The Web Site for Critical Realism The WSCR Archive

[Although the following discussion paper has been edited in a few spots in order to make the format and punctuation of the notes more internally consistent as well as to add a bit of missing punctuation, it is otherwise a mirror of the original, which can be found on the author's Web site. The paper is archived here with the permission of the author, who retains the copyright.]

Critical Realism and Information Systems: An Exploration


Most writing on information technology draws on an 'objectivist' or determinist tradition which identifies technology as a major, if not the most important, force in transforming economic and social organization. Furthermore, it does not adequately theorise the specificity of the information dimension of I.C.T. [Information Communication Technology]; thereby failing to differentiate it from general technological change (Coombs, Knights and Willmott, 1992: 51-52).

The central argument of this paper is that attempts to 'adequately theorise the specificity of the information dimension' have failed to tackle the central problem of the agency/structure couplet that is a constant preoccupation of social theorising. The paper also seeks to explore some of the issues in application of that form of theorising known as 'critical realism', here coupled with the tradition of historical materialism. The paper discusses a number of approaches to information systems theory with a view to outlining, within the constraints allowed, some key weaknesses. It then looks at responses to the question of information from within the Marxist tradition, concluding that these take an over-deterministic perspective. A potential alternative, drawing principally on the works of Archer and Callinicos is outlined. However, before proceeding to the body of the paper, we need to consider briefly what we mean by 'critical realism'.

In many ways the use of the word 'critical' is fraught with difficulties. In some senses it is redundant, as all good social theorising is in some sense 'critical'. The use of the word in this context is taken to assume a position which stands outside the existing arrangement of things and seeks to suggest alternatives. However, in this it runs the risk of being aligned with Critical Theory, that body of thought associated with the work of, in particular, Marcuse. However, critical realism stands sharply opposed to such traditions in insisting on a realist ontology which posits the existence of elements of the social world which exist despite or irregardless of our current state of knowledge. In this, it builds principally on the work of Roy Bhaskar and his investigations in the philosophy of science (Collier, 1994; Norris, 1995). This body of ideas has been built upon by scholars in many fields and is beginning to make an appearance in the field of organizational studies (Reed, 1997). The most developed example of the application of critical realism to social theory is in the work of Margaret Archer, and this paper will draw heavily upon this (Archer, 1995). A central tenet of the approach taken by Archer is 'analytical dualism', the need to treat structure and agency as analytically distinct categories in order to examine the inter-relationships between the two. It is necessary, therefore, to look at the ways in which writers on information systems have treated this issue. The contention will be that they have tended, in practice, to focus on agency and that this leaves the impact of structure vague and ill-defined.

One group of writers might be termed the 'critical mainstream'. I include in this grouping writers such as Zuboff, Kling and Orlikowski (Zuboff, 1988; Kling, 1987; Orlikowski, 1996). These writers have questioned the 'objectivist' tradition by arguing that information systems are crucially social systems and they have provided a wealth of case material illustrating how the simple application of technology is insufficient to bring about significant change. In particular, Zuboff's coining of the term 'informate' has been influential in bringing issues about information and its use onto the agenda. In much of their work the structural context, notably hierarchies within organizations, plays a clear role in moulding the use of information, but this structural dimension remains poorly defined. It is the 'formative context' in Orlikowski, the 'web' of social relations in Kling, all hinting at forces beyond the organisation itself but failing to specify them. This leads to their work often having a slightly unreal, detached feel about it, exemplified by this from Zuboff:

This fails to tackle the very real pressures that organisations are under from the broader system of economic relations, articulated through a series of markets. This failure to consider broader structures is also found in the work of those whom I will term 'critical theorists', some of whom emerge out of a theoretical tradition, Marxism, which is held by many to over-emphasise structural dimensions. We can see in some of these writers a desire to escape from these supposed weaknesses and to seek support from other theorists who, they would argue, offer a more nuanced account.

One such theorist, whose emphasis on the disciplinary power of the knowledge/power couplet seems attractive in the context of information systems, is Foucault. The problem here is that the agency/structure divide is dissolved into the disciplinary power of vague concepts of knowledge which, wrenched out of their historical context, bear down upon human agents, shaping their destiny almost beyond their will. Ironically a theory which emanates from the extremes of structuralism ends up by removing any distinction between different classes of agents. We are reduced, in the words of Porter's provocative and compelling piece, to the struggle of 'all against all' (Porter, 1996). This comes out clearly in the work of Knights and Murray on the development of information systems in the financial services sector (Knights and Murray, 1994). Here the very real pressures of the market are subordinated to the social construction of that market by managers struggling against each other for career advantage: "it is clear", they argue, "that particular distinctive views of the market are mobilised by different managers ... In part, this mobilisation is an attempt to legitimise specific divisional and career interests by couching them in projects which are claimed to be coincident with the corporate objectives" (Knights and Murray, 1994: 187). There is little doubt that this is part of the reality of information systems in an organisation, but the end product tends to be a focus on individual struggles which gives us very little analytical purchase. The problem is that Knights and Murray end up facing both ways at the same time, with the pious conclusion that, "It remains to be seen if organisations can move beyond the forms of conflictual user relations generated in the companies we studied" (252), but no indication of how this might happen. Given their analysis, in which all the pressures surrounding organisations and individuals turn out to be socially constructed, with no prioritisation of casual weight and a focus on individual career aspirations, it seems difficult to imagine change occurring.

The motor of change in another approach, that espoused by Alvesson and Willmott, is the knowledgeable individual (Alvesson and Willmott, 1996). The treatment of information systems by these authors is very limited, forming a small part of a larger work and sharing distinct weaknesses with regards to the wider IS literature with the work of Knights and Murray (Mutch, 1996). However, there are some useful concepts which could be developed from their application of the work of Habermas if we were to remove the excessive weight placed on the communicative ability of the individual expressed in such work. As Archer argues,

What we find, therefore, in the works considered so far, is that in practice the agency/structure dilemma is solved by an almost exclusive focus on the agency dimension. However, there is one more very important approach to the problem to consider, which is the notion of 'structuration', developed by Giddens and applied to the information systems domain in the work of Walsham (Walsham, 1993). Structuration attempts to overcome the problems of attributing causation to either agency or structure by conflating the two and treating social structure as something which is 'instantiated' through social practice. What this means is that we are precluded from an "examination of their interplay, of the effects of one upon the other and of any statement about their relative contribution to stability and change at any given time" (Archer, 1995: 14).

What we find, then, is that most existing accounts give too little weight to the impacts of structure on the use of information, an emphasis which in its turn paradoxically gives little attention to the impact of individual psychology on that use. Rather, we find a conflation round the concerns of social psychology, which

However, before going on to consider how we might integrate such insights into an approach to information systems, we need to consider the potential contribution from a body of work which might be considered to redress the balance in favour of structural explanations, that of historical materialism. What we find is a very patchy and unsatisfactory treatment. One area of weakness in such Marxist accounts as exist is in their definition and treatment of the concept of information. The word tends to be used interchangeably with both data and knowledge and the focus shifts to the control and use of the technology which produces it. Witheford, following Negri, attempts to distinguish between information and communication (Witheford, 1994: 102-3). The former is equated to dead labour expropriated and abstracted to become a prison for living labour whose emancipation, leaning on Habermas, is seen as resting on communication. However, Witheford notes that the distinction is a difficult one to sustain, although this caveat seems to be based on the communicational opportunities provided by technology. However, this does not tackle the difficulty posed by seeing information as a thing rather than as a relationship between knowledge and data. Thus, the capture of knowledge from a worker and its transformation into data does not necessarily that the worker's stock of information is diminished. Information, unlike other resources, is not a static entity. It is not consumed on use but can grow, and grow in a dynamic fashion, dependent on how it is received (Eaton and Bawden, 1991). This feature is not one recognised in many discussions of information technology, which confuse information with the raw data on which it is based, a confusion which can be traced back to the founding theories of information science. The work of Shannon and Weaver was concerned with, ironically, communication, and sought to break information down into its component parts for secure and accurate transmission (Goontilake, 1991). It has been criticised for a failure to focus on meaning and has given rise to an emphasis on the production of data abstracted from its context.

However Carchedi would go still further. He argues that computers have their roots in "the mechanical and formalized way of reasoning" inherent in the production of knowledge under capitalism (Carchedi, 1989: 446). This, he argues, is because computers are machines, they necessarily have a "mechanised and formalized way of reasoning [which] is inherent in the computer and reveals the capitalist nature of this machine" (Carchedi, 1989: 444). This form of reasoning then determines the associated labour: "the computer forces people to reason (to use their minds) in a certain way and only in that way" (Carchedi, 1989: 444). This parallels an argument made by Melucci:

Carchedi goes further to extend the process to communication itself. "For the first time in history," he argues, communication must take place through a machine and through the language of the machine, something which implies a great mechanization, impoverishment and standardization of language. communication and social contacts" (Carchedi, 1989: 445). Over and above the way such machines are used, therefore, they represent capitalism in the very nature of their being. "It is," argues Carchedi, "because of these primary features that the computer and the mental labour inextricably associated with it are ultimately incompatible with a socialist society" (Carchedi, 1989: 446-7). We need to review these statements, for it seems to me that they rest on a series of misconceptions and logical fallacies.

One must say that it is possible to find support for the type of reasoning that Carchedi deplores. There is certainly considerable support for the notion of information as both a thing and moreover a thing heavily quantitative in nature in the management literature, influenced as it is by a tradition of 'management science' (Locke, 1989). Thus we can find in Simon the contention that "Nowadays, with computers everywhere, we can think of information as something almost tangible: strings of symbols which, like strips of steel or plastic ribbons, can be processed- changed from one form to another. We can think of white-collar organization as factories for processing information" (Simon, 1976: 45). Such arguments lay at the root of much of the work on artificial intelligence. However, the relative failure of that work should cause us to take stock. Whilst work in this area has produced some apparently spectacular triumphs in areas such as chess, in truth what it reveals is the limitations of computer reasoning. It also indicates just how far away we are from having an effective theory of the mind (Dennett, 1993). The failures of AI have turned some of its advocates into a closer examination of how meaning is constructed in everyday life (Winograd and Flores, 1986). We need to consider then how the mental labour associated with computers intersects with such daily communication. One would wish to make three points here. One is that there is a danger in arguing that because the laws which govern the working of certain types of machines resemble certain types of thinking which one might argue of characteristic of at least part of a mode of production that these laws are reducible one to the other. A stronger argument might be that the rules governing machine behaviour have an influence on other modes of thought, in the way that a fascination with science had a powerful influence on modern art (one thinks of Le Corbusier's 'machines for living'). However, we see the emergence of computing in the mathematical thought of Turing and its early application in military contexts. Now, to be sure, both are occurring within capitalism, but the direct connection to an economic mode is tenuous. Secondly, the argument is to ignore the layered nature of computers. That is, at the basic, machine level, we have nothing more than pure on/off states. The reasoning that emerges from this cannot be reduced to these states, as we ascend through ever more complex layers, not just from assembler to high level languages, but increasingly to much richer media. Simply because pictures, video, text and hyperlinks are all stored and manipulated using a very mechanical form of reasoning does not of itself impose that form of reasoning at the level of the interface.

The final argument is about the impact of the final product on the form of reasoning employed. Now, it is certainly true that there are impacts. One is certainly a tendency to view all data as being quantitative in nature. However, one has to question whether this is the independent determining impact of the machine, or whether it is not a reinforcing mechanism for tendencies that are already present. Does it not make the obsession with measuring and quantification much easier to chase? For the notion that computer force people to think in a "certain way and only in that way" is removing any capacity for human response. That human response is grounded in a set of meanings which are produced and reproduced in daily life. The products of computers enter into this circuit of production and reproduction as an important element, but not as an unchallenged one. Such products will there clash with other ways of seeing and it is these contradictions that we need to examine. A theoretical approach which seals off the possibility of examining such clashes by deciding a priori what the results of such investigation will be is of little help. Rather, the notion of information as a process involving an approach to data and a link with knowledge could be considered more useful. Boland focuses on the 'in' of information to argue that information is neither a 'thing' nor is it 'power', both valuable correctives to fashionable notions. "Information," he argues, "is not a resource to be stockpiled as one more factor of production. It is meaning, and can only be achieved through dialogue in a human community. Information is not a commodity. It is a skilled human accomplishment" (Boland, 1987: 377).

Now, there is a danger with such definitions, a danger that they exclude both the structural constraints on the production of such meaning and the human limitations which prevent the full accomplishment of this achievement. It is in this area that Archer suggests that we have been too broad in our definitions of agency. We tend to use it as a portmanteau term rather than, she argues, recognising the greater value of a tripartite division into persons, agents and actors. Persons are individual human beings, emergent from, but not reducible to their biological make up. A focus here enables us to attempt to integrate the insights being gleaned by those working on materialist concepts of consciousness and language which emphasise their basis in the biological structures of the brain with the work done in cognitive psychology (Dennett, 1993; Pinker, 1995; Demasio, 1995; Huber, 1991). However, as Archer argues, there are other aspects of human agency which have properties beyond a simple aggregation of individual capacities. Agents are, she argues always collective categories which can play passive or active roles and which condition the life chances of those who belong to them. Finally there is the category of actors, emergent from collective agents and enabled or constrained by the characteristics both of those groups and of the structured situations in which they find themselves. Here, therefore, might be a useful way of looking at individual reactions to information, allowing us to isolate those constraints of individual psychology from the reactions of groups and the strivings of individual actors. Interestingly, Alvesson and Willmott suggest a potentially useful model based on the work of Forester on communications. This model takes account of two dimensions: the inevitability or otherwise of distortion and the source of the distortion, be that individual or structural. This gives the following matrix:




Cognitive limits

Division of labour


Interpersonal manipulation

Structural legitimation

For Alvesson and Willmott, the first box is of little interest to critical theorists: but might one suggest that it is, in fact, of central importance in understanding how individuals respond to information? For the further the boundaries of this box extend, the less possible is it that adjustments in other areas can affect it. Therefore, it behoves us to treat individual psychology as a serious variable. Whilst such an approach might seem inimical to the claims of classical Marxism, the work of Callinicos suggests that this is far from the case. A robust, non-teleological Marxism must, he argues, recognise that "the intentional activity of human beings is not structurally determined" (Callinicos, 1989: 82). However this in no way prevents the appropriate consideration of structural factors: it is just that now we have a better concept of the relationships between agency and structure. This brings us to a further critical point derived from Archer and that is of the importance of the temporal dimension to analysis. If structure is developed over time as the result, intended or otherwise of the actions of human agents, then it follows that no account which lacks a historical dimension will be able to do justice to the structures within which agents now find themselves (Pollert, 1996).

This is a necessarily very brief and tentative exposition given the constraints available. It would appear that our current theories of the use of information in organisations are lacking in two areas: they tend to neglect issues of individual psychology and they underplay the structural constraints within which individual actors create meaning. It is suggested that a development of the ideas presented by Archer, in particular the tripartite nature of agency, the inter-relationship between structure and agency and the need to incorporate the temporal dimension, offer useful ways forward. Such a development might help to integrate the insights from the widely disparate set of sources used here. There are many insights scattered across a wide body of literature based on research traditions which often fail to recognise the existence of other traditions. However, what remains is to develop the ideas presented here both in a theoretical and concrete sense.

Works Cited

Alvesson Mats and Willmott Hugh (1996) Making Sense of Management. A Critical Introduction Sage London

Archer Margaret (1995) Realist social theory: the morphogenetic approach Cambridge University Press Cambridge

Boland R. J.and Hirschheim R. A. (1987) Critical Issues in Information Systems Research John Wiley Chichester

Boland Richard J. (1987) 'The In-formation of Information Systems' in Boland and Hirschheim pp363-380

Callinicos Alex (1987) Making History Polity Cambridge

Carchedi Gugliemo (1989) 'Between Class Analysis and Organization Theory: Mental Labour' in Stewart Clegg (ed)

Ciborra Claudio (ed) (1996) Groupware and Teamwork. Invisible Aid or Technical Hindrance? Wiley Chichester

Stewart Clegg (ed) (1989) Organization theory and class analysis: new approaches and new issues De Gruyter Berlin

Collier Andrew (1994) Citical Realism Verso London

Coombs Rod, Knights David and Wilmott Hugh(1992) Culture, Control and Competition; Towards a Conceptual Framework for the Study of Information Technology in Organizations Organization Studies 13/1 pp51-72

Damasio Antonio R. (1995) Descartes' Error. Emotion, Reason and the Human Brain Picador London

Dennett Daniel C. (1993) Consciousness Explained Penguin Harmondsworth

Eaton J.J. and Bawden D.(1991) What Kind of Resource is Information? International Journal of Information Management 11 pp156-165

Goontilake Susantha (1991) The Evolution of Information: Lineages in Gene, Culture and Artefact Pinter London

Huber George P. (1991) Organizational Learning: the Contributing Processes and the Literatures Organization Science 2(1)

Kling Rob (1987) 'Defining the Boundaries of Computing Across Complex Organisations' in Boland and Hirschheim pp307-362

Knights David and Murray Fergus(1994) Managers Divided Wiley Chichester

Locke Robert R. (1989) Management and higher education since 1940: the influence of America and Japan on West Germany, Great Britain and France Cambridge University Press Cambridge

Melucci Alberto (1996) Challenging codes: collective action in the information age Cambridge University Press Cambridge

Mutch Alistair (1996) Marxism, Managers and Information Paper at 14th International Labour Process Conference

Mutch Alistair (1997) Information Literacy: An Exploration International Journal of Information Management 17(5) pp377-386

Norris Christopher (1995) Truth, Science and the Growth of Knowledge New Left Review 210 pp105-123

Orlikowski Wanda (1996) Evolving with Notes ppOrganizational Change around Groupware Technology in Ciborra(ed) pp. 23-60

Pinker Stephen (1995) The Language Instinct. The New Science of Language and Mind Penguin Harmondsworth

Pollert Anna (1996) Gender and Class Revisited; or, the Poverty of 'Patriarchy' Sociology 30(4) pp639-659

Porter Sam (1996) Contra-Foucault: Soldiers, Nurses and Power Sociology 30(1) pp59-78

Reed Michael I. (1997) In Praise of Duality and Dualism: Rethinking Agency and Structure in Organizational Analysis Organizational Studies 18(1) pp21-42

Walsham Geoff (1993) Interpreting Information Systems in Organisations Wiley Chichester

Winograd Terry and Flores Fernando(1986) Understanding Computers and Cognition. A New Foundation for Design Addison-Wesley Reading

Witheford Nick (1994) Autonomist Marxism and the Information Society Capital and Class 52 pp85-125

Zuboff Shoshana (1988) In the Age of the Smart Machine. The Future of Work and Power Heinemann London

Copyright © 1999 Alistair Mutch

The Web Site for Critical Realism The WSCR Archive