Centre for Philosophy, Technology and Systems, Free University of Amsterdam, Netherlands and University of Salford, U.K. +44/0-161-295-2913, A.Basden@salford.ac.uk
Keywords: Information systems evaluation, Habermas, Dooyeweerd, Aspects, Critical Systems, Philosophical underpinnings.
In Knowledge and Human Interests (1972), Jürgen Habermas suggests there are three types of science, each with its own 'interest'. Empirical-analytic sciences seek technical control and operate by experiment and deduction, historical-hermeneutic sciences seek consensus in the practical situation and operate by interpretation (e.g. of texts), and critical sciences seek emancipation (especially from hidden presuppositions) and operate by self-reflection. Three major perspectives on information systems align with these three types of science: hard systems thinking, which sees information technology as a way to gain control, soft systems thinking, which acknowledges interpretation and 'appreciation' (Checkland, 1981), and critical systems thinking, which sees information technology as a route to emancipation (Flood and Ulrich, 1990).
But why should we stop at emancipation as the ultimate perspective for information systems? There are both theoretical and practical reasons for asking this question. The theoretical reasons have to do with inherent weaknesses or inconsistencies in the underlying Habermasian philosophy or in its outworking in critical thinking, which have been discussed by a number of people, such as Fuenmayor (1990), Mouw and Griffioen (1993), Mingers (1992), but we will not address them in detail. Instead, we will consider practical reasons, which come from our experiences in real life use of information technology (I.T.), especially from an intuitive view thereof.
Even though I might be glad to be emancipated from the domination of Microsoft, by using a different word processor (Protext) for composing this paper, I am not primarily aware of being emancipated. Rather, I feel enabled. The word processor allows me to both set down thoughts already structured as a linear stream and undertake non-linear dynamic reinterpretation of what I have already set down. I find certain features important: immediate spell checking, macro insertions, auto-correction (tailored to expand my own favourite abbreviations), as well as features I take for granted (e.g. ability to move around the text quickly), and a feature of myself: ability to type with ten fingers. I feel similarly enabled, as I employ Deluxe Paint to create an animation, but by different features like ease of altering the palette of colours, moving among the animation frames and defining animation trajectories.
These two examples are, of course, of application of technology by someone who happens to be in control, where emancipation might not be an issue. But, when technology impinges on me without my control or consent, I don not think primarily of emancipation, but rather of the need for more responsibility in commissioning, design and use of technology. I find this view prevalent among 'ordinary' people.
Whilst emancipation is an important issue it is only one amongst many, and often not the major one. So, a whole perspective of information systems centred on emancipation and related issues like critical self-reflection, would seem rather too narrow, almost as narrow as the perspectives centred on control and interpretation that preceded it.
This paper suggests that a wider, richer perspective is required, and possible. Usually, a paper of this kind would outline flaws in existing work before proposing the new ideas as solutions. But we see critical systems not so much flawed as limited, and limitations are better discussed in relation to a fuller proposal, so this paper is structured in reverse: first alternative framework, then philosophical underpinnings, and then short critique of critical systems thinking (and also hard and soft). Discussion of philosophical underpinnings is necessary because the proposal rests on a very different philosophical stream and possesses a different flavour from most. The framework is new, and much research is still required to knock into shape, test it and then either discard or refine it. It is hoped that this paper will stimulate some of that research.
The framework proposed here is founded on a radical pluralistic philosophical framework by the Dutch philosopher, Herman Dooyeweerd (1955), that eschews the roots of Western thinking found in ancient Greece (and from which even Habermas starts in his 'General Perspective on Knowledge and Human Interest' (1972:301)). As explained later, it restores intellectual dignity to intuitive awareness and everyday activity.
This framework is centred on the notion of multi-aspectual human functioning. All our experience, activity and knowledge as human beings is because we function in a number of distinct aspects. Using some information technology artifact is no exception to this. For example, in the analytical aspect we make distinctions, in the formative aspect we plan and create, in the lingual aspect we represent in symbols, and so on (we meet the aspects later). We do not start from technology nor even its development methodology. Rather, we start with the idea that an information technology artifact is being (or has been or will be) inserted into a working situation. So our central concern is with this change or intervention and the difference between using the artifact and not using it. We consider first the primary users' perspective then widen to those of other stakeholders.
The artifact we use might be as small and individual as a word processor, or it might be a total ERP system in a large corporation, or even a global system like the Internet. Whatever it is, we must address three questions:
These three questions form the three main sections of a framework which, though it might not provide answers, nor even ready-made methodologies, can guide and enrich our analysis.
In the early 1990s the ELSIE knowledge based system (Brandon, Basden, Hamilton and Stockley, 1988) had been in use for a few years by a few hundred quantity surveyors to assist in providing budget estimates. Castell, Basden, Erdos and Barrows (1992) undertook a study of, among other things, the benefits (or other impact) to the surveyors and changes to their ways of working. It found three distinct categories of benefit and change - Features, Tasks and Roles - which were also found in the results of a wider study by the U.K. Government at the time (DTI, 1992). Moreover, the categories were found to have philosophical justification, in Hart (1984).
Features. By its design the artifact has a number of facilities and characteristics. Features are those that are important to the users. The spell checker etc. mentioned earlier are examples. Features also include facilities and characteristics of the human user (such as my ability to type with ten fingers) and of the working context as a whole. However, because we are concerned primarily with change occasioned by insertion of the artifact, most of the features we consider during any analysis are those brought in by the artifact itself.
Tasks. However, as Hart (1984) points out, features only become meaningful when they are employed by human users to aid their humanly-meaningful functioning; this is the task category. The distinction between task and feature is one of meaning. Compare "phoning my wife" (task activity) with "dialling a number" or even "using the phone" (feature activity). It is the same activity, seen from the perspectives of the human being and the technical device respectively. At the task layer we think in terms meaningful to whole persons, while at the feature layer we use terms related to parts or devices (brain, mind, hands, number, dial). Benefits
Roles. However, tasks are carried out in fulfilment of roles, which have to do with the relationships between the people involved and their environment. In the example given, the task of "phoning my wife" finds its meaning in my role as husband. Tasks without roles are meaningless; roles exist even when no tasks are being carried out, though they expect tasks. Hart characterises the difference as that we could put a task in our diaries ("3.0 pm. Phone my wife"). The roles layer is important when we consider social and organizational issues.
Benefits (or their inverse, detriment) of artifact use are found at each layer. Table 1 shows a selection of benefits of KBS technology found during the studies by DTI (1992) and during use of ELSIE; for a fuller table and discussion, see Basden (1994).
Table 1. Three Layers of Benefit of KBS Usage.
The taxonomy becomes a model by adding relationships between the categories: features enable or hinder (human) tasks, while tasks fulfil (human) roles. 'Enable' is much wider than the 'control' sought by empirical-analytic sciences. Use of newly inserted features makes some tasks more effective, efficient or harmonious (or less), enables new tasks to be undertaken and makes other tasks impossible. So the task profile changes, which, in turn, can lead to changes in role. See Fig. 1.
Fig. 1. Features, Tasks and Roles.
For example, in using ELSIE the quantity surveyors found the accuracy and completeness of the knowledge base (two features) speeded up the task of estimating a budget (from, typically, two weeks to two hours), so that an initial rough figure could be obtained during phone conversation with a client. Then, using ELSIE's features of what-if, overriding and explanation with the client present, the client's requirements would be clarified and the budget figure refined. This meant that the introduction of ELSIE changed the task profile from a single budgetting task to two tasks: gain initial figure, then clarify user requirements. However, it was observed that the roles changed too - from novice and expert to two equal partners working towards a shared goal.
It might be thought, especially from the critical systems perspective, that the surveyors would resist losing their power status as experts, but in fact they entered the new role gladly, saying that it enabled them to provide a better service.
The main utility of the taxonomy is in clarifying thinking and discussions during technology planning, design or evaluation. Such discussion often displays a mixture of categories, and this taxonomy has proved very fruitful in overcoming the confusion so caused.
However, as a model it is, on its own, still rather weak because the nature of the two relationships (enabling and fulfilling) is not well understood. While we question whether such relationships can ever be understood to the precision desired by the empirical-analytic sciences, it is likely that better understanding can be achieved, to improve prediction, planning, design and evaluation. Carroll and Campbell's (1989) Task Artifact Cycle, which emphasises that insertion of an artifact into a situation changes the very tasks it was trying to aid, is an attempt to elucidate the features-tasks relationship, but it still lacks crucial detail. Initial suggestions for how we might understand these relationships are made later, section 2.3.
One thing the three layer model lacks is the ability to differentiate between success and failure. To provide this we introduce Dooyeweerd's aspects.
One of Dooyeweerd's (1955) main proposals was that there are fifteen distinct aspects in which a person functions. Some (Hart, 1984, de Raadt, 1997) question the precise number but do not propose any major modifications to Dooyeweerd's suite, which is shown in Appendix 1. The meaning of a single aspect is quite broad; for example the formative aspect covers culture, history, technology, creativity, achievement of goals, planning, formulation of artifacts, formulation of ideas, methodology, technique, and so on - everything in which human formation is central. (The reader does not need to understand all aspects; the text below will explain what is needed.)
Each aspect has its own kernel meaning that is fundamentally irreducible to that of other aspects, but together they form an ordered spectrum of Meaning. (We use the terms 'later-earlier' for this order rather than 'higher-lower'.) Thus the total meaning of any instance of artifact use can be 'diffracted' into its aspects using certain types of abstractive critical inquiry (Clouser, 1991).
When an individual carries out any task s/he is functioning in a number of aspects (usually all of them). Most tasks are qualified by certain aspects which are of greatest importance to that task. For example, writing a letter is a task qualified by the lingual aspect, and creating an animation, by the kinematic and/or aesthetic aspects. (Strictly, we should distinguish three types of qualification (Stafleu, 2000) but we need not do so here.)
But most human activity is multi-aspectual, even when it is qualified by one or two important aspects. For example, as I write my letter, I am also functioning in the social aspect (being polite rather than rude in my letter), the juridical (in trying to give the recipient and others what is their due), and, in a different manner, the biotic (in that I undertake life functions like breathing, digesting food, etc.) The "different manner" is because of two different types of relationship between the aspects, explained in section 3.2: dependency and analogy. In general we can say that our functioning in dependency-related aspects is normally taken for granted and we only become aware of them in breakdown situations (as discussed for example by Heidegger). But our functioning in analogy-related aspects often guides and gives flavour to our functioning in the qualifying aspects; thus our relationships with the recipient, of which there can be several (social, economic, etc.), will affect the main task of writing the letter.
Each aspect has its own distinct set of laws, ranging from largely determinative laws in the earlier aspects (e.g. laws of arithmetic in the quantitative aspect, laws concerning energy and mass in the physical aspect) to normative laws in the later aspects (such as laws of syntax, semantics and pragmatics in the lingual aspect, laws concerning social interaction and social institutions in the social aspect, laws of harmony, surprise and fun in the aesthetic aspect).
The laws of the aspects pertain whether or not we are aware of, and obey, them. Therefore all our activity has repercussions, which derive from each aspect in which we function in carrying out that activity. Since our activity is multi-aspectual functioning, we can expect distinct types of repercussion: social repercussions, economic repercussions, juridical repercussions, and so on. If we function in line with the laws of an aspect then there will be positive repercussions from that functioning, while if we function against the laws of an aspect then there will be negative repercussions. For example, in I am rude in my email then the recipient might be less disposed to grant my request. So, for any activity, there will be a profile of repercussions, positive, negative, weak or strong, which can be depicted as in Fig. 2, where bars to the right and left express positive and negative repercussions of the author's use of Deluxe Paint for creating animations.
Fig. 2. Diagram of Aspectual Repercussions.
The ideal state would be positive repercussions in all aspects (all bars extending to right); this yields what is denoted by the Hebrew word, 'shalom', a general well-being that affects not only we who so function but all around us too. Any negative repercussion detracts from this shalom.
This is the basis of a means by which we can differentiate benefit from detriment. It has still to be developed into a workable method, but it has advantages of combining determinative and normative aspects of the situation, of multi-aspectuality which stimulates a rich yet critical analysis of human activity situations, and of a philosophical grounding that avoids arbitrariness. Also, it can be extended to address a number of normally tricky issues in I.S. use.
Because the aspects transcend all situations and all stakeholders, we can use the above type of analysis to address unintended impact, indirect impact, long-term impact and multiple stakeholders.
In this third section we come back, full circle, to where we started: the artifact that is inserted into the working situation; see Fig. 3. The artifact and the working context must be designed, which means identifying what features these should have. Not all the facilities implemented in artifacts are features, only those (potentially) relevant to tasks. Studies by Kivijärvi and Zmud (1993) showed that the more effort devoted to design the more successful the artifact is likely to be. Too often some of the most useful features are not included in early versions of an artifact but added later after working habits have already been set. Too often, features of the new working context are not considered in advance, but left to emerge as people go through the painful process of adaptation. Then, as the artifact expands, new stakeholders require new features. It would be better if features were considered and included right from the start.
Fig. 3. Completing the Circle: Aspect-Oriented Design
Perhaps an aspectual approach might be able to help identify which features are likely to be needed, especially over the longer term. For each aspect certain features become important - spelling and grammar checkers for lingually qualified tasks such as writing papers, trajectory designers for spatially qualified tasks, currency conversion for economically qualified tasks, etc.. Though the precise shape of, and need for, such features must depend on the individual users and situation, it might be possible, with research, to draw up general categorisations of features that aid functioning in each aspect. Then, by identifying the aspects that are important to the task profile (current and future), it might be possible to speed up the initial selection of features and ensure none is forgotten.
Such a categorisation might also enrich our understanding of the enabling relationship between tasks and features. It was said in section 2.1 that the three layer model on its own was too weak to provide useful predictions, but an understanding of what features can aid functioning in each aspect and how could help us discern the types of repercussions possible in given situations. This would presumably make a contribution to the field of information systems evaluation.
Dooyeweerd's philosopical framework (1955) arose from his concern that, as Whitehead (1937) put it, "All else", in Western thinking over the last 2500 years, "is a mere footnote to Plato". As Habermas (1972) has noted, the concept of theory "presupposed a demarcation between Being and Time". Dooyeweerd believed this presupposition to be untenable. He undertook a thorough analysis of theoretical thought over the past 3,000 years, with more detailed examination of the key thinkers like Kant, and showed how it has been unable to bring together coherence and diversity, or meaning, being and time, or law and entity, or theory and practice. He showed that while proposals might enjoy acclaim over a few decades or even centuries, most eventually proved unsustainable.
Dooyeweerd agrees with Habermas that the theoretical attitude of knowing is not neutral. But, unlike Habermas, who sought to uncover them by means of critical self-reflection, he proposed that all theoretical thinking is founded not on reason but on religious commitment (Clouser, 1991) - commitment to a way of seeing the Cosmos that is not open to theoretical explanation (though it can be analysed). This provided a new way of understanding presuppositions and their power that is at a deeper level than offered by Habermas. Dooyeweerd made a study of the religious ground motives that have provided the presuppositions for theoretical thinking over 2,500 years (Form-Matter, Creation-Fall-Redemption, Nature-Grace and Nature-Freedom) (Dooyeweerd, 1955, 1975). He showed how three are dualistic and how this makes them unable to provide sustainable answers and always end in fundamental antinomies that would, for example, always drive theory and practice away from each other.
However, Dooyeweerd did not only demolish; he accepted the challenge of constructing something to replace what he wished to remove. Started from the presuppositions of the non-dualistic ground motive, that of creation, fall and redemption, he critically examined the basic assumption that Being is the fundamental property of all we experience in temporal reality, and proposed instead that Meaning is the fundamental property. All Being is, in fact, Meaning. Note that this Meaning is an of an ultimate kind, set into the fabric of the Cosmos by its Creator, rather than meanings assigned to symbols as we function in the lingual aspect, which Mingers (1995) discusses, though the two are not unrelated. By adopting radically different presuppositions, he could conceive ontology, epistemology, theory, practice, the subject-object relationship, and indeed the very notion of entity, in novel ways. This in turn allowed him to postulate a pluralistic ontology, which we have met as his suite of aspects above.
Though, as mentioned above, the aspects are irreducible, they are linked by two types of relationship. One is dependency, which defines the ordering among the aspects, in which laws of later aspects depend on the proper functioning by laws of earlier aspects. The other is analogy, by which in each aspect there are echoes of all the others. For example power relations (critical theory) are of the juridical aspect of justice, but something akin to power can be found in the physical aspect (force), analytical aspect (power of argument), formative aspect (achievement) and so on. This 'intertwined irreducibility' is Dooyeweerd's starting point for addressing the philosophic issue of unity and diversity and explaining both the diversity and coherence that we experience in life.
Even though he would acknowledge that his particular suite of aspects could be refined, Dooyeweerd made an ontological claim for his notion of aspectuality: that there are distinct aspects which are intertwined, and that their laws pertain whether or not people have discovered them, are aware of them or obey or transgress them. This is the philosophic basis for there being repercussions of all activity - positive when we obey and negative when we transgress - and the framework's ability to handle unintended consquences and to transcend even multiple stakeholders. This should not be seen, however, as essentialism, which presupposes Being, because Dooyeweerd presupposes Meaning and Law. Moreover, as Mingers (1992) points out, positing an external reality need not end up in objectivism. Dooyeweerd escapes objectivism by showing that objectivisation is not an absolute but merely part of functioning in the analytical aspect; see below. So, even though the laws pertain and there are repercussions, these are not always determined.
Western liberalism sees Law as constraints on freedom, because it presupposes self-dependent Being. In contrast, Dooyeweerd sees Law as enabling, and intended by the Creator to lead to a joyful, responsive, full shalom for the whole of the Cosmos. Law enables and, where it is normative defines, the contribution to this shalom a particular aspect can make. It is because of this that Dooyeweerd maintains that if we operate in line with the laws of all aspects then things will go well, but if we transgress any then things will suffer. This is the philosophical underpinning of the proposal above for differentiating beneficial from detrimental repercussions.
Starting from this idea of Law, Dooyeweerd has an interesting notion of the subject-object relationship that integrates the three English language uses of 'subject'. Entities function by responding to laws of certain aspects, and this responding is what it is to be both subject to those laws and a subject, in the sense of being subject in a sentence. Being 'subjective' refers to being an analytical subject, i.e. making one's own distinctions etc. While humans can be subjects in all fifteen aspects, most animals can only be subjects up to the sensitive aspect, plants to the biotic and non-living things to the physical. However, all entities can be objects in any functioning. For example, a mineral can be an economic object but not an economic subject.
An entity's Being, therefore, is closely tied to the aspects. Kant drove a wedge between Ought and Is, and generations of technologists since then have focused on one to the detriment of the other, on technology without ethics. Dooyeweerd reconnects them because both Ought and Is are founded in Meaning, the former in the normativity of the aspects and the latter in subjection to laws of aspects.
Any stakeholder can stand in a subject relationship to functioning with the I.T. artifact or an object relationship or both. Subject-stakeholders can, by their response to relevant aspectual laws, have some influence on that functioning. In information systems, the main subject-stakeholders are the client, designers, developers and users. Object-stakeholders merely receive the repercussions of the functioning. As we discuss below, this gives us an understanding of emancipation.
Because he made an ontological claim for his notion of aspectuality, at first Dooyeweerd sided with the Realists against the Nominalists. But he soon moved away because Realism, by making the Greek presuppositions, ends up postulating the detached observer and the divorce of theory from practice. To Dooyeweerd (as to Habermas) the observer is part of the observed. He shares with Bhaskar's (1986) critical realism both an external reality and the validity of interpretation. But, while Nominalism hypostatizes (absolutizes) naming and Interpretivism, interpretation, Dooyeweerd claims these are merely subject-functioning in the analytical and lingual aspects, and all such functioning is non-absolute, as we shall now discuss.
If there is both an external reality and interpretation, how can we have knowledge that we can be sure of? Kant's attempt to answer this ended up by claiming that we cannot ever know the 'Ding an Sich', the thing in itself. Most philosophy since then has accepted this, but tried to find a foundation for sure knowledge. Phenomenology suggested we must identify hidden presuppositions and then escape them. Wittgenstein, originally a phenomenologist, reacted against it, into 'language games'. Habermas saw the solution in critical self-reflection.
By eschewing the primacy of Being, Dooyeweerd might offer a useful answer. The philosophical problem here is how our knowledge of a thing relates to the thing itself, the relationship between epistemology and ontology, which he examined in depth. He analysed not only knowledge but the activity of knowing, not only the status of theory but the doing of science.
If all human activity is functioning in the aspects then so is knowing. Epistemology is not of one type; each aspect has its own distinct type, so that each requires different research methods and criteria for what good knowledge is. We can see a reaching for this in Habermas' three types of sciences. It also allows us to suggest that each type has an arena in which it is valid, rather than think, as many do, that newer forms (critical) are superior to older (empirical-analytic). What all sciences share is the isolation of an aspect in order to study its laws, which is a special type of analytical functioning, higher abstraction (Clouser, 1991), that generates theory. It is theory, rather than knowing as a whole, that is unable to grasp the 'Ding an Sich'. Intuitive awareness and everyday knowing are multi-aspectual and, Dooyeweerd argues, are truly able to grasp the 'Ding an Sich' in spite of the protestations from theory.
This is because, to Dooyeweerd, nothing in the Cosmos or our experience is ever self-dependent, neither Being, which is always grounded in Meaning, of which the aspects form a spectrum, nor the aspects themselves, which are always referring beyond themselves, both to each other and also to their Creator. This means no aspectual functioning is absolute, that is self-dependent. (This includes the process of discerning what the aspects are, which is why Dooyeweerd would not make an ontological claim for his own suite of aspects. It also explains why his ontology was able to be pluralistic without fragmentation.) Reductionism, to Dooyeweerd, is absolutization of one aspect, reducing others to it. Our functioning in that one aspect might be enhanced but our overall multi-aspectual activity becomes impoverished.
Over 2,500 years we have absolutized the analytical aspect (in which lie theory and reason) as the 'true' way of knowing, superior to everyday knowing. But, if no aspect can be absolutized then theory itself is fundamentally limited in scope, and only one way of knowing. For the highest quality knowing, all aspects must be involved. Intuitive awareness and everyday activity, being of this kind, are of a different kind from theory, which is uni-aspectual and isolative, and thus not to be seen as inferior to theory. In this way Dooyeweerd restores dignity to the latter and provides grounds for properly addressing intuitive awareness and everyday activity, such as we began this paper with.
With this perspective we can understand the fundamental difference between on the one hand the research and development of technologies and methodologies as such, and on the other hand, the use of technical artifacts. The former are specialised activities, often uni-aspectual in nature, heavily influenced by the theoretical attitude. The latter is multi-aspectual, and should be classed as everyday activity in which intuitive awareness is key.
We can now set out a philosophical underpinning for the three categories of roles, tasks and features. Hart's (1984) arguments for them were a development of Dooyeweerdian ideas, so we can expect they will integrate well with the underpinning of success and benefit and of theory and everyday activity outlined above.
In systems thinking, a system has subsystems and can itself be seen as part of another system: epistemologically, wholes can be parts and parts, wholes. But Dooyeweerd distinguishes wholes from parts ontologically, a whole being an entity that is capable of meaningful independent action (Hart, 1984), while a part has little meaning outside the whole to which it belongs. For example, an animal is capable of meaningful independent action in the sensitive aspect, but its heart is only involved in that aspect while part of the animal. Remove it from the animal, and it is only a pump, qualified by the physical aspect. Hart calls wholes 'functors'.
Hart sums up his argument in the phrase "Functors functioning in relationship" - and this gives us our three categories. The central category, tasks, focuses our discussion and consideration on functors (e.g. people) themselves, so what they do (tasks) is seen in terms of the meaning those wholes have. The roles category focuses our attention on the relationships between functors (and with environment). The features category focuses our attention on what is needed to help the functors in their functioning. The I.T. artifact itself is not to be seen as a functor; like the animal's heart, if we isolate from its context, it is merely a physically qualified entity and loses its task-level meaning. Hart argues that the features-task relationship carries a degree of necessity (e.g. calling my wife requires a phone).
Both tasks and roles are multi-aspectual, but features are often related to a single aspect. This underpins the suggestion, in section 2.3, that features can be related to specific aspects. It also enables us to see what contribution theoretical knowledge can make to information systems usage. Usage as such is multi-aspectual because of tasks and roles, and therefore should be seen as everyday activity, and theory should have no pride of place. Features, on the other hand, are often developed from theoretical knowledge. For example, ray-tracing in 3D animation software was developed from physical theory of light, and non-technical features of the working context might be addressed using theory from psychology or social science. So the contribution theory can make is in the design of facilities that might be features in a given situation.
However, to give theory too much credence in considering tasks and roles risks allowing an element of reductionism to creep in, and an overall impoverishment of the process. Theory can indeed be applied to information systems usage, but one has to be very careful to avoid this reductionism. Unfortunately, it is not always the case that sufficient care has been exercised in academic discussion. The problem goes beyond positivism to reductionism and elevation of theory of any kind, even critical theory.
We can now attempt a very brief critique of part of critical systems thinking. While Fuenmayor (1990) has argued that the critical approach is not necessary, and Mouw and Griffioen (1993) have pointed out fundamental inconsistencies in Habermas' ideas, it is not our concern here to discuss flaws in the underlying theory.
Rather, we will address the practical questions of how the Dooyeweerdian framework can throw light on several facets of the critical systems jewel and bring them to life, and whether emancipation, the central focus of critical systems, is too narrow. There are areas of agreement between Dooyeweerd and the critical approach that suggest they are not inimical to each other, and so this might be feasible. Both share the belief that all theoretical thought is founded on presuppositions, which should be uncovered. Both recognise the social element in information systems usage. Both recognise an external reality and intersubjective interpretation.
One facet is self-reflection, which is proposed by critical theory as the means of uncovering hidden assumptions. At the very least, if assumptions might be related to several aspects, Dooyeweerd's suite of aspects could be used as a taxonomy to stimulate the search for assumptions. More specifically, Dooyeweerd would point to the pistic aspect as where the most basic presuppositions and 'interests' lie.
A major facet is emancipation which, following Mingers (1992), we will see as emancipation from material conditions. These can include adverse impact of information systems. A Dooyeweerdian analysis would first seek to understand under what conditions emancipation is appropriate. Though perhaps oversimplified, we could see emancipation to be needed where a person is an object-stakeholder, receiving negative repercussions of functioning with the artifact, without playing any subject-stakeholder role by which they could influence the functioning. As discussed above, negative repercussions often occur when aspects have been ignored during design or commissioning of the artifact. If so, then an aspectual analysis of the situation could help identify where emancipation might be needed, of what type it is, and what action could be taken to achieve it.
A third facet is power relations, which is claimed by the critical approach to be the core of any solution to the problem of emancipation. But why should we believe this claim? Why not, for example, propose as a solution "Just don't worry about it!" or "Just laugh it off!" or "Ensure you make friends with the boss"? Dooyeweerd can supply at least one reason to take the claim seriously by identifying the qualifying aspect of emancipation and any proposed solution. Emancipation is not just freedom of any kind: it is freedom from unjust constraints and conditions, and hence links most strongly with the juridical aspect, of 'what is due' to each stakeholder. By the principal that a problem in an aspect should be first addressed by a solution from the same aspect, we can see that power relations are an appropriate solution because they too are of the juridical aspect. The three suggestions above are of the sensitive, aesthetic and social aspects and, while they might have a part to play, do not provide a sustainable solution on their own.
However, the very strength of critical systems approach is also its weakness. As we illustrated in the introduction, emancipation, is not the only valid perspective to hold on information systems, and often not even appropriate. If we try to force all our consideration of information systems through the mesh of emancipation then we rob the situation of its diversity of meaning. To do so is a kind of reductionism. Therefore we can see that a Dooyeweerdian critique would not sideline critical systems but take us beyond it.
This is also how the weaknesses of the hard and soft systems approaches can be seen. Hard systems elevates the aspect whose theme is control (the formative aspect which also includes technology). Soft systems can be said to elevate the aspect interpretation (the lingual). What we have proposed here is a way of avoiding all such elevations and placing each type of systems thinking in perspective among the aspects in which we function.
In practice, however, good quality practitioners in any of the three systems approaches might produce good information systems. This is not because of higher specialist skill but because such practitioners tend to consider a wide range of aspects tacitly, even while focusing their attention on one or two main ones. Thus, for example, Peter Checkland and his colleagues make a success of Soft Systems Methodology because they tacitly consider many aspects. In this we see again the difference between theory, which is uni-aspectual, and good practice, which is multi-aspectual. We could call this 'wisdom', which is precisely what de Raadt (1991) has done in his book, 'Information and Managerial Wisdom', which describes a systems approach that makes use of Dooyeweerd's aspects.
Though Habermas believed ontology to be in ruins, Dooyeweerd has suggested a fresh approach to ontology that frees it from being driven inexorably towards objectivism, and allows interpretation and other non-deterministic aspects to integrate with the more deterministic aspects.