The problems with extant theory suggest some directions for the development of a robust portfolio-level theory of IS as it relates to organisational and societal structures. The finding that IS innovations change industry structures rather than entrench competitive advantages (Clemons and Row, 1988; Kettinger et al., 1995) is one possible starting point. A good IS theory (in contradistinction to competitive advantage theory, which was predominantly business-oriented) would deal with a range of social phenomena that so far lack a broad analytical explanation. Those phenomena include the increasing encroachment of standardised IS structures on social behaviour through the implementation of standardised data and process definitions in a range of systems. As standard IS structures become more widespread, so commercial and government organisations come to look more and more alike, at least in behavioural terms.

The agency-structure relationship has been a central concern in sociological theory for a long time. Are social structures ‘real’ when it is clear that they are constructions that must be affirmed by human agents acting with some degree of individual autonomy? Putative answers, all of interest, and all shedding light on complex social issues have come from theorists as diverse as Marx (1981), Giddens (1984), and Bourdieu (1980) among many others. But the point at issue here is that these theories do not deal with the impacts of structures reified in formal information systems. Such structures clearly allow for voluntarism in principle, as people may choose to ignore IS constraints, but they cannot then achieve their transactional goals. Yet active resistance to the influence of standardised structures clearly becomes more difficult the more widely adopted they are. IS structures are in this perspective more rigidly defined, and more formally constraining, than structures that depend on rules interpreted and enacted by people for their enforcement (Giddens, 1984).

An analysis of the possible social impacts of such structures would be the central concern for theory development in this area. While it is perhaps a little premature to identify the issues in advance of the theory, three possible areas of interest can be mentioned. These are, first, the likely lengthening of IS change cycles as the number of organisations dependent on the same standardised structures continues to increase. Second is the capacity for standardised structures to be used to create, intentionally or otherwise, people who are system ‘outsiders’ in some way (the history of Nazi Germany’s use of IBM technology is an extreme but nevertheless instructive illustration of what was already possible in this regard fifty years ago [Black, 2001]). Third is the change in social risk relationships – while the adoption of standard IS structures reduces the number of possible points of failure or breakdown, it simultaneously raises the stakes for any breakdown that does occur.

As the Y2K experience demonstrated, IS structures have considerable inertia once installed, making them highly resistant to change; much more so than structures maintained by human behaviour. The argument in this paper is that the possibility of developing an explanatorily powerful theory linking IS with organisational and societal structures is therefore a real one, that IS is the discipline best placed to develop such a theory, and that for the reasons discussed earlier, this would have a range of benefits for the discipline as a whole.