Printer Friendly

Generative Conversations: Applying Lyotard's Discourse Model to Knowledge Creation Within Contemporary Organizations.

   I don't think there is any discourse without efficacy. (Lyotard, 1985, p.


The goal of generative conversation heuristics is to provide support and guidelines for participants who wish to collaborate in creating new knowledge within organizations. The structural forces governing conversations often limit and regulate what individuals can say within organizations. These forces emanate from the rules which guide the formation of concepts, objects and themes within conversations (Foucault, 1972). Foucault (1972, p. 107) describes discourse as `the group of statements that belong to a single system of formation'. There is a system of rule elements and relations that enable individuals to make statements. Statements are seen as articulations that exist because of a series of conditions that function at a certain time and place. A discourse is a group of statements that share the same formative system. This approach sees conversation as emerging from a system of local rules and relations that enable individuals to say what they say. The issue for Foucault is not what things said mean, but how and where statements are able to come into existence and remain in circulation and disappear. Within an organization there is not one single discourse but many, depending on specific contexts and systems of formation.

Foucault (1972, p. 15) distinguishes between two types of knowledge: `connaissance' and `savoir'. Traditional knowledge (`connaissance') concerns `the relation of the subject to the object and the formal rules that govern it' (Foucault, 1972, p. 15). It is at the level of know-how, at the level of a discipline or body of knowledge. Formative knowledge (`savoir') refers to `the conditions that are necessary in a particular period for this or that type of object to be given to connaissance [traditional knowledge] and for this or that enunciation to be formulated' (Foucault, 1972, p. 15, brackets mine). Foucault (1972, p. 38) defines rules of formation as the `conditions of existence (but also of coexistence, maintenance, modification, and disappearance) in a given discursive division' (body of knowledge). Systems of formation are divided into four interrelated areas:

(1) the formation of objects;

(2) the formation of statement modalities;

(3) the formation of concepts;

(4) the formation of strategies or themes.

Each of these divisions is organized into three elements that interact in the formation process. Lyotard (1988, p. 29) also argues that conversations are regulated and governed by certain stakes. It is these patterns that have to be changed if conversations are to become generative and creative. The modern industrial organization (through its communication structure, methods, and culture) bounds and limits the type and nature of conversations that can be held within it. This is feasible in an industrial economy where the stakes are predetermined as profit, efficiency, and production of goods. In knowledge-based companies where the production of new knowledge becomes the stake, the limiting of conversations needs to be minimized. Lyotard (1984, p. 17) argues that even in modern institutions `the limits of the institution imposes on potential language "moves" are never established once and for all (even if they have been formally defined)'. Lyotard (1984, p. 43) identifies `two different kinds of "progress" in knowledge: one corresponds to a new move (a new argument) within the established rules; the other, to the invention of new rules, in other words, a change to a new game'. These are comparable to the notions of first- and second-order change (Watzlawick et al., 1974). First-order change emanates from solutions that are logical within the current context of rules. Second-order change occurs when the context of rules itself is changed. In order to support the creation of new knowledge within organizations we need to reverse the normal process of regulative conversation. Regulative conversation assumes certain stakes and restricts the kinds of moves (phrases and statements) allowed. Such an intervention must free individual subjects to collaborate in the formation of new stakes, patterns and themes.

This paper aims to provide heuristics for generative conversations that create new knowledge and change the patterns produced by the formative system.


Lyotard (1984, p. xxiv) adopts a postmodern perspective in which he argues that `The society of the future falls less within the province of a Newtonian anthropology (such as structuralism or systems theory) than a pragmatics of language particles. There are many different games -- a heterogeneity of elements. They only give rise to institution in patches -- local determinism' (my italics). Jackson (1991, p. 299) argues that there are issues raised by postmodern philosophy (and I would argue the emergence of knowledge-based organizations) that have `important implications for systems thinking and practice'. Four specific issues are highlighted:

* Logic and order -- the feasibility of these in systems is questioned by postmodernism.

* Progress -- performance and emancipation are considered potentially dangerous traps.

* Power -- central to a postmodern view, is ignored or simplified by systems thinking.

* Language -- assumed transparent by systems thinking, assumed deceptive by postmodernist philosophy.

This paper explores some of these issues by providing heuristic support for generative conversations.


In his analysis of the condition of knowledge in postindustrial society Lyotard (1984) emphasizes the pragmatics of language and how they affect knowledge. He focuses on the effects of utterances on senders, addresses and referents. Lyotard (1984, p. 10) following Wittgenstein (1953) focuses on language games as `modes of discourse' where `each of the various categories of utterance can be identified by rules specifying their properties and uses to which they can be put'. Wittgenstein (1953, p. 5) makes it clear that language games are pragmatic and systemic: `I shall also call the whole, consisting of language and the actions into which it is woven, the "language-game".' The pragmatic nature of language games is again highlighted when Wittgenstein (1953, p. 11) states that `The term "language-game" is meant to bring into prominence the fact that the speaking of a language is part of an activity or a form of life.'

Lyotard (1993, p. 21) positions himself against `an ultimately empiricist notion of language use in Wittgenstein's writings. In these terms people make use of language. They play at it.' The argument against such instrumentalist assumptions is the fact that people `do not know all the rules of the language games' (Lyotard, 1993, p. 21). Gadamer (1981, p. 50) agrees in arguing that `Language is not an instrumental setup, a tool, that we apply, but the element in which we live and which we can never objectify to the extent that it ceases to surround us.' Lyotard (1985, p. 51) suggests that it is not that people play language games, rather games `make us into their players'. He argues against an anthropocentrism that assumes human subjects are in control of language games. Subjects are instead positioned within language games (Lyotard, 1984, p. 15). It is only from these positions that subjects can make moves (utter phrases or statements) in the game. Conversations have stakes and `when the stakes are attained we talk of success' (Lyotard, 1988, p. 137). The conflict and agonistics of language games are `not between humans or between other entities; rather, these result from phrases' (Lyotard, 1988, p. 137).

In using language games to analyse contemporary knowledge, Lyotard (1984) defines two principles. Playing and being played falls into the `domain of general agonistics' (Lyotard, 1984, p. 10). If we are to understand social relations (and generative conversations as social relations) we require `not only a theory of communication, but a theory of games that accepts agonistics as a founding principle' (Lyotard, 1984, p. 16). This is true even of collaborations that have the declared intention of new knowledge creation. The emergence of new knowledge is secondary to the game. Humans are positioned in language games but are not indifferent, `but behavioral or strategic -- in other words, agonistic' (Lyotard, 1984, p. 57). In a note on the above Lyotard (1984, p. 100) references Gilles-Gaston Granger (1960, p. 142) as stating that `probability reappears here, no longer as the constitutive principle of an object, but as the regulating principle of a structure of behavior' (my italics).

Complementing this is the second principle that `the observable social bond is composed of language "moves"' (Lyotard, 1984, p. 10). Statements and phrases are moves that position participants in conversation games, which are governed by stakes and rules. The games, their stakes, rules and moves are `the minimum relation required for society to exist' (Lyotard, 1984, p. 15). Inquiry into the social bond is itself a game since `it immediately positions the person who asks, as well as the addressee and the referent asked about' (Lyotard, 1984, p. 15).

Lyotard (1984, p. 10) points out that language games:

(1) have rules that are `the object of a contract, explicit or not, between the players'. These rules, however, `do not carry within themselves their own legitimation', and are `not necessarily invented by the players';

(2) are defined by their set of rules -- `if there are no rules there is no game, that is even an infinitesimal modification of one rule alters the nature of the game, that a move or utterance that does not satisfy the rules does not belong to the game they define';

(3) that conversations are games -- `every utterance should be thought of as a "move" in a game'.

Conversations have three pragmatic poles and can be thought of as a `pragmatic triangle' (Lyotard, 1985, p. 71). There is a sender who says something, an addressee who listens or receives, and a referent -- that which is being spoken about. The ways in which the pragmatic positions (poles) can change define the language game. Each conversation has a `specific pragmatics' -- meaning that any utterance (move) in context has a necessary `effect on the world' (Lyotard, 1985, p. 52). The effect is primarily in the repositioning of the participants; an addressor becomes an addressee, or visa versa. A move (phrase or statement) may keep the addressor and addressee positions the same but change the referent and thus the content direction of the conversation. A common conversation pragmatic in organizations is the manager--subordinate positioning in which the subordinate is expected to remain in the position of addressee while the manager decides the referent.


Lyotard's key philosophical work is The Differend (Lyotard, 1988). He reconceptualizes conversation by focusing on the phrase as unit of analysis. Phrases present a universe consisting of instances: an addressor, addressee, a referent and a sense. The addressor and the addressee are not independent of the phrase -- it is not a message passing between them. The phrase is a system, a constellation that `is defined by -- as it, in fact, defines -- the situating of its instances (addressor, addressee, referent, sense) with regard to one another' (Lyotard, 1988, p. 193). The phrase is `is not a grammatical -- or even linguistic entity ... but a pragmatic one' (Lyotard, 1988, p. 193). Phrases `which are moves in language games' (Lyotard, 1993, p. 21) may include gestures, music and signals. The system of instances that a phrase indicates is as much context as it is text. When considered this way any differentiation between text and context is meaningless.

The instances of a phrase can be simplified as follows (Lyotard, 1988, p. 14):

* Referent: what a phrase is about, the case, a pointer to `reality'.

* Sense: what is conveyed, expressed and signified about the case.

* Addressee: that to which the sense of the case is addressed.

* Addressor: that from which, or in the name of which, the sense of the case is addressed.

The interrelationships of the instances are arranged in the phrase universe. There may be none, one or many of each of the instances in a phrase. In presenting his thesis Lyotard (1988, p. xii) outlines a structure in which a phrase -- as a constellation of instances -- is constituted according to rules. These rules make up regimens such as `reasoning, knowing, describing, recounting, questioning, showing, ordering, etc.' One cannot translate a knowing phrase into a questioning phrase. However, phrases from different regimes can be `linked one onto the other in accordance with an end fixed by a genre of discourse' (Lyotard, 1988, p. xii). This second level of rules (genre) links phrases (pragmatic constellations) in teleologies. Genres of discourse have goals; there is something at stake -- `to know, to teach, to be just, to seduce, to justify, to evaluate, to rouse emotion, to oversee' (Lyotard, 1988, p. xii). Phrases are linked in terms of these ends.


Phrases have to be linked (silence is a phrase) -- `to link is necessary; how to link is contingent' (Lyotard, 1988, p. 29). The necessity of linking is ontological, the necessity of there being a next phrase `is a presupposition for "objects", for their "witnesses" and so on' (Lyotard, 1988, p. 66). This is the cornerstone of Lyotard's approach to discourse; we have to link phrases -- human reality depends on it. A `differend' occurs where phrases cannot be linked, where world-views encounter each other. It is where different language games or genre meet. It is `a case of conflict, between (at least) two parties, that cannot be equitably resolved for lack of a rule of judgement applicable to both arguments. One side's legitimacy does not imply the other's lack of legitimacy' (Lyotard, 1988, p. xi). The differend is that state where feelings are not yet expressible in phrases. Often a new link between phrases needs to be created. It is important that the differend is not `smothered' by litigation; we should rather search for idioms that can express them (Lyotard, 1988, p. 13). In organizations the differend may be a creative source of knowledge. This source of new idioms and knowledge is wasted if it is not respected, and if strangled could lead to alienation of people and mere compliance. Such is the case in strong cultures where there is a `monopoly on procedures for the establishment of reality' (Lyotard, 1988, p. 4).

In such situations silence often occurs as an alternative phrase, and may be a comment on the addressor or addressee's competence with regard to a certain case (referent). It may be an indication that the situation does not exist or cannot be signified (Lyotard, 1988, p. 13). Silence does not indicate which instance or instances of a phrase (addressor, addressee, referent or sense) are lacking ability and negated. If we apply Lyotard's (1988, p. 14) notion of silence as a phrase to silences in the organizational conversation we can recognize one or more of the following:

(1) The addressee lacks the competence required in order to be spoken to about the case.

(2) The addressee is not considered worthy of being spoken to about the case.

(3) The case does not exist.

(4) There is nothing to say about the case since it is senseless, or inexpressible.

(5) The addressor lacks the competence to talk about the case.

(6) The addressor is not considered worthy to talk about the case.

(7) The addressee does not recognize the authority of the addressor.

The differend asks for expression, and the pain of silence or the pleasure of expression reflects this lack or creation of a new idiom (Lyotard, 1988, p. 13). An organization's conversations die or are kept vibrant through the suppression or resolution of the differend. Generative conversation can be viewed as an attempt at differend resolution, through the generation of new idioms that fuse (link) normally separate conversations and phrases. In generative conversations the different knowledge and expertise that are brought together make fertile ground for differends. The key challenge is to find ways

of linking phrases so that new themes and stakes can emerge. The ethical paradox is that in order to ensure that no genre dominates, some structuring (domination) needs to occur. This is the fundamental paradox of the postmodern attitude where `the justice of multiplicity: it is assured, paradoxically enough, by a prescriptive of universal value' (Lyotard, 1985, p. 100).


In providing structural support for generative conversations our purpose is to free up the possibility of participants making creative phrases (moves) and links. This is a bottom-up approach that lets the links and relations develop themes and stakes. It focuses on relations between phrases and lets the system (conversation and its stakes) emerge from these relations. This approach is in contrast to traditional regulative systems approaches that start with purpose and deduce parts and relations. This open, bottom-up approach encourages participants to explore `the possibility of no longer being, doing, or thinking what we are, do, or think' (Foucault, 1984, p. 46). I have arranged the heuristics into five steps.

Step One: Becoming Aware of Games

The strategy, goals, and objectives that govern organizations also regulate phrases and linkages. Participants become trapped in regulated conversations that to some extent shape how they think, speak, and act. There is a way out of this trap. When practising generative conversations it is necessary to first explain the concept of language games to the participants. This enables participants to see the nature of regulative conversations and to avoid becoming trapped in them. Watzlawick et al. (1974, p. 99) identifies Wittgenstein (1956, p. 100) as the first to point out the fact that once we become aware of the game we are in we `can no longer naively go on playing'. Once we are aware of the regulative nature of normal organizational conversation we can take steps to free ourselves. Our strategy is to replace one game with another. Generative conversation is a game in which our ability to link phrases is the stake. Instead of letting traditional organizational goals unconsciously govern our linking, we consciously link phrases in any way we wish. Generative conversation is practised when we are attempting to create new knowledge; it is creative, divergent, and builds new relations between previously separate bodies of knowledge.

Step Two: The Rule

There is only one rule in generative conversation and that is: always link to the previous phrase. There are four possible instances that may be linked to: the addressor (AD), the addressee (AS), the referent (R), or the sense (S). Some examples of linking possibilities are as follows:

* Apply another S to the same R.

* Apply the same S to a different R.

* Link a new R to the current R.

* Unpack the detail of the S or R.

* Describe the containing R.

* Describe the history or future of the R.

* Link the R to the history or future of the AD or AS.

* Redescribe the S or the R through metaphor.

* Switch the S from the R to the AD or AS.

Step Three: The Guides

* Generative conversation is a game in which we play with ideas, not against each other.

* It is useful to appoint a facilitator at the start to monitor the application of the linking rule.

* There is no rush; regulative conversation occurs at speed.

* Allow at least three seconds of silence between each phrase.

* Watch the pull of habit and pattern. Be aware of the tension to link in a certain way.

* Keep a notebook to jot down ideas so that they are not forgotten.

* Questions can form part of the conversation but must obey the linking rule.

* Make use of creative misunderstanding.

* Listen -- breath -- think -- link.

* Remember silence is a phrase.

The following acronym serves as an overall guide to linking:
   Listen to the whole phrase. Inhale -- take a few breaths. Nurse the current
   theme. Kerb your initial reaction. Invent. New moves and patterns. Gently.

Step Four: Setting a Broad Context

Goals and focus regulate conversations. Generative conversation begins with linking and lets themes and focus emerge as they may. It is a process of relating that supports the possibility of thinking and saying something new. It can be used as an underlying practice for any creative group process. A systems method may guide the group but all interactions are based on the linking rule. When generative conversation is practised as a stand-alone approach, the starting context of the conversation needs to be as broad as possible. By being too specific in defining the issue of concern we may exclude possible phrases and linkages.

Step Five: Capturing New Stakes and Themes

If generative conversation is used to support other systems methods then natural themes will emerge from the structure of the containing approach. When stand-alone is used as a method to open up thinking then a phase of reflection is needed. The initial stake of any generative conversation is the ability to link to the previous phase. In practice the conversation develops lives of its own as themes first emerge, and then begin to regulate further linking. It is within these themes that the seeds of new knowledge (in the form of new objects and concepts) reside. Once these themes and their stakes are identified they can be explored using contemporary systems methods. The stakes are redescribed as a new system's purpose or measures of performance. They are then explored using Checkland and Scholes' (1990) idea of a root definition (a system to do x by y in order to achieve z) and accompanying activity system or they may be modelled as viable (Beer, 1985) or dynamic systems (Forrester, 1961).


The twin idea of games and play are key in Gadamer's understanding of language and dialogue. `The basic constitution of the game, to be filled with its spirit -- the spirit of buoyancy, freedom and the joy of success -- and to fulfil him who is playing, is structurally related to the constitution of the dialogue in which language is a reality' (Gadamer, 1976, p. 66). It is in such a spirit that we can use Lyotard's ideas in guiding generative conversations. We need to play at and be played by generative conversations always watching for the emergence of new themes. We treat our conversation as if it were an open system. This bottom-up approach appropriates some postmodern philosophy (the differend, phrases and linking) into a systems practice, while preventing the systems idea from regulating our conversation. In approaching generative conversations as play -- as a meta-game of knowledge -- the metaphors of system, language game, and differend support `the possibility of no longer being, doing, or thinking what we are, do, or think' (Foucault, 1984, p. 46). As in any play or game each conversation will be particular with different movements and linkages of phrases.

The systems idea can serve as a guide for linkages. Viewing generative conversations as open systems may help us in linking our phrases and knowledge without smothering differends that arise. The idea of a system is then temporal and emergent, a way of viewing a conversation that enables linkages. This does not discount the agonistic (behavioural or strategic) principle underlying conversations, but also does not assume anthropocentrism. Generative conversation views situations of multiplicity as open systems, thus enabling pragmatic linkages. We need to view any consensus that emerges from conversations as `only a particular state of discussion, not its end' (Lyotard, 1984, p. 65).


This paper has explored some of the implications of postmodern philosophy on systems thinking and practice. The focus has been on the space between the knowledge-forming system and the individual subject. The heuristics developed help us to engage some of postmodern philosophy's important implications for systems thinking and practice as identified by Jackson (1991, p. 299). By assuming that phrases position the pragmatic efforts of subjects we can link power, language, and the emergence of localized logic and order in the process of generative conversation through the use of (open) systems ideas. In supporting the paradoxical ethic that no language game should dominate or limit a conversation we recognize and engage power at a micro level. By defining progress as new moves within a game or the establishment of a new game and stakes we avoid the potential traps of prescribed performance or emancipation.



Beer, S. (1985). Diagnosing the System for Organizations, Wiley, Chichester.

Checkland, P., and Scholes, J. (1990). Soft Systems Methodology in Action, Wiley, Chichester.

Foucault, M. (1972). The Archaeology of Knowledge, Routledge, London.

Foucault, M. (1984). In Rabinow, P. (ed.), The Foucault Reader, Penguin, London.

Forrester, J. (1961). Industrial Dynamics, MIT Press, Cambridge, MA.

Gadamer, H. G. (1976). Philosophical Hermeneutics, University of California Press, Berkeley, CA.

Gadamer, H. G. (1981). Reason in the Age of Science, MIT Press, Cambridge, MA.

Granger, G. G. (1960). Pensee Formelle et Sciences de l'Homme, Aubier-Montaigne, Paris.

Jackson, M. C. (1991). Modernism, post-modernism and contemporary systems thinking. In Flood, R. L., and Jackson, M. C. (eds), Critical Systems Thinking: Directed Readings, Wiley, Chichester, pp. 287-301.

Lyotard, J. F. (1984). The Postmodern Condition: A Report on Knowledge, University of Minnesota Press, Minneapolis.

Lyotard, J. F. (1985). Just Gaming, University of Minnesota Press, Minneapolis.

Lyotard, J. E (1988). The Differend: Phrases in Dispute, University of Minnesota Press, Minneapolis.

Lyotard, J. F. (1993). Political Writings, University of Minnesota Press, Minneapolis.

Watzlawick, P., Weakland, J., and Fisch, R. (1974). Change: Principles of Problem Formulation and Problem Resolution, Norton, New York.

Wittgenstein, L. (1953). Philosophical Investigations, Prentice-Hall, Englewood Cliffs, NJ.

Wittgenstein, L. (1956). Remarks on the Foundations of Mathematics, Basil Blackwell, Oxford.

Warren Topp School of Engineering Management, University of Cape Town, South Africa

Warren Topp, Correspondence to: Warren Topp, School of Engineering Management, Room 522, Menzies Building, University of Cape Town, Rondebosch 7700, South Africa.
COPYRIGHT 2000 John Wiley & Sons, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2000 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Topp, Warren
Publication:Systems Research and Behavioral Science
Geographic Code:6SOUT
Date:Jul 1, 2000
Previous Article:Ensuring Delivery Through Organizational Design.
Next Article:Toward a Humanized Systemic Organization: A Confucian Perspective.

Terms of use | Privacy policy | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters