We have been examining the structure of science from the point of view that it is complex. However, many things. are complex that are not science. ln order for us to understand how science has achieved its hegemony in our technological civilization we must understand both the ways in which it resembles other systems, and the ways in which it is unique. In the case of science, most philosophers of science will agree that it is because science deals with Popper's third-world of reality, the world of pure ideas. But I will show that this is not sufficient. Popper's world 3 includes many things that are not science and excludes some things that are.

First, science began with, and to a great extent still is concerned with, Popper's world 1, or the world of physical objects. It is all well and good to develop a theory of prime numbers, but unless they, concern solid objects; or theories about solid objects then they become the province of the philosopher or poet and not the scientist. Ever since Thales began speculating about the structure of the world, a great mass of theories, methods. and attitudes has accumulated which have proven to be successful in explaining to us the nature of world 1. Call this mass a "hard core", or a "Paradigm" if you wish. This resides in the non-autonomous part of Popper's world 3. This is the same world that includes economic, social, and political systems. There is a difference between. these systems and science, that cannot be explained away simply by calling science a world 3 system.

I am going to designate the non-autonomous world 3 the 'pragmatic world 3' even though I am sure the term will rankle many Popperians. It is a world created by men for men.. Its value lies in the way it regulates men's actions. Scientists solving real world 1 or world 2 problems act in. ways that are constrained by the set of theories, attitudes and methods proscribed by the scientific constructs assumed within the pragmatic world 3. Scientists attempt to explain the first and second worlds In terms of the autonomous world 3. The others do not. That is an essential underlying difference.

To add to the difficulties. of the scientist. he can interact directly only with the pragmatic world 3. That is, he can change or modify the pragmatic but he can only discover the autonomous. Even the word "discover" is too strong. We can never really know if we have or have not discovered, we can. only guess that we may have come close to the truths of the autonomous world 3. The greatest difference between the scientist and the politician is the scientists faith that if he keeps working and improving his methods he can. come closer to making his pragmatic third-world structure an accurate reconstruction. of the autonomous third-world.

This mechanism I am going to Call "Paradigmatic complexity" because it is. a special application of complexity theory (as developed by Bonner, Simon and others) to the structure of science and scientific change. We begin with the subject of the last section, that is, that paradigms, .In my sense rather than in Kuhn's form a hierarchical control program to constrain the activities of scientists in order to increase the freedom of science as a field of endeavor. This is quite a different kind of constraint than that suggested by Kuhn which led his. critics to picture the Kuhnian scientist as a rather bored automaton following proscribed methods under dogmatically assumed theories and attitudes. The ideal scientist in my world is one who attacks his work like a demarcationist using methods, theories, and attitudes that have been shown in the past to be successful, possibly in a multitude of disciplines. I seem to think he comes closer to the kind of man who is attracted to science to begin with. He is a puzzle-solver in the Kuhnian tradition but with an open mind and a critical eye. Assuming this kind of man to be the rule rather than the exception in science can generate a theory as to how science grows and changes over time.

In this thesis I have been criticizing Thomas Kuhn, but for his ambiguities, and not his basic ideas. In general most of the criticism of Kuhn's thesis has been based on a misunderstanding of his fundamental ideas, probably caused by these ambiguities. Most of this misunderstanding was based on Kuhn's thesis of two kinds of scientific activity, normal and extraordinary. Most of his. critics rejected this because whether an activity was normal or extraordinary was more dependent on the observer than on the activity. I hope to clarify this with a description of the two kinds of change that illustrates both the ways in which they differ, and how they are related.

The first kind of change I want to discuss is what we might call normal change. Consider a scientist in his usual role of solving problems through science. He applies the theories. methods, and attitudes; of his program, and we can loosely define his program as those assumptions that he and his contemporaries find useful, and when successful he has broadened the role of his Program to include the kind of problem he was working on. Problems seldom restrict themselves to a single program and so there is likely to be some other related scientific activity that is affected by his success. Thus, he might very well have broadened not only his own program, but a number of related ones.

This may seem obvious. Most scientific activity follows this line, of improving the validity of a research program by applying it to real problems. However, In a very important sense it is. anti-Popperian. Popper and his followers believed that scientific progress could only be made by submitting theories to possible refutation. Yet they too attempted to subsume it into their over-all description of scientific activity. Popper in his concept of "Verisimilitude" and Lakatos in his "progressive problemshift". We might take a moment to discuss. these important ideas.

Both begin with the assumption that we can never be sure that we have found a scientific "truth". This, as I have made explicit before, is because science exists in the "pragmatic world 3" while scientific truths, if they exist at all, do so in the "autonomous world 3", a world we can never experience. we can only approximate. verisimilitude, according to Popper, is "truth likeness", it is the faith we have that our pragmatic approximation is similar to the abstract truth we can never completely know. Something has greater verisimilitude when we have faith that it does indeed come closer to the abstract truth. Of course, to Popper this is a rational faith because in each case we have reached our consensus by thoroughly testing all of the theories in question and one has passed its tests in such a way that it leads us to believe that it is a better approximation than any other. In his view our faith is still rational even if later we are to find out that what we had formerly believed to be the laws, of mechanics have suddenly changed. Popper's verisimilitude is highly dependent on the rationality of his refutationist approach. Lakatos, In an attempt to overcome objections to refutationism divided scientific activity into two different kinds of approaches. The first he called the negative heuristic, or activities performed to protect the 'hard core--the programs underlying assumptions. The second he called the positive heuristic. lt includes all activities performed to increase the empirical content of the program. If a scientific research program is operating primarily in the positive heuristic, that is, if, in a series of theories, each one predicts some novel, hitherto unexpected fact, then a theoretically progressive problemshift will result. If some of those theories are corroborated, then lt becomes empirically Progressive. Once it has achieved both, it is considered Progressive.

Popper's. verisimilitude and Lakatos' progressive problemshift are different descriptions of the same kind of scientific progress. Each description, though different, complements the other. For Popper it is, the attempt to develop a pragmatic third-world model that simulates autonomous third-world concepts. In the case of: Lakatos, his primary interest is in increasing empirical content. If we want to develop an understanding of science as it is in real life we must have a model that will include both verisimilitude and progressive problemshift. We must have a theory of scientific change that recognizes science as a complex system.

We must turn our attention back to our discussion of paradigms and in particular to the concept of a paradigm as a hierarchical control program. A paradigm is the complex system of theories, methods, and attitudes that scientists assume to be sufficiently accurate that they need not question them in the performance of their normal work. At the lowest level it consists of individual theories, attitudes, and methods that have each been tested and found to add empirical content to the entire system. The next level is the set of interrelations that link; these together to form a coherent body of scientific knowledge. At the highest level, these emerge into a model of reality that attempts to emulate Popper's autonomous world-3.

When a new idea first interacts with this complex paradigmatic system, it does so by solving or helping to solve a problem that a scientist is actively working on. Having accomplished that, it has to fit in with the set of interrelationships that tie the paradigm together. Like adding a piece to a puzzle. It must fit and lt must add to the picture. Finally, and most important, it must merge with the paradigmatic model of reality by being tested against the physical world 1, through the psychological world 2. Only then can it be considered part of the paradigm.

Perhaps these ideas can be clarified by an examination of the mechanisms that make It work. Popper said, "most of us today strongly believe --and have reason to believe-- in the Copernican model of the solar system (as revised by Kepler and Newton): as a certain arrangement of planets, moving in near ellipses around the sun, accompanied by their moons. What is_ the real reason why we believe in this model? Popper insisted that they consist ln "the story of critical discussion", critical evaluations of observations, of all of' the theories since Anaximander, even including Heraclitus' hypothesis that a new sun was born every day. At any point in time during this long period the prevailing set of theories was united by what Popper calls "the reasonableness. of belief," By this he meant that each one of these theories was put to the critical test by the scientists of their day. lt is because most of those Involved In science were persuaded that each of these theories represented the "best sense of reasonableness known to me", that they were accepted as part of the paradigm. Thus, they not only had to pass individual tests they also had to relate to each other in a rational manner. In this way one theory can be judged superior to another theory without proclaiming the truth of either. This is a picture of normal change, the gradual evolution of science ln general, of the scientific paradigm.

By way of contrast, as long as we understand that the paradigm is; just an attempt at modeling reality and does. not represent truth, that it is part of the pragmatic third-world and not the autonomous, then the possibility that the paradigm departs from reality by a considerable margin cannot be rejected. The best example is Newtonianisim. Part of the Newtonian paradigm was that its truths were absolute. This led Kant to the development of a priorism, which would resulting a stifling of scientific thought. However, Einstein refuted that absolutism. Of course Newtonian mechanics was reinterpreted as a special case of Einstienian, but the change from Newtonian absolutism to Einstienian relativism was a major and abrupt paradigmatic change, the kind of change that Kuhn called "Revolutionary". The mechanisms for this kind of change are quite different than those of normal change.