Science is a systematic means of obtaining knowledge about the world. One of the basic facts underlying this endeavour is the observation that the world exhibits order. This observation seems to be shared by all cultures. An important step in understanding the order of the natural world consists in systematically describing its phenomena. Of the innumerable and endlessly variable phenomena, some exhibit their own order and can be classified into groups. Modern science, which originated in the 17th century, tends to depict this order in a special way, namely by positing laws of nature. Laws of nature are general regularities that hold between classes of events. Such regularities form the basis of scientific predictions and scientific explanations and are an essential part of scientific theories. Even those theories which scientists refer to as 'models' posit regularities of some sort or another. Thus, one of the core activities of science consists in the classification of events and the discovery of general regularities among such events in order to explain them and to predict their behaviour.

Many scientific explanations are reductionist; that is, an explanation of the behaviour of a system is sought in terms of its components and the laws governing their behaviour. This is why the scientific approach has often been called analytic. In a reductionist research strategy, systems are analysed into their components, their configurations, and their interactions in order to understand the system as a whole. Such strategies are a very powerful systematic means of directing research. Even when reductive explanations are unsuccessful, they usually lead to interesting results and foster some knowledge of the systems in question. Whether or not this reductionist approach can always be completely successful is a controversial questions. There may be systems whose behaviour cannot be fully understood in terms of their component parts in principle (see section 1.3).

Human beings are prone to error, prejudice and superstition. However, human beings are also capable of learning from error, and science employs systematic means to do so. Thus, science is not only systematic in displaying orderly structures of nature, but also in how it establishes knowledge claims and improves their precision. A large variety of critical methods designed to locate and contain different kinds of errors have been developed. As science is a human endeavour, it cannot entirely eliminate error, but it can minimise the probability of errors and evaluate their magnitudes.

In order to produce knowledge, science employs a curious mix of speculative and critical elements. The speculative element is necessary in order to find general regularities because they are not easy to identify and cannot simply be read off the phenomena themselves. To explain observed regularities, science often posits barely observable or even outright unobservable entities. At first, their very existence may even be highly dubious. Well-known examples of such initially speculative entities are atoms and genes. Such entities, and the properties attributed to them, must then demonstrate their existence either by revealing themselves indirectly in a broad variety of different phenomena, or by becoming observable through the development of new means of observation.

Since antiquity, the unaided observation of phenomena has been one means of making discoveries and controlling theoretical knowledge claims. But modern science has invented additional means to these ends, such as increasingly powerful instruments which mediate observation. Moreover, not only can science provide the human senses with artificial aides, but it has also discovered means of observation for which we lack sense organs altogether. A classical example is radio waves, which allow astronomers to explore the depths of outer space. The other principal means of knowledge acquisition is the scientific experiment. While all cultures have used some sort of experimentation, or 'trial and error' procedure, for improving their technologies, the systematic use of experiments for the acquisition of theoretical knowledge is a fundamental novel aspect of modern science. Roughly, experiments can be classified into two classes. First, the so-called 'explorative' experiments discover novel kinds of phenomena or hitherto unknown connections between different phenomena. In this way, they can suggest new avenues for further empirical and theoretical investigation. Second, experiments are used to test specific hypotheses about general regularities. For example, by systematically varying the different factors influencing the behaviour of a given system, and recording the system's response, experimenters are capable of distinguishing between genuine causal relations and mere correlations.

For several centuries, there has been a general belief in the existence of a specific scientific method which ensures the reliability and excellency of scientific knowledge. This idea dates back to some of the pioneers of modern science such as René Descartes and Isaac Newton. They suggested that scientific knowledge could only be gained by adherence to a set of absolutely binding rules, later dubbed 'The Scientific Method'. Since the late 19th century, however, a different picture of the advancement of science has emerged. This new account is mainly based on detailed historical research into scientific research processes. According to this new picture, science proceeds from the knowledge that it has already produced. In particular, eminent solutions to research problems function as exemplars for the identification and solution of other problems. Thus, in many cases, the advancement of science can be seen as a self-amplifying process in which existing knowledge forms the basis for new knowledge. In each particular field, continuous research traditions typically emerge. But the productive potential of some body of knowledge to guide a research tradition may eventually exhaust itself, and fundamental changes may be necessary in order to secure further progress. These changes take place in scientific revolutions when a fundamentally new point of view is generated which can transform the conceptual foundations of a discipline. Most notably, chemistry, biology and physics have seen such revolutionary transformations in the 18th, 19th, and 20th centuries respectively. In this new picture of science, the reliability of scientific knowledge is guaranteed by the precise focus and depth of research. If there is a hidden mismatch between nature and the theories underlying research, this mismatch will reveal itself in the research process. Eventually this will necessitate substantial corrective changes in the theories involved. In this way, the research process itself is capable of detecting and locating errors in the theory-nature fit.

The public character of scientific knowledge and its in-built mechanisms of self-correction and expansion distinguish science from most traditional forms of knowledge. Many cultures have developed highly sophisticated systems of knowledge, especially in astronomy, natural products, medicine, and mathematics, to mention just a few. But traditional knowledge has often been restricted; for instance to the leaders of certain religious elites, making broad dissemination of this knowledge impossible. More importantly, there does not seem to have been a systematic mechanism for effectively checking the reliability of knowledge and securing its growth into new areas. These undoubtedly unique qualities of scientific knowledge should not lead to an uncritical dismissal of traditional knowledge. In some areas, especially in medicine, there are still stocks of traditional knowledge not yet understood or even considered by science, but which nevertheless remain extremely useful, especially with regard to practical application. For many cultures, much traditional and popular knowledge has been and continues to be essential to survival. Furthermore, it should not be forgotten that traditional knowledge has contributed to the very development of modern science and an interaction between the two can be productive for all parties. However this does not warrant the conclusion that we could do without science, as some anti-science movements suggest. All countries can ignore science only at their own peril. It can be an important weapon in our fight against ignorance, poverty, superstition, and diseases and should be recognised as such.

The growth of science over the last 400 years displays an amazing increase in diversity. Even a very rough classification of the sciences specifies several hundred different special disciplines. The advantage of specialisation is obvious: it makes in-depth knowledge of a particular domain of phenomena possible. However, there are also some disadvantages of specialisation. The more the fragmentation into specialities and sub-specialities of science progresses, the more difficult communication between these specialities becomes. And as much research into urgent problems has to draw on resources from different disciplines and sub-disciplines, the fragmentation of science can inhibit progress. Fragmentation also creates communication problems between science and the public.

However, there are two main tendencies in scientific development which counteract the tendency toward ever-increasing diversity. The first tendency is the development of increasingly comprehensive theories - physics and biology provide telling examples. Seemingly disparate fields of science become connected by these very general theories. Thus, over and above the increase in diversity among scientific fields, a conceptually unifying net of theories tends to develop. A second unifying tendency of modern science is the internally driven inter-disciplinary research that leads to a growing overlap in the fundamental sciences such as biology, chemistry, and physics. Especially in research in molecular science, this inter-disciplinary overlap is clearly visible. New knowledge gained about as yet unknown molecular mechanisms will strongly influence the future of health research, the environmental sciences, and research on novel materials, all of which will continue to have a deep impact on society.

1.2   The universal value of fundamental science       

For the most part, there are two different sources of the problems that science tackles. One source is science itself. Because of its systematic character (see section 1.1), science generates its own questions, both with respect to content and with respect to method. The attempt to explain some phenomena systematically, to establish knowledge claims systematically, and to improve their precision within their domain - each of these generates certain questions that must be addressed. This is the domain of what is called fundamental (or basic) science. Fundamental science can be summed up as the generation of new knowledge. It tackles questions that are generated by the science system itself. The developmental dynamics of fundamental science is thus mainly driven from within science. This process depends on various resources and can be influenced by them (see section 2.3). In particular, new technological resources like better experimental apparatus, measuring instruments with higher resolution, or faster computers can make intrinsically interesting problems accessible which were previously out of reach, thus opening up new frontiers for research.

By contrast, the other important source of scientific problems is the social environment, or some subsystems of it, in which scientific research finds itself. Any given society has many problems to solve, be it on a national or an international level. For many of these attempted solutions, society turns to science. Obviously, some of these problems may be tackled with the scientific resources already at hand. In these cases, the term 'applied science' is fully adequate. By applied science, we simply mean the use of already existing scientific knowledge which is sufficient for solving a given problem. Here, there is often a smooth, but often time consuming and costly, transition between applied science and the commercial development of new products.

At this point, the following questions arise: why should society finance fundamental science? Isn't fundamental science simply a playground for scientists that is otherwise useless for society? Don't we have enough pressing problems that well-educated people like scientists should be trying to tackle without wasting time and resources on fundamental problems largely driven by the scientist's curiosity? Or, to put it more directly, isn't fundamental science a waste of money that even the industrialised countries, let alone the less developed countries, can no longer afford? Aren't those politicians who tend to cut expenses for fundamental science in times of low public budgets on the right track?

In spite of these reservations, there are some compelling reasons why fundamental science is imperative for both the industrialised and less developed countries. First of all, the knowledge needed to solve many of the most pressing problems the world faces does not yet exist. Thus, certain societal needs and desires directly trigger fundamental science. In the 20th century, questions posed to science by society have led to new areas of research and ground-breaking discoveries. And even though the questions are generated from outside of the scientific arena, they do trigger fundamental scientific research. In other words, scientific problems arising as a result of societal needs can lead to research that is focused on the discovery of novel knowledge in a certain domain. The first large-scale example of this was probably the development of the atomic bomb, but research on cancer, nuclear fusion, novel materials and various environmental problems exemplify the same pattern. It is quite likely (and will even be necessary) that the proportion of fundamental research driven by societal needs will increase in the future (see section 2.5). This is because, on a finite Earth with an increasing population, increasingly complex problems are arising which do not fall under the rubric of any particular science. Again, environmental problems provide a host of telling examples. In these circumstances, the know-how to solve these problems cannot simply be drawn from a well-established discipline and then applied to the particular set of problems at issue. Rather, it will be necessary to consult several disciplines, each of which is inadequate by itself. This must include co-operation and contributions from the social sciences. Interdisciplinary co-operation can lead to new solutions, and in this way new interdisciplinary approaches are continually evolving.

Second, apart from this problem-induced form of fundamental research, a case can be made for supporting curiosity-driven fundamental research, in which the application of results to real world problems is not the primary intention, and may not even be foreseeable. First of all, fundamental science simply stimulates the human creative potential. But the real reason why fundamental science is invaluable is found in the countless examples drawn from the history of science in which knowledge produced for its own sake later resulted in socially invaluable technological potential. The isolation of penicillin, for example, was the result of years of basic research into the nature of mould which at first did not appear to have any practical benefits or economic applications at all. Quantum mechanics, to take another example, is a physical theory that was invented because physicists saw various flaws in its predecessor, Newtonian physics, especially because the theory could not explain why normal matter is fairly stable and does not collapse into nothingness. For the lay person, the stability of a stone is a trivial fact of experience which does not need an explanation. It is simply taken for granted. At first glance, scientists who consider this fundamental question seem to be wasting their time. But in fact, this seemingly innocuous and irrelevant question led to innumerable technological applications in material science. Some of these applications have dramatically changed our world, most notably the invention of the transistor. Another striking example of the development of a powerful new technology that arose out of an attempt to answer a basic scientific question is the discovery of restriction enzymes. This discovery opened up the possibility of applying molecular genetic analysis and introducing specific genetic changes to any kind of organisms (see section 1). Genetic engineering is now widely recognised as a key technology for the pharmaceutical and biotechnology industries, as well as an indispensable research tool in almost all of the life sciences. Or consider a whole discipline like botany. Strictly speaking, botany is a fundamental science concerned with investigating the nature of plants. Many aspects of botany, however, have direct importance for human welfare and development. Such fields as forestry and horticulture are closely tied to fundamental botanical studies and others, such as pharmacology and agronomy, still depend on basic botanical knowledge. Thus, in the long run, the possibility of solving problems by scientific means depends on the existence of fundamental theories, methods, and insights that are provided by fundamental scientific research alone. But we should not overlook the fact that the transition of fundamental science to technology is far from automatic. Additional intellectual and institutional means are necessary in order to put the fruits of fundamental research to sustainable practical use.

Thus, with respect to the potential benefits, fundamental research can be seen as a long-term investment. As an investment, fundamental science has the unusual feature that the potential benefits of research are often unforeseeable. A large part of research in fundamental science may never yield any economic returns, but when fundamental science does turn out to be economically rewarding, the benefits can be immense. This element of unpredictability in the economic returns of fundamental science makes goal-directed action and priority-setting complicated and difficult (see section 2.3). The reason for this unpredictability lies in the very nature of fundamental science itself. It arises because fundamental science is mainly aimed at gaining novel knowledge in some field, whereas applied science is oriented towards a goal and tends to work with phenomena which are already well-known. No research foresight strategy can completely eliminate this unpredictability. However, it is clear that this kind of investment must be supported mainly by institutions which have a special responsibility for the long-term prospects of global society. That is why applied science supported by industry is often subsidised by short-term investments directed at very specific product- or service-oriented goals. Market proximity and profitability is affecting research and development financed by the private sector.

Third, scientific knowledge gained through fundamental science is often needed for long-term purposes which have no immediate economic returns. This holds especially for the use of fundamental science in long-term planning, both on a national and an international scale. It is important for governments and inter-governmental panels to be aware of long-term prospects such as climate change or demographic tendencies, which may include a large range of projected consequences in such areas as the health-care system, insurance policies, etc. Again, this holds for all countries, independently of their degree of industrialisation. Hypotheses on these issues must often be based on data collected over long periods of time. This is especially obvious in the case of extrapolation on climatic trends, such as estimates of the prospects of global warming or occurrences of global phenomena like El Niño, which may in turn have consequences for public health, such as an increase in the rate of diseases like malaria.

Fourth, it is extremely important to note that not all of the benefits of fundamental science come in the form of economically exploitable discoveries. A considerable share of fundamental scientific research is done by young people today, in particular by graduate and post-doctoral students. Some of these young scientists will pursue academic careers. But others will seek employment in industry or in government positions in order to work in research and development, in laboratories providing different kinds of services such as food quality control, medical testing, policy decisions, etc. In other words, society needs scientific experts in all kinds of positions outside fundamental scientific research. But fundamental scientific research is clearly the best form of training for scientific experts. This is due to the fact that it is providing those engaged in it with a sound understanding of state-of-the-art scientific theories, the technical skills needed, as well as the systematic approach required in many areas of the workplace and other societal activities. Thus, fundamental research has an important educational role, which is frequently underestimated. This point will be taken up again in Section 1.5.

Lastly, the cultural aspect of fundamental science should not be overlooked. Science can provide us with an extremely rich picture of our world, from its most minute details right up to the largest objects in the universe. All cultures have developed a desire to know the world in which they live, and science is a particularly strong method of fulfilling this desire. Fundamental science is imperative for every nation, industrialised and industrialising alike.

'Complexity' designates a set of loosely connected scientific ideas having to do with the phenomenon that certain systems exhibit, in spite of being governed by relatively simple laws, a number of unexpected properties. An offshoot of the theory of dynamical systems, complexity has become a subject of considerable scientific interest over the last few decades. Parts of it have become widely known under the fashionable rubric of chaos theory. This shift in focus towards complexity is not so much a consequence of some dramatic new discovery or revolutionary development, since certain essential details have been known for quite some time. Instead, it was primarily the result of advances in computer technology which have allowed scientists to address previously intractable problems. Nevertheless, whether this shift towards complexity constitutes some sort of scientific revolution is controversial.

Complexity research claims a high degree of generality: it is supposed to apply to extremely heterogeneous areas. The unifying idea that binds these heterogeneous areas is that of a complex system. Complex systems typically show different patterns of behaviour than simple systems. Research into complex systems can only meaningfully begin once the simple systems (i.e. parts of complex systems) are more or less understood. There are several areas of science where simple systems were indeed deciphered in the course of the first half of the 20th century. Interest then moved on to more complicated systems. There are several source of such a general idea of a complex system.

The first source was the research into dynamical systems in the context of classical mechanics that started as early as the end of the 19th century. The solar system exemplifies such a dynamical system. The solar system appears to be fairly simple because the planets look as if they would revolve around the sun forever. Yet, mathematical research has shown that it is far from clear that the revolution of planets around the sun will indeed go on forever. For instance, it is quite imaginable that one of the planets gains so much energy from the other planets that it leaves the solar system altogether, leaving the others in states of lower energy such that they circle the sun in lower orbits. The question that results is whether the solar system is stable or not.

A second source of complexity research is computer science and this in a two-fold sense. Computers are the primary tool used for complexity research. For instance, for dynamical systems, exact solutions are hardly ever obtained. This is the reason why the field lay mostly dormant for about half a century. Most of the insights of complexity thinking are gained through the use of computer models of the most diverse kinds. Computer models represent a given situation by abstracting from everything that does not seem relevant to those aspects of the system's behaviour one is interested in studying. Thus, seen as naturalistic representations of a given system, computer models appear as gross misrepresentations. Still, in successful models this does not prevent them from mimicking exactly the relevant dynamical aspects. For instance, in the very active field of artificial life research, most characteristics of the real physiological set-up of actual animals is completely ignored. Only a few traits reminiscent of real animals, like the production of offspring or certain rudimentary forms of locomotion or predation, enter the picture. If they are cleverly chosen, they prove sufficient to mimic certain aspects of the dynamics of a population of real animals. For instance, in a computer model that simulates a population with replication, mutation, and competition among individuals, spontaneous emergence of parasites may occur along with some new phenomena these parasites might generate.

Computer science stimulates complexity research in a second sense. Computations themselves provide a comparatively perspicuous model for the distinction between the simple and the complex, which is at the core of complexity research. Still, it should be recognised that there is no definition of the central notion of complexity that is both wide enough to cover all of the paradigm cases, and narrow enough to exclude trivialities. Further sources of complexity research include cybernetics, information theory, the theories of automata, of autopoiesis, and of molecular self-organisation, as well as systems theory, non-equilibrium thermodynamics, and synergetics.

The area of complexity research emerging from these various sources attempts to be a new, unified way of contemplating nature, human social behaviour, life, and the universe itself. It is an interdisciplinary approach fuelled by sophisticated mathematics, mathematical modelling, and computer simulation. It is inspired by observations that have been made on complex systems in the most diverse fields: meteorology, climate research, ecology, economics, physics, embryology, computer networks, and many more. These systems exhibit behaviour that is strikingly different from that of more simple systems. Typically, the behaviour of complex systems cannot be guessed or calculated on the basis of knowledge of their parts and their composition in the system. In fact, the components of the system interact in a way which severely limits predictability. The limits to predictability come in degrees. Some of these limits can (and will) be overcome by greater computing power and better algorithms. Some limits are of a deeper nature, but could be overcome if we had unlimited computing power and exact calculations. But some limits are of an absolute nature and could not be overcome by any possible means. Thus, complex systems exhibit so-called emergent properties and laws. In other words, they exhibit properties and laws shown only by systems of that degree of complexity that come as a surprise given the knowledge of the system's components and their composition. Complexity research is, therefore, seen by many of its proponents as anti-reductionist (see section 1.1), since new levels with new laws emerge which could not have been predicted by the analytic procedures characteristic of reductionistic research strategies.

One of the key processes responsible for the surprising behaviour of complex systems is self-organisation. This is the emergence of an orderly behaviour of some or all components of the system; in other words, some co-ordination among them. The crucial point is that this co-ordination is not brought about by some force or influence acting on the system as a whole, but by the interaction of the components which leads to this collective effect under certain circumstances. Self-organisation is a paradigm case for the development of order out of disorder. Typically, the emergence of new order happens in systems that are neither too orderly (like crystals), nor too disorderly (like turbulent fluids). Metaphorically speaking, it happens in systems 'on the edge of chaos'.

As a result of self-organisation, complex systems may exhibit spontaneous transitions into new states without apparent macroscopic causes. The reason for this is either that minuscule outside influences may bring about huge effects, or that the system's intrinsic instability drives it in some direction. Of particular interest are complex adaptive systems that occur in various sciences such as economics (for example the economy of a certain region), ecology (the ecosystem of a pond), biology (the immune or nervous system of an organism, the development of an embryo), and artificial intelligence (computer networks), to name just a few. These systems adapt to changes in their environment in an often extremely surprising way. In these cases, the idea of complexity research is that there must be some general common principles governing this sort of adaptive behaviour.

One of the most intriguing features of complexity is the fact that very complex behavioural patterns of a system can be generated by the rather simple mathematical rules the system follows. Many dynamical models begin by replacing the continuous flow of time by a set of equidistant points in time. The system's behaviour is then modelled as a series of discrete states. This series is generated through the repetitive application of a fairly simple rule starting with some initial state. Even if that rule is fairly simple, extremely complex and surprising behavioural patterns which do not seem to be built into the design of the rule can result. Typically, these transformation rules are non-linear. Non-linearity is a precise mathematical concept that can be explained as follows. A system is linear if that system's behaviour can somehow be described by a proportion. For instance, if a system's response to a disturbance is linear, then the response will increase with an increase in the disturbance and it will decrease with a decrease in the disturbance. If a behaviour over time is linearly dependent on its initial conditions, then small changes of the initial conditions will result in small changes of the system's behaviour over time. In non-linear systems, these properties that make the behaviour of linear systems easy to predict do not hold. Small changes may have disproportionally huge effects. The so-called butterfly effect catches this point nicely. Due to the extreme non-linearity of the global weather system, the disturbance caused by a single butterfly in Africa could result in a tornado in North America thirty days later. Thus, the properties of complex systems mentioned above, such as unpredictability, the emergence of new properties and laws, and self-organisation, are all related to the non-linearity of those systems.

Sometimes, highly enthusiastic pronouncements about complexity research are heard. It has been claimed that complexity even offers an entirely new world view which has the potential to settle some large issues; for example, how did the world become so complicated? Why is there so much order and structure in a world with so much instability? Why does innovation seem to thrive at the boundary separating order and disorder? It remains to be seen whether these promises will be fulfilled.

Science is a social enterprise dependent on communication and co-operation among scientists. Communication has a two-fold function in science. It is necessary both in order to avoid wasteful duplications of research efforts (this is accomplished through the quick dissemination of research results), and to ensure that systematic criticisms of any claim to scientific knowledge can be made through independent evaluation. Science's specific claim to knowledge includes its objectivity, and objectivity implies intersubjectivity. This means that the validity of the results of scientific research should be independent of factors such as gender, ethnicity, age, and nationality, as well as any other distinguishing characteristics of the researchers involved. Thus, according to the nature of science, there should not be any national barriers which hinder the dissemination of research results and their critical evaluation. Furthermore, because the properties of the systems and objects which are studied by many different fields are of a universal nature, such as properties of matter, principles of life, etc., the world-wide exchange of data, knowledge, and ideas is to the advantage of researchers around the world.

As a matter of fact, few enterprises in the world are as thoroughly internationalised as science. In many scientific disciplines, the leading laboratories and institutions are scattered over different parts of the world, but they exchange personnel, ideas, and research materials. There are innumerable international science organisations, including international field-specific unions of scientists, which unite national science organisations. These international unions share an umbrella organisation, ICSU (International Council for Science), together with official national representatives. Research results are published in a growing number of international journals. In most institutions, the credentials of scientists are judged according to how well their work is represented in such journals. Scientific consensus, where it exists, transcends any national, cultural, or continental borders. Where it doesn't exist, the lack of consensus usually has nothing to do with national mentalities or cultural differences, at least not for the last 50 years or so. But even earlier, there are many known cases of scientists trying to collaborate with colleagues in an enemy state during a state of war. Science allows people from very different cultural backgrounds to communicate and share ideas in the interests of the common good.

Another reason for the international co-operation in science which has steadily continued to develop over the last few decades is simply the sheer size of many large-scale projects. The size and expense of these projects has made it simply quite unfeasible for many nations to maintain scientific research activities in a growing number of fields; unfeasible, that is, unless they enter into co-operative arrangements for the construction and operation of expensive scientific facilities. Perhaps the most well-known example of such co-operation is the planning and construction of the international space station. But right here on Earth, high-energy particle accelerators, the human genome project, and many forms of global environmental research are all examples of enormous projects which require international co-ordination and co-operation.

The most important reasons for international co-operation have been recognised over the last 20 years. The detrimental effects of human activities have become so widespread and intensive that they are affecting the environment on a global scale. Such changes in atmospheric composition, in land cover, and in the oceans, as well as related climate changes and diminishing biodiversity are now beyond dispute and are collected under the rubric 'global environmental change' or 'global change'. Understanding and addressing global change requires a truly international scientific effort of unprecedented co-operation and interdisciplinarity. In response to this challenge, UNESCO and ICSU have sponsored or co-sponsored the World Climate Research Programme (WCRP), DIVERSITAS, a programme on biodiversity science, the International Geosphere-Biosphere Programme: a Study of Global Change (IGBP), and HDP, a programme that addresses the Human Dimensions of Global Environmental Change, to name just a few. Moreover, these programmes co-operate where suitable, especially in the interfaces between the natural and the social sciences. There have also been many international co-operative conventions, mainly as a response to global environmental degradation. These include the Montreal Protocol on Substances that Deplete the Ozone Layer (1987), the Basel Convention on the Transboundary Movement of Hazardous Wastes and Disposal (1992), conventions on Biological Diversity (1992), Climate Change (1992), and Desertification (1994), again to name just a few.

All nations share a vital interest in monitoring global change, and they also have much to contribute to the common understanding of the Earth system. Satellites can scan the globe, but much of the information needed by researchers must be obtained locally from the land or the oceans. Whilst laboratories, data banks, and computers produce impressive analyses, local observations and insights are needed to bring theses analyses to life. Only studies focused on regional and local conditions can adequately assess the real implications of global environmental changes on a global scale. For instance, the Global Change System for Analysis, Research and Training (START), co-sponsored by ICSU and UNESCO (among others), is the international scientific community's response to the need for research on regional environmental change. START promotes interdisciplinary research at a regional level by developing research networks. The purpose of these networks is to assess regional impact and to provide regionally important information. The START initiative also helps build endogenous capacities in the developing regions of the world so that they can participate effectively in the various scientific projects of the global change research programmes. Thus, regional research networks are instrumental in mobilising resources to augment existing scientific capabilities and infrastructure - something of particular importance for developing countries because the developing countries influence global change and are especially sensitive to it. Further objectives of START are the enhancement of communication between researchers and the strengthening of data and information-system capabilities supporting the regional research networks.

The growing need for international collaboration and the novel characteristics of many of these enterprises raise several policy issues which deserve careful consideration. First of all, it is important to recognise that international co-operation on fundamental scientific research projects has not grown proportionately with the internationalisation of emerging science and technology issues facing the global community today. Nor has international co-operation increased proportionately with the spread of scientific competence and the advent of new information and communication technologies. Given the large incentives and potential benefits that large-scale collaboration can provide, why isn't there more international collaboration in scientific and technological research? How can the potential for free riding be reduced without jeopardising knowledge distribution? Who will take the initiative and who will manage participation in international projects involving sovereign countries?

Because nearly all scientific research is still funded, organised, and implemented at the national level, international co-operation presents a major challenge to the global community. In order to help build coherent efforts out of basically national research projects, the international programmes have adopted a 'value-adding' approach aimed at knitting together the contributions from individual projects in order to address larger issues. This includes building consensus on research priorities and plans, and co-ordinating the use of expensive infrastructure in order to achieve effectiveness and resource efficiency. In addition, this approach supports the creation and development of international, inderdisciplinary research networks, common experimental protocols, standardised methodologies and data. It also supports the comparison of models, the integration and synthesis of global change research, and the timely and appropriate dissemination of knowledge to the policy and resource management sectors.

This last activity is a major challenge facing the international global change programmes. Global change itself is a manifestation of current unsustainable development undertaken by societies around the world. The transition to sustainability must be based on a sound scientific understanding of the global systems and their effects on human beings.

There appear to be enormous opportunities for constructive change in global scientific research management. The increasing use of new information and communication technologies will facilitate information transfer and allow international research networks to develop. There seems little doubt that, faced with global challenges, these enhanced technologies will lead to a growing clamour for reform aimed at increasing integration of resource allocation, research planning, and information distribution. One of the most significant challenges facing the international community in the coming century will be the development of management and policy mechanisms for the co-ordination of the international scientific community in order to assure that the division of labour reflects abilities to contribute to co-operative scientific projects rather than national interests. Of course, a primary aim should be to develop realistic and practical policies to reduce the increasing disparities and disadvantages that already exist.

The new information and communication technologies have modified the techniques of scientific investigation. They have provided researchers with the tools for simulating, controlling, recording, and analysing huge quantities of data. In economic terms, they have helped reduce the costs of scientific research by making data collectively available - data which is often very expensive to gather or produce. ICSU has established a Committee on Data for Science and Technology (CODATA) which is concerned with the organisation, management, quality control, and dissemination of scientific and technical data. Collective data sharing is leading to new institutional configurations and establishing electronic relations between researchers around the globe. In short, science is developing into a test-bed for many complex technical, economic, social, and organisational issues concerning international communication and information distribution.

But the tensions emerging out of these new developments should not be overlooked. Science needs unrestricted access to data world wide. The private sector, however, has a strong interest in protecting data in some areas. But in other areas, it has an equally strong interest in ensuring the free collection of data of various kinds. On the other hand, individuals have a desire and right to protect their privacy. Actions have been taken by the World Intellectual Property Organisation, the European Union, and the United States aimed at introducing new intellectual property laws. Data bases are not covered by copyright because they do not meet the test of creativity in the arrangement of data. Sectors of the information industry, however, believe that a new copyright clause is needed to protect their investment in creating databases and to guard against piracy. Extending these rights could impose serious constraints on science and education, undermining the ability of researchers and educators to access and use scientific data. It would make it more difficult for scientists to compile global or regional data bases, or to re-use and re-combine data for publication or instructional purposes. The trend toward commercialisation of scientific data is a cause of great concern for all developing countries because it counteracts access to scientific knowledge and data which is indispensable.

In addition, because of rising costs, the mainstream scientific literature which is produced in a few countries has increasingly become inaccessible to scientists, students, and even libraries in a growing number of countries. Scientific publishing has become a big business and a commodity only for those countries who can afford it. This increases the importance of efforts aimed at increasing participation in science publishing by all countries. The potential benefits of modern information technologies are not being sufficiently put to the service of the global scientific community. Instead, they are being used to the economic advantage of just a few enterprises. In 1992, ICSU, in co-operation with UNESCO, established the International Network for the Availability of Scientific Publications (INASP), in order to address many of these challenges. Its objectives are to support and strengthen existing programmes for the distribution, publication, exchange and donation of books and journals, to encourage new initiatives to improve the availability of quality scientific literature, and to help establish the sustainable exchange and distribution of scientific publications.

It is difficult for scientists in less developed countries to compete with colleagues in more affluent nations, as they often lack adequate resources. Laboratory equipment, journals, software, and other necessary items tend to carry Western price-tags. The same is true of page charges in many scientific journals. In addition, the dominance of the English language in the international scientific community conveys an advantage to scientists from English-speaking countries and from Europe. All of these factors lead to a marginalisation of the developing world within science, and consequently to a very low representation of developing countries in the leading international journals. To make matters worse, a further consequence is that assessments of scientific productivity solely relying on citation analysis drastically underestimate the research output of developing countries, as the scientometric institutes mainly index international journals. South-South co-operation is one strategy for addressing these problem. For instance, the Third World Network of Scientific Organisations, in collaboration with the Third World Academy of Sciences and with the participation of the Centre for Science and Technology of the Non-aligned and other Developing Countries, has produced many useful publications including Profiles of Institutes for Scientific Exchange and Training in the South. The African Energy Policy Research Network (AFREPREN) is a good example of a successful African network coordinated and run entirely by Africans. The main conclusions to be drawn are that there are still many challenges which need to be met before the ideal of science as truly international can be realised, and international co-operation must be used more effectively, especially to counteract monopolist trends and to ensure a wider and more democratic distribution of the resources and products of scientific activity.

It is evident that human resources constitute the ultimate basis for the wealth of nations. Whereas capital and natural resources are passive factors of production, human beings are the active agents who accumulate capital, exploit natural resources, build social, economic, and political organisations, and carry forward national development. Furthermore, as the 1987 report of the World Commission on the Environment and Development Our Common Future emphasised, development must become sustainable to ensure that it meets the needs of the present generation without jeopardising the ability of future generations to meet their own needs. But the first prerequisite for sustainable development is education. Only an informed public and a trained workforce can introduce the new sustainable production and consumption pattern which is required. What are the primary skills that people must have in order to carry forward the development of their country with the additional constraint that this development must be sustainable? What knowledge do they have to have? What resources must they be able to use? On what can they rely? UNESCO and ICSU maintain that one of the key components that make the development of these abilities possible is science, and more specifically, on top of a solid general education, a proper science education. What are the grounds for this conviction?

A look at recent history may be helpful. Japan and Germany have been enormously successful in re-building their countries after World War II. Among the developing countries, some have been much more successful than others in developing their economy, especially in Southeast Asia. Clearly, as stated above, it is human beings who drive development. But what do the countries mentioned have in common? Two things stand out. First, all of these countries have placed great emphasis on secondary education with a strong science component. Secondly, research and development activities in these countries were supported by a variety of measures. These factors played an important role in the positive development of these countries as a pool of trained individuals is of prime importance for developmental activity. Fundamental sciences form an essential part of any curriculum for training these individuals.

Why is this so? Why is training in mathematics, physics, chemistry, and biology so helpful? Through this training, students acquire skills that derive from the special nature of science, more specifically, from its systematic nature (see section 1.1). Of course, this training provides students with the ability to address scientific questions as they arise within the realm of science. But perhaps even more importantly, it also provides the intellectual skills needed for addressing some of the general questions the global community is facing today. For most of today's environmental and developmental issues, the sciences are essential for detecting and analysing problems, identifying solutions, and ensuring sound policies and actions. At the same time, the complexity of problems makes interdisciplinarity and integrated approaches, which include contributions from the social science, important methodological tools.

On the one hand, science trains systematic thinking by forcing the articulation of appropriate questions and the identification of the resources needed for their answers. This systematic training is applied to concrete problem-solving tasks. On the other hand, the systematic approach must not petrify the thinking process. Innovative, critical thinking is required as not all solutions to given problems can be arrived at by applying well-established methods. In science, the prime authority is the better argument, not custom, social authority, or convention. However, the need for creative innovation does not imply an anything-goes mentality. If science has developed ways to cope with a particular set of problems successfully, then any would-be innovation must be backed up by very good arguments, otherwise it will be ignored or discarded.

Science students must first learn some mathematics and informatics because these are essential parts of science which serve many functions: theories are frequently articulated in mathematical language. Testing theories and hypotheses involves statistics. Given large quantities of data, filtering out the interesting ones requires mathematical screening methods. Writing computer programmes makes use of algorithms, and much more. Mathematical thinking trains the ability to abstract: to learn to distinguish the unimportant from the relevant in a given context, and to discard the unimportant. In every science problem, a clear articulation of the situation to be dealt with and the questions to be answered is required. What are the relevant aspects of the situation at hand? What are the resources for identifying the relevant? Is it a well-confirmed theory, working hypothesis, or sheer prejudice? What is it that I want to know about the given situation? Is it possible, on the basis of the information about the given situation, to answer these questions? Do I need more information, and if so, what do I have to do in order to get it? Do I need additional observations or experiments, or can I rely on existing information? Where is this information accessible? Am I able, given my resources, to perform the necessary steps, be they informational, experimental, observational, calculational, or theoretical? What theories, hypotheses, or models have to be used to answer the main questions? Are these theoretical assumptions secure enough for the given purpose? How is the relevant information about the given situation to be extracted from the theoretical assumptions to be used?

These are the kinds of questions that students are trained to deal with during their science education. They are also the kinds of questions asked and answered in the process of productive scientific research, which usually begins at the postgraduate level. Yet, there are also other abilities which come into play during active research into partly unknown territory. First of all, in spite of the fact that initial ideas come from individuals, productive research in the natural sciences nowadays mostly occurs in groups. This, then, involves all the social skills necessary to deal with such a situation. Second, in many research endeavours, a narrow scientific background will not suffice because knowledge from many areas must be drawn upon if the problem at hand is to be successfully solved. Thus, perspectives from other disciplines have to be integrated, which necessitates the establishment of interdisciplinary communication. As the cultures of different scientific disciplines are fairly different (subdisciplines may differ considerably - even within a single discipline), productive interaction between people with different backgrounds and outlooks is often required.