Abstract
Attempts to produce an adequate and long-lived subject indexing system for information systems research have failed. In this paper we seek to address this by proposing an approach by which the terms expressed in research literature, such as those in the information systems literature, can be systematically and meaningfully categorised. The approach is significant in that it draws upon rigorous and philosophically compatible bodies of work in two areas. Firstly, we draw on work addressing the nature, existence, and categorisation of literary expression found in research papers (Roman Ingarden’s ontological analysis of the scientific work of art). Secondly, we draw from qualitative research methods addressing how meaningful categories can be analysed from text and related to each other (grounded theory). The resulting approach has the potential to be applied in many scientific disciplines beyond information systems, and to form the intellectual core of an information tool in e-research.
Table of Contents
Roman Ingarden developed a number of conceptual and methodological frameworks for ontological analysis of texts, which are documented in his books The Literary Work Of Art (1965) and The Cognition of the Literary Work Of Art (1968). While Ingarden’s primary focus was on mainstream literature, he also considered scientific works along with a number of other literary forms as borderline cases of the literary work of art. We are presently involved in a project, a significant aspect of which involves the analysis of papers reporting information systems research in academic journals. A broader description of this project and a discussion of the rationale for using Ingarden’s frameworks can be found in Lamp and Milton (2003, 2004).
The issue of applying Ingarden’s framework to scientific works is significant because, while his work has been extensively applied to mainstream literature (e.g. Thomasson, 1996), there are no reports in the archival literature relating to developing his ontological analysis of scientific works into a technique which can then be applied more generally.
A threshold matter that must be considered at this point is whether or not articles publishing information systems research can be considered to be scientific works in the sense intended by Ingarden in his analysis. Ingarden published in Polish and German and the versions of his work that we are using are translations into English published in the Northwestern University Studies in Phenomenology and Existential Philosophy series. As Ingarden himself comments about translating scientific works ‘a “good” translation is not impossible, though it may often be difficult’ (Ingarden, 1968). In this context the comments of Ingarden’s translators are relevant:
… it must be noted that ‘scientific’ is used here in a much broader sense than usual, in connection not only with the natural sciences but also with any serious field of study, just as the German wissenschaftlich is used. (Ingarden 1968)
This interpretation is confirmed by Cassell’s German-English English-German Dictionary (Betteridge, 1978), which offers ‘scholarly, scientific, learned’ as translations of wissenschaftlich. Accordingly, it can be validly asserted that articles publishing information systems research can be considered to be scientific works in the sense intended by Ingarden.
To illustrate Ingarden’s framework based on scientific works, we use Broadbent et al (1999) as our exemplar. This paper is from the journal MIS Quarterly. While MIS Quarterly is often rated as the most significant information systems journal (e.g. Peffers and Ya, 2003; Katerattanakul et al, 2003a, 2003b; Bharati and Tarasewich, 2002; Mylonopoulos and Theoharakis, 2001; Walstrom and Hargrave, 2001), the choice of this particular paper is not significant; it was chosen simply because it was conveniently at hand. To provide a context and assist with understanding the analysis in the later sections, the abstract of the paper is reproduced here:
Business process redesign (BPR) is a pervasive but challenging tool for transforming organisations. Information technology plays an important role by either enabling or constraining successful BPR. This paper explores the links between firm-wide IT infrastructure and business process change. IT infrastructure is the base foundation of the IT portfolio, which is shared throughout the firm in the form of reliable services, and is usually coordinated by the IS group. IT infrastructure capability includes both the technical and managerial expertise required to provide reliable physical services and extensive electronic connectivity within and outside the firm.
Exploratory case analysis of four firms (two in retail and two in petroleum) was used to understand the ways IT infrastructure contributes to success in implementing BPR. The finding was that all firms needed a basic level of IT infrastructure capability to implement BPR. The firms that had developed a higher level of IT infrastructure capabilities, before or concurrent with undertaking business process redesign, were able to implement extensive changes to their business processes over relatively short time frames. The higher level of infrastructure capability was provided in the form of (1) a set of infrastructure services that spanned organisational boundaries such as those between functions, business units, or firms, and (2) the ability of the infrastructure to reach particular constituencies inside and outside the firm to transfer information and process complex transactions.
The more extensive business process changes were more innovative and radical, crossing business and functional unit boundaries, and resulted in more significant business impact. The practical implication of the study is that before embarking on any form of BPR, managers should complete a business audit of their IT infrastructure capabilities, as these capabilities have an important impact on the speed and nature of business process change.
Examining the abstract, one can see many features that are in common with many other works in the information systems literature. Specifically, the paper considers BPR, a strategic and tactical tool in which information systems play an enabling role, it applies both qualitative and quantitative techniques, it examines complex case studies in four organisations, and it discusses many infrastructural and organisational impacts on the success of implementing BPR and associated enabling technology. Consequently, we argue that by applying Ingarden’s framework to this paper we will be able to gain insight into applying the methodology to other information systems literature.
Scientific works are asserted by Ingarden to consist almost exclusively of genuine judgements (1968), the most significant ontic items of which are:
-
the states of affairs described;
-
schematised aspects; and
-
the represented objectivities.
In the following sections we examine the concepts in Ingarden’s framework in some detail.
Ingarden (1968) asserts that literary works of art contain no genuine judgements, they contain quasi-judgements. The literary work of art concerns a portrayed world, in which assertions, or statements by portrayed persons, can only be considered within the context of the portrayed world. In contrast, he asserts that the role of the scientific work in the transmission of cognitive knowledge requires that its context is that of states of affairs in the real world. Consequently, because there is a real world to which judgements refer, he calls the judgements in scientific works genuine judgements.
Ingarden states that genuine judgements are assertions that may be true or false, but they lay claim to truthfulness. For example, a paper may report ‘The management style of company A was undemocratic’, which is a result perceived as true by that particular researcher, and yet a second researcher may report a different result. Despite their essential contradiction, both statements are genuine judgements. Ideally, genuine judgements allude to means of confirmation that may be found in experience, or are contained in literary proofs based on reasoning and written in conceptual language. Failure to provide means of confirmation weakens the paper and reduces its functional value (Ingarden, 1968). Broadbent, et al (1999) shows many examples of the first form of confirmation.
From March to September 1992, a small team further examined process approaches concurrently with a detailed study of the capabilities of the firm’s current systems and infrastructure.
This quote is a genuine judgement about the efforts of the CostCo business process reengineering team, as collected by the authors of the paper; that is, it reports events actually experienced by the person reporting to the authors of the paper, and therefore claims authority based on reporting an actual contemporary experience. On the other hand, consider:
Business process redesign (BPR) is a pervasive tool for transforming organisations (Grover et al. 1993) and [is] ranked as one of the most important issues for information systems (IS) executives since the early 1990s (Brancheau et al. 1996; Index Group 1994; Watson et al. 1996).
This quote is a genuine judgement regarding the views of IS researchers on BPR; that is, it reports a state of affairs reported in the IS literature and therefore claims authority based on previous accepted research.
Genuine judgements whose authority is based on literary proofs are relatively unusual in information systems research and are not found in Broadbent et al (1999). An example of this form of genuine judgement is the following definition from Smith (1998):
Broadbent et al (1999) also contains questions such as those in the following extract:
Important questions to consider include:
To what extent does the firm have at least the 10 core infrastructure services (see Table 1) together with the seven boundary-crossing services in place?
What is the reach in terms of who can be seamlessly connected?
What range of services are available: only the ability to access information or the capacity to perform complex business transactions across multiple systems?
Such questions could be reworded as assertive statements, reporting evidence or argument supporting the affirmative or negative, and hence should also be considered genuine judgements (Ingarden, 1965).
Objects represented in a literary work are derived, purely intentional, objects projected by units of meaning (Ingarden, 1965). They are intentional because an author has written them with a purpose. For literary works of art, the purpose is to tell a story or generate a particular aesthetic effect; for scientific works, it is the transmission of cognitive results (Ingarden, 1965). In both cases the objects are derived, because we cannot enter the mind of the author. Finally, they are projected because it is only through language (in this case written language) that can we understand what is intended.
Consider the following extract, also from Broadbent et al (1999):
CostCo has a robust network with numerous LANs in place at its head office, in large cities, and other major sites. ‘About 2,000 PC users have whatever multihost connectivity is required for their business needs We have moved from computer-centric to network-centric computing’, noted the CIO. Business units utilised these networks as a basis for a new distributorship, retail and electronic funds transfer, and point of sale (EFTPOS) systems.
This extract concerns the existence and nature of the network in place at CostCo — it describes one aspect of CostCo. The physical nature of the network is described along with the scope and purpose of the network and the attitude of management towards its use. An outline of the functionality offered by the network in place is also described.
However, what is represented by this extract does not stop at the network but extends further to linkages provided for EFTPOS, the retail activities of CostCo and other related activities, and the general management policy framework within which the specific network policy lies, even though none of this is directly given to us. This is also typical of scientific and other literary works because there is seldom enough room to completely describe a state of affairs. Equally, this description does not delve into the details of the precise networking protocols, hardware, operational requirements and other minutiae of CostCo’s network. If this depth of analysis were provided, then the extract would no longer be talking about the network, but of the components of the network and how it was managed.
For these reasons, literary works necessarily consist of incomplete descriptions, termed schematised aspects, which contain fulfilled (explicitly described) components and unfulfilled components that, while not explicitly described, may not be indeterminate. The reader may fill these out from aspects held in readiness from previous experiences. Prompting the most appropriate aspect is influenced by the word choice and represented objectivities selected by the author. For example, by using two words identical in meaning, but different in word sound, the reader may be influenced towards different aspects (Ingarden, 1965). Substituting ‘many’ for ‘numerous’ in the preceding quote does not alter its meaning, but may influence the reader’s choice of aspects between perceiving CostCo’s LANs as an unordered collection or as an ordered and therefore managed collection.
Because these aspects are based in perception, and aspects of the same object that are experienced by different individuals must differ in various respects, it is not possible for the reader to actualise with complete accuracy the same aspects intended by the author (Ingarden, 1965). The degree of this type of perceptual error in a scientific work is reduced, as schematised aspects are intended only for assistance in the transmission of cognitive knowledge. The use of decorative or evocative aspects is unnecessary, and may hinder the essential aim of a scientific work — accurately transferring knowledge as intended by the author (Ingarden, 1965).
A literary work of art describes people, animals, lands, houses and other items. This represented world is not the real world — the represented objects within it may not exist in the real world or may behave differently to such objects within the real world. As a reader reads a passage of words and phrases (meaning units) containing a represented objectivity, he or she relates directly to the state of affairs that the represented objectivity is helping to clarify. Consequently, a particular represented objectivity within a scientific literary work causes us to direct ourselves to corresponding states of affairs. Because we are dealing with a scientific work rather than a literary work of art, this directional ray passes through the content of these represented objectivities so that they refer to objectively existing states of affairs, or to objects contained within them rather than to some fictional creation (Ingarden, 1965). See Figure 1 for a representation of this. It is through this directional meaning ray that the represented objectivities claim to determine objects in the real world as they are in themselves and thereby claim to be genuine judgements (Ingarden, 1968).
In a scientific work, clarity in writing directly affects the transparency of represented objectivities. Where readers have difficulty in relating to the state of affairs beyond the text then the represented objectivities are not ‘clear’. Ideally the represented objectivities are transparent; that is, the way that they inform the reader regarding the particular state of affairs under discussion requires little conscious interpretation. In the extract above, the IT infrastructure (the state of affairs being discussed) is correlated with the represented objectivities ‘computer-centric’ and ‘network-centric’. For readers of MIS Quarterly, these would be readily understood concepts and hence transparent. It should be noted that if that was not the case, and the authors went on to define the concepts, then the concepts would become states of affairs for which other, hopefully more transparent, represented objectivities could be found to describe them. In the absence of such explanation from the authors, the reader re-reads the sentences, concentrating on the word meanings and syntactic interconnections until the represented objectivities become clear and unequivocal (Ingarden, 1968).
Summarising, any scientific work will contain genuine judgements, states of affairs, schematised aspects, and represented objectivities. Effective subject indexing in any academic field requires the terms in these ontic categories be identified and the relations between them defined. The terms in these categories are meaningful to some group(s) of people and form the basis of the subject indexing. Clearly, there are potentially scores of specific terms that fit within each of the identified categories. Further, the relative importance of each term is not evident a priori. Thus, in order to operationalise Ingarden’s high-level categories, we need an approach that captures the essence of the activity. We seek terms that (a) are meaningful to a group of people, (b) exhibit cognitive economy (Rosch, 1978), and (c) are discovered through a process that is repeatable. The principle of cognitive economy is that categorisation should provide a great deal of information about the item categorised with only minimal cognitive resources expended. In the following section we outline an approach based on both Ingarden’s categories and the grounded theory method.