A number of different approaches have been taken to account for the nature of different disciplines and their development. A review of the literature identified that three of the approaches were largely considered independently of other approaches; these were theory, social processes and research methods and standards. More holistic approaches have, however, also been used to explain disciplinary nature and progress, by considering two or more of theory, social processes, research methods and standards, topics of knowledge, symbol sets for communication, the impact of local factors and the degree of professionalism. Literature that deals with theory, social processes and research methods and standards to explain the nature and development of disciplines will each be examined in turn below, followed by literature that takes a more combined approach.
Interestingly, some of the approaches to explain disciplinary development have waxed and waned in popularity at particular times in the past 50 years. In 1959, Popper argued that it was only through the generation of theories that scientific progress could occur. The importance of theory to a discipline has been recognised up to the present.
Kuhn (1970:182) used the word ‘paradigm’ in different ways, including ‘universally recognised scientific achievements that for a time provide model problems and solutions to a community of practitioners’. He argued that scientific progress arose as a result of new observation or experience that necessitated a ‘reconstruction of prior theory’ and resulted in a paradigm shift. ‘Normal science’ represented the body of theory, practice and methods of inquiry that were accepted by a group of researchers, typically expounded in textbooks of the discipline. Wernick and Hall (2004) analysed the textbooks of a discipline allied to IS, software engineering, to examine the underlying belief system of authors from that discipline, to find that it was pre-paradigmatic with a common core of knowledge supplemented by competing sets of beliefs. It has been claimed by many researchers that there is limited theory in IS (for example, Grant 1991; Keen 1991; Paul 2002), while few textbooks have been published that provide an overview of the discipline. Hirschheim and Klein (2003), however, saw a ‘generalisation deficit’ (p. 257) in IS, rather than a lack of ‘theoretical knowledge’ (p. 268). Kuhn saw the presence of paradigms as a sign of maturity in a discipline, as they gave researchers a basis for choosing problems as well as guiding them in their investigation. Despite frequent discussion of Kuhn’s work in articles about its development, IS has also been classified as pre-paradigmatic (Culnan 1987; Seddon 1991). Consequently, from this perspective it appears that IS might not have achieved the state of normal science—at least not by 1991.
Kuhn’s (1970) analysis of physical optics before the time of Newton is illuminative. As that discipline had ‘no common body of belief…each writer…[built] his field anew from its foundations…[and] there was no standard set of methods or of phenomena’. Kuhn saw early fact gathering that was not guided by some ‘theoretical and methodological belief that permits selection, evaluation and criticism’ as a ‘nearly random activity’. He saw the result of undirected research as a morass of ‘mere facts’ that was too complex to be integrated with theory. Senior IS researchers have recognised the problem in IS for decades. For example, ‘We seem to randomly generate research projects with the outcome that we have a scattering of results which presents a severe problem of pattern recognition’ (Dickson et al. 1982). More recently, it has been claimed that IS is characterised by the problems it studies more than by a body of knowledge, or theories (Paul 2002). It could be that some IS research is guided more by the ease with which data can be gathered rather than by other criteria. It appears that, from Kuhn’s perspective, IS could be at a very early stage in the development of a discipline.
Elias (1982) referred to the nature of theory in traditional physics as ‘law-like theories’ rather than ‘process theories’, which could be more appropriate in other sciences. The ideas regarding theory types might be interesting ones to apply to IS, even though not all researchers in the area—including Fitzgerald (2003)—would refer to the discipline as a science. The development of theory in IS has been acknowledged as difficult (Fitzgerald 1993; King 1994; Paul 2002; Fitzgerald 2003), despite a well-known proponent of the philosophy of technology, Bunge, seeing technology as ‘philosophically productive’ (Ihde 2004:120), with technological systems putting ‘forth…philosophically significant theories’ (Bunge 1979:172). Just one part of the difficulty could be that the origin of IS in technology and its past link with computer science has given those within and outside the discipline an expectation that law-like theories are appropriate for it. While the nature of IS stresses organisational issues rather than technical ones (Avison and Fitzgerald 1991; Galliers 1992; Hirschheim 1992; Fitzgerald 2003), the more recent emphasis on ‘interpretivism’ has not produced more general theory than the positivist approach (Hirschheim and Klein 2003). Another researcher (Zahedi 2004) has proposed a list of theories for IS, but argued for their greater consolidation. Fitzgerald (2003:226), when discussing theory in IS, distinguished between rules (or laws), evidenced guidelines and normative guidelines, where the last is ‘an interpreted view of something a practitioner developer might consider doing, under appropriate circumstances, but…[that] would not necessarily lead to success’. Fitzgerald held that only the last kind of theory was possible in IS. Furthermore, the location of management information systems (MIS) in business schools in the United States and, to a degree, in Australia, also suggests that law-like theories and scientific method might not be the only, or even the most appropriate, approach for IS. So a mismatch between expectation and achievement as well as the complexity of process theories could account in part for the limited production of theory in IS. Regardless of what kind of theory is produced in IS, or what it is called, there is support for its development (Paul 2002; Fitzgerald 2003).
A decade after Popper referred to the role of theory in scientific progress, a very different view of the development of disciplines emerged. At that time, the role of social conditions in the production and assessment of scientific knowledge was recognised (Whitley 1984a).
Even though Kuhn’s views have been referred to in previous discussion on the contribution of theory to the development of a discipline, he is associated more with a different approach. In 1970, at a time of burgeoning science and higher education sectors in many Western nations, Kuhn published a seminal analysis of the social process of science, which is still referred to in the IS literature and elsewhere decades later (for example, Wernick and Hall 2004). Kuhn’s treatise, The Structure of Scientific Revolutions, influenced the change in attitude to science and the nature of the development of disciplines. Kuhn emphasised the social mechanisms that created a scientific discipline (Ariav et al. 1987), such as conferences, journals and academic departments, which have also been referred to as ‘mechanisms of control’.
Hirschheim and Klein (2003) saw the control of rewards and punishments by academics from other disciplines as driving IS research to become more theoretical and less applied. Although such pressure could increase the acceptance of IS as a discipline, Hirschheim and Klein considered that a less applied orientation reduced the relevance of IS to practitioners, and therefore its viability.
A link between knowledge and power was proposed by Foucault (1977), who also recognised the significance of social issues on the nature and development of a discipline. Foucault would see the status of IS as a political issue, rather than the achievement of ontological or epistemological positions. The supporters of this view see IS as becoming a discipline only once sufficient ‘status has been conferred by institutional practices…[including] the ability to form departments, appoint chairs, organise conferences [and] edit journals’ (Introna 2003), or the achievement of mechanisms of control. The nature of academic leadership is another way that social issues can impact on the direction of a discipline and its perceived status (F. Land, personal communication, 23 January 2006). It can be seen that one way to evaluate the status of IS as a discipline would be to examine whether it had the mechanisms of control normally associated with a discipline.
Other researchers have considered research methods and standards in the development of a discipline. This work has led some researchers to examine the relationship between disciplines, particularly reference disciplines. Elias (1982) argued that ‘high-status sciences’ retained their position by imposing their methods on other sciences. He considered it inappropriate for scientific method to be imposed on newly emerging sciences, particularly as it was developed for, and by, other disciplines. As the emphasis of one discipline could be on physical objects while in another discipline it could be on organisational issues, the scientific methods of some disciplines could be irrelevant to other disciplines.
Different disciplines rank more highly than others in the public and academic mind. As each has its own ideology and values that colour the knowledge it produces, Elias (1982) saw interdisciplinary collaboration as ‘exceedingly difficult and almost impossible in many cases’. He argued that only low-status disciplines would take heed of interdisciplinary criticism. Moreover, Elias contended that modelling a low-status discipline on a high-status discipline or its characteristics, in an attempt to gain kudos for the field or researcher, usually failed. If this last proposition is true it could have ramifications for IS because of its close relationship with its reference disciplines. The use of theory and research approaches from reference disciplines could reduce the viability of IS as a discipline while it attempts to improve the rigour of its research. Furthermore, although it has been lamented that the IS literature is not read by those in its reference disciplines (Keen 1991), this characteristic could be a typical of any discipline.
The concept of ‘restricted and unrestricted science’ (Rip 1982) is relevant to an analysis of the nature of IS. In restricted sciences there is considerable control of the ‘knowledge object’, which allows a researcher to tightly restrict the behaviour of the object being studied, whereas in unrestricted science the reverse is true (Rip 1982). Rip argued that the high status of restricted sciences encouraged researchers from unrestricted sciences to become more like a restricted science by importing restrictedness. Signs of a restricted science include use of sophisticated instruments, standardised procedures and empirical generalisations that give increased credibility and allow research assistants or research students to undertake routine work.
There is evidence that points to IS as being unrestricted, despite technical IS research appearing more restricted. Criticism of IS research approaches indicated that many of the instruments that had been used were not sophisticated, research procedures were far from standardised (Straub and Carlson 1989; Boudreau et al. 2001) and a wide range of approaches from reference disciplines were considered appropriate (Ahituv and Neumann 1986; Culnan and Swanson 1986).
Although greater standardisation of some research procedures has been seen more recently—at a time when it has been suggested that IS is now itself a reference discipline (Baskerville and Myers 2002)—other researchers have pointed to the difficulty in reaching consensus on the most appropriate methods for IS (for example, Hirschheim and Klein 2003), and challenges associated with the diversity of methods in use (Frank 2006). Even though the view of the development of IS by practitioners is an external perspective, and this chapter has restricted itself to the internal view from IS academics, there will also be an internal perspective on the relevance and quality of IS teaching to future and current practitioners and other students.
Research education itself could be seen as a standardised procedure where it is specialised to a particular discipline. Relatively recent figures indicated that IS researchers in the United States had gained their highest degree across a broad range of disciplines (Walstrom et al. 1995), revealing that the research training process for IS academics trained before 1995 in that country was far from standardised. It appears that a similar diversity was found in the education of Australian IS researchers until 1996, but that this diversity has narrowed since then. An anonymous reviewer of this chapter contended that the diverse educational backgrounds of the ‘fathers of the field’ of IS are significant when examining the current nature of IS.
In addition to theory, social issues and research methods and standards, other approaches were identified in a review of the literature as having been proposed as contributing to an understanding of the nature and development of a discipline. These were more joint approaches, which combined two or more other approaches.
There is majority support for examining the body of knowledge along with the social processes, when considering the development of a discipline (Becher 1994), as ‘we cannot…artificially separate the…substantive content from…social behaviour’ (de Solla Price 1970). Becher (1987) examined the nature of three different disciplines by examining their tacit knowledge (which derived from the body of knowledge) along with their linguistic behaviour (a social process). Like Tardif (1989), Paul (2002) and Hirschheim and Klein (2003), Fitzgerald (2003) considered that a discipline required a core body of knowledge. Fitzgerald and Paul, however, saw a body of knowledge as being more than an agreed set of topics; it included the set of laws, rules or evidenced guidelines—that is, theory. Fitzgerald (2003:226) postulated that IS had ‘the trappings of a discipline…[such as] mechanisms of control [which are social processes], but without the core body of knowledge or agreed theory’. Consequently, Fitzgerald viewed IS as a ‘perspective’ rather than a discipline.
Hirschheim and Klein (2003) saw the IS body of knowledge as incorporating some social processes and theory, when they proposed that its four components were technical, theoretical, ethical and applicative knowledge. Hirschheim and Klein saw the development of applicative knowledge, which ‘required…[the application of] theoretical knowledge to specific circumstances’ (p. 266), as being necessary to reach understanding and consensus in IS. They considered that the limited extent of applicative knowledge in IS threatened the viability of the discipline. The discourse needed to develop the body of knowledge for IS, particularly applicative knowledge, would increase communication and the relevance of IS—internally to the academic discipline and externally. In later work, Klein and Hirschheim (2006) saw IS at risk because it was made up internally of several ‘communities of practice and knowing’ (CoP&K), each with different values and legitimacy criteria, and because there was limited connection from academic IS’ internal CoP&K to the outside. Discourse would, however, benefit the development of the discipline by helping to derive a common language across groups impacted on by IS, reducing the state of fragmentation in IS and overcoming its significant communication gaps (Hirschheim and Klein 2003).
Shinn (1982) considered concepts that related social processes, research procedures and theory development when he examined the intellectual and social structure of a range of disciplines, in particular looking at the intellectual division of labour. He found some disciplines to be highly formalised, with a dichotomy between the gathering and collation of findings on the one hand, and experimentation, theory and hypothesis on the other. One would expect that the more restricted a science (Rip 1982), the more formalised were its intellectual and social structures. As expected from its degree of restrictedness, IS is not as formalised in this way as are some other disciplines. Where interpretivist research is undertaken in IS, the data gathering and the theory building could be interleaved. Even in IS positivist investigations it is likely that the chief investigator/s will be involved in all stages of the process. The latter characteristic is dependent partly on the limited success of IS in attracting research funds (Ridley et al. 1998), which relates in turn to the perceived status of the discipline. Consequently, relatively few academic IS researchers have funding to employ assistants to carry out some of the research tasks.
The work of Chubin and Connolly (1982) allows further understanding of the combined pressures that have acted on the IS discipline. The authors argued that ‘research trails’ became institutionalised by offering potential rewards such as legitimacy and access to resources. On the other hand, ‘tinkering’ with new ideas or novel developments is normally opposed. Research trails are likely to use the epistemologies, research strategies, theory and perhaps even the topics of existing established research of reference disciplines. In IS, those who follow the existing research trails that were established by the reference disciplines might be more likely to be rewarded with tenure, promotion and access to research grants, as rigour is easier to demonstrate. Efforts to establish appropriate independent research approaches and traditions for IS could be seen as tinkering, as it is more difficult to claim that work is rigorous if it does not follow established traditions. Elias (1982) has, however, argued that greater independence of a discipline leads to its development and, it is assumed, eventual research rewards. It is possible that the path to the development of IS could involve breaking with some traditions established by the reference disciplines.
Whitley’s theory of scientific change (1984b), which viewed disciplinary development as a social process in combination with other approaches, has been applied to many disciplines, including IS (Banville and Landry 1992; Checkland and Holwell 1998). Whitley categorised some sciences as highly professionalised, with high task certainty, routinisation of activities and division of labour—a categorisation that echoed the work of Biglan (1973), Kolb (1981), Rip (1982) and Shinn (1982). Kuhn’s (1970) ‘normal science’ fell into this category; however, other sciences were not highly professionalised, with high task uncertainty, decentralised control of work process and limited routinisation of tasks. Where a discipline was not highly professionalised, local contingencies had a high impact, such as the influence of local political pressure. Ruscio (1987) also found that local factors resulted in substantial variations among universities for the same discipline. Non-professionalised disciplines could account for Ruscio’s finding.
In disciplines that are not highly professionalised, researchers investigate disparate problems that are likely to vary in nature and approach to those of concern to practitioners. Researchers work in flat non-hierarchical groups, or independently, rather than in highly structured teams with a clear division of tasks. Information systems appears to fit the mould of a discipline that is not highly professionalised. Furthermore, if local contingencies are likely to have a high impact on IS, it would be expected that considerable variation would exist in the nature of IS research between different universities and regions. There is evidence of considerable variation in the nature of IS research between nations, IS curricula and IS research education.
Whitley (1984b) suggested that three conditions needed to exist for the establishment of distinct scientific fields. These were the need for
scientific reputations to become socially prestigious and to ‘control critical rewards’
the establishment of standards of research competence and skills
a unique symbol system to allow exclusion of outsiders and unambiguous communication between initiates within the field.
Whitley’s first condition is a social process and relates to mechanisms of control. Scientific reputations are established, and critical rewards are obtained, through publication records and success at attracting research funding (Mingers and Stowell 1997. There were more IS publication outlets available in 1995 than in 1980 (Cule and Senn 1995), and many more in recent years (Hirschheim and Klein 2003). In general, these publication outlets are now administered by fellow IS researchers. Access to funding, however, remains tenuous while external funding decisions are made by individuals outside the discipline—as has happened with the allocation of Australian Research Council (ARC) grants (Ridley 1997) during much of the development of the IS discipline in Australia.
Whitley’s second condition—the establishment of research skills (and standards)—appears to be one component of a core body of knowledge, just as in the preceding discussion it has been seen that theory is also a component of the body of knowledge. Continuing debates regarding the quality of IS research and appropriate epistemologies and methodologies (Benbasat and Weber 1996; Boudreau et al. 2001) are signs that activity is taking place regarding Whitley’s second criterion for the establishment of a field, but that it has not been resolved. More recently, however, there have been some signs of increasing consensus.
Whitley’s third condition—the existence of a unique symbol set—appears to be another component of a core body of knowledge. Whitley’s third criterion for the establishment of a field is hard to meet as long as reference disciplines remain important to IS. Because so many reference disciplines inform IS research (Walstrom et al. 1995; Baskerville and Myers 2002) and the symbolic systems of each vary and compete, a dedicated and accepted IS symbol system has yet to emerge. The Framework of Information System Concepts (FRISCO) report that was produced in 1996 to clarify important IS definitions (Verrijn-Stuart 2001) is one demonstration of attempts to satisfy the third criterion.
Interestingly, however, two other components that were identified in earlier discussion of the literature as contributing to a discipline’s body of knowledge fall outside Whitley’s three conditions for the establishment of a distinct scientific field. Theory (or laws, rules and evidenced guidelines) is not included within Whitley’s conditions for the development of a distinct field; nor is an agreed set of topics.
In earlier work, Whitley (1984a) conceptualised seven stable categories to classify variations in the degree of mutual dependence between researchers of a field as against variations in the degree of task uncertainty. These seven categories can be used to differentiate the nature of one discipline from another. The categories have been applied to IS by researchers for two decades.
Researcher mutual dependency was defined as ‘dependence upon particular groups of colleagues to make competent contributions to collective intellectual goals’, while task uncertainty referred to ‘the extent to which work techniques are well understood and produce reliable results’ (Whitley 1984a:781). Where task uncertainty is low, there is an ‘established set of research techniques’ that ‘can be acquired through formal training programmes’ where ‘success is easy to determine’ (Whitley 1984a:781). Of the seven categories, IS has been classified as a fragmented adhocracy (Culnan and Swanson 1986; Banville and Landry 1992; Culnan et al. 1993; Swanson and Ramiller 1993; Checkland and Holwell 1998; Hirschheim et al. 1996; Hirschheim and Klein 2003; Kanungo 2004). Fragmented adhocracies display high task uncertainty with low researcher mutual dependence, so researchers from these disciplines make diffuse contributions to fluid goals that are contingent on local pressures (Whitley 1984a). Another characteristic of fragmented adhocracies is their openness to the general public (Whitley 1984a), as they tend not to have unique symbol sets that exclude the uninitiated. It is suggested that the characteristics of IS as a fragmented adhocracy work against it becoming a distinct scientific field.
It has been seen in a review of the literature that the establishment of theory, social processes, research methods and standards, a unique symbol set and a set of key topics have been used to explain the nature and development of disciplines in the past. A core body of knowledge appears to subsume theories, research methods and standards, the existence of a unique symbol set and a set of key topics. An examination of the relationship between the impact of local pressures and the degree of professionalism has also been used to help account for the nature of disciplines.
Although little literature was found that related teaching issues to the state of IS in universities, as an internal academic perspective must include teaching issues, any framework developed will need to be capable of encompassing this area. It is argued that for completeness, the relevance and quality of teaching need to be considered. The set of key IS topics must denote then, not only research topics but relevant teaching topics, and will be one means of achieving interaction between the internal and external perspectives of IS. Like research, teaching quality is also concerned with methods and standards.