(Pre-publication version of article to be published 1991 in the J. of Applied Systems Analysis, Vol. 18, 39-55)

by Kristo Ivanov

Umeå University, Inst. of Information Processing

S-901 87 UMEÅ (Sweden). Tel. +46 90 166030 — Fax +46 90 166126

This paper presents a summary of some features of soft systems methodology—SSM, and of critical systems thinking—CST as they have been experienced from the point of view of the field of applications of information technology. It highlights the manner in which CST completes SSM in the context of the design of computer support in the form of HYPERSYSTEMS, and evidences some problematic aspects of the two approaches that push the practitioner into philosophical issues. One concluding hypothesis is that further developments of systems practice must be sought at the interface between formal science, political ethics, analytic psychology, and religious thought. For tutorial purposes, a great amount of literature is related to these issues.

KEYWORDS: Critical social theory, liberating systems, cooperative work, constructive systems, hypertext, soft systems, expert systems, dialectical pragmatic systems, power, ideology.

SSM has been comprehensively presented in an extensive work [14] but the following will be edited out of a more recent, and therefore presumably more developed, presentation of SSM that is related to informations systems [17]. It will sometimes be completed and followed by my comments.

Information systems can be seen as an organized attempt at meaning—creation—from—data by a social institution. The creation of meaning is seen as requiring a "semantic information theory" that may be grounded on the welcomed theory of "speech acts" in modern philosophy. An SSM idea that is considered crucial is that a set of activities be linked as to form a purposeful whole that could itself be regarded as a kind of system, a "human activity system". Such systems can be adequately clearly described only in relation to a particular world view, or Weltanschauung. Methods—models—systems that are relevant to debate real-world activity are developed (in a logic—driven stream of analysis), and compared with real-world action in a problem situation, in order to structure a debate about change. The problem situation itself is simultaneously explored (in a social stream of analysis) as a culture, with social and political characteristics. It feeds both the former choice (design) of relevant systems and the debate about change among participants/designers/users/end users/(ever-changing)-organizational members. This view of social reality is that of an ever changing outcome of a social process in which human beings continually negotiate and renegotiate, and so construct with others their perceptions and interpretations of the world and the rules for coping with it, rules that are never fixed once for all.

Future computer projects need to be complemented with such a technical—social—political process for continuous thinking and rethinking of organizations tasks and processes [activitities], together with the rethinking of the enabling information flows. When iterations of the process produce models which are widely agreed to be relevant in a company [problem] situation, then such consensus activity models can be converted into information flow models, and the more traditional methods of information system design can be initiated. Lately there have been attempts, ideas and projects to design multimedia computer support of similar processes and their iterations [35; 61]. The activity models can be transformed into traditional information flow models [19], as known in the field of information systems analysis [57], by asking of each activity: (a) what information is required in principle to do this activity, in what form, with what frequency, from what source?, and (b) what information is generated by doing the activity? Even if prototyping as related to continuously learning systems apparently makes the idea of stable information flows obsolete, it is still true that some provisionally stable structure of the information system must be available in order to be periodically revised. Co-constructiveness requires constructiveness. Learning systems are in the best case expected to increase the frequency of revisions or updatings, or then, to create structured databases of continuous opinion polls.

It is now time to reflect on the SSM approach summarized above. Lately there has been an articulated criticism of SSM at the interface with critical systems thinking. It will be covered later but for the time being we can note a kind of summary of this criticism that was presented in a recent paper [39]. It is stated there that the soft systems approaches were unable to deal with the fundamentally conflictual nature of social systems. Influential students of the method of social science in the context of analysis of radical, antiorganization approaches noted that the home of the twentieth century's critical theorists was the so called Frankfurt school. As alternatives or complements to SSM are in fact mentioned the "critical systems heuristics" that is directed at practical action of a critically normative nature, and the "theory of communicative competence" that provides a clear and pure theory of social interaction and of the means of societal reproduction, a rational reconstruction of the formal conditions required for communicative competence, i.e. the ideal speech situation. It is noted that the term "critique" has a long history. It was first used to describe the art of informed judgement appropriate to the study of ancient texts, whether the Classics or the Bible. The appeal to critique gradually displaced the criterion of truth, from revelation towards clear and rational, or critical, thought. The critical approach proposed by the critical heuristics of social planning is seen to remediate the failure of the critical theory of communicative competence in bridging the gap between theoretical justification and political or strategic action. It seeks to bridge this gap between theory and praxis by introducing the notion that any theory of society must be critically normative, and hence practically oriented. This means treating the model of communicative competence as an ideal which cannot be practically achieved, but can be striven for. This means striving not at an objective solution but merely at a critical solution. Critically normative reflection, however, must not remain extrinsic to systems thinking and system practice, or simply "added on" to instrumental-strategic reasoning.

It is not easy to follow this rather abstract criticism of an approach like SSM which has the intrinsic merit of being quite empirical. It will be therefore fruitful to meet SSM's problems first on its own grounds. SSM already in its early programmatic declarations was quite sensitive to its political interpretations. It was therefore stated [14, pp. 281ff.] that the social systems theory implicit in SSM covers the area in which influential surveyors of sociological paradigms had located the critical theory of the so called Frankfurt school as represented by Jürgen Habermas. It was claimed that the differences between the two approaches stem from critical theory's more overtly political stance. SSM had not yet any theory of how the structure of society, especially its stratification, might limit fundamentally the range of debate about change. To critical theory's "communicative competence" corresponded SSM's consensual debate that has as criteria of success its usefulness to the actors (instead of its validity for the analyst).

One trouble with this position is that the choice of actors, and especially customers or problem owners, must certainly be "ethically valid" for the analyst who will espouse SSM. An analog point could be made for SSM's easy going admission that stages 5 and 6 of the methodology (comparing conceptual models with reality, and implementing "feasible and desirable" changes) will be inhibited by society's structure. It is further admitted that there are limitations on our creative freedom and on our powers of rational criticism since we and our values are products of existing institutions and past traditions, "but this does not mean that the use of SSM methodology cannot in fact be emancipatory for the actors concerned".

Even disregarding for the time being the questionable WHY-NOT?-flavor of the double-negation above, that I will address later, it should be noted that the position taken by SSM with regard to institutions and traditions displays a "Lockean" bias in the tradition of British empiricism, in that ethical man is assumed to be born as a blank tablet to be written on by institutions and traditions. In other words, we are far from "categorical imperatives". The issue is also the choice of which institutions and traditions are to be fostered, since some of them may be assumed to be better than others in certain ignored respects.

All this, eventually, undermines the trustworthiness of an ethics that tends to get reduced to a communicative competence that, in turn, gets easily reduced further to a kind of politics of negotiations and re-negotiations of perceptions and interpretations of the outer world.

One most conspicuous claim by SSM, in spite of all emphasis on the unlimited negotiability of social (as opposed to natural) reality, is that the negotiations about the method (SSM) are explicitly constrained: "But the way one finds out about [social reality] may in principle be reasonably stable: hence the importance of methodology rather than findings, of process rather than content....The unquestioned prime value embodied in 'a systems approach' is that continuous, never ending learning is a good thing" [14, p. 285, my emphasis]. What is conspicuously missing in the context of the claim is a discussion of what is to be meant by a good learning as opposed to the statement that learning is good. It is therefore symptomatic that both the words knowledge and learning, not to mention ethics, are missing in the subject index of the SSM basic textbook (ibid.). To the extent that at least meaning is considered, it is selfassuredly observed that the social scientist will be reduced to studying not exactly social reality but only the logic of situations, producing findings of the kind "in situation A, a likely outcome is B", without any guarantee that this will hold in any particular situation: "And over the years, with the growth of human knowledge, the 'logic of situations'—which will involve actors' attribution of meaning—will gradually change" (ibid., p. 71). Ethical relativism in SSM is thereby enhanced by the missing dimension of purposes, in spite of the avowed standpoint against "determinists, dictators, or demagogues" (not to mention dogmatists) (ibid., p. 285). It could be instructive to compare the ethical relativism of this unproblematic reference to the logic of situations with other qualified attempts at the origins of moderns social systems theory [20, p. 183ff.].

SSM authors have recently criticized the idea of purposeful systems, and have substituted (e.g. work process) activities for purposes [16] while paradoxically referring to purposeful activities of work processes. Our suggestion [26] is that the attempt to bypass the assumedly old-fashioned question of purposes, as also evident in later developments of soft systems methodology [16; 18], is rather a symptom of the inadequacy of available theories in accounting for the political and ethical dimensions. It leads to paradoxical appeals to concepts such as purposeful activities, effective/efficacious/efficient systems or activities [1, p. 714], culturally feasible or politically acceptable solutions, or agreed-accomodated purposeful activity systems.

SSM's criticism of systems teleology [16] has been formulated in terms of assuming that this teleology implies depictive mirroring views of systems out there in the world, instead of implying a constructive process of model building that captures possible perceptions of the world. This charge against SSM appears to be related to other charges which have been made against dialectical systems theory for allegedly implying that an independent observer (analyst) is able to arrive at an "objective" knowledge of the clients' Weltanschauungen [70, p. 53]. This criticism unfortunately, seems to be a straw-man argument. It may have been motivated by a misunderstanding that arises out of the SSM's unconscious unintended empirical-realistic bias in the dichotomies of idealism-empiricism [22, p. 41, and chap. 12] or idealism-realism [21, chap. 14] that are well recognized in the dialectical-pragmatist systems approach, particularly in the context of "Hegelian inquiring systems" [23]. This may be the reason why the critics themselves so often refer unproblematically to the "real world" without detailing the issue of perception as related to teleology. The dialectics between purposes and activities appears in fact to have been well acknowledged in the context of the criticized social systems teleology, even if the ultimate import of this dialectics could not be ascertained [23, pp. 249ff.]

It may be the case that SSM looks for a new balance between political negotiation and ethics. But the whole ethical context may in fact be too complex to allow us to get away with its renaming and formal reduction to "enacting a social process in which humans seek to maintain or modify relationships according to applied standards which are themselves the product of the previous history of the relationship maintaining/modifying activity" [16, p. 382]. An equally problematic earlier version was the conception of social reality as being "the ever-changing outcome of the social process in which human beings, the product of their genetic inheritance and previous experiences, continuously negotiate and re-negotiate with others their perceptions and interpretations of the world outside themselves" [14, pp. 283f.]. Earlier analogous approaches admitted at least the challenging connection to the field of (natural) law [47, pp. 4.45f., 5.26f., 5.34, A11.10].

Under the label of apparently a-theoretical and non-purposeful multimedia structured discourses, opinion polls, negotiations, or decision support we will then find an unconscious revival of the sociologistic schools of undemanding conversational "interpersonal relations" [58, pp. 135-138] that will be difficult to distinguish from misunderstandings and abuses of an "experiencing of relations...[that] is or is expected to be satisfying in itself" [1, p. 13, quoting Geoffrey Vickers]. This recalls what elsewhere has been addressed as the "Don Juán syndrome" [48, pp. 135ff]. Negotiations with co-workers on the basis of shared information space [2, p.365f] and effective communication that allows both ambiguity and clarity [77] then will not entail purposeful organization of actions under the legitimate constraints of institutions and traditions [48]. They will rather express the fact that "the administration of justice gives way, in a therapeutic society, to a complicated process of negotiation....Justice is fixed by means of ...bargains among interested parties" [58, p. 174]. Such bargains could include certain kinds of negotiations between labor and capital [6, cf. Ivanov, 1986, pp. 143ff.] that constitute the pride of influential Scandinavian approaches to systems development [27].

SSM appears ultimately to be impaired theoretically by what may be important shortcomings besides the question of teleology versus activities considered above. SSM does not consider that activities consume resources. Resources, without theoretical support, are stretched to include even information resources. Activities, however, co-produce in an environment that cannot be defined except in terms of social actors like decision makers. Activities may then be correlates of subsystems in that they must be related to other subsystems in face of the problem of design separability. Because of this the above mentioned SSM information flow models must probably be structured in terms of morphological—functional—teleological classes [23, chap. 3] that require an explicit relating of Weltanschauung to activities, and may destroy the possibilities of the flow metaphor itself, and of the semantic information theory in terms of the theory of "speech acts" that was welcomed by SSM. Fundamental concepts like methods—models—systems are not penetrated in that they are often used as synonyms. Comparisons are made with so called real—world action (in a problem situation) in spite of the disclaimer that the real world can only be controversially perceived and negotiated but not described. Fundamental categories of social actors that are correlates of purposes, e.g. participants—designers—users—end users—(ever changing) organizational members, are used sometimes [17] in an unclear intermingled way in the context of loosely structured negotiations. On other occasions the richness and nuances of social-political roles have been overtly recognized [14, p. 294; 15] but it is this looseness of the structure of practical negotiations through an "open, participative debate" that may be at the heart of the SSM-problems. This is not to say that SSM cannot be or has not been helpful in many practical non controversial consultancy situations where political and ethical evaluations are not paramount, if one only knows how to determine what is not controversial. This is rather a challenge for its evaluation and improvement with a particular focus on basic concepts and primitives or categories.

It is at this point that recent critical systems thinking—CST related to early seminal work [85], may contribute by means of sharpened critical categories that attempt to counter the basic paradox of the ideal models of practical discourse that correspond roughly to SSM's negotiation towards wide agreement and consensus. It is a kind of attempt that has some features in common with the hypersystems approach mentioned below.

Before we go into CST let's remark that it seems to rely heavily on a "Kantian twist" of the influential critical tradition in philosophy. The formulations of CST-LST variants, however, have often been programmatic and somewhat lofty. It is, for instance, asserted [62], that critical systems—CS— are emancipatory in the context of the researcher's ideology and praxis, and that the explicit focus is on the ideology of the individual researcher in determining methodology and outcomes, implying also an extended understanding of praxis. It is a systemic idea of the relationship between theory and praxis in which the two are inseparable [as in pragmatism?]. It uses working methods derived from either the positivist of interpretive camps. Positivist techniques will reflect a perception of agreement on problematic areas, between all those who are perceived as being involved or affected. Ownership of the perception lies with the researcher. Interpretive techniques of interpersonal exploration will reflect a perception of disagreement between people who are also perceived to have power relationships with each other, that will not obstruct this exploration. CS' conceptualization of situations as a series of questions with different contexts, results in "methodological partitioning" taking center stage in the debate, and providing a far richer source of information than if he had been working with just one methodological approach. CS represent the second epistemological break in systems science — the first having occurred when SSM moved away from positivist approaches toward interpretive ideas.

In the context of liberating systems theory—LST— [32, presents a bibliography], it has been proposed [78] that LST developers must answer the question of (1) For whom? Consultants-customers-users-owners-clients, (2) To do what? Goals-activities, efficiency and effectiveness, viability and adaptation, learning and development, all leading to liberation and social justice. The activities will be to support problem owners in their efforts to address their own problematic situations, i.e. in "management" activities that include creating fundamental social change, e.g. destroying oppressive social systems. It is remarked that there is a "serious gap" towards dimensions such as beauty, the aesthetic, the erotic, the spiritual, but this is surprisingly stated to be "worthy of a separate project". [82, instead, suggests that the aesthetic project should not be separate.]

In the context of "interpretive systemology" [37], it is remarked that the opening of possibilities brought up by interpretive discussion should be translated into a process of enlightenment which is rich in political consequences, depending upon the institutional preconditions for practical discourse among the general public. The striving is to gain a pluralist dialectical understanding of social phenomena, the unconcealment of that which normally is taken for granted. Surprisingly and paradoxically, however, this will be done without any disturbing pragmatic intention of the research, since that would be very likely to coincide with e.g. the interests of those in power, having the ultimate effect of stopping the unconcealing drive.

Let us now go back to a recent presentation of CST itself. In a summarizing account [86] it is remarked that ideal models of practical discourse paradoxically presuppose what they are supposed to produce, namely, cogent rational argumentation—the ability and will of all participants to argue in a compelling way and to rely on nothing but the force of better argument. This paradoxical "utopia" seems to be a typical Kantian heritage as represented by the late developments of the tradition of critical social theory. It has certainly been the object of criticism that per se could have been devastating [72, is a valuable example]. It has, however, the merit of stimulating, as in CST, the design of conversation-negotiation systems, and the dealing with argumentation-negotiation break-offs. They are akin to what sometimes has also been called "conversation killing" [59] even if conversation killing may be conceptualized in vastly diverse terms [23, pp. 104f, 119, 172f].

In order to support the identification of justification break-offs and the challenging of boundary judgments or normative presuppositions in systems design, CST groups twelve boundary questions in four classes, each comprising three kinds of categories: social roles, role-specific concerns, and key problems. The four classes ask for the normative OUGHT of (a) the sources of motivation: client, purpose—measure of performance, (b) sources of control: decision maker, components, environment, (c) sources of expertise: designer, expertise, guarantor, and (d) sources of legitimation: the affected people's witnesses, their emancipation, and their world views—Weltanschauung. The twelve OUGHT questions above are then to be contrasted with the pertaining answer to the corresponding IS question, laying open the normative basis of the planning system and its evaluation.

A paradoxically recursive question appears at this point in that the identified or postulated boundary OUGHT judgements remain dependent for their justification on a discursive process of consensus formation—a rational discourse—among the involved and the affected by the system. The normative content can be justified only through the voluntary [informed!] consent of all those who might be affected by the consequences. In order to foster reflection and the discursive dialogue process about the normative implications of systems designs, CST proposes that Kantian "polemical" employment of alternative boundary judgements be used to challenge the expert's own normative presuppositions. Affected citizens can secure for themselves an advantage of argumentation by imposing the burden of proof upon the involved expert who may then be embarrassed for being unable to prove the superiority of his assumed normative judgement. Such an approach will hopefully mediate the conflicting demands of democratic participation of all affected citizens, and those of rational compelling argumentation on the part of the involved planners and experts.

It is then seen that CST attempts to "push" further the struggle for conversation-negotiation by fostering the rational conditions and presuppositions of this process. This is done in the same spirit as early critical-constructive approaches to development of computer applications which aimed at "allowing a gradual learning, self improvement of the information system" [47, pp. 5.37ff.]

It may be instructive to compare the above presentation of CST with the latest program for LST as formulated by the same author [87] on the basis of his earlier "critical heuristics". A critical solution to the problem of the antinomies of pure reason as envisaged by Kant is to seek to lay open the conditions that are presuppositions for our knowledge and understanding, reflecting on the fact that such conditions are never totally given to us through experience and logic. Critical heuristics attempts something analog in the domain of practical (ethical) reason. Rather than concentrating on the theoretical goal of demonstrating the in-principle possibility of "objective" generalizable justification of normative validity claims, it will focus on the task of helping practitioners achieve at least a critical solution. This is done by developing methodological tools and models of rational argumentations, and helping us in the everyday task of dealing critically with situations of imperfect rationality. Churchman's process of "unfolding" in the dialectical pragmatism of his systems approach fosters certainly an ethical discourse but it requires support in its "justification". Dialectical pragmatism is namely strong in driving the process of critical self-reflection but relatively weak in justifying its standard of critique, while the opposite happens in scientism. We need an alternative paradigm of rational criticism, one that would truly free itself from the unreflected presuppositions of scientism and thus might guide us toward a nonscientistic model of critical systems science. The key to such an alternative paradigm lies (for the espoused critical social theory of Habermas) in the language-pragmatic or communication-theoretic turn of practical philosophy. Its implications are a program of communicative rationalization of society. Functional rationalization by empirical-analytic science is then complemented and controlled by the dimension of communicative rationalization that is attained by the institutionalization of practical discourses. [Complementing, however, in this case may not imply integration but rather eclectic juxtaposition]. Such discourses seek to secure value transparency and consensus regarding disputed normative assumptions. This is done by testing the generalizability of the implied norms of action or, at least, the degree to which the propositions in question depend on non generalizable normative assumptions that benefit specific interests. The underlying utopia is the emancipatory idea of a community of free and self-responsible citizens coming together and seeking to achieve consensus on matters of public (nonprivate) concern by means of argumentative, oppression-free will-formation and democratic majority vote. [Pragmatism, however, denies the sharp separation of public and private]. The key strategies toward LST share a common inspiration by Kant's critical philosophy. The underlying concept of critique is one of emancipatory self-reflection with respect to the conditioned nature of our knowledge and understanding. The built-in emancipatory utopia of a community of free, communicatively competent and self-responsible citizens does not preclude a critically handled methodological pluralism, but it is supposed to preclude mere methodological eclecticism and ethical relativism. It gives a systematic place to moral judgement, as well as to the practice of democracy. So long one of the latest presentations of CST.

Before going into details it may be noted that some of those who have criticized SSM's shortcomings from a critical point of view [54] tend to make altogether too unproblematic claims for the rescuing capability of the critical approach. It is, for instance, stated that the only possible justification for implementing the results of a soft systems study must be that results and implementation have been agreed upon after a process of full and genuine participatory debate among all the stakeholders involved or affected. Soft systems thinkers should therefore be critical of all social arrangements which prevent the kind of open, participative debate which is essential for the success of their approach, and which is the only justification for the results obtained. As for Stafford Beer's viable systems model—VSM—what is required is a democratic milieu in which to operate properly: the philosophy of SSM demands communicative competence as the foundation of the process it orchestrates. The goal is man's emancipation from slavery.

This is obviously all right, but the satisfactoriness of CST in these same respects is not clear and accountable unless one happens to subscribe already to Marxism and to Habermas. The tone of the "shoulds" is strongly ethical, not to say moralistic, but the ethical content of Marxism and of Marx-influenced schools of thought is, to say the least, problematic for many people [13, pp. 60-62; 75; 83].

If we leave for the moment these more encompassing issues we may revert to the detailed text of the CST-presentation above. One initial problem that may be observed in the CST approach above is that the idea of "burden of proof" has a political power component that may be also an ethical one. What if those powerful agents who organize the negotiations and who ought to feel embarassed by polemical questions do not feel embarassed? What if they, instead, state for example that "We find that it is too expensive to bear the burden of proof that you want to impose on us" or "It is too expensive to elicit the voluntary consent by all those who might be affected by the system and its consequences", etc. This was recently acknowledged by CST [33, p. 24], but the proposed answers run into serious difficulties like the following one.

One aspect of this issue is what has been elsewhere called the problem of the "WHY—NOT?" strategy, or the strategy of easy questions and difficult answers [53]. The more difficult the answers become because of easily formulated questions, and the lesser the number of knowledgeable potential answerers becomes, the easier will be to disrupt the possible wisdom that has been accumulated in history and tradition. This turns out to be a fundamental debated shortcoming of Kant's thought in what concerns the ability to integrate religion with ethics, to integrate history and tradition with the Enlightenment ideal of an "invisible Church" that along an endless approximation converges at the infinity into a moral society or ethical community [30, p. 49ff.; 80, p. 69f.]. That may be one reason why CST refers to logically compelling polemical arguments [86, p. 281] but not to historically or psychologically-emotionally compelling complementary components such as amply considered, for example, in the pragmatist tradition (William James) or in the rhetorical tradition of antiquity. On the contrary, we may hear that a rational discourse or an ideal speech situation is the one that has the best chance to liberate us from the "historical compulsions" of the past, or to correct "false consciousness" and deliberately distorted forms of communication, or simply "psychological distortions" due to "individual bias", even if unfortunately it may fail to correct psychological (but not logical!), social and physical constraints that have been inherited [4, pp. 216f, as quoted by Gregory, 1989, p. 279; 46, pp. 1209ff.] [39].

This appears to be the Kantian background of the ideal discourse situation where on the one hand certain kinds of WHY—NOT? are deemed necessary for the development of the invisible Church while the questions so posed undermine the legitimacy of the "expert" visible-Church positions that happen to have been conquered in the historical process of sweep-in and endless approximation. It all bears a flavour of "Singerian inquiring systems" in the context of the social systems theory that inspired CST [23, pp. 194ff]. It may also be the background for why teology, and in particular ecumenical efforts and so called natural teology [65, pp. 435ff], may have something to teach us about the conditions for dialogue as well as for ideal and practical discourse where deep emotions and values beyond the cogency of logic are taken into account. All this may be relevant to the effort to oppose the Kantian attempt to dissociate knowing from the reality of the intelligible world, and to subordinate teology to ethics (ibid., p. 189, 191). It may also help to understand why the "social construction of reality" implied in the ideal discourse and possibly implemented in constructive computer applications may occasionally recall the charge that Kantian Protestant Puritan reason is a "system of external espionage" constrained by the laws of an external logic as opposed to an internal psychology. "The external technicism of production in Anglo-Saxon countries — as first programmatically conceptualized by Francis Bacon — stands in the most intimate spiritual continuity with this internal technicism of the ruling of impulses" [79; 80, pp. 84f]. It would be not surprising, then, to see the external technicism of the social construction of reality being promptly implemented in computer applications that put forth claims of having an ethical potentiality. They would be embodyments of an external technical logic that apparently revives prestige words like democracy and participation but works in practice mainly as systems of external espionage. Constructiveness, then, would be trying to obviate the shortcomings of our ethical conceptions, whether Kantian-Protestant or not.

This motivational background of the social construction of reality as appearing in CST ideal models of practical discourse and in the constructive design of computer applications may also illuminate the difficulty that these approaches have in grasping the meaning of the conflict-creating gaps between logic and emotions, between reason and power, and in bridging them. It will be recalled that Kant's critical conception of ethics may be regarded as having contributed to the split of its content on the one hand in logic, and on the other in psychology, akin to the split between law/duty/justice, and love [67, p. 144; 89, p. 81f]. It is therefore a little late, in the modern context of computer science conceptualized as an embodyment of logic, to reinstate ethics in logic by means of Kantian approaches. A superficial sense of togetherness—cooperation offered by ethical constructive communities may have mainly the effect of alleviating the tragic loneliness of the systems hero that has been conditioned by Kantian suicidal misunderstandings of the relations between "logic, ethics, and the ego" [88, chap. 7], and has not been consoled by Marxian or Freudian theoretical complements.

The inability to face emotional, and therefore also political, conflict is evidenced by liberating consultants' often recurrent claim that "so many as possible" involved or affected people be swept participatorily into the constructive ideal discourse. At the same time, however, there is a certain resistance in sweeping in and accounting for the participation of any real "deadly opponent", as it has been recommended in early approaches to constructive quality-design of information systems [47, chaps. 4-5; 50]. In other words, "so many as possible" harmless dialogue partners or supporters, yes; but "deadly opponents", no. Deadly opponents, if not downright malevolent, are at least odd, illogical or crazy. In principle they should participate, especially in jobs far away from the consultants' home-base, but that would be too dangerous or expensive in time and emotional efforts, considering the higher purposes of the assumedly noble endeavour: worldwide democracy through progressive marketing of profitable ideas. So goes often the argument, even in constructive quarters, when the matter boils down to concrete daily development work. The question here is whether CST, seen as a qualification of SSM, can do anything about it.

Participation risks to get reduced to constructive multimedia information technology that is supposed to streamline continuously structured opinion polls in organizations. It is to be hoped that they will profit at least of the lessons from survey techniques, interviewing and psychological testing techniques, or economic and social statistics: validity, reliability, and all the rest [3; 66].

There are other ambitious currents of critical systems thinking which, in the Marxist-Frankfurt tradition, address some of the above issues. They tend , however, to reduce the ethical dimension to the social and political one. The power-ideology —PI— thinking [71], for example, attempts to address the question of origins of the conflicts that SSM and constructive approaches seem to take for granted. These approaches are insightfully envisaged as mainly attempting to fix conflicts or differences in perceptions by means of so called continuous learning that is grounded on relatively simple means-ends schema. It does not even incorporate the pragmatist lessons on morphological-functional-teleological classes in the study of purposeful activities. The emergence of interests is seen as a function of discursive constructive processes aimed at reaching decisions and means of acting on the basis of such decisions [63]. The PI-thinking notes that in so doing one forgets that those processes themselves, the ensuing decisions and possible actions, are all circumscribed by an initial set of unequal and differential conditions of the dominant social system: "Indeed, one wonders what, in the first place, the social struggles are all about in such uncritical conceptions of power" [71, p. 38]. The constructive-relativistic answer is, of course: "Well, let's then have an additional learning dialogue about that".

PI-thinking, however, is soon confronted with the question of the lack of a concept of truth (not of good). Critique presupposes criteria for distinguishing falsity from truth, where provisional truth cannot be equated to the result of constructive negotiations. If true interests cannot be distinguished from expressed preferences, prevailing power relations become the ultimate arbiters of true interests and truth.

It is interesting to note that this kind of problems enables an understanding of why critical theory needed to recur to psychoanalysis. "Related to this is the view that a theory of truth must somehow provide a conception of reason and rational action in terms of which certain forms of consciousness can be said to be ideological and judged to be irrational" (ibid., p. 39). Because critique seeks the true meaning of an ideology in relation to a historical context, it lacks the grounds on which to assert a priori criteria of its own truth, that must be considered as historically conditioned. "Because critique cannot develop formal a priori criteria of what counts as ideology, its strength lies not in a body of theoretical statements from which empirical states of affairs might be inferred but in a theory-dependent method that guides research into the meaning of a form of consciousness by relating it to its contexts of interests and realities. The philosophical implications are that for critical theory not to undermine its own claim to a relative rationality it must criticize a form of consciousness 'immanently'. That is, 'criticism gains its right to impute ideological meaning to a text insofar the text is irrational with regard to its own criteria of adequacy'" [71, p. 40, my emphasis, quoting M. Warren about "Nietzsche's concept of ideology"].

That was not clear. CST claims that the implications for systems theory are that these insights can be read as an attempt to broaden the inquiry beyond the "structural" aspects of systems control. This is done by focusing on the process of organizational power relations and the functioning of ideologies, including the significance of forms of subjective consciousness or the ideological formation of human subjectivities (ibid., p. 46). To me this seems to be rather vague, but it is certainly akin to the spirit of the so called human action theory and action control theory, based on Soviet psychology, that have lately been adduced in the context of interactive computer systems and can be seen as an attempt to qualify constructiveness of computer applications [68].

In summary, it will be noted that PI-thinking senses some important limitations of SSM regarding the consideration of the nature of social control imposed upon a social system, seen as the "product of conscious actions of human beings as makers or victims of history". Nevertheless, the frequent reference to consciousness, forms of consciousness, reason, criteria of truth, rationality, relative rationality, and irrationality, points at psychological issues that once upon a time required the leaning of the Marxian-Kantian heritage on psychoanalysis. It will, in fact, be noted that the psychology of Marx's historical time was either hypertext-like associationistic-connectivistic, or mainly a Hegelian socialization or politization of the science and psychology of Kant, the first master of "forms of subjective consciousness" who still related to ethics and religion.

Today we need no more to keep ourselves circumscribed to this heritage. There is the option of Jungian analytical psychology, conditioned by Kant's phenomenology or phenomenalism, Nietzsche's cultural criticism, William James' pragmatism and by psychoanalysis itself. It is true that such conception of human (ir)rationality may be accused of implying that ideology is a "naturalistic", instead of being a "historical" phenomenon [70, pp. 4ff.]. By naturalistic it is then meant that it is essentially rooted in unhistorical forms or in the universal character or essence of human nature, in the psychical structure of the individual, in the innate predispositions of the human mind, in the non-logical preconceptions or prenotions inherent in the human intellect, or in the unconscious impulses, instincts, and human passions and desires. By contrast, the historical conception views ideology as produced and reproduced through human practice, reflecting the historical development of men's social relations.

Disregarding the embarassing fact that PI itself apparently regards reason and rationality as rather "naturalistic", and has serious troubles with its so called criterion of truth, it must be noted that analytic psychology's conception of prenotions allows for both a historical-political dimension and an ethical-religious one [69; 74], even if these aspects have not yet been studied as much as for psychoanalysis. If, under such premises, ideology is still to be regarded as naturalistic, so much worse for ideology.

There is a growing evidence that the answer to these questions must be conquered at the interface between philosophy and religion, in particular Christianity. The striving for progressive amplification of the ethical discursive community and the construction of growing social relations, family—nation—humanity, may well be conceptualized for instance in Christian terms in the context of studies of human action that bridge the Kantian disjunction between theoretical and practical reason [7; 29, p. 230f].

It will be noted that at least of couple of systems theorists who have been very concerned with systems ethics [24; 85] have relied very heavily on a Kantian philosophical basis. SSM is phenomenologically influenced, and critical social theory was originally influenced by the Kantian revolution developed further through Hegel and Marx, and supported by psychoanalysis. It will also be noted that a recurrent persistent theme in most, if not all attempts to develop a social systems approach has been the reliance on the basically social idea of "so many as possible", "communication", "emancipation", etc.

These key words relate mainly to Jürgen Habermas' critical social theory. It has been remarked [56, p. 1130, my trans.] that there is undoubtedly a persistent deficit in the anthropology constructed by Habermas, which can be related to his moralism of Protestant inspiration. It marks a deviation in comparison with the first generation of the Frankfurt school in what concerns its treatment of the emotional sensory world, even if the theory of communicative action would make place for the idea of "bodily attraction". For the rest, Habermas seems to relativize progressively the psychoanalytic model, and to turn more and more towards the idea of "reconstruction". It indicates an ex-planation [mise à plat] of what is implicit, i.e. a systematic explanation of a pre-theoretical knowledge that is mastered only at the level of practice. The emancipatory interest and the utopic-prophetic dimensions of neo-marxism inherited from earlier critical theorists have come to be played down. Instead, Habermas seems to approximate further a rationalism that is inspired mainly by the tradition of Enlightenment, and a "realism" that emphasizes the growing convergence with Max Weber. In effect the exigence for emancipation has been confirmed to be of a rational nature, and to be conveniently rooted in the "structure of language". Above all, Habermas emprehends the reconstruction of the materialistic and historical genesis of reason, whose "procedural" nature is engendered by language itself, since "with the first pronounced phrase, a will of universal and unconstrained consensus expresses itself without ambiguity".

It seems to me that it is less attractive to subscribe to this account of Habermas' development that, for once, is less loaded with prestige words. In any case it is possible to reflect on the fact that critical social theory tends more towards abstract reason and towards language (as an obvious ethical tool?) than towards ethics itself which, as a concept, does not appear to have a central place in the discourse. The message has a somewhat prophetic tone, revealing a new participatory dialogue-prophecy that has not yet been implemented within the university itself, not even in the limited "laboratory" conditions of its own institutional setting. Habermas, in fact, apparently ridicularizes the "old mandarins" or at least questions the nineteenth century's belief in the totalizing power that the inner dynamics of research work would have [44]. The new mandarins of the participatory prophecy, however, apparently do not realize that the university, on the basis of reason as substitute for religion, can be seen as one of the most extensive and intensive historical attempts to start implementing an ideal discourse in society as a whole. The inner dynamics of research corresponded to the inner dynamics of reason. In all its social complexity the university would serve as a forerunner and model of a progressively "democratic" society based on reason. [5, pp. 123f.; 81].

Prophecies recall the concern for the ethical dimension. Nevertheless it has been difficult to clarify its theoretical status, and whether it was subsumed under the social-political dimension, and if so, on what grounds.

Let's recall possible questions about Kantian ethical thinking. I am not a professional philosopher, even if philosophical pragmatism hints at the necessity of everybody attempting to be one. In any case the following lines will be pragmatically oriented towards supporting one main intuition, in spite of Kantian reason having been "canonized" by the philosophical community. It is the hypothesis that both SSM and CST share some basic assumptions about the problematic subsumption of ethics under politics. It is also the hypothesis that this is not good, and that it leads us into a kind of vicious circle, in that it tempts systems scientists to develop themselves and their trade by pulling up themselves, as it were, by the hair.

Since Kant's work is extremely complex it would be prohibitive, especially for non professional philosophers, to risk to get caught up and seduced by plunging into it. To search for possible shortcomings in Kantian thinking by following the threads of Kant's own original work [55] implies a temptation to espouse his own concealed presuppositions incurring furthermore in the WHY-NOT?-problem mentioned above. It implies also a superficial non-hermeneutical view of "what Kant actually meant". Because of this it seems convenient to complete occasional readings of and about Kant, as his thought has been considered relevant in the systems context [23, chap. 6; 24, chap. 6; 85], with an outline of the criticism that historically has been directed against Kant's work.

It may be particularly informative to choose a summary of this criticism as it has been surveyed by somebody who nevertheless defends Kant's ethical positions [13]. The opposition to Kant, the launcher of modern critical philosophy, came from several diverse quarters as represented by the key schools of idealism (G.W.F. Hegel, Karl Marx, Benedetto Croce), phenomenology (Georg Simmel, Max Scheler), empiricism and neopositivism (Betrand Russell, Hans Kelsen), realism (Jacques Maritain), and others (J.G. Hamann, J.G. Herder, F.W.J. Schelling). In spite of much emphasis on critical philosophy and much reliance on Kant we have heard very little about this criticism of critique, with the possible exception of the wholesale acceptance of Marxist criticism and wholesale dismissal of the ignored but conjectured positivist-empiricist criticism.

A study of the opposition to Kant is, however, strongly recommended to all of us who wish to advance systems thinking in the political-ethical aspects, and in close contact with the formal sciences that are embodied in the ongoing so called computer revolution. Now, back to the announced questions.

1) Kant formulates many fundamental and sharp distinctions that are very controversial, and have very definite consequences in defining "gaps" that consequently are hopefully to be bridged by (paradoxically Kantian?) systems approaches. But: "distinctions are the weapons of the elite"? [23, p. 270]. Or, is it lack of religiosity that is the main weapon? [10, p. 93]. Some such distinctions that have become "classical" and are apparently implied in CST are, for instance: theoretical-pure and practical-ethical reason, ethics and religion, gnoseology and ontology, reason and will, noumenon-spontaneity-freedom and phenomenon-causality-necessity, virtue and happiness. Do we understand why we are willing to accept these distinctions?

2) Kantian thought emphasizes the notion of science seen as universal and necessary knowledge (of "nature"). By so doing it may have unintendedly favored misunderstandings leading to prestige-inflation of natural and formal science, such as in scientism, not the least in the context of systems thinking. Which should be the place of "lesser sciences" like psychology and anthropology, compared with logic and mathematics, in critical thinking? What place would CST grant to "a metaphysical psychology to un-Locke our ailing world" that develops the meaning of all those perceptions that allegedly contribute to false consciousness? [42.]

3) In opposition to Hegelian thought which tends to believe that all political institutions have a moral value, fostering a reduction of ethics to politics, Kantian thought finds that both law and politics must be judged by ethics. Such an ethics, however, was conceived in terms of a concept of freedom (cf. liberation) that is metaphysical and relates to the person independently upon social-political institutions. This is what differentiates Kantian thought and original revisionistic social-democratic thought from so called liberation philosophies (Hegel, Marx, Storicism). Such liberation philosophies have been accused to deny transcendence. They "divinize" man and, in spite of all emphasis on communicative competence, close definitively the door to discussions about theology and religion. In fact I have not seen these words being mentioned in the CST literature, and CST has still to develop the meaning of liberation.

4) Kantian ethics calls for an appreciation of moral imperatives that could enable an ethical conversation, when other imperatives as contained in holy books have been "explained away". It is, for instance, possible to conceive of a computer program that not only works along the Kantian categories of systems definitions, but also counters every fundamental proposal for action in an "activity system" with the question of whether one would like to see the maxim of his action enpowered to become an universal law. Or rather: "Would you act as if the maxim of your proposed action had to be erected by your will to a universal law of nature [in its broadest sense of "form"]?"

Furthermore it will be observed that Kantian thought considers several duties that may also be seen as enabling ethical conversation, i.e. fostering both "future generations" and "conversation pushing" as opposed to conversation killing. These duties are dintinguished as being duties towards oneself, (e.g. concerning suicide and lies), and towards others (like charity, gratitude, and sympathy). Should such important aspects of an ideal speech situation be considered in future computerized CST-models of hypersystems? Or will this suggestion be accused of Kantian liberalism, or for attempting a reversal towards early critical theory's reliance on psychoanalysis and naturalistic thought? What about C.G. Jung's analytical psychology instead of Freudian psychoanalysis for taking care of feeling and perception? [42.] Does Kantian "empiricization" of psychology and anthropology, as contrasted to the universalization of reason and logic, contribute to CST's and to the computer revolution's emphasis on the latter to the detriment of the former, as well as contribute to SSM's poverty in the treatment of "perceptions" of the world? Why are such words as charity, gratitude, love seldom if ever mentioned? And finally:

5) Does the (unavoidable?) failure to implement (self-doomed?) Kantian wishful thinking on "transcendence" of ethics into religion throw us into the arms of political ethics, an ethics in the form of constructive-ideal speech situation, akin to a vague concept of "divinized" Democracy that exposes us to demagogical technicalities and Faustian bargains in the computer-network revolution?

In summary: while CST refers unproblematically to "an untenable pre-Kantian understanding of rationality" [33, p. 8], others [41, in the introduction, p. 6f, 13] recall that thinkers like A. N. Whitehead sought to return to pre-Kantian modes of thought, circumventing the Kantian critique by correcting some errors in pre-Kantian philosophy: both Whitehead and Jung are postmodern in that they both reject tenets of the modern worldview (Weltanschauung), yet without returning to a premodern approach. They retain the formal commitment of modernity to rational empiricism, but they reject some of the substantive presuppositions of modernity.

In the search for promising development of the ideas above one way could be to explore computer implementation of the systems concept that has been proposed in terms of the idea of co-constructive hypersystem [36]. The core of the hypersystem idea would be to keep track, by means of adequate "pointers" or links between simple or composite nodes of hypertext computer software, of the relationships between various affected social actors' instantiations of the "primitives" (basic social systems concepts) during the process of solution or learning. The question about which social actors or agents will be thus swept-in, and in what kind of communication or self-reflection, may be discussed in terms of error-accuracy [47; 50], of the classical issue of "power and ideology" [71], or "metaphysical psychology" [42], but also in terms of ethics and religion.

The primitives mentioned above could be the categories of the original dialectical social systems approach [24, pp. 79ff.]: (1) Clients, purposes, measures of performance, (2) Decision makers, components, environment, (3) Planners, implementation, guarantor, and (4) System philosophers, enemies of the systems approach, significance. They could be also CST's categories reviewed above, or SSM's so called CATWOE-categories: (1) Customers, (2) Actors, (3) Transformation processes, (4) Weltanschauung, (5) Ownership of the system, (6) Environment. At an apparent level of meta-systems it has also been suggested that various paradigms of information systems development be described and interpreted in terms of the categories: (1) Key actors (the "who" part of the story), (2) Narrative (the "what", or the key activities), (3) Plot ("why" did the action take place, akin to causes and purposes), (4) Assumptions (the fundamental beliefs or Weltanschauung, or epistemological-ontological assumptions) [46]. Several authors have recently liked to create their own variants of the original categories, without affording to ground their categorizations in basic considerations of controversies in philosophy and in scientific method. The difference, of course, is that certain sets of systems categories (notably Churchman's) establish demanding and commiting relations between categories, fostering ethical disciplined thinking. So, for instance, one would not talk, as in SSM, about "environment" without a commitment about who is/should be the decision-maker, and which are the clients' purposes.

Reference is often made to a learning process. Such a process is rather a dynamic continuous learning-follow-up of the particular systems problem. There will be communication, opinion polls, dialogue-negotiation, conversation or discussion, but they will be mainly or initially about certain primitives (primitive concepts) and their relations, a certain model or structure, with particular functions or goals, and they will strive to involve at least certain particular actors or role bearers whose absence would hide the problems of power and of emotions [47; 48].

The question is whether the hypersystem idea in the outlines of its particular computer implementation in terms of co-constructive computer applications [34; 35; 36] or other similar approaches [64] will be able to meet social-political and ethical realities that have been exemplified in other comparable contexts [26; 33; 38; 48; 49; 50; 71]. The so called learning process that is based on the dogma that learning is good, but where neither the meaning of (good) learning nor purposes or constraints are discussed in order to build the ground for some political-ethical evaluation, opens the doors for endless expensive consulting, negotiating, innovating, and purchasing of computer equipment that works as a communication-negotiation "shell" [9; 40]. If we welcome this new technology, what about the old knowledge? [51.] The consultant may be akin to "the analyst as facilitator" in the paradigm of social relativism [46]. Working for the clients or problem owners may be just an unintended contribution to make our world more frenetically commercial as one grinds out a vast array of products and services to assuage the hungry appetites of world citizens, wherever they may live in the new global democratic economy [12]. The social construction of reality may eventually run into the SSM negotiation process mentioned above, into a search for wide agreements and consensual activity models. Such constructiveness can easily turn out to be the expression of a basically technical idea, a generalization of the very idea of technology which, either supported by constructive computer applications or not, claims to incorporate the human sciences [25, p. 25]. It may therefore easily appeal to social engineers who believe that they will build up revolutionary social ideas into technical equipment. Therefore the challenge may be to turn the analyst-as-facilitator in the spirit of social relativism either into the analyst-as-labor-partisan in the paradigm of radical structuralism or, (preferably?), in into the analyst-as-emancipator or the analyst-as-social-therapist in the paradigm of neohumanism [46]. That might be all right, if one only knew a good meaning of these terms, and no other alternatives were available.

The success of any computer implementations will then probably depend to a great extent upon the soundness and the implementability of the proposed categories. They, in turn, rely on the soundness of the Kantian, neomarxist and pragmatist political-ethical conceptions, which have been objects of an extensive criticism that is apparently ignored in our contexts and in recent developments of constructive social systems theory. In this context one may wonder whether the forms of poetry or of dramatic dialogue have any relation to categories, and whether particular categories in a Kantian sense, including logic, place any fundamental restrictions on the human power of expression in its dependence upon content. The rhetoric tradition of argumentation [31] may inspire some alternative approaches [45, is just an attempt]. Compare, about song: "The dimension disclosed by the tones can certainly be called 'inner life', but it is not the inner life of the subject as opposed to the object; it is not the inner world of the self but of the world, the inner life of things. This is precisely why the singer experiences inner life as something he shares with the world, not as something that sets him apart from it....Music prevents the world from being entirely transformed into language, from becoming nothing but object, and prevents man from being nothing but subject." [43, quoting the musicologist and philosopher Victor Zuckerkandl].

In the best case it may be found out that the importance of our technical implementation efforts in hypersystems is akin to the importance of Robert Pirsig's motorcycle in fostering critical reflection on philosophical issues, like the purpose or meaning of a trip, that per se may improve future system thinking [73]. In the worst case it may be found out that the importance of the efforts "has less to do with their direct usefulness to organizational decision making and more to do with their symbolic value", promoting an image of leading-edge consultancy and facilitating the marketing of other products [38, p. 282f.] Even in this latter case there are possibilities for legitimate meaningful research in regarding the computer as an image in a psychological Rorschach-projective test [84] or as a mathematical-psychological construct [52].

My growing conviction is that provisional answers to many of the above questions must be sought along the developments of dialectical pragmatism and dialectical systems theory [23; 24]. Critical systems thinking [85] may help us to sharpen the political and ethical dimensions of systems thinking, but only if pushed to the interface between philosophical ethics, analytical psychology, and reconquered reinterpreted religions. I am thinking, in particular to the religion closest at home, i.e. Christianity, not because it should be the most "attractive" but because it most likely conceals our deepest presuppositions and has been the most studied in the context of science [8; 11; 28; 60; 76].

Computer applications actualize the basic shortcomings of Kantian-Marxian normative thought regarding the gaps between formal science, human science, ethics, and religious thought. Such applications might in the future try to incorporate some of the suggestions in this paper by "imposing" interactively to the computer users ethically loaded questions that, if possible, should also be aesthetically, e.g. poetically, attractive. Such questions, to be interactively related to each other and to various agents' answers, could deal, for instance, with systems primitives or categories, boundary judgements, categorical imperatives. Instead of working mainly as systems for external espionage, for consensus and negotiation, or for marketing, the system of questions would be especially designed to awaken the constructive ethical concerns of those who succeed in getting involved in the technical equipment. We would, at the same time, be on guard against the "technological imperative" at least as much as we do against ethical imperatives. It may be the case that anything that is worth doing in these respects must be done without any technical equipment.