Given the ongoing debate on the explicit and implicit foundations of theories and methods within the IS discipline, this chapter examines the circumstances that appear to favor a close connection between Information Systems (IS) research and philosophy (phenomenology in the specific case examined). The question addressed is whether references to philosophy are typically made when a researcher is confronted with issues and topics that she considers crucial and which she believes inappropriately treated or neglected by current theories and practice. To answer this question the influence of Heidegger on Ciborra’s work is explored. The method adopted is the review of Ciborra’s publications and his references to Heidegger; meaningful issues were identified and significant passages of both authors compared. It appears that Heidegger’s influence has been substantial in terms of support for both Ciborra’s critical analysis and innovative solutions for the IS-related world. Given the multiplication and growing pervasiveness of information technology and IS and the consequent growing need for policies based on sound research, the “escalation” to philosophy (and especially the engagement with phenomenology), though demanding and difficult, will probably become increasingly useful.
TopIntroduction: Setting The Question
Philosophy, let alone phenomenology, is not explicitly mentioned in the Information Systems (IS) related disciplines mentioned by Baskerville and Myers (2002) in their article where they advocated a closer connection among different research branches interacting in a ‘knowledge network.’ In fact, they claimed that IS had become mature enough to be a reference discipline itself, able to contribute to the advancement of learning in several fields such as medical informatics or electronic commerce. Possibly for this very reason—the newly acquired awareness of the significance of IS—Lee and Baskerville (2003) the following year addressed (in depth) a crucial issue for any scientific specialty, the question of ‘generalizability’ of the results of a research endeavour. In their essay they point out that several articles published in leading journals relate a reduced sample size, or a limited number of cases, to a limited generalizability of findings. In order to show that the concept is misapplied and misguides both authors and reviewers, they conduct their analysis considering the relevance for IS research of the problem of induction raised by the eighteenth century Scottish philosopher Hume, and well known to epistemologists: no ‘proof’ can be reached through data unless one assumes the ‘Uniformity of Nature’ (UN). Uniformity cannot be proven by induction since what has to be explained (UN) cannot be the basis for its own explanation (UN itself). The authors used Hume’s argument to show that in statistics it is not possible to generalize the estimates of a given ‘sample’ to the characteristics of the ‘population’ so that quantitative descriptions are not generalizable beyond the domain that the researcher analyzed; of course, larger sample sizes produce more ‘reliable’ results in that—within a given population—other samples based on the same criteria are more likely to show converging results. But ‘reliability’ of samples is quite a different concept from ‘generalizability’ of a sample to the population. For the same reason, results of qualitative research cannot be extended to other settings beyond the one observed. The authors conclude that a theory is generalizable to another context only if it survives an empirical test in that new context. This consideration, they comment, questions the possibility of generalizing to a comprehensive IS theory—for example the determinants of IS success, as in DeLone and McLean (1992) whom they cite—from results of different research studies conducted on different instantiations of information technology (ERP, Internet, packages, etc.). One possible solution is the adoption of ‘action research’ where theory can be both tested on a new setting and enriched by the interaction with the new context. Even though in their article Lee and Baskerville do not state explicitly to draw on phenomenology, their indication of action research is very close to a phenomenologist’s position whereby “the positing of entities outside experience is ruled out as meaningless” (Moran, 2000, p. 13). In fact, Lee (in a 2004 article co-authored by Martenson) proposes a kind of action research named ‘dialogical action research’ that explicitly refers to phenomenology:
Unlike other forms of action research, dialogical AR takes notice of and addresses heterogeneity in the forms of knowledge held by the scientific researcher and the real-world practitioner, where the heterogeneity is related to what phenomenology would call ‘the scientific attitude’ taken by the researcher and ‘the natural attitude of everyday life’ taken by the practitioner. (Martensson & Lee, 2004, p. 508).
In the Hume-Lee and Baskerville case, the IS researchers used a philosophical conception both to criticize current practices and assumptions in the IS discipline (they uncovered misleading constraints in the evaluation of research, e. g. quantitative research is not more generalizable than qualitative research) and to propose solutions (they indicated an approach to be used by researchers to pursue both scientific significance and practitioners needs, namely action research). Certainly, there have been several other cases where the recourse to philosophy has had a significant role in IS research, the most famous probably being Dreyfus’ (1992) criticisms of the computational model of the mind adopted in artificial intelligence research that he developed from the years sixties onwards drawing on phenomenology.