a particular example from the field of politics. To the non-Marxist the
word "imperialism" implies colonial exploitation of any kind, with the result
that in recent years it has been applied to the Soviet Union's domination of
Eastern Europe. To the Marxist, however, the word has a specialised technical
meaning concerning the tendency of capitalist states to colonise when faced with
a lack of opportunity at home. It is a notion initmately linked with the
dialectical materialist conception of historical development of capitalism. The
problem this causes is that for each school of thought a web of secondary
meanings builds up around the word, so that in the end communication between
the two becomes impossible.
Here we mean "impossible" in its literal sense: because, since every language
has such complexities of meaning attached to its terms and thus is value-laden -
there is no possible neutral language in which to translate and mediate. A
value-free language would be necessary but does not exist. This means that
there can be no real understanding between scientists in different disciplines and
of course inevitably has its implications for the accounting researcher.
In part this is because he himself has such a language, and in crossing
disciplines comes across different languages which he can understand only to a
lesser extent than the practitioners in that discipline. Similarly, when he
compares ideas from two disciplines (for example, if he looks at theWeberian
notion of power and compares it with that of the political scientist) for cross-
fertilising to accounting research, he can never be certain to what extent the
concepts of power are harmonious or contradictory.
The interdisciplinary researcher faces all these problems, but also one more
which seems crucial. He is, inevitably, not a "neutral" interdisciplinarian: that
is, it is extremely unlikely that he has a training in more than one discipline
equally. Rather, he has been trained in one discipline (eg. accounting) and often
only at the advanced stage of work will he explore the findings of other relevant
disciplines. Also the very decision as to which discipline to look at then
becomes problematic. With very little knowledge of other social sciences, how
does he choose which is useful, interesting, relevant to him, best? To give the
most obvious example here, does he look to psychology, to social psychology,
or to sociology (or even anthropology) for this analytical framework? What
criterion has he for ever knowing that he has made the right choice?

The result of all this is that he is either not aware of one particular problem
which is crucial, or he trusts that it is not important to him and quietly ignores
it. That is, the problem that even within one field of study there are - in the
social sciences in particular - different views of how study in that field should be
Several Paradigms
These are what Kuhn (1962) has called "paradigms" different approaches,
different languages. Many social sciences are at the stage where there are
competing paradigms. To cite a previous example again, psychology has its
behaviourists and their opponents: in sociology the functionalists and their
opponents: and, closer to home, in accounting a school (perhaps little-
articulated but nevertheless pervasive) which appears to believe that all
problems can be solved given sufficiently sophisticated mathematical
techniques, and their opponents (among whose numbers we count ourselves).
Lest we be accused of an unreasonable criticism, a recent example may
highlight our concerns. The research by the "Human Information Processing"
school of researchers [see the American Accounting Association Committee
Report 1978, -2 and Libby and Lewis (1977) for summaries] is seemingly well-
respected. Yet it is essentially mechanical in orientation: the models used derive
from cognitive psychology. No creativity in the decision-makers is allowed in
the models considered: input is pumped in at one end and output analysed at the
other. The perspective is reductionist in that the interactions of decision-makers
and longitudinal interactions are ignored. There is, too, a presumption that the
individual can be left to his own devices within the total organisation, and that
problems will sort themselves out thereby (see Tinker, 1977). Our criticism is
not so much the lacunae in this kind of approach, but the indifferences of such
researchers to holistic problems of organisations, and hence indifference to the
many other researchers in other areas who have pointed out these problems.
Nowhere in the review articles referred to above (never mind the original
research papers), is the problem mentioned or the partiality of the approach
As Hopwood has pointed out, there has been a gradual "opening-out" of
accounting researchers' perspectives as a whole in the last few years, from the
entirely closed-system approach through the individual-accounts approach to the
total organisation approach (now well exemplified by Gordon and Miller 1976).
But despite this, it would seem that the overwhelming amount of research still
being carried out is based on psychology and social psychology (see for

example, Rockness 1977, Collins 1978, Ruland, 1978). We might usefully
ask whether this might be because any appeal to sociology as a significant
discipline for interdisciplinary research might be perceived as ideologically
questionable; that is, it appears to be much easier to pretend that psychological
research can be "value-free" than sociological research. Frightened by
ideological implications, it is only perhaps to be expected that accountants will
restrict their research to those disciplines where they can safely ignore possible
political implications.
For example, a common and very useful concept to the sociologist over the
years has been that of role and the allied concept of status. These have been
distinguished as follows:
"When the social analyst refers to a social position which is defintely
institutionalised (e.g. mother, physician) he is more likely to use the term
'status'. By contrast, he is more likely to use the term 'role' when referring to a
social relation which is less insitutionalised (e.g. peer relations in play group)."
(Goode 1960).
Cicourel (1970) comments:
Thus statuses are defined as the class of roles which is
One immediate implication of this for accounting researchers concerns the
results of accounting systems in organisations. One of the principal
communications channels between superior and subordinate in an organisation
is the budgetary system, whereby the subordinate is responsible to the superior
for achieving targets' set forth. The result of such an institutionalised system is
an interaction between the two based on the roles set out for them: the one to
criticise or praise the other, the other to respond as his role demands (see Lowe
& Shaw 1968, 1970).
It seems clear, therefore, that the accounting system, if designed in this way,
would have a reinforcing effect on the interrelationship between the two,
continually being instrumental in redefining the relationship. A budget system
is used by superiors to reinforce their managerial style: and the neutral budget
will be perceived differently by different subordinates depending on the use made
of it by the superior. Generally, then, the budget will affect the relative status
of the participants in the interaction: and equally their relationship will affect
the future budget set. Thus we have a complex interaction over time, which

continually affects both the level of the budget and the relationship perceived by
each participant. But where is the research which considers the implication of
this for the long-run structure of the organisation? Might we not suggest that it
would be too politically sensitive to analyse the function of a budget in this
way, so the research remains undone? (A more outrageous hypothesis is
available: namely, that the reinforcement of hierarchy in the organisation is the
principle function of the budget and the top management continues it at least as
much for this reason as for the more explicit reasons to do with "planning",
"motivation" and so on).
Changing Paradigms : A Crucial Issue in Social Science Research
We now wish to move on to a more specific issue, and one which we feel is
becoming more important as time goes by, the problem of change. Two facts
are fairly well established: firstly, that the rate of change in society is
accelerating (there is a great deal of literature, both "popular" and "scientific" to
underline this point): second, that social science models, unlike those of the
natural sciences, do not have change built into them as a specific analytical
variable. We need only look at the changing labour relations atmosphere or the
relative downgrading of the status of white-collar workers to recognise that there
is a constant shifting of relationships between elements of an organisation.
Thus we might suppose that the contextual characteristics considered by
accounting researchers when investigating the function of accounting systems
might also change. An example might again be helpful here. Hofstede (1967),
in a work which has become accepted as a classic in its field, considered as part
of his research the effect of participation in standard-setting on motivation to
achieve the standards. For this, he used a scale (adapted from Likert) ranging
from "decision taken by me/them without consultation" to "my/their opinion
not asked; decision not explained to me/them" to measure participation. That
is, the intertemporal problem was bypassed: answers were only solicited on the
basis of the situation at that time. Now one might suppose that a changing
situation might have changed or in some way affected the findings. Thus if
three years previously there had been an unhappy experience with participation
one might expect present participation to be less than welcome. Also, if the
extent of participation was increasing, one might hypothesise that the attitude
to the present level might be very different from that where the extent was
decreasing. It should be emphasised that this is not intended as a criticism of
Hofstede; rather as a suggested extension which might improve (or worsen!) his
correlations. In other words, there is some significant information intrinsic in

the change variable which has not been included, and as a result of which the
results must be treated with caution.
At this stage, then, it is being suggested that in contextual accounting research
not only should the context as it is be considered alongside the specific
accounting system variables as they are, but we should investigate the reasons
why the contextual variables themselves arose. Only in this way can the
context be understood sufficiently well to enable it to be discussed in the light
of another variable such as the accounting system. Of course, it is quite likely
that the dynamics of the accounting system themselves were part of the process
which produced the existing context. The context should be better understood
for other reasons too. For example, research in these areas has principally been
in the USA, the UK and Holland (Hofstede 1967). To what extent are the
cultures and institutional structures which influence human attitudes sufficiently
similar to use the results of one of these studies as a predictor of what would be
found in another country? Again, we do not know. But we might suppose that
a different atmosphere in Dutch industry from that in the USA (for example,
there might be less managerial labour mobility, and greater tolerance for
suboptimal performance: we do not know) might make the results non-
comparable. [See, for example, Turner and Lawrence (1965) who imply that
job enrichment is only helpful in small-town plants, not city plants, even
within the same country].
As an illustration of how the change problem has been approached in the social
sciences, we now move on to outline the discussions which have taken place in
another place in sociology over a ruling paradigm which has some inbuilt
defects, and a suggested collection of new approaches which are challenging that
paradigm. Before we do we should like to make an obvious disclaimer. We are
not sociologists and we approach the two methodological views as outsiders,
with what is inevitably a lack of great depth. We do so, however, to
demonstrate a useful approach which we think does have promise for accounting
The "old" paradigm in sociology goes by the name of structural-functionalism;
the new has, it appears, developed through different schools which are variously
named the "phenomenological" approach, "ethnomethodology" and "symbolic
interactionism". If the approach is itself subtle, the differences between the
approaches are even more subtle, and irrelevant we think to our purpose here:
we shall therefore follow Silverman's (1970) example and call them collectively

the "action" approach. This is an unfortunate label in some ways since one of
the prime functionalists, Parsons, also claimed that his was an action approach
(denied by his critics). We shall ignore this latter problem.
Structural Functionalism
Functionalism appears to have arisen in anthropolgy, as a reaction against an
evolutionary approach to the explanation of societal characteristics (Worsley
1970). The social scientist was faced with explaining social order:
Functionalism begins with Hobbes' problem of order and proceeds to
ask some important and interesting questions. 'How is it...that
society manages to work and to survive continual changes of
personnel? How do people with different genetic make-ups and
personality types learn to co-exist with one another and even to enter
into more or less stable and predictable forms of relationship?
(Silverman 1970).
Rather than attempt to explain a social state in terms of the society's history,
anthropologists such as Radcliffe-Brown suggested that an institution within a
society could best be explained by its function in society:
The function of any recurrent the part it plays in the
social life as a whole and therefore the contribution it makes to the
maintenance of the structural continuity.
(Radcliffe-Brown 1952).
Thus "the function of a recurrent activity in terms of its contribution to the
maintenance of this structure." (Worsley 1970).
Society is conceptualised as "a system of social relationships and clusters of
such relationships (institutions)" (Filmer et al. 1972). Moreover, the units
which make up society are no longer "people": they are norms, which build up
clusters to become roles, which in turn build into institutions (Homans, 1964).
In this way, the action of people is explained in terms of the norms which
govern their behaviour, the consequent roles in which they are cast by society,
and the institutions which are interpretable as the interaction of people acting
their social roles. Thus, society as a whole is analysed as a complex system, in
which the elements are kept in equilibrium by the forces of other elements of
the system: "Action is a product of system properties" (Filmer et al. 1972).

Talcott Parsons, without doubt the most influential sociologist of the last two
decades, considered that his own brand of functionalism had within it a theory of
The actor is articulated to the social system through processes of
socialisation and social control. The former is treated by Parsons in
terms of the internalisation of roles by the actor - that is, the
internalisation of the appropriate normatively oriented expectations
attached to locations within the organisational structure of the system.
(Filmer et al. 1972)
The appeal of such an approach is obvious. It is holistic, obedient to
Durkheim's exhortation to the sociologist to avoid psychologism and
Since their (social facts') essential characteristic consists in the power
they possess of exerting, from outside, a pressure on individual
consciousness, they do not derive from individual consciousness, and
in consequence sociology is not a corollary of psychology.
(Durkheim 1927).
It is, moreover, pleasingly available to analysis research. Moreover, it allows
the analyst to understand social action differently from the way in which it is
interpreted (or at least explained) by the actors:
Malinowski...insisted that the native Trobriander's account of
Trobriand society must be adequate, that the sociologist's account of
institutions is a construction not available to the untutored awareness
of the native informant.
(Macintyre 1967).
Some Criticisms of Structural Functionalism
The functionalist approach, then, is concerned with a macro theory of social
order. However, the influence of functionalist kinds of thought is more
pervasive than this might imply. As Mouzelis (1975) points out, it treats
human beings in a "passive, puppet-like manner." He continues:
For Parson's work, despite its "action" label, stresses the way in
which the social system's core values and their institutionalisation
into norms and roles shape an individual's activities through

socialisation, internalisation and social control. If one looks, for
example, at the way in which conventional sociological theory studies
such organisational roles as that of the foreman, the worker and the
manager, the direction of influence is always from the "organisation"
as a system to the individual - never the other way around.
This example illustrates that the functionalist viewpoint affects the way in
which we view individuals and their behaviour. The viewpoint is all-pervasive,
right down to the individual and his role. The very concept of "role" is taken as
given: there is no explanation within the methodology for how the role came to
exist in the first place, or how it was shaped in the precise way it was.
Moreover, there is no concept of change: of how someone might change his
role through effort, or through some kind of structural process in which others
begin to perceive him differently. Role becomes a prison within which the real
human being is trapped - and society is seen as a greater prison in which the
prisoners act out the functions assigned to them by the system.
A very significant point can be gleaned from the above, namely that although
the result of the approach was to take a social situation and ask "what is the
nature of this social situation and what are the consequences of it?", it does not
ask the primary question, how did the situation come about? (see, for example
Homans 1964).
Two other criticisms are possible along these lines.
First, that the approach assumes equilibrium, and thus the method has no room
for the explanation of how change comes about in a society. Let us be quite
clear about this. There is no denial within the approach of the fact that society
is dynamic: that transactions are taking place between people. But these people
are "trams": they change only in accordance with the roles they occupy. This
ignores the point brought out well in cybernetics that there is a concept of
"change of change", that the direction in which movement is taking place might
itself be altered, and that the alteration might come from within, through some
conscious effort of an actor. (The systems approach is criticised by many
authors as being synonymous with structural functionalism although we would
argue that this is a basic misunderstanding of the totality of the systems
approach. Principles such as equifinality allow for a kind of "consciousness",
"free will" with and dynamic homeostasis). Thus we cannot explain change by
a functionalist perspective.

Second, even though social norms may be internalised, they are not necessarily
expressed in behaviour. A norm becomes in effect a conscience: a super-ego.
But people do not always act in accordance with their consciences. As Wrong
(1967) has pointed out, they may act and feel guilty afterwards. Thus in the last
analysis a Parsonian approach cannot explain many actions.
A final criticism of the approach (and one which applies equally to its
alternative) is that it is based on the myth of "value-free sociology": that it
assumes that without predisposition the analyst can observe events and interpret
them in some kind of objective manner. 'Today, all the powers of sociology,
from Parsons to Lundberg, have entered into a tacit alliance to bind us to the
dogma that "Thou shalt not commit a value judgement" especially as
sociologists.' (Gouldner 1973).
Accounting and the Equilibrial Assumption
The equilibrial assumption permeates disciplines other than sociology. It exists
in economics, in approaches such as "equilibrium analysis", in the ceteris
paribus assumption, and the implication that factors of production have
common value systems. [But see Michael Aglietta (1979) and his careful
analysis and critique in which he proposes that change and regulation replace
equilibrium as the central concepts in economics theorising]. It exists also in
accounting research. There is a fundamental assumption, in accounting
research, that organisations are wholes, and that goal congruence is a desirable
"norm" towards which we are striving. It is as a result of such a value system
that we research bias in budgeting, for example, as a deviant action. We may
come to the conclusion that it is inevitable, but that is not to deny that we see
it as essentially disturbing to the natural and desirable order of things.
We would suggest, therefore, that any accounting teacher or researcher who does
not wish to take an approach such as this should make his assumptions
explicit. He should: state that he considers equilibrium to be "desirable", he
should make explicit what he precisely implies by equilibrium; and he should
be able to defend this rationale. In particular, he should be aware that his choice
of equilibrial or non-equilibrial assumptions is in fact a political choice: because
the acceptance of equilibrium is an acceptance of the status quo. We are not
suggesting that the researcher has no right to support the status quo. Rather,
we suggest that as a conscientious and honest academic seeking truth, he should
face up to the reality of his assumptions and be prepared to defend them.

Let us now turn to the alternative approach of the sociologists (which is not
necessarily the only alternative approach) and see in what way it seeks to solve
the problems created by rigid structural functionalism.
The Action Approach
We shall in general approach the action method through phenomenology, the
philosophical school founded by Husserl. Schutz later used Husserl's ideas and
modified them to create fresh methodology for sociology. We begin with
Husserl's ideas in an inevitably simplified form and move as quickly as possible
to their implications for social science.
Husserl starts from a similar point to Descartes, considering the "stream of
experience of the thinking ego" (Filmer et al. 1972). Starting from
consciousness, which he seems to reify, Husserl starts from rejection of science.
I am not the outcome or the meeting point of numerous causal
agencies which determine my bodily or psychological make-up...I
cannot shut myself up in the realm of science. All my knowledge of
the world, even my scientific knowledge, is gained from my own
particular point of view...The whole universe is built upon the world
as directly experienced and if we want to subject science itself to
rigorous scrutiny and arrive at a precise assessment of its meaning and
scope, we must begin by re-awakening the basic experience of the
world of which science is the second-order expression.
(Merleau-Ponty 1967).
The conscious, then, is paramount; it is seen as a relationship between a subject
and object; and the notion of intentionality is an attempt to describe this
The intentional consciousness is directed towards the "life-world": this is
intersubjective- that is, as being in the life-world we are related, and share an
inherent sociality of consciousness, experiencing the same world. This, say
Schutz, suggests that this has two implications: that people take it for granted
that if they changed places they would have the same experiences of the
common world; and that for all practical purposes, we assume that "in spite of
our unique biographical situations, the difference in our systems of relevances
'can be disregarded for the purpose in hand.'" (Filmer et al. 1972) and we
interpret things in a similar manner.

Thus, the approach is very much concerned with the subjective experience of
reality. It is concerned with the meaning which actors place on their situations:
and through the interpretations of these meanings, the way in which reality
comes about through the interaction of actors. Moreover, the reality of the
world is continually reaffirmed through men's actions: "meanings are not only
given, they are socially sustained" (Silverman 1970). When we go into
situations with others, we respond not to their activities but to their presumed
intenti ons.
The social objects...that people create not only have specific
meanings to their founders, but more importantly, are sustained and
changed by the manner in which programmes of action are interpreted
by the present participants in the context of social science.
(Filmer et al. 1972).
In this way, reality is changed by the action of men. As people interact with
one another they define a reality, and thus the reality comes into being: and
since life consists of a whole complex series of interactions between people, the
world is being constantly redefined through interactions, and thus is constantly
being changed by them. In this way, too, the action approach claims to have
explained the dynamics of soceity through the use of constructs involving real
people with intentions and consciousnesses, rather than prisoners inside
socially-conditioned roles. It should also be mentioned that there is a further
definition which cannot be ignored: one's self-definition. This is itself
continually redefined by interaction.
There might be a temptation to dimiss such an approach on the grounds that it
is intuitive rather than scientific, and on the grounds that it is reductionist,
abandoning the insights which have been gained through holistic interpretative
methods. This is considered unjustified by its proponents. As Silverman says
"(it) directs our attention less to unique experience, which one can never fully
grasp, than to the logic that underlies social experience as a whole...It seeks to
grasp not the idiosyncratic meaning of the other, but the emergent rules which
actors negotiate in order to invest experience with meaning." (Filmer et al.
1972, p. 171). Thus, as he says, it is not so much a commonsense sociology
as a sociology of commonsense.

A further answer to the change of reductionism is given by Wagner (1964):
Reductionist theories explain the behaviour of parts in terms of their
individual biological or psychological action
theories...are concerned with explanations in terms of interpersonal
human action.
It should be said that such action approaches do have critics. In particular two
criticisms should be briefly mentioned. The first is that the process is in the
end unscientific because it is concerned only with the interpretations of reality
by actors rather than the realities themselves. Thus, it might be argued that we
cannot have any kind of "reality" to analyse and test because it is ulitmately
unknowable by the method and in constant flux. A second criticism from
Allen (1975) suggests that there are two problems which phenomenologists
cannot answer: the first is that they are concerned only with the subjective
interpretation of reality whereas there is an objective reality which, interpreted
or not, will affect the actors: and his second point is that the approach avoids
the issue of "how do interpretations of the world acutally occur?"; that is, what
is it that governs how an actor will understand his world and interpret the life-
world before him? It might also be said from this point of view that the
approach is trite: that it is concerned too much with what people construct of
their reality and not enough with the underlying reality itself. If this is what
Allen is saying then the criticism cannot it seems to us hold water: for the
approach specifically defines that "objective reality" as being the product of
social interactions, so that "objective reality" is simply a product of social
action. Social action, of course, is precisely what the approach seeks to
The Implications for Accounting Teachers and Researchers
An accounting system exists within a contextual framework; it exists by
courtesy of that framework, to serve its information needs. The accounting
system is only of value so long as it affects human behaviour. No accounting
information should be produced which does not in some way affect human
behaviour, and it is the kind of behaviour which is produced and which is
desired which should form the criterion for good accounting information. It is
equally true that accounting information is itself the result of human action.
We now suggest that some of the implications of the kind of approach implied
by the action theorists was demonstrated in the illustrations given earlier in this

paper on the deficiencies of present accounting research methodology. The
context in which accounting systems operate have been assumed to be static
and in equilibrium: and although in small "bursts" theorists have accepted that
"accounting affects people and people affect accounts" the longitudinal
implications of this continued interaction have been largely ignored.
Accounting is often called the "language of business", although such a phrase
is often considered too trite to repeat in academic treatises. But this is as true
as it is trite, and in this we get a clue to the profound importance of these
issues. Being a language, accounting is a medium which actors use in their
transactions to define reality. In a key passage Berger and Luckmann said:
The common language available is grounded in everyday life and
keeps pointing back to it even as I employ it to interpret experiences
in finite provinces of meaning. Typically therefore I "distort" the
reality of the latter as soon as I begin to use the common language in
interpreting them, that is I "translate" the non-everyday experiences
back into the paramount reality of everyday life.
From this we can see that language is a limited tool; that it cannot express the
totality of experience, and the imperfections both in the richness of the
language itself and in the use of it by actors cause the language to be a cause of
the interpretation of interactions and thus itself. This means that, since
accounting information is used in just the same way as a true language to
express a point of view or to act as an intermediary in interpersonal
negotiation, or to express the user's meaning, and so on, so the language of
accounting itself comes to affect the reality of the situation and thus,
ultimately, the reality of the total organisation. To paraphrase Allen's ideas:
there is an "objective" set of accounting information, which is a set of figures
on a piece or pieces of paper: but there is also a "social" set of accounting
figures, which are the same numerals but used in defining a situation in such a
way that they become part of the interaction process.
The implication of this, then, is that there is an essentially dynamic aspect to
the use of accounting information which has not yet been articulated by
researchers: in effect, the way in which the "social" accounting set defines the
reality of the organisation through human action.

We have suggested in this paper that there has been a lack of awareness on the
part of accounting researchers of the wider implications of contextual
frameworks. In particular, there has been a tendency to start from accounting
and "move outwards" just far enough to claim to be able to understand some
contextual links of cause-and-effect. We are suggesting here that this is
essentially approaching the problem from the wrong direction: that the problem
is not so much a justification of the accountant's role and an enhancement of
his understanding of the context in which he works (so as to make some kind
of marginal improvement in his work), but as an understanding of the total
organisational and societal problems of which the accounting systems is a part.
In other words, we should move inwards to develop ideas for redesign rather
than taking the myopic view that accounting is a kind of essential force in
itself which can be improved by incremental but limited jumps.
In conclusion, in considering the accounting system's role for negotiation of
sociological variables in the organisation, our concern is not with the
accounting system per se but with the total organisational system as an
institution of society.
Our ultimate aim in systems design must not be to
create some kind of optimum accounting system in itself, but rather one which
develops criteria within the context in which it operates. It must never be
forgotten that the sole purpose for an accounting system is its usefulness to its
users: and hence the research should concentrate on understanding those users in
a holistic sense with the accounting system as a significant but limited variable
rather than any kind of converse. No doubt the intention of accounting
researchers often may have been as we describe, but unless the explicit
methodology is reconsidered, we cannot expect the output of the research to be
similarly desirable.
Abcll, P. (1971) Model Building in Sociology. Weidenfeld and Nicholson.
Aglietta, M. (1979) A Theory of Economic Regulation. New Left Books.
Allen, V. (1975) Social Analysis. Longman.
American Accounting Association (1978) Report of the Committee on Human
Information Processing. Committee Reports Volume 1978, 2.

Berger P. L. and Luckman T. The Social Construction of Reality. Penguin.
Cicourel, A. V. (1970) "Basic and Normative Rules in the Negotiation of Status and
Role" in Recent Sociology No 2 ed H. P. Dreitzel. Macmillan.
Collins, F. (1978) "The Interaction of Budget Characteristics and Personality Variables
with Budgetary Response Attitudes", Accounting Review April.
Durkheim, E. (1927) Les R'egles de la M├ęthode Sociologique (eighth Ed) (Alcan)
quoted by Homans op cit.
Filmer, P., Phillipson M., Silverman, D. & Walsh, D. (1972) New Directions in
Sociological Theory (Collier-MacMillan).
Goode, W.J. (1960) "Norm Commitment and Conformity to Role-Status Obligations",
American Journal of Sociology LXVI Nov.
Gordon, L.A. and Miller, D. (1976) "A Contingency Framework for the Design of
Accounting Information Systems" Accounting, Organisations and Society I, 1.
Gouldner, A. (1973) For Sociology Allen Lane, the Penguin Press.
Hofstede, G.H. (1967) The Game of Budget Control, Tavistock.
Homans, G.C. (1964) "Bringing Men Back In" in The Philosophy of Social
Explanation ed A. Ryan Oxford.
Hopwood, A.G. (1974) Accounting and Human Behaviour, Accountancy Age.
Kuhn. (1962) The Structure of Scientific Revolutions, Chicago.
(1970) "Reflections on my Critics" in Lakatos and Musgrave (eds) Criticism and the
Growth of Knowledge, CUP.
Libby, R. and Lewis, B.L. (1977) "Human Information Processing Research in
Accounting: the State of Art" Accounting, Organisation and Society, 2, 3.
Lowe, E.A. and Shaw, R.W. (1968) "An Analysis of Managerial Biasing: Evidence
from a Firms Budgeting Process", Journal of Mangement Studies, September.
(1970) 'The Accuracy of Short Term Business Forecasting: An Analysis of a
Firm's Sales Budget", Journal of Industrial Economics, Summer.

Maclntyre, A. (1967) "The Idea of a Social Science" in The Philosophy of Social
Explanation ed. A. Ryan, Oxford.
Merleau-Ponty, M. (1967) "What is Phenomenology?" in J. Kockelmans (ed)
Phenomenology, Anchor.
Mouzelis, N.P. (1975) Organisation and Bureaucracy, RKP.
Radcliffe-Brown, A.R. (1952) Structure and Function in Primitive Society, Cohen and
Rockness, H.O. (1977) "Expectancy Theory in a Budgetary Setting: An Experimental
Examination" Accounting Review, Oct.
Ruland, W. (1978) "The Accuracy of Forecasts by Management and by Financial
Analysts", Accounting Review, Apr.
Ryan, A. (1970) The philosophy of the Social Sciences, Macmillan.
Silverman, D. (1970) The Theory of Organisations, HEB.
Tinker, A.M. (1977) "Predictive Ability and the Poverty of Empiricism" Paper
presented to seminar at University of Manchester, March.
Tumer, A.N. and Lawrence, P.R. (1965) Industrial Jobs and the Worker, Harvard.
Wagner, H. (1964) "Displacement of Scope: A Problem of the Relationship Between
Small-Scale and Large-Scale Sociological Theories" American Journal of Sociology,
Worsley, P and others (1970) Introducing Sociology, Penguin.
Wrong, D. (1967) "The Over-Socialised Concept of Man" in System, Change and
Conflict ed N.J. Demerath and R.A. Peterson, Free Press.