1. Modelling based on equations
Global or world modelling may be understood as the attempt to represent rigorously the economic, political, social, demographic and/or ecological issues and their interdependencies on a global scale. The models map these relationships as explicit mathematical equations which may be "run" forward in time to study their dynamic behaviour. They can thus be used to simulate future developments under a variety of conditions. Such modelling may be considered as the most sophisticated approach to dealing systematically with the nature of, and solution to, world problems.
Global modelling is seen by some as the key to understanding the global problematique. Indeed it was the group that initiated this approach that first formulated the term "global problematique". The accomplishments of this group through the 1970s have been reviewed, by some of those involved, in an exceptionally honest book: Groping in the Dark; the first decade of global modelling (Donella Meadow, et al., 1982).
The authors distinguish two types of model: (a) mental or verbal models; and (b) mathematical or computer models (which may be based on mental models). Mental models are complex, shifting and often unverbalized, and when they are verbalized can be understood as implying different mental models from those intended. Computer models express precise mathematical relationships such that conclusions can be calculated from initial assumptions, especially when the situation is so complex that it cannot be encompassed in words or simple equations. The authors perceive such computer models as providing the necessary guidance for policy decisions in response to the global problematique. Some may even perceive such guidance to be both necessary and sufficient.
Global modelling has continued to proliferate in the 1980s, both geographically and in terms of issues and methodologies. An overview of global modelling in 1985 was provided for UNESCO by Heinrich Siegmann of the Wissenschaftszentrum Berlin (UNESCO, 1987). Some 29 models were then considered relevant to the study.
Global modelling is considered to be still developing as a research field. According to Siegmann, hardly any model should be considered completed. The global models represent the confluence of modelling streams from three disciplines: political science, systems dynamics and econometrics. Global economic models have been put to use by policy-making institutions in order to aid in short-to-medium term projections. The modelling time horizon has in general become shorter. The issues addressed have become more specific.
Modellers regretted the cessation in 1981 of the Global Modelling Conference sponsored by the International Institute for Applied Systems Analysis. The conference had provided a consolidating infrastructure for global modelling, a role which UNESCO considered providing prior to its own programming and budgetary difficulties from 1987.
2. Alternative approaches to modelling
Following the appropriation of the term "global modelling" by those designing models based on mathematical equations, it might be assumed that no other forms of "modelling" of the global problematique are possible. Mental and verbal models are less satisfactory for the reasons noted above, but interesting models of systems can also be explored using analog methods.
It is interesting to note that a number of disciplines use other kinds of models in order to grasp the nature of complex systems. In the case of chemistry, molecular structures made up of many thousands of atoms are displayed graphically under conditions where the real complexities of the system do not lend themselves to mathematical analysis.
This example points to a key issue, namely availability and adequacy of information in constructing equation-based models. For such an equation to be formulated, precise information must be available through which a mathematical function can be defined. If the information cannot be formulated in precise terms, then nothing can be incorporated into the model.
Unfortunately, very few non-economic problems can be formulated in precise terms. It is, however, possible to use a different kind of precise mathematical relationship to provide some means of reflecting fuzzy perceptions. A network of perceived relationships between problems can be progressively identified. Such a network model can be held and explored on computer (see Section TZ). Such global modelling is based on graph theory (and related disciplines), a quite distinct branch of mathematics. It has, for example, been used in artificial intelligence research to model (and interrogate) systems of attitudes held by an individual.
3. Contrast to this Encyclopedia
Global modellers have not envisaged this possibility, which is the inspiration behind the project of which this Encyclopedia is a product. It is important to recognize the fundamental difference between these two approaches:
- assumptions concerning the objectivity of the "real world" vs. assumptions that "subjective" and "erroneous" assessments of "facts" and attitudes may be just as real to some constituencies and, as such, may have real consequences;
- dependence on quantitative ("hard") data vs. registration of qualitative ("soft") data;
- insights only emerge from mathematical analysis vs. insights emerge from relative positioning of nodes in a network;
- complexity can only be managed to the extent that it can be reflected in a set of equations vs. complexity can be navigated (in the same way that a complex road map is navigated);
- data sets must be reliable and complete to be of any value vs. recognition that data sets are necessarily partial and unreliable.
- the structure of the model does not involve an on-going dialogue with the constituencies whose views are represented there vs. the model is only useful to the extent that it reflects the current, and potential, biases of the views reflected therein.
4. Weaknesses of an equation-based perspective
From the perspective of conventional global modellers, the kind of information collected for this Encyclopedia is completely superficial since it does not analyze the relationship between economic and social factors and express them in mathematical terms so as to define a model which can be explored under different conditions.
There are a number of weaknesses in this position to which this project provides a partial response.
(a) Competing models: The early belief that it would be possible to produce a single satisfactory global model has been severely eroded by the emergence of some 30 competing models reflecting different assumptions, and producing conflicting conclusions. The authors of the above review are quite explicit about the disagreements between such competing approaches. What is strange is that there is no attempt to produce a model to model the disagreements between the models. The existing models are closed perceptual systems that are not designed to acknowledge alternative perceptual frameworks. They are scientific constructs representing only a portion of reality. This project is a step in the direction of modelling the conflicting perceptions about the many aspects of the global problematique. It attempts to encode that disagreement.
(b) Elitist: Whilst the Club of Rome's first model, described in Limits to Growth, attracted world-wide attention, the attention accorded to subsequent models has been limited to an increasingly smaller elite. Global models have been slow to integrate political factors in response to early criticism and failed to incorporate the many non-quantifiable issues to which the international community is obliged to respond (such as problems of human rights, corruption, alienation, drug abuse, etc.). Even many semi-quantifiable problems, such as threats to species and food chains, have not yet been incorporated into global models. There is an increasing gap between the world of the global modeller and the world in which many in the international community perceive themselves as living. As such, modellers face increasing difficulty in legitimizing their efforts to the wider public which must approve any major policy decisions based on their models. Such a factor should be integrated into a realistic global model. For this reason this project is based on entries with which such constituencies identify, using their language wherever possible.
(c) Inability to encompass "surprises": Global models are unique because of the breadth of the theories they endeavour to embody. Whereas the human brain is well suited for finding complex relationships, it is only such computer simulations that are adept at tracing numerous dynamic relationships simultaneously over time. Unfortunately the theories and simulations are based on data series established and accepted in the past. New events, or configurations of crises, not envisaged by the theories and for which no theoretically acceptable data exists, cannot be rapidly incorporated into such models. The models may even have to be totally redesigned to handle such data which cross, interlink, and call into question, existing categories. As in warfare, there is a limit to the degree to which reliance can be placed on simulation of the evolution of a battle. Unforeseen situations require another form of data.
(d) Specialization: In response to the competition between global models, modellers have been encouraged to specialize. Indeed the authors of the review point out that many observers have argued that a policy-oriented model will be most useful when: it focuses on a specific, well-defined problem; the problem reflects the objectives and needs of a specific client; the sponsor and the client are the same (Meadow op. cit. p.97). The authors also draw attention to the growing domination of economists in modelling initiatives. Seigmann stresses that world modelling continues to be scattered across a variety of fields. "This affects both the consolidation of world modelling into a scholarly discipline, and the establishment of its usefulness to policy-makers and other possible sponsors of global modelling."
Models now tend to have boundaries that are more narrowly defined, and to have shorter term horizons and more disaggregation (ibid p.100). Such models can only be said to be global in a very limited sense, reinforcing fragmentation of perspectives and begging the question of how such perspectives are to be interrelated to respond to a network of problems (the original inspiration of global modelling). This project is an effort to offer an alternative to this tendency.
(e) Scope: Following from the previous point, according to Seigmann even world modellers often disagree on what comprises a world model and what objectives global modelling might have. Global models may create a false impression of adequacy in dealing with those issues for which they are designed, precisely because they are incapable of reflecting their inadequacies in dealing with other issues they may assume are irrelevant. This may completely undermine the value of the insights they offer. In this project, especially in relation to world problems, the emphasis is on registration of the issues which appear in the literature. The intention is to provide a framework within which different approaches to the data may be explored, enabling different experiments on its use for very different kinds of constituency.
(f) Uncertainties and scepticism: Seigmann argues that the future of global modelling, over twenty five years after Harold Guetzkow's Inter-Nation Simulations, and more than a decade after Limits to Growth, remains beset by uncertainties. Its utility to academics and policy-makers is not yet assured. He considers that scepticism regarding its usefulness may even have grown. He sees these problems and uncertainties as attributable to the cross-pressures involving world modellers themselves, the academic research environment, the potential clientele of world modelling in policy and decision-making, and the general public.
On the depth versus breadth of a model Seigmann's notes: "This trade-off affects one's modelling aspirations as well as one's standing in academia and with potential clients in policy-making. Ordinarily, limited computer power, time constraints, limited resources and the task of coping successfully with complexity, require a compromise between modelling in sufficient levels of detail, yet having a broad scope. Global modelling thus becomes an exercise in brinkmanship. The specialist will tend to dismiss the model as irrelevant, if he finds it too simple -- which it usually is -- in his special area of expertise. The generalist, ie particularly in policy-making, will similarly reject the model as lacking in realism, if he finds certain sectors or regions not included." Global modellers may then be rated "ineffective in reaching out to planners and civil servants due to the failure to be comprehensible and convincing."
(g) Comprehensibility to a wider public: With the exception of Limits to Growth, the results of global modelling have attracted little attention from the informed general public. On the question of popularization, Seigmann notes: "Global modellers must therefore have an interest in delivering a "marketable" product. The broader the public discussion of model "results", the better the chances for continuing, extending and refining their work. That way, global modelling can create its own following (eg environmentalism) acting in its interest as a political pressure group. The popularization of global modelling, of course, entails considerable risks of two different kinds. First, in the frequent trade-off between scientific appropriateness and popularity, the latter might obtain too much weight. The modeller might tend to emphasize those aspects that are likely to spur the biggest headlines. The media might concentrate on those aspects of the model and its findings that they can sell best...Those models with the most spectacular "predictions"...will tend to be drawn into the limelight....A second kind of risk exists if the model, its "results", or the whole discipline, are being used -- deliberately or not -- to legitimize courses of action which might not be born out of "hard" facts."
Again it is interesting that there appears to be no recognition by such modellers that the status of a model needs to be reflected in the model itself. In the above quotations the model is treated as the product of an interest group seeking to impose its perspective. The model fails to encode the dynamics relating its perspective to others, whether other modellers, other disciplines, or the wider public.
By contrast, in the case of the project giving rise to this Encyclopedia, the dilemma created by conflicting perspectives is central to the whole approach. It is assumed that yet another formulation of "the truth" or "the methodology" would be as significant as the multiplicity of other such extant "answers" to the challenge of the times. The task is therefore defined as one of providing an appropriate framework through which conflicting and incommensurable answers and (perceptions of) "facts" can be related, at least to some degree.
At the data and analytic level, graph and network theory is seen as providing a way forward where more rigorous, and possibly over-demanding techniques cannot travel. In terms of comprehension of more complex dynamics and patterns by wider public, the role of metaphors has been emphasized. Given the current advances in graphic software techniques, there is no reason why these two approaches should not be "married" in some way that would be of immediate significance to policy-making. Organized in this way, specific equation-based models could be integrated as required.
4. Evaluating long-term developments using global models
In a follow up for UNESCO of the Seigmann study, Sam Cole (1987) concludes that conventional "global modelling is relevant, because understanding long-term developments at the macro-level are important in defining a perspective for shaping the future. It is also important because the intellectual investigation of alternative policy options through scenario studies can reveal problems which have been overlooked but which are worth considering." Cole's report, intended to establish the relevance of such modelling for UNESCO, has "no simple answer to the question of how global models may be integrated in the overall goals of UNESCO in the fields of education, science, culture and technology." The reason may be that the conventional modelling approach is quite inappropriate to the registration of most of the real, but only partially quantifiable, issues to which an organization such as UNESCO endeavours to respond.