Introduction

The 1970s have seen the development of considerable enthusiasm for "network" building, whether among individuals or among groups and institutions. Much hope has been attached to this "alternative" vehicle for action following the failure of  "coordinating bodies" and "organizational systems" to respond to the perceived needs without imposing unwelcome forms of order. Recommendations to create a network are widely felt to be low-key, low-threat options in a variety of sensitive situations. As such they may also serve as convenient ("cosmetic") tokens of action where "effective" action is not considered possible.

The purpose of this paper is to examine some of the assumptions underlying the enthusiasm for networks and their operation in practice. It is hoped that such an exercise will identify some of the pitfalls of the network option and identify possibilities for improvements. 

Minimal requirements for network emergence

The creation of networks is facilitated by the following factors which should be contrasted with their equivalents in conventional organizations.

1. Minimal commitment: A member of a network is seldom obliged to make any major commitment to the network or to other members individually. Any strong commitments may be made on an ad hoc basis, but they may also be avoided. Pressures to respect network obligations are mitigated by the member's self-arrogated right to reserve response to such pressures. Participation tends to be undemanding.

2. Diffuse membership: It is characteristic of many networks that the membership boundary is unclear. Particularly when the network does not have a single controlling centre, portions of the network may relate closely to bodies not perceived by other portions to be part of the network. Such bodies may perceive themselves to be part of the network and may be so perceived by those to whom they relate. Membership is often not of the card-carrying variety and is more a question of degree of involvement as perceived by others over a period of time. Consequently some are considered members who do not perceive themselves to be, and others are not so considered although they may well perceive themselves to be. A network of interacting bodies may of course exist even though it is not recognized or labelled as a network. This is often the case with "invisible colleges" (1).

3. Minimal organization: Since networks are frequently created to avoid conventional modes of organization, and since alternative modes tend themselves to be lacking or suspect, little can be done to "organize" a network.

Responsibilities can seldom be allocated, since an "allocator" is not recognized and there is little obligation to respect such allocation anyway. A degree of organization is introduced through agreement that a particular body should process information for the network. Such activities exert a pressure on other members which results in a minimum amount of organization. However this may effectively be equivalent to the action of a newsletter or journal on its readership even if some "readers" are stimulated to correspondence, others to write articles, and others to participate in "readers clubs".

4. Minimal expectations: It is characteristic of many networks that members do not necessarily have high expectations concerning the action of the network. Frequently networks are conceived as auxiliaries or complementary to action which members may undertake individually through other (conventional) structures. Or alternatively networks may function where expectations are reasonably low because it is recognized that major or sudden progress is not possible, particularly through conventional structures.

5. Diffuse concerns: Whilst some networks have very specific concerns, the members of others have a wide- range of preoccupations which overlap or reinforce each other in a complex manner. Of particular importance are those cases where the domain of interest of the network is highly complex, transdisciplinary and involving a variety of possible responses (research, political action, personal life-style change, etc.). The concerns of the network as a whole may well be extremely elusive to the point that members recognize each other less in terms of a shared attitude to present concerns, and more in terms of a shared response to potential future concerns.

6. Minimal organization of preoccupations: It follows from the previous point, and from point 3, that the concerns of a network are seldom well structured. The complex subject domain may resist conventional efforts to organize it and members may themselves resist efforts to order their perceptions of it within any particular framework. Where an effort is made to use some framework, this tends to be viewed as an administrative convenience minimally related to the non-explicit substantive ordering of the domain. It follows from this that conceptual integration tends to be a major difficulty (even if its desirability is not rejected for reasons analogous to members rejection of the organizational coordination or integration, which gave rise to the network in the first place). Efforts to use the network model to structure the substantive concerns have not paralleled its use to structure the relations between members.

7. Minimal collective learning: Since a network maintains no central repository of written records, collective learning (if any) tends to take an oral form. This can be powerful in its own way but fails to build up a body of knowledge (as opposed to lore) which can be drawn upon on suitable occasions.

8. Minimal activity: It follows from the above points that the network, as a network, is often characterized by minimal activity or productivity, verging on total passivity or inertia. But again the requirement that a network be "productive" or "active" may well be rejected by members in favour of "being" (as opposed to "doing"). As with the traditional "old boys network" its significance emerges from its existence, not the specific activities which it may facilitate from time to time. This is not to deny that a network may suddenly be activated in response to some specific situation (e.g. a crisis, an election, etc.), although in becoming "active" its members may prefer to create one or more conventional (ad hoc) structures through which to work.

9. Unpredictable potential transformation: As implied by the previous point, most of the above characteristics need to be qualified by the fluidity of networks and the attitudes of members towards them. Networks can change and evolve very rapidly, to the point of manifesting characteristics contrasting markedly with those noted above. It is not clear what factors contribute to, or trigger, such changes.

Unpleasant networking realities

In continuing this study, it should be noted that the purpose is to highlight the weaknesses of network activity not its many strengths which have been adequately lauded elsewhere (2, 3, 4).

Clearly combinations of the weaknesses noted above may result in a network of minimal significance, if only to those who tend to perceive themselves as members. Such activity as there is may then be characterized by :

• regular contact between key members only
• irregular or no contact with some members
• member contact (if any) with central elites and rarely (if at all) with other members
• fragmentation of the network into sub-networks
• member activity only in response to stimulus or to occasions, namely not selfactivating or continuous  member dependence on continuing encouragement, whether verbal or in the form of some financial support (namely " activated" members as opposed to "self-activating" )
• limited ability of members to process communications from other members and to integrate them into some larger framework
• reliance on forms of communication which in themselves hinder integration and collective learning (or action) :
• presentations, or exchanges of documents, in a "show-and-tell" spirit, to impress others of the importance of particular isolated activities
• publication of collections or compilations of documents which require that the reader perform the task of integration which the contributors avoid
• presentation of results as the work of individual member bodies rather than as an integration of their thinking
• member interaction designed to improve respective individual contributions but not to integrate them
• inability to focus (or build) on issues raised by individual contributions, or on the lacunae which emerge between them
• different skills and perspectives remain alien (or occasionally hostile) to each other and do not lead to the production of a framework which exemplifies their complementarity.

Difficulties such as these are due to many factors which will become better known in the future. However, insofar as the network is designed to reinforce what the members are doing individually anyway, it comes to be evaluated against the ability of the member to act without the network. This loses sight of what the network can achieve as a whole. This is examined below by considering the  "communication units" and "comunication frameworks" within the network.

Networking operations

A. Communication units: By this is meant the physical unit for information transfer. For example:

• books in a library network
• bibliographical records or abstracts in a documentation network
• transactions in a financial network
• papers (or verbal presentations) in a research network
• event announcements in an "alternative" network.

These examples suggest distinctions such as:

• some networks exist only to ensure the transport of the units, and their significance lies in their ability to do so throughout the network
• the units transferred in some networks are of value to the members as indicators of action they can perform elsewhere independently of the network.

In the case of the research network, it is supposedly the facts and concepts contained within the communicated units (i.e. the papers) which are of value to members. But in such a network, presumably the idea is not only for members to "feed" each other so that they can act better elsewhere independently of the network. The contents of the communicated units are supposed to be processed, evaluated, and reordered into more useful patterns within the network.

The question is whether, by emphasizing the use of papers as communication units in a research network, this obstructs the communication and integration, of the ideas that they contain. By embedding a useful idea in a (lengthy) paper, it may be easily overlooked and filed with the paper under "waste paper" The same is true of a verbal presentation, a debate, or any form of dialogue. The essential logical units upon which collective learning and progress depend quickly drift into oblivion under current procedures. Against this it may be argued that key concepts are retained, despite the enormous wastage considered acceptable. Or alternatively, if emphasis is placed on the learning process, then whether or not ideas are "lost" is irrelevant, since similar ideas will be rediscovered on the next occasion that the process is activated.

B. Communication framework: By this is meant the setting within which the communication units are exchanged. For example:

• a " visit ", with its many opportunities for discussion and exchanges of opinion
• a face-to-face group meeting, namely a form of "multilateral visit"
• a newsletter, bulletin, etc.
• a journal or compilation of papers - a data network.

Clearly when the prime purpose of the network is to transfer the units between the members, no special difficulty arises. But when, as in the case of a research network, the network as a whole has to process, evaluate and integrate the logical units contained within the physical units, then further questions must be asked concerning the communication framework:

• does it ensure storage and retrieval of logical units (as opposed to physical units), or is there extensive leakage of logical units into oblivion
• does it provide facilitative processes to ensure the juxtaposition and integration of logical units, not only of a similar kind, but also where dissimilar units have to be maintained in a dynamic balance or state of complementarity
• does it ensure that all resources assembled at any one time (e.g. at a meeting) interact appropriately, or :
• is much time devoted by all to polite attention to a speaker repeating (familiar) arguments already circulated in writing
• do some participants feel inhibited, intimidated or unable to interact effectively because of the momentum established by the articulate minority (however incorrect the viewpoint promulgated)
• do some issues, of major interest to a minority of participants, remain undiscussed because of the agenda setting procedures and convenient time constraints.

Containing and focusing network dynamics

The previous section clarifies the central problem associated with those networks which have potential, as a network, to move to a new level of significance. The problem is one of containing and focusing the wide variety of ideas generated so that they interact appropriately to permit the emergence of new insights of a more comprehensive and more integrated nature.

But "containment" would appear to imply a networking philosophy which is contrary to that which prevails. It implies a level of discipline which the first section (above) shows to be uncharacteristic. And yet in some kinds of networks a major effort is made to minimize "leakage" or maximize "coverage" (e.g. interbank, bibliographical, etc.). The looser inter- personal or inter-institutional networks would tend to view this as a step towards "coordination" with all its attendant ills. And indeed the problem is neither so challenging, nor so potentially rewarding, in the case of networks :

• which allow themselves to be coordinated from a central point - for these are merely loose or disguised hierarchies; or
• which are solely concerned with the " transport, of communication units between network members - for this is primarily a hardware and standardization problem, even if all communication passes via a central clearing point.

The challenge lies more with networks whose members could interact as much (or more) with each other as with (or via) any central point. A clue to a remedy would seem to lie in the complementary attributes of tension/compression which are characteristic of hierarchical systems but are absent from interinstitutional networks (5). Such networks, as shown above, tend to be "flabby and , sloppy" However the "jackboot" characteristics of hierarchical systems, to which the tension/compression attribute contributes, are equally unsatisfactory.

The problem would thus seem to be that existing networks are "untensed", whereas hierarchical systems have an undesirable form or degree of tension/compression (for some purposes at least). The lack of tension in networks is particularly evident in the tendency for "distance" to be established between those ideas (or, more irrationally, those advocating them) which are antagonistic to one another or perceived in some way as incompatible. The normal consequence is for no relation to be established between them - or worse still, their advocates ignore each other, refuse to dialogue, or even adjust the interaction lines within the network so that no further interaction is possible. This is how a network keeps itself "cool". It is also how it renders itself irrelevant because the resulting lines of communication tend to favour "conceptual incest", or some organizational equivalent.

A desirable level of tension may therefore be introduced by maintaining "confrontation" (compression) relationship between opposing ideas (or even between the bodies supporting them). The difficulty is that such relationships tend to break the network apart, as noted above. But this tendency may be opposed by, compatibility, (tension) relationships between mutually supporting ideas (or even their corresponding advocates). The challenge is to balance the confrontation and compatibility relationships within an appropriate structural configuration of a non-hierarchical variety. (N.B. They are not balanced in a hierarchical structure). The result would be a tensed network.

Relatively little is known about such tensed networks and that is limited to general principles derived from the study of structure in the abstract (but from a design viewpoint). An attempt has been made to show the relevance of such studies to the elaboration of a rich variety of alternative forms of organization - whether of groups or of concepts (5).

Exploring this avenue further should show how the key networking problems of leakage, storage, integration and focus of logical units can be resolved within the framework of such tensed networks. Of special interest is that the "energy level" of the network increases the more it is tensed, namely the greater the number and variety of incompatible elements that can be balanced within the configuration by compatibility relationships.

Tensed networks and social reality

The moment there is a question of interrelating incompatible logical elements, the gap between theory and reality is highlighted. If the elements are "incompatible" they cannot be integrated within a theoretical framework dependent, as most are, on logical compatibility. The problem of interrelationship between such elements tends of necessity to be of no theoretical interest.

An analogous problem exists between institutions. Unless they are compatible, in the sense of having the same legal, ideological, or substantive basis, they cannot be integrated within an institutional framework dependent on such compatibility. The problem of interrelationship tends to fall outside normal institutional concerns, and is "dumped" under "public relations". And, in fact, networks have been developed to overcome the problems to which this closure gives rise.

Despite these two tendencies, social reality contains incompatible elements of both a theoretical and an institutional kind, with the one often reinforcing the other. And it is their incompatibilities which are significant in the dynamics of that reality. Tensed networks thus constitute an interesting bridge between compatibility and incompatibility and could possibly provide a more adequate reflection (or model) of social reality. It is important to draw attention in this way to the limitations associated with the conventional fixation with the need to advocate monolithic theoretical or institutional frameworks from which incompatibilities have been hygienically removed -despite the high probability of their persistence in society. There is a special irony in the tendency of some networks to recommend such frameworks when their members are unable themselves to tolerate the monolithic character of such structures.

Facilitating networking

The communication frameworks (see above) used by networks do not lend themselves readily to containing and balancing network dynamics within new configurations - as suggested by the preceding sections. One form of communication which is ideal for this purpose is computer conferencing. Its relevance to a research network has been described elsewhere (6), together with the importance of, intermediate communication interfaces - where computer technology is not appropriate. A study should be made to compare the costs and advantages of using this approach with those of assembling individuals at a meeting.

However, even when individuals are physically assembled at a meeting there are many problems of ensuring the best utilization of the intellectual resources so mobilized. Such meetings deteriorate only too easily into communication frameworks characterized by leakage, fragmentation and lack of focus of logical units. The possibility of using computer conferencing to enhance face-to-face meetings has been explored elsewhere (7). The technique was used in this way in 1976 at a Congress of the International Society for Technology Assessment (8).

Further work required

(a) Tensing Networks: in search of clues: There are interesting constraints on the manner in which networks can be tensed by the introduction of confrontation elements. In order for there to be a balance between such counteracting (compression) elements, some degree of symmetry is required. In fact, in the absence of symmetry it is difficult (although not impossible) to tense a network satisfactorily. A good guide to reflection is the problem of tensing a (fishing) net. Even if it were torn in many places, it would still be overly simple because it is planar. As such it could only be tensed by pulling externally on its perimeter - which would constitute a model of external dependence. Of greater interest is the closed net (e.g. a string shopping bag). Here the network can be tensed against itself by inserting enough (compression) struts between adjacent knots until the resulting spheroid can no longer be crumpled. However, in order to do this effectively attention needs to be given to the allocation of the struts. This is where symmetry enters the picture.

Just how much attention needs to be given to symmetry is not clear. But by exploring symmetry constraints, ideas emerge concerning some of the ways in which networks can be tensed, even if such formulae are only ideal types to which no natural network will conform exactly (cf. the relationship between naturally occurring crystals and crystal symmetry classes).

In the spirit of the search for clues to answers, rather than in the hope for immediate answers, an attempt has made to identify the range of symmetry forms which help to understand more about how networks can be tensed. This is done for both 2-dimensional and 3- dimensional forms, since understanding of the one is a guide to understanding of the other. This approach is a development of earlier work (1) and was originally prepared in connection with an analogous problem with networks of concepts (9).

(b) Network self- representation: Some means is required to represent the variety of issues, questions, assumptions, concepts, etc. with which the network is concerned. The representation should provide an integrative overview, preferably in (wall) chart form, reflecting areas of compatibility and areas of confrontation. (A description of such a meeting aid will appear in a forthcoming issue).

(c) Network diseases: To give greater clarity to thinking about the manner in which networks can fail, it would be useful to examine the varieties of network " disease " (See, for example, pp. 486-489).

(d) Network function / dysfunction maps: It should be possible to condense insights concerning how a network functions (2, 3, 4) onto a single sheet in the form of a "map" This should also indicate the various ways in which a network can fail or be drawn into some state of imbalance. Such maps could constitute a valuable guide to working with networks.

References

1 . Diana Crane. Invisible Colleges; diffusion of knowledge in scientific communities. University of Chicago Press, 1972.

2. Anthony Judge. International organization networks; a complementary perspective. In: Paul Taylor and A J R Groom (Eds) International Organizations; a conceptual approach. Frances Pinter, 1977, pp. 381-413.

3. David Horton Smith with Anthony Judge. Inter-organizational networking. Transnational Associations, 30, 1978, 11.

4. Gerald E Klonglan et al. Creating Interorganizational Coordination (Project report; an orientation; instructor's guide). Dept. of Sociology and Anthropology, Iowa State University, Ames, Iowa 50011, USA, 3 vols, Sociology report nos 122A, 122B, 122C (Submitted to Defense Civil Preparedness Agency, Washington DC, May 1975).

5. Anthony Judge. From systems-versus-networks to tensegrity organizations. Transnational Associations, 30, 1978, 5, pp. 258-265.

6. Anthony Judge. Facilitating the networking processes of a transnational university using computer conferencing. Transnational Associations, 30, 1978, 4, pp. 205-214.

7. Anthony Judge. Enhancing communication at a large conference/festival (using computer conferencing). Transnational Associations, 29, 1977, 12, pp. 532-540.

8. Karl L. Zinn. CONFER at the ISTA Congress. Transnational Associations, 29, 1977, 10, pp. 412-417 (see also pp. 418-422).

9. Anthony Judge. Viable need patterns and their identification through constraints on representation in 3-dimensions. (Paper presented to the workshop on human needs, Berlin May 1978, of the Goals, Processes and Indicators of Development project of the United Nations University, Human and Social Development Programme).