Theory and Reality

by Peter Godfrey-Smith

  However, Feyerabend massively overextends his argument, into a principle that cannot be defended: "Hence it is advisable to let one's inclinations go against reason in any circumstances, for science may profit from it" (197 5, 156). Feyerabend claims that because some principle or rule may go wrong, we should completely ignore it. The claim is obviously crazy. The policy of catching trains that are scheduled to take you where you plan to go is a policy that may go wrong. The train might crash. Or if you caught a different train, you might meet the love of your life on the way. All that is possible, but no rational person regards these mere possibilities as sufficient to discredit the rule that it is best to take trains that are scheduled to go where you want to go. All applications of everyday rules of rational behavior presuppose judgments about which outcomes are probable, or typical, and which are far-fetched or unlikely. Sometimes we might be able to put exact numbers on the chances of these different outcomes, but often we cannot. Sometimes we rely on informal judgment to guide us. In rational behavior, nothing is guaranteed, but some policies and rules can be justified despite this. (The modern Bayesian theory of belief and action, which will be discussed in chapter 14, is based upon this idea.) And what is true for everyday behavior is true also for science. The mere possibility that a rule might lead to bad consequences proves very little. We need more than a mere possibility before we have grounds to doubt a principle. Science "may profit" from all kinds of strange decisions, but "may" is not enough.

  So we have a mixture of good and bad argument in Feyerabend's treatment of these issues. It would be foolish for a philosopher to ignore the strangeness of having one's favorite principles of theory choice come out in favor of the seventeenthcentury Aristotelians and against Galileo. The example is so important in the history of science that it takes a brave philosopher to ignore it. It is by no means clear that well-developed philosophical theories do make the wrong call, though. Recall Laudan's rules discussed in section 7.4 above; they would probably steer us rather well here. These rules would at least tell us to pursue the Galilean program, on account of its rate of increase in problemsolving power. And in time, it would become rational to accept the Galilean view.

  7.6 Pluralism and the Ramblings of Madmen

  For Feyerabend, science benefits from the presence of a range of alternative ideas and perspectives. Let us look at Feyerabend's ideas about pluralism and diversity in science.

  In his 1970 paper about Kuhn and Lakatos, Feyerabend proposed two general principles that should guide science. We should keep in mind that "rules are made to be broken" in Feyerabend's view. But the principles are worth discussing.

  The first rule Feyerabend called the "principle of tenacity." This principle tells us to hold onto attractive theories despite initial problems and allow them a chance to develop their potential.

  That is a start, but if everyone followed this rule, nothing would ever change. So Feyerabend adds a second principle, the "principle of proliferation." This principle tells us to make up new theories, propose new ideas.

  Kuhn says the proliferation of new ideas should wait for a crisis. But why not put forth new ideas all the time? Thus we reach Feyerabend's ideal picture of science: we have a population of people happily developing their theories and also trying to think up new ones. Some pedestrian work is needed to help develop the existing ideas, but this should not interfere with the imaginative work.

  Feyerabend claimed to be following the tradition of the philosopher and political theorist John Stuart Mill. In his classic paper On Liberty ([18591 1978), Mill argued that society benefits from a diversity of ideas and ways of life. The constant proliferation of new theories creates a "marketplace of ideas," in which many options are explored and the best prevail. Some of Feyerabend's descriptions of ways in which science might benefit from unusual sources of ideas-"the ramblings of madmen" (1975, 68)-were intended as descriptions of inputs into this marketplace. (See also Lloyd 1997.) Feyerabend argued that we can often only perceive the limitations of our current perspective if we try to step outside it, at least temporarily. Achieving a novel, external perspective on ideas that we usually assume uncritically is often the beginning of progress. What we usually consider "established facts," including observable facts, are often laden and contaminated with prejudices and outdated ideas. Any source of an external vantage point is to be valued. Alternative theories, even theories with massive problems, can provide this kind of external vantage point.

  These ideas about the need for external vantage points in challenging familiar assumptions are interesting. And the idea of a marketplace of ideas is a powerful one. But Feyerabend's account seems to have left out something very important. And this omission undermines his version of the "marketplace of ideas" doctrine.

  What is missing in Feyerabend's picture is some rule or mechanism for the rejection and elimination of ideas. Feyerabend gives a recipe that, if it was followed, would lead to the accumulation of an ever-increasing range of scientific ideas being discussed in every field. Some ideas would probably become boring and might be dropped for that reason. But aside from that, there is no way for an idea to be taken off the table. So a question immediately becomes pressing: what are we supposed to do when we have to apply one of these theories to a practical problem? What do we do when the bridge has to be built? Which ideas should we use? Not the most "creative" ones, surely! Feyerabend never gave a satisfactory answer to this question.

  If what we really want from science is to have a community full of lively, imaginative discussion, then Feyerabend's recipe for scientific behavior is an appropriate one. Science then would be very similar to art. But if part of the role of science is to guide us in solving practical problems, then Feyerabend's recipe seems completely misguided. If science must be applied to problems, there must be a mechanism of selection in science, a mechanism for the rejection of some ideas. Proliferation of alternatives is part of science, and another part is selection among alternatives.

  In the last few years, for example, the government of Thabo Mbeki in South Africa has shown an interest in radical ideas about the causation of AIDS. According to these ideas, the virus identified by mainstream science as the cause of AIDS, HIV, is regarded as either relatively unimportant or altogether harmless. In reply to the storm of criticism that resulted, Mbeki has sometimes said that he is simply interested in an open-minded questioning of theories and the exploration of diverse possibilities. Surely that is a properly scientific attitude? This reply has been rightly criticized as disingenuous. Science needs the invention of alternatives, but it also needs mechanisms for pruning the range of options and abandoning some. When the time comes to apply scientific ideas in a public health context, this selection process is of paramount importance. Then we must take from science the well-supported view that AIDS is caused by a virus transmitted through body fluids, and we must guide policy and behavior with this view.

  Michael Williams, in the encyclopedia article I mentioned earlier, says the following about how Feyerabend looks in retrospect: "While some of his views may strike some philosophers as overstated, their general spirit has some claim to be seen as today's conventional wisdom." I would give an almost opposite summary. While a few individual pieces of Feyerabend's view have become something close to conventional, the "general spirit" of his work involves a principle that is still unconventional because it is clearly false. This is the principle that we should think of the social role of science in the same way we think of the social role of art. On the contrary, imagination and creativity are one side of science but not the only side.

  7.7 Taking Stock: Frameworks and Two-Process Theories of Science

  In this section I will discuss a general theme that has appeared several times over the last few chapters. This section is a "time out" from the chronological story central to the book.

  The theme I will discuss has reverberated continually through twentiethcentury philosophy and continues to be important. I will introduce it as a distinct
ion between views about scientific change, or conceptual change more generally. The distinction is between one-process and two-process theories. If we are trying to understand scientific change, should we recognize two "levels" in science, with different types of change occurring at these two levels? More precisely, should we see scientific change as involving (z) changes made within the boundaries provided by a general framework and (z) changes at the level of the frameworks themselves?

  The alternative (or rather, one alternative) is to give a unified account, in which there is no qualitative distinction made between two levels, layers, or kinds of change.

  The opposition between Popper and Kuhn can be used to make the contrast vivid. For Popper, scientific change always involves the same process the cycle of conjecture and refutation. This is what we see at the level of small changes to details and also at the level of fundamental changes in worldview. For Kuhn, in contrast, there are two qualitatively different kinds of change in science; changes within paradigms and change between paradigms involve fundamentally different processes. A paradigm (in the broad sense) is a clear example of a framework in the sense relevant here. Change within a framework is guided by principles supplied by the framework. Because frameworks supply these principles, moves between frameworks are more problematic, hard to describe, and often disorderly.

  Let us move to another contrast that is fairly clear and vivid: the contrast between Carnap and Quine. Carnap, in his later philosophy, used the term linguistic framework and distinguished moves made within these frameworks from changes made between them. The principles that are fundamental to a framework will appear as analytic sentences if they are explicitly stated. Moves made within a framework involve the assessment and testing of synthetic claims. For Carnap, many alternative frameworks are possible, and people can switch between them. These switches, however, involve a different kind of process from moves made within them. Moves made between frameworks are not sensitive to particular factual results; they are sensitive to a kind of pragmatic assessment of the overall usefulness of the framework. If one framework seems not to be working, then we try another. Carnap's frameworks are "thinner" than Kuhns; they involve just basic linguistic and logical rules, not scientific principles.

  Quine, in "Two Dogmas of Empiricism" and elsewhere, argued against this two-layered view of language and knowledge. For Quine, all changes made to our belief system, whether large or small, involve the same kind of holistic tinkering with the web of belief. We accommodate experience by making as few changes as possible and keeping our worldview as simple as we can. There is no distinction between changes within and changes between frameworks.

  How might we decide between a one-process view and a two-process view? Within twentiethcentury philosophy, many people were persuaded by Quine's holism. These arguments were based on very general considerations and not on the history of specific episodes in science. Quine's most powerful argument is usually seen to be his claim that there is no way to mark out the distinction between changes within and changes between frameworks in a way that is scientific and does not beg the question.

  Kuhn, however, had no problem distinguishing normal science from revolutionary change in actual scientific cases. He saw two processes as a clear fact of history. And it could be argued that Kuhn was led to important insights via his recognition of this distinction. Recently, Michael Fried man (zoos) has argued that Kuhn was right and Quine was wrong on this point. If we approach actual scientific episodes using the idea of frameworks and the distinction between two kinds of change, we will be able to make more sense of how the sciences evolved. Quine's claim that we cannot make good sense out of a distinction between two kinds of conceptual change seems to be based on an overly stringent conception of what we would have to do in order to recognize two kinds of change.

  Is Quine's denial of the two-process view justifiable? Thomas Ricketts, in a z98z. defense of Quine, tries to say in more detail how Quine might resist a two-process view. Suppose a scientist actually pins a set of basic principles on the laboratory door and insists that these principles constitute his framework. The principles on the door might change, the scientist says, but then they will change by a special process. Surely Quine will have to accept a two-process view in that case? Perhaps not. A scientist might pin some principles on the door and say they are different, but if Quine is right about the holistic nature of testing, the actual processes by which the scientist makes modifications to his beliefs will all be of the same kind. A pragmatic process of making adjustments to avoid tensions and unexpected observations is all that goes on, whether the ideas being changed are pinned on the door or not. There is no sense in which the moves made within the framework are "guided by facts" and those made to the framework are "merely pragmatic." All changes made to any belief result from the same kind of tinkering with the total network.

  So in working out whether the two-process nature of scientific change is merely an illusion or not, we are led once again to fundamental questions about testing and confirmation. Still, I would have to say that despite these problems, the idea of a two-process theory of scientific change certainly looks like it has been useful to people working on specific cases. Alongside Kuhn we have Lakatos and Laudan, who have different kinds of two-process views. As I said earlier in this chapter, we can use the different theories of Kuhn and Laudan to make distinctions between different kinds of scientific fields-some fields are guided by Kuhnian paradigms, and some involve ongoing competition between research programs. Other fields might have a mixture of the two. These certainly look like useful distinctions. And Quine and his allies would concede that usefulness is the ultimate arbiter in a case like this.

  This whole area has been the topic of lively discussion recently, and there is a lot more that could be said. The ghost of Immanuel Kant hangs over the whole discussion, as Kant was the first philosopher to develop a detailed view in which an abstract conceptual framework acts to guide empirical investigation ([1'781] 1998). According to Kant, the basic framework that we bring to bear on the world is fixed and universal across all normal humans. We can never escape the framework (and would not want to if we could). Through the twentieth century, many philosophers found the idea of a "conceptual scheme" or framework appealing but insisted that these schemes can be changed and are not universal across cultures. Some of the radical ideas discussed in the next chapter can be seen as combining Kantian ideas about the role of conceptual schemes with a relativistic view in which alternative schemes are possible.

  Not all philosophers of science can be neatly categorized as having "two-process" views or "one-process" views. Feyerabend is an interesting case. He recognized the psychological power of linguistic and cosmological frameworks, but he insisted that the imaginative person can resist the bounds of a framework. Popper, also, rejected the whole idea of frameworks as constraining thought and knowledge. He called it "the myth of the framework." Feyerabend did not see frameworks as mythical, but he thought that their bounds could be resisted and overcome.

  Another interesting response to these issues is seen in Peter Galison's work (1997). Galison argues that what we often find in science is that fundamental changes in the different elements making up a scientific discipline are not in step with each other. Whereas Kuhn described a process in which there is simultaneous change in theoretical ideas, methods, standards, and observational data, Galison argues that within physics, fundamental change in experimental traditions tends to occur nonsimultaneously with fundamental change in theory. This is because of the partial autonomy of these different aspects of large-scale science. (Instrumentation is yet another tradition, with its own rates and causes of change.) So a big theoretical shift will be made more manageable by the fact that we can expect other aspects of the same field not to be changing at the same time. Disruptions happen more locally than they do on the Kuhnian model, and there are more resources available to the field to negotiate the transitions in an orderly way. The history of a scientific
field shows "seams" of several different kinds, but these seams do not line up with each other. The structure as a whole is made stronger as a result.

  Galison's picture shows us that there are many options for thinking about the relation between different kinds of scientific change. We should not think that there is a simple choice between the one-process camp (Popper, Quine, and Feyerabend in a sense) and the two-process camp (Carnap, Kuhn, Lakatos, Laudan, and Friedman). The situation is more complicated. And different kinds of frameworks have different roles-we should not think that Kant's universal conceptual framework has the same role as one of Laudan's research traditions! These are very different kinds of things. I have also not said anything about the distinction between two-process views that see people as modifying their frameworks and those that see people as jumping between them. (Maybe in the end there is no difference.) In any case, the introduction and criticism of two-process views of conceptual change has been a recurring motif in the last hundred years of thinking about science and knowledge.

  Further Reading

  Lakatos's most famous work is his long paper in Lakatos and Musgrave, Criticism and the Growth of Knowledge (1970). Another key paper is Lakatos 1971. Cohen, Feyerabend, and Wartofsky 1976 is a collection on his work.

  Feyerabend's most famous book is Against Method (1975), but his earlier papers are also interesting (collected in Feyerabend 1981). His later books are not as good, though Science in a Free Society (1978) has some interesting parts. The Worst Enemy of Science (Preston, Munevar, and Lamb zooo) is a collection of essays on Feyerabend. Horgan 1996 contains another great interview.