Andreja Jonoski and Hamish Harvey

Inclusive decision making and the NDDSS

The direct involvement of all "stakeholders" in the decision making process has been identified as crucial in any water-related decision making process (World Bank, 1993). While often accepted as desirable, even necessary, the implications of this inclusion are far reaching indeed, and it is widely regarded as difficult or impossible to realise.

All stakeholders is in general a very large group of participants indeed, with a wide variety of backgrounds and interests, who will bring an enormous range of values to the negotiation table. Genuine inclusion requires that these stakeholders do in fact have an influence on the decisions made, and that the traditionally powerful interests do not, in the final analysis, swamp their views. This not least because stakeholders will stay away in droves if they perceive that their input does not have any such influence.

The traditional decision making process, in which experts advise decision makers, whose decisions are then delivered to the general population as ready made "solutions" with which their compliance is expected (especially if they have been "consulted"). Where attempts at inclusion have been made, they have generally attempted to modify this traditional process. Such attempts have generally started from one, inherently false, assumption: that if there are various interests to be included in the decision making process, then it is possible to foresee and describe them in advance.

A new, inclusive decision making process must overturn this assumption. It must acknowledge the fact that it is never possible to know in advance what range of different values will be held by the newly enfranchised participants, and it is certainly not possible to encode those multifarious values and value gradients and from that encoding produce a single, optimal solution. Indeed it is not even possible to enumerate in advance all possible solutions; the complexity of environmental management problems is such that the decision making process should be open to new solutions and variants throughout. Instead of assuming this prior knowledge, a process of collaborative learning and negotiation towards a decision is envisaged.

This restructuring of the decision making process is summarised in Figure 1. The traditional decision maker becomes a "persuader", whose engagement is one of collaboration and negotiation with - and persuasion of - all others engaged in the process. The "knowledge provider" is a new component understood as the totality of the sociotechnical infrastructure which provides and manages the knowledge "content" brought into circulation. Experts are at least as important in this new paradigm as in the old, but their effort, instead of supporting only the institutional decision makers, becomes more important in supporting the knowledge provider and thus ultimately a much wider range of participants in the decision making process.

The proposal advanced in detail in Jonoski (2002) and summarised here is to view these problems as one subset of a more general set of problems of knowledge circulation, and thus as problems which can only be resolved through the introduction of new kinds of sociotechnical tools and arrangements. These constructs are referred to as Network Distributed Decision Support Systems (NDDSS), where the word "network" refers first to the fact that these new systems have to be developed and deployed over electronic networks, primarily the Internet. However the electronic network is merely a technology which enables the realisation of an essentially sociotechnical system, which in turn enable the creation of networks of people and institutions engaged in processes of knowledge circulation.

The NDDSS is conceived of as a (virtual) environment structured in three major functional components: the fact engine, the judgement engine, and a collaboration and negotiation platform.

Fact Engine

The fact engine component of the NDDSS is the part which is administered by the knowledge provider. A central part of the fact engine will be recognisable as a modification of the already familiar modelling system. Existing modelling systems will however require significant modification to be effective in this new mode of application. In particular, the user interface now familiar from fourth generation modelling systems will be of use only in setting up models; the users of the NDDSS will require that facts from these models are presented in such a way as to best support the judgement forming process. In general facts are presented to the user in order to support the user's judgement processes, and as such these user interfaces can properly be regarded as part of the judgement engine, and no longer part of the fact engines themselves at all.

The concept of the NDDSS reinforces the importance of genuinely integrated modelling, providing linkages of knowledge across domains and spatial and temporal scales. In addition, a full NDDSS must provide access to knowledge quite different in nature from numerical models, however well integrated. These may include databases (of aerial photographs, for example), expert systems, and (hyper)texts.

Judgement engine

The judgement engine is concerned with assisting users in making their own judgements about a particular intervention in the system, based on the interaction between their beliefs and interests and the received facts from the fact engine component. Because of the diversity of the participants involved in the process, provision must be made to present these facts in a customised way; in particular interfaces which provide access to facts without demanding specialist scientific knowledge are required.

The judgement engine is envisaged as primarily facilitating the formation of judgements of value. Diversity in values and value gradients among participants must be respected and tendencies towards restrictions of certain values — or towards the reduction of all values to one (usually monetary) — discouraged. Techniques developed to support decision making processes within the existing decision making paradigm (such as, for example, multi-criteria analysis) have their place within the NDDSS, but great care must be taken to ensure that these tools are used in such a way that the participants in the process remain free to make judgements about options and outcomes in the context of their own values.

Where these tools will come into their own is in the role of aggregating judgements. It may be expected that the individual participant will provide judgements about several issues of his or her concern; these judgements must be aggregated into an overall position towards the proposed courses of action. Further, in order to facilitate the processes of negotiation and collaboration in which the judgement forming processes are situated, the judgements of all users need to be aggregated into a "social landscape" against which each individual's position can be evaluated.

A primary requirement of the judgement engine is the provision of transparency in the judgement forming process, in support of processes of second order reflection among users of the system. It is through secondary reflection, and the awareness it engenders of ones own and others' values, that individual judgements can change, and that a negotiation process can have a satisfactory conclusion.

Collaboration and negotiation platform

The collaboration and negotiation platform is required to support human interaction. In the NDDSS concept supports an iterative, emergent approach to decision making where interactions between participants may lead to options — and therefore decisions — which could not have been foreseen. The collaboration and negotiation platform must be designed in such a way that it will support democratic approaches to negotiation and limit as much as possible the opportunities for abusing the process.

As individual participants make judgements on possible courses of action and with the judgement engine assisting by aggregating those judgements into a social landscape, alliances may form based on shared interests and values. This process can be facilitated by providing measures of various kinds of "closeness" between a given participant and other groups or individual participants within the "social landscape" of judgements made by all participants mentioned above.

These processes of collaboration, cooperation, and negotiation take the form of a highly non-linear discourse, and are not at all amenable to formalisation. This component of an NDDSS cannot be designed as an encapsulation of a given set of procedures to be followed. Individual to individual communication entails some synthesis of email, instant messaging, weblogs and web fora, enriched with support for the management of content including scenarios, specialist analyses of fact engine outputs, tools for visualising positions in a social landscape, and so on.

Aquavoice: prototyping the NDDSS as a multiple player game

With the theory of how a Network Distributed Decision Support System might work established, there remains the problem of how to realise such a new approach to decision making. A new sociotechnical arrangement can only be achieved through an iterative process of development, where the technical elements are progressively refined to achieve the socially desirable results. Here games play a critical role.

Some of the strengths ascribed to games as learning vehicles in other articles in this newsletter can be recognised as similar to those established as requirements for an NDDSS. Computer based simulation games of the SimCity genre make use of the ever increasing processing power of commodity computing hardware to provide game play based around real (or rather game) time simulation of actual or hypothetical realities. The game play depends heavily on the emergent properties of connected simple simulations. These games can provide an environment which supports learning about complex systems and emergent properties in general, and about particular classes of system. The popularity of the game SimCity and its many offspring support a certain optimism that a market already exists for alternatives to those quintessential simulacra, the "first person shoot-em-ups" set in a simulated world with only the most passing connection with reality.

The arguments for multiple player games meanwhile can be seen as recognition of their strengths in promoting knowledge circulation and second order reflection. The use of realistic, even if hypothetical, scenarios enables such games to help bring experts' latent knowledge into circulation, while the removal from a real situation in which real money must be spent allow the freer exploration of novel ideas. Bringing together a variety of people with different interests draws in a variety of knowledges, as not all scientific.

Aquavoice, a prototype NDDSS, utilises this synergy between concept of the NDDSS and the recognised strengths of games by being itself a game. The structure of the NDDSS is provided around a realistic scenario, and players of the game take on the roles of likely stakeholders. In doing so it provides real, immediate value, supporting learning about both the problem space of the game scenario, and about the processes of collaboration and negotiation which (should) underpin decision making in such scenarios.

At the same time, the theory and implementation of the NDDSS concept can tested at its current development stage in a realistic scenario. The way in which players respond in the game and make use of the tools available provides invaluable feedback for consideration in future iterations.

Conclusion, and a cautionary note

Returning to the more general issues surrounding the use of games for learning and education, it is perhaps worth sounding a note of caution. With group role playing games involving a facilitator the players themselves bring a wide range of experience to the game, so there is a high likelihood of valuable and safe lessons being learned by the participants. Unattended learning by playing a computer game is more problematic; the simulation models used are necessarily to some degree incomplete and incorrect. The demands of playability constrain the level of detail at which the system can be modelled, and thus short cuts must be taken. Those short cuts can embed beliefs of the game developers, or have unexpected emergent properties which are properties of the model and are not representative of the supposedly modelled system. In short, games can teach unintended lessons, and those lessons may be counterproductive. Analysis of these situations is necessary (primarily through the study of the processes of signification and communication using theoretical concepts of general semiotics) particularly if steps are made from games towards real-world systems.

These concerns notwithstanding, games support a powerful mode of learning, and deserve perhaps more attention in the fields of water and environmental management than they have received to date. They also, as we have discussed in this article, provide a valuable platform on which to test new tools intended to support genuinely inclusive decision making processes.

References

Andreja Jonoski (2002) Hydroinformatics as sociotechnology: promoting individual stakeholder participation by using network distributed decision support systems. (PhD Thesis. TU Delft/IHE Delft), Sweets & Zeitlinger B.V., Lisse, The Netherlands.

World Bank (1993) Water Resources Management. A World Bank Policy Paper. World Bank. Washington.